Citation
Chemical engineering education

Material Information

Title:
Chemical engineering education
Alternate Title:
CEE
Abbreviated Title:
Chem. eng. educ.
Creator:
American Society for Engineering Education -- Chemical Engineering Division
Place of Publication:
Storrs, Conn
Publisher:
Chemical Engineering Division, American Society for Engineering Education
Publication Date:
Frequency:
Quarterly[1962-]
Annual[ FORMER 1960-1961]
quarterly
regular
Language:
English
Physical Description:
v. : ill. ; 22-28 cm.

Subjects

Subjects / Keywords:
Chemical engineering -- Study and teaching -- Periodicals ( lcsh )
Genre:
serial ( sobekcm )
periodical ( marcgt )

Notes

Citation/Reference:
Chemical abstracts
Additional Physical Form:
Also issued online.
Dates or Sequential Designation:
1960-June 1964 ; v. 1, no. 1 (Oct. 1965)-
Numbering Peculiarities:
Publication suspended briefly: issue designated v. 1, no. 4 (June 1966) published Nov. 1967.
General Note:
Title from cover.
General Note:
Place of publication varies: Rochester, N.Y., 1965-1967; Gainesville, Fla., 1968-

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
01151209 ( OCLC )
70013732 ( LCCN )
0009-2479 ( ISSN )
Classification:
TP165 .C18 ( lcc )
660/.2/071 ( ddc )

UFDC Membership

Aggregations:
Chemical Engineering Documents

Downloads

This item has the following downloads:


Full Text

























Maurice A.^ Larson








A WARDLECTUR


















UNITED SATE Statement of Ownership, Management, and Circulation
(Requied by 39 USC 3685)
I Pubamtin Tila 2 PubEaton Number 3 FUln Dat
CHEMICAL ENGINEERING EDUCATION 1 0111- 9 0 9/8/98
4 Iss Frequency 5 Number ol ssuae Publshed Annually 6 Annual Sutscnpton Pnce
4 See Attached
Rate sheet
CHEMICAL ENGINEERING EDUCATION, Room 319, Chemal Eng neern Dept. Carole ocum
University of Florida, Gainesville, Alachua, Florida 32611-6005 Telepho.
352-392-0861
a Compt.e M'tIng A-;a.v of Haduaaea oa Ganeas Bueinas 00.u a put.lhr (Nonpr krj
Chemical Engineering Division, American Society for Engineering Education
11 DuPont Circle, Washington DC 20030

9a Full Names and olete mlng Maddres of Publear, Edlter, and Managing Edlor (iOa leave blak)
Pubusher (Name I d cm plet eallno addlea)
ASEE-Chemical Eng. Division, 11 DuPont Circle, Washington DC 20030

Edotor (NeMe mn~are eaIAog address)
Tim Anderson, Chem. Eng. Dept., Univ of Florida, Gainesville FL 32611-6005


Manaun Edits (ad m and .m.ea maa y -, -es)

Carole Yocum, chem. Eng. Dept., University of Florida, Gainesville FL 32611-6005

I0. Ownm, (m o non leann l Ie Oepuanioaa, oaeaday at aaona. aie Oa'e nmea td madn si olhe amanlin emsediat ollwd by iaa
naes )ad addreaasses of aa s slooInd m o nag I piepaonr onnaa olt eMeal a ount of stock llanaOnedbyat nran g' (he
n r eandaddses oln o naa~aa ors 1 a ,aownetmd bypa npaltnars oreneas, o er n aS namve a nda ds as eols oe o

Full Name Complete ailing Address
Official publication of Publisher Any mail addressed to owner should go to

as listed above Editor or Managing Editor listed above,







11 Knin Bondhida, Monrgan and Oher SC unmty Hoada Owning or
oldng 1 P it or Mnm ol Total Amount ol Bondls Mo. a, or
mm1rSotamea ena. hecbai0 ioneom
tell Narem_ CoRpe a baa ing A 55ram.










.12. Tax an St etu anFns .1ro 6y atnor r a vl0 aor- (o o nd aiat i rs ati..
o H Nol Canngs.d Duing Prseding 1E Monel
O Hea Changed During Peceding 12 Monlh (Pubuisher lamu subril epanaon ol change w*i tNis stalemnl)
PS Fam 3526, Snptwemer ins (aSam a alni Roem)

13 Pumbln a Tre 14. InDalte fo CIulafo Da a Belo
CHEMICAL ENGINEERING EDUCATION Summer 1998
mIS. an .snim u elaa, Avetn Ne. CO.pI ECmn Imue AotuI Nse COpl:i o Single Iu
enland Nature o Circultlon rlng ing 12 Months Pubihed Nr lo Fiong

a. Total Numbr olCopaes (Nlpress) 2075 2060


b Paid andtor nd Counter Sales (No l aied) 0 0
b nld 1 ) Sals T gh Dnle ma im S1ret Vemndm,

Requested --- --------------------------------
Cimatn (2) Faole equesma Ma Subip "on a Inde 1978 2003


c.Totm Paid andor Requlad culaton 1978 2003
(SumoHS ll0anzdlSl))

d Free oldlibmo by Mail
(Saplb. corpsman end oher a) 35 35

e. Fm Ditarbul Outside ma Mal (Cimera o oter mea) 0 0

f. Tot Fme Olstnbuaiaon (mo 1Sd ande) 35 35


g.To Distnbueon (Sm of Isc a IS )0 2013 2038

(1o) onre us, Leftner, s poad 62 22
h ConIM not
(2) Retums rom NeL iAgents 0 0

I Toa (Sum o 5g, I5h(l), and shl(2)) 2075 2060

Percent Pai o ander Rquested circmaon 98.26% 98.281

16 Publcrlln il emnl Owneahip Winter '99 uo ulcon
o PIsainton required Will be pnnltd in a eth- Isue of thi pulleaon,

17 Sigatr andTeoE akor. Pblit., .eBuries manager, owne Da
aInt6 a. s)o/ei 9, n ? diRtu_ 9/1/9


I erfy mai al eIormgas on furnihed on ah fornm m ue end Impleta i understand theat one who lumihes fals or imsleadng ,mloralln n this lor
or ,who oanru maral or Inlermal a r mqu lted on thelor maey be bje to ci a ano aa (limludSng nes and Impriona nt) ad/or san)csati
Onadring ml pe damage and nvil penaliu).


Instructions to Publishers

1. Complete and il ne oncopy ol this form with your postmaster annually on or before October 1 Keep a copy of he completed lfom or
your records.
2. In cases where the tiockholder or secdly holder is a trustee,. alude in la 10 and 11 the name of the person or corporation for
whom the trustee Is acting. Also include the name and address of indiliduals who am stoclholdeni who o or hold 1 percent or
mor of the total amount of bonds, mortgages, or other secuies of the publishing corpoaton. In e, 11, II none, check the box Use
blank sheets if more space is required.
3. Be sure to punish all crculaton lormato called for n Item 15. Free cduala on must be sho, n tems 15d, a. and
4 If the publcaton had saondclas authoraton as a general or requaeler publaton, this Slatemnen o Ownersip, Management, and
Cemulation must be published, it must be printed In any issue in October or. if lha publlcaon Is not published dunng October he it
sue printed aler October
5. In item 16, indicate the dale of the lssue in which this Statement of Owership will be published
6. iaem 17 must be signed.
FaHuma to fle or publish a isatemena ol ownership may lad to suspension of seondms authriaion

PS Form 3526, September 1995 (aReIr)












EDITORIAL AND BUSINESS ADDRESS:
Chemical Engineering Education
Department of Chemical Engineering
University of Florida Gainesville, FL 32611
PHONE and FAX: 352-392-0861
e-mail: cee@che.ufl.edu
Web Page: http://www.che.ufl.edu/cee/

EDITOR
T. J. Anderson

ASSOCIATE EDITOR
Phillip C. Wankat

MANAGING EDITOR
Carole Yocum

PROBLEM EDITORS
James O. Wilkes
University of Michigan
LEARNING IN INDUSTRY EDITOR
William J. Koros
University of Texas, Austin

-PUBLICATIONS BOARD
CHAIRMAN *
E. Dendy Sloan, Jr.
Colorado School of Mines

PAST CHAIRMEN *
Gary Poehlein
Georgia Institute of Technology
Klaus Timmerhaus
University of Colorado

MEMBERS
Dianne Dorland
University of Minnesota, Duluth
Thomas F. Edgar
University of Texas at Austin
Richard M. Felder
North Carolina State University
Bruce A. Finlayson
University of Washington
H. Scott Fogler
University of Michigan
David F. Ollis
North Carolina State University
Angelo J. Perna
New Jersey Institute of Technology
Ronald W. Rousseau
Georgia Institute of Technology
Stanley I Sandier
University of Delaware
Richard C. Seagrave
Iowa State University
M. Sami Selim
Colorado School of Mines
James E. Stice
University of Texas at Austin
Donald R. Woods
McMaster University


Chemical Engineering Education


Volume 33


Number 1


Winter 1999


> EDUCATOR
2 Maurice A. Larson, of Iowa State University

> DEPARTMENT
6 Washington State University

> AWARD LECTURE
12 Do Changes in the Chemical Industry Imply Changes in Curriculum?
E.L. Cussler


> CLASSROOM
18 Discontinuities in ChE Education, Stephen Whitaker
58 Permeation of Gases in Asymmetric Ceramic Membranes,
Carlos Finol, Joaquin Coronos

> CLASS AND HOME PROBLEMS
26 Non-Adiabatic Container Filling and Emptying, Noel de Nevers

> RANDOM THOUGHTS
32 FAQS, Richard M. Felder, Rebecca Brent

> LABORATORY
34 Evaluation of Computer-Simulation Experiments in a Senior-Level Capstone
ChE Course, Scott R. White, George M. Bodner
46 Demonstrating Simultaneous Heat and Mass Transfer with Microwave Drying,
Cheri C. Steidle, Kevin J. Myers
50 Medical Surveillance and the Undergraduate Thesis, lan A. Furzer
54 Laboratory Experiment in Biochemical Engineering: Ethanol Fermentation,
Alberto Colli Badino, Jr., Carlos Osamu Hokka
66 Introducing Process-Design Elements in the Unit Operations Lab,
Christine L. McCallum, L. Antonio Estivez

> CURRICULUM
40 Process Analysis: An Electronic Version, George B. DeLancey
62 A Joint Chemical/Electrical Engineering Course in Advanced Digital Process
Control, Joseph J. Feeley, Louis L. Edwards

> ASSESSMENT
72 Ranking Graduate Programs: Alternative Measures of Quality,
John C. Angus, Robert V. Edwards, Brian D. Schultz

LEARNING IN INDUSTRY
84 From the Classroom to the Workplace: Motivating Students to Learn in
Industry, A. Christian Fricke

> 11 Positions Available
> 11,65 Book Reviews



CHEMICAL ENGINEERING EDUCATION (ISSN 0009-2479) is published quarterly by the Chemical Engineering
Division, American Society for Engineering Education, and is edited at the University of Florida. Correspondence
regarding editorial matter, circulation, and changes of address should be sent to CEE, Chemical Engineering Department,
University of Florida, Gainesville, FL 32611-6005. Copyright 1999 by the Chemical Engineering Division, American
Society for Engineering Education. The statements and opinions expressed in this periodical are those of the writers and
not necessarily those of the ChE Division, ASEE, which body assumes no responsibility for them. Defective copies replaced
if notified within 120 days of publication. Write for information on subscription costs and for back copy costs and
availability. POSTMASTER: Send address changes to CEE, Chemical Engineering Department., University of Florida,
Gainesville, FL 32611-6005. Periodicals Postage Paid at Gainesville, Florida.


Winter 1999









Educator


MA URICE A.


LARSON

of
Iowa State University



"Mr. Crystallization"


Widely respected and recognized the world over
for his research contributions, Maurice Larson is
a distinguished and accomplished chemical engi-
neer who has been a influential figure in chemical engineer-
ing education and practice for nearly forty years. In the
area of crystallization, which he was instrumental in
changing from an art to a science and into a thriving field
of research, he is looked upon as a father figure. His
contributions have been both pioneering and original in
the context of their times.
Mauri was born on a farm in the Loess Hills near Missouri
Valley, Iowa, the second child in a family of two girls and
one boy. He attended a one-room country school that had
twenty students in eight grades. (Interestingly, of the six
students in his class, three ultimately received PhDs in engi-
neering!) When he was nine, his family moved to another
farm, this one near Ayrshire, Iowa, 150 miles away. The
school there consisted of all twelve grades and had about
220 students.
Mauri graduated in a class of twenty-one students in 1944.
In 1946, he enlisted in the Army and was later discharged
just in time to begin study at Iowa State College in the fall of
1947. He initially planned to study chemistry but, influenced
by a newspaper article he happened to read, he enrolled in
chemical engineering instead. Subsequent to his graduation
in 1951, he accepted employment with the Dow Coming
Corporation in Midland, Michigan.
Mauri had always had in mind going on to graduate school


Mauri as a farm boy in 1936,
as an enlistee in the Armed Forces in 1946,
and as a distinguished professor in 1998.


and, after marrying Ruth Gugeler, an ISU chemistry gradu-
ate and a co-employee at Dow Coming, he decided the time
had come to do so. He enrolled in graduate school at Iowa
State and secured employment as a teaching assistant. He
found he greatly enjoyed working with students and was
soon was promoted to instructor. His thesis research was in
the area of phosphate fertilizer chemistry and production.
After receiving his PhD in 1958, Mauri was appointed
Assistant Professor in the department, and his research broad-
ened from fertilizer technology to process dynamics and
control. He organized and taught one of the first undergradu-
ate process control courses in the U.S. and was also one of
the first to organize a course around the seminal text Trans-
port Phenomena by R. B. Bird, W. E. Stewart, and E. N.
Lightfoot. He and his graduate students published in the
areas of fertilizer technology and process dynamics, and
nationally he was elected Chair of the Division of Fertilizer
and Soil Chemistry of the American Chemical Society.
In 1959 a new graduate student, Alan Randolph, arrived at
Iowa State with industrial experience in the crystallization of
ammonium sulfate. Alan was concerned about the limit cycle
in crystal size distribution that occurred in large continuous


Copyright ChE Division of ASEE 1999


Chemical Engineering Education








crystallizers. Mauri's interest in both process dynamics and fertilizer technol-
ogy resulted in a perfect interest match with Alan, and a result of their collabo-
ration was the landmark paper "Transient and Steady State Size Distributions in
Continuous Mixed Suspension Crystallizers"-an important publication that
set the agenda in industrial crystallization research for the next thirty-five years.
"Mauri and I met in 1959 and we immediately realized we had similar
research interests," says Alan Randolph, now an Emeritus Professor of Chemi-
cal Engineering at the University of Arizona. "I was fortunate to be one of his
first PhD students, and he and I have been lifelong colleagues and friends ever
since that time. Mauri is clearly responsible for much of the early fundamental
work in crystallization. He stimulated the work of many others and the field has
benefited greatly from his work," Randolph continues. "A good example of his
supportive approach is an occasion when he was visiting our family in Trona,
California. Mauri and I went on a trip to see the Panamint Mountains and our
old blue Rambler was not able to make it up one of the steep grades. I shouted to
Mauri, "Get out and push!" Amazing as it may seem, he actually pushed the car
uphill-illustrating that Mauri is not only imperturbable, but he is also a true
friend who knows the meaning of support!"
Mauri's subsequent research at Iowa State was concerned with many aspects
of solution crystallization. In addition to analyzing continuous crystallizer
stability, he developed methods to collect data on and realistically describe
nucleation and growth, secondary nucleation effects of additives, and growth
dispersion. Alan and Mauri were the first to provide a firm mathematical
foundation to the analysis of crystallization by introducing and expanding the
use of Alan's population balance models. Their early studies on population
balances formed the basis for several entirely new areas of crystallization
research and theory. Several of Mauri's twenty-six PhD and twenty-seven
MS students have been continuing contributors to crystallization funda-
mentals and practice.
Many if not all fields of chemical engineering have pioneering names indel-
ibly associated with their growth. By any account, Maurice A. Larson is one
such name. No book on crystallization, no review on the subject, indeed no
really worthwhile paper in the specific areas of his research, would be complete
without reference to his contributions, in particular to the book he coau-
thored with Alan, The Theory of Particulate Processes, Analysis and
Techniques in Continuous Crystallization (A. D. Randolph and M. A.
Larson, 1st ed. 1972; 2nd ed. 1988; Academic Press). "It has become one
of the standard references for the field," says Dr. L. K. Doraiswamy,
Anson Marston Distinguished Professor and Herbert L. Stiles Professor of
Chemical Engineering.
Mauri's international standing in the area of crystallization is firmly estab-
lished. His activities have included collaboration with experts from the United
Kingdom, Australia, India, China, The Netherlands, Poland, Czechoslovakia,
and many other countries. For twenty-five years he has also been active with
the European Federation of Chemical Engineering Working Group on Crystal-
lization. To strengthen his research, he spent a year at Stanford on a National
Science Foundation Faculty Fellowship, a year at University College London
as Shell Visiting Professor, and a year at the University of Manchester Institute
of Science and Technology (UMIST).
His interest in undergraduate education and opportunities for undergraduate
study led him to initiate several chemical engineering student international


"His vision and

example have

shaped the

culture of the

department-

a balance of

teaching,

research, and

service ....

ISU

undergraduates,

alumni, faculty

members

across campus,

and

crystallization

researchers

around the

world hold

Mauri in the

highest regard."


Winter 1999












4 Gordon
Youngquist pre-
sents Mauri with a
framed copy of the
program prepared
for a special AIChE
session honoring
Mauri's 70th
birthday and his
many contributions
to the field of
chemical
engineering.


exchange programs, notably with Bradford University, Ham-
burg University, and University College London. The most
successful has been the summer-study program at Univer-
sity College London, which he organized in 1972 during his
sabbatical there. The program is still in full force and is now
shared by several additional American universities including
the University of Wisconsin, Georgia Institute of Technol-
ogy, Case Western Reserve University, and Virginia Poly-
technic Institute and State University.
John Garside, Pro-Vice-Chancellor and Professor of Chemi-
cal Engineering UMIST remarks, "I have known Mauri now
for well over twenty years. He took a faculty-improvement
leave at University College London in 1972 when I was on
the faculty there, and I subsequently spent a year working
with him at ISU in 1976-77. At UMIST, we use Mauri and
Alan's book, The Theory of Particulate Processes, for both
undergraduate and graduate courses. It is an inspiring book
and has had an important impact on teaching in this field
throughout the world. Their work has been so influential that
use of population-balance modeling is now accepted as a
matter of course when dealing with crystallization pro-
cesses. Before their work, the population balance was
almost unheard of."
Crystallization has been a traditional area of chemical
engineering education and research, having evolved from its
empirical beginnings to the highly sophisticated approaches


Mauri's office door has always been open to
students and colleagues alike. (1997) V


S4 Mauri is surrounded by well-wishers on the
occasion of his retirement. Each of them has at
one time or another served as Chair of the ISU
ChE Department: (left to right) Terry S. King,
Charles E. Glatz, Maurice A. Larson, R. C.
Seagrave, and George Burnet.


practiced today. Maurice Larson has been associated with all
stages of recent development of this subject and has been
a trendsetter and a pioneer in nurturing and shaping it.
Indeed, it is inconceivable that any account of crystalli-
zation today would be complete without a reference to
Mauri's dominating role.
In 1991, Mauri organized an informal group known as the
Association for Crystallization Technology. This organiza-
tion brings together 70-100 researchers and technologists
from industry and academe for three days every year for the
sole purpose of discussing the science and technology of
crystallization. Mauri's extension programs in the form of
workshops in crystallization have always succeeded in bring-
ing together chemical engineers from many parts of the
country, indeed the world, and have resulted in reports and
valued suggestions for future research.
Mauri has been vigorously involved in many national
programs and activities of the American Institute of Chemi-
cal Engineers. He was recognized as an AIChE Fellow in
1982 and served as Vice Program Chair for the
organization's 1992 Miami Beach meeting. He has also
been selected to receive the 1998 AIChE Separation
Division's Clarence G. Gerhold Award.
In the American Society for Engineering Education, he


Chemical Engineering Education










received the Western Electric Fund Award in 1970, served
as an officer of the Chemical Engineering Department Heads
Group in 1973-74, and was Co-Chair of the 1987 Chemical
Engineering Division Summer School. He has served on the
editorial boards of the AIChE Journal and the Journal of
Separation Process Technology, and received the U. K.
Wilton Park Award in 1978 and the Iowa Governor's Sci-
ence Medal in 1989.
In addition to his research work, Mauri has always been
concerned with the teaching of both graduate and under-
graduate courses. Over the years he has taken an active
interest in practically all matters pertaining to education and
student affairs at ISU. The high national reputation that
undergraduate education in chemical engineering at Iowa
State enjoys today may be largely credited to his efforts.
Evaluated by his students as one of the best teachers they
have encountered, they cite his ability to make his classes
rewarding and enjoyable. The are challenged by his dedica-
tion and thoroughness. E. L. Cussler, ChE Professor at the
University of Minnesota, remembers, "Mauri was a marvel-
ous mentor for young faculty. When I was starting, he found
me and encouraged me when I was still surprised to be
treated as an adult. I and many like me owe an enormous
debt to Mauri for his generosity and time in encouraging
and shaping our own careers." Mauri's combined interests
of teaching and research resulted in his promotion to Asso-
ciate Professor in 1961 and to Professor in 1964. In 1977 he
was awarded the highest rank in engineering at ISU-that
of Anson Marston Distinguished Professor.
Throughout his career, Mauri has also been active in de-
partmental organization and administrative services. He has
served on many university, college, and departmental com-
mittees and has always contributed to curriculum develop-
ment, both at the college and the department levels. He
served as Department Chair from 1977 to 1983, and many
who have visited the department testify to its increased stat-
ure as a result of his leadership. As an educator and adminis-
trator, Mauri influenced the growth of engineering at ISU
through the breadth of his interests and his ability to fashion
a common theme from diverse views.
Near the end of his term as Chair, Mauri initiated a cam-
paign to raise funds for a new addition to Sweeney Hall, the
chemical engineering department building, to provide ex-
panded research space and an updated teaching laboratory.
More than $1.5 million was raised from friends and alumni
and this, with an appropriation of over $5 million from the
State Legislature, resulted in a new wing in 1994. As a
result of Mauri's campaign, one of the donors also gave a
million dollars to endow the Herbert A. Stiles Professor-
ship in Chemical Engineering.
A special session was programmed at the 1997 AIChE


Annual Meeting in Los Angeles to recognize Mauri's distin-
guished career and to honor him on the occasion of his 70th
birthday. Dr. Terry King, former head of the ISU Chemical
Engineering Department and currently Dean of the College
of Engineering at Kansas State University, declared

"Mauri's determination to provide for the building
needs of the future for chemical engineering, even
when support for such a request was not initially
evident, set a pattern for what is now being
achieved by the ISU College of Engineering as a
whole. He melded a partnership of alumni,
industry, and the state to create enthusiasm and
funding for a building project from whose labora-
tories have come many important results being
presented today in technical sessions worldwide.
That partnership was novel for ISU at the time,
and it has been enormously successful, resulting in
our new construction. His vision and example have
shaped the culture of the department-a balance
of teaching, research, and service. As a result,
ISU undergraduates, alumni, faculty members
across campus, and crystallization researchers
around the world hold Mauri in the highest
regard."

Larson has served as a teacher, researcher, ISU supporter,
colleague, and an inspiration to his friends for over four
decades. He and his wife Ruth are the parents of three
children-Richard, who was lost as a teenager in an acci-
dent, Janet, who has a horse ranch in New Mexico, and John,
who is employed in San Francisco. Mauri and Ruth could
not get farming out of their systems and thus have a
substantial number of acres of farmland near Ames. Ruth
retired in 1996 after working in the ISU Department of
Animal Science and the couple has used their free time to
travel extensively.

"Mauri Larson is the quintessential scholar, an
outstanding professional, and one of the nicest
people I have ever known,"

states friend and collaborator Ronald W. Rousseau, cur-
rently Professor and Chair of the School of Chemical Engi-
neering, Georgia Institute of Technology. Dr. James R. Fair,
McKetta Chair Emeritus Professor from the University of
Texas-Austin makes things crystal clear when he states,

"The achievements of Mauri Larson have been
notable. One of the foremost has been his rise to a
pre-eminent stature as an expert in crystallization.
He is considered to be a foremost authority in this
field and his many publications bear this out. The
record is clear-and well documented-on this
issue. He really is Mr. Crystallization. 0


Winter 1999









o department




Washington



State



University





So, what can 50 years of chemical engineering education mean
in the context of 2,000 years of calendar time, especially when
generated from a relatively small program in the rural town of
Pullman, Washington, located in the middle of wheat fields?
Surprisingly, Washington State University's chemical engineering
department is a unique example of one of the oldest chemical engi-
neering programs in the country that is still a thriving concern for the
university and related industry in today's world. Having evolved
from being part of the chemistry department in 1917 into a sepa-
rate department in 1950, ChemE now focuses on education and
research in advanced gas and chemical processing, hazardous ma-
terial cleanup, and bioengineering.
The department has graduated more than 1,100 students, the major-
ity of whom have taken significant roles in industries such as oil and
chemicals, pulp and paper, pharmaceuticals, food, microchip manu-
facturing, environmental protection, and bioengineering. Alumni can
be found in such diverse industries as petroleum refineries, pulp and
paper mills, nuclear and synthetic fuel processing facilities, and food
processing plants. They are hired by such companies as Dow Chemi-
cal, Westinghouse, ARCO, Boeing, Weyerhaeuser, Kaiser Aluminum,
Intel, Battelle Pacific Northwest National Labs, as well as other Hanford
contractors and smaller companies in the region.
"Graduates from WSU's chemical engineering program are highly
marketable," reports Richard Zollars, department chair. "Starting sala-
ries in the mid-$40,000s are common, as qualified candidates in these
fields are highly sought after by industry and agencies."
Marc VanderSchale and Brian Erickson The small size of the program and connections with industry work in
remove CO2 from air in its favor. With about 100 undergraduate, 30 graduate students, and 10
packed-column gas absorber. permanent faculty, class sizes of 25-35 afford seminar possibilities
Copyright ChE Division of ASEE 1999
6 Chemical Engineering Education











Washington State University's

chemical engineering department is a

unique example of one of the oldest chemical

engineering programs in the country that is

still a thriving concern for the university

and related industry in today's world.



and close interaction between faculty and students. Because of its select nature,
the program attracts students with above-average scholastic standing, and as
many as one-third of all those certified in the chemical engineering program are
in WSU Honors Program.
Given the above collegiate atmosphere, the faculty, research, industry partnerships,
and educational activities take on a distinctive quality and focus.

THE UNDERGRADUATE PROGRAM
The undergraduate curriculum allows for broad education in the sciences and sound
basics in the discipline, but with a flexibility that allows students to individualize
studies. In the upper-division, after students certify, the following courses provide the
basics in chemical engineering:
Sophomore Year
ChemE Process Principles; Process Simulation
Junior Year
Transport Phenomena; Fluid Mechanics/Heat Transfer; Thermodynamics; Sepa-
rations; Kinetics
Senior Year
Chemical Engineering Lab I and II; Control; Design I and II; Seminar
Another seven or eight electives allow juniors and seniors to customize their focus
in bioengineering, environmental, or other allied fields. They also have opportunities
for multidisciplinary study and research.
For example, Becky Russell, a recent graduate who knew she wanted a career that
impacted people, explains "I decided to make chemical engineering my preparation
for medical school. The skills I learned in technical problem solving and biomedical
applications have served me well."
Not all is serious study, however. Students are encouraged to join the College
Ambassadors group and to participate in the AIChE student chapter or other engineer-
ing societies for networking and leadership growth. WSU's student AIChE chapter
recently won first prize at its regional conference for proposing a national competition
around student-built chemically controlled cars. The group will put this to the test at
the national conference in Miami in mid-November (after the deadline for this
article). President Wendy Anna and other officers were able to raise funds from
industry to defray the cost of several students' travel to Florida for this purpose.
As the nationwide need to recruit and retain engineering students escalates, WSU's
chemical engineering department continues to step up its efforts for scholarships and
other incentives. "There's a good chance that students will qualify for a scholarship if
they achieve a 3.4 GPA or better and show promise while enrolled," reports Zollars.
Because of the generous support of alumni and corporations, about one-third of the
Winter 1999


ABET Accredited


E-mail student services
thorntd @che.wsu.edu

Website at
http://www .che.w-su.edu


FACULTY

Ivory. Cornelius
PhD, 1980: Princeton
Bioprocessing,. separatiors,
modeling

Lee, James
PhD. 1978: Kentucky
Bioprocessig, miring

Liddell, KNona
PhD, 1979: Iowa State
Hazardous wasres.
electrodepositron

NMahalingham, R.
PhD. 1968: New\castle-Upon-
Tyne. England
Hazardous wastes, shock
reactions

Miller, Reid
PhD. 1968, Califora. Berkeley
Thermodynamics

Petersen, James
PhD, 1979: lowa State
Bioremediation

Peyton, Brent
PhD. 1992: Montana State
Bioremediation. etiremophiiic
bioproce 4ing

Thomson. William
PhD. 1969. Idaho
Materials, kinetics, catalysis

Van Wie. Bernie
PhD. 1982. Oklahoma
Bioprocessing. biomedical
engineering

Zollars, Richard
PhD. 1974. Colorado
Collordal/interfacial phenom-
ena, reactor engineering










Ethics
are taught in
one senior class
by using the
Dilbert
Ethics Challenge
game.
Teams ponder
how to handle --
case studies. .-


undergraduate students receive some form of financial aid.
With the exception of athletics, chemical engineering's alumni tradition-
ally give the highest per-annum donations of any other unit at WSU.

GRADUATE/RESEARCH PROGRAM
While virtually every undergraduate student who wishes to may get
involved in faculty research, the research program is the heart of the
graduate program. Currently, more than $2 million in grant activity is in
progress, with healthy prospects for more.
Emphases in bioengineering, environmental restoration, and hydrocar-
bon processing result in such projects as biotreatment of hazardous con-
tamination, diagnostic medical devices, and converting natural gas to useful
products. The current roster of graduate-student research covers such topics
as electrodeposition, isoelectric focusing, development of biosensors, pro-
tein production in plant cell systems, bioremediation of chlorinated sol-
vents and heavy metals, oxidative coupling of methane, etc.
The Center for Multiphase Environmental Research (CMER), under the
direction of Professor James Petersen, conducts interdisciplinary research
addressing important environmental problems for industries and govern-
ment agencies. It seeks opportunities to transfer this technology to
industry and, in the process, to educate the next generation of environ-
mental professionals.
CMER faculty are drawn from the civil, environmental, chemical, me-
chanical, and biological systems engineering departments. The Center and
chemical engineering faculty currently play a part in three such projects
with large government grants: remediation and recycling of creosote-
treated piers at Navy ports and two D.O.E. projects to clean up toxic
metals in soils and aquifers.
As a result of her quality work on hazardous waste treatment through
CMER, doctoral student Juli Sherwood last year won the university's
Harriet Rigas Award from the Association of Faculty Women as WSU's
outstanding doctoral woman student, citing her for her research, teach-
ing, and leadership.
Another recent graduate student success story is illustrated by Sherman
Xu, who within five years gained both his master's and doctoral degree,
valuable research expertise, and a job with the Amoco Research Center in
Illinois. His research explored a more productive procedure to convert
natural gas to ethylene-the feed stock for many of today's plastics.


ABOUT-THE UNIVERSITY
;: -WSJT is ala-d-grant research university
dedite t.o :excellence in undergraduate
educialif. F-ounded in Pullman in 1890. it
h-as 2'_- f students at four campuses, sev-
r-;Wairl e ing Centers- and other sites
t- .ut ,e state. WSt'Ps -ine colleges
4i10 4t5.0a t i.fg graduate ma-
t_-idistafite Extended De-
A tm T ouAny-
s resi-

A s ---fl4i n-ps- _an residece
t-^^%I:ts rue ot 1* muost "wired00
-, ti nPd iyhsisgod access to

AUg 4 s tu d -fitn e August
S through
iS--i4 -""lu---i m"ri i sessions.




SChemical Eng ingeerion
U 1 4& so mmitted to
A l -learn-

td prepare
profewsioas and life-

cu0i0iWblrht network
A r th~ aff-stdies -cholar-
rtbjinternshipsand project
ut~rttle Approximately, 110 perma-
e An avenra& 02000 un-
L;-"to jyiobtaio tt an -20



iiLtipdox ther Nw
squae
AnMknb Teaehins esearcb
tawortoy recently open iand holds state-
oftheat toting and alysis facilities.
Tkt~1ap stst W rncid-a skills
wi ;I pieatin onri-
sentat-i itlotlwr spieciaatdent needs.
-Ti 4etC p eru~ik and raain izdviduals
rgreadgr-n itnnthugh the
Mdntyiand Women's Engieering Pro-
gransm

Chemical Engineering Education











< O.H. Reaugh
Lab dedication:
Student Carrie O'Rourke,
Chair Dick Zollars,
lab director Bill Thomson,
WSU President Sam
Smith, and alumnus
O.H. Reaugh.

Reaugh, Thomson,
and student
Matt Fountain
tweak the new
equipment. V


"With my WSU credentials, I
found myself very competitive in the
job market," says Xu. "I'm grateful
to my superior graduate professor
who gave me the 'big picture,' and
allowed me to find my own answers.
His close rapport with industry pro-
vided me opportunities to see practi-
cal uses of our research."


FACILITIES
The department's premier research lab-the O.H. Reaugh
Oil and Gas Processing Laboratory-was dedicated by its
namesake this fall in a ceremony (attended by his family,
university leaders, students, industry partners, and friends)
that epitomized the momentum that the department's history
brings to its future.
Orland Harry Reaugh, a 1933 graduate of the program,
became a petroleum engineer and a leader in independent oil
production in Oklahoma, Kansas, New Mexico, and Illinois.
He has since become a major benefactor of WSU's chemical
engineering department, which now has the only such alumni-
endowed and -named laboratory within the university. Reaugh
provided $250,000, ARCO another $50,000, and matching
donations are being sought to complete the lab's $500,000
endowment. The income will ensure a lab that is equipped
with state-of-the-art instrumentation.
Lab director William Thomson says the facility "allows us
to conduct research that is fast disappearing from the univer-
sity scene-innovative studies to find less expensive and
ecologically friendly ways to add octane to gas, bring natural
gas to remote locations, and to create compact, economical
hydrogen fuel cells.
More than 9,500 square feet of lab space are located in the
Winter 1999


A Jim Petersen and graduate
student Lin Sha look at
grimy creosote being eaten
by bacteria and turned into
CO, in the lab.


Snew Engineering Teaching and Re-
search facility containing analytical
equipment such as spectrometers,
chromatographs fitted with auto-
samplers and various detectors,
anaerobic incubators, dynamic x-ray
diffractometer, laser light scattering, and electrophoresis de-
vices. The George T. Austin endowment for undergraduate
lab equipment, augmented by funds from alumnus Gene
Voiland, help keep learning technologies current.
Computing equipment and workstations with parallel pro-
cessing features are provided for all graduate students and
researchers. Keeping such quality computing current is al-
ways a concern of technical educators, however. Typically,
alumni donations help replace one-third of the undergradu-
ate machines every year in the 15-station computing lab.
Full-scale commercial versions of Pro Vision, MatLab,
Mathematica, MathCAD, spreadsheets, word processors, and
other programs are used on these machines.

INDUSTRY CONNECTIONS
A hallmark of WSU's engineering college is its continuing
strong connection to industry. Companies and agencies pro-
vide internships, scholarships, collaborative research, and
even job exchanges. WSU in turn provides on-site or webbed
distance learning, qualified potential hires, and R&D.
One collaborative project is Prof. Bernie Van Wie's work
with the Spokane Interdisciplinary Research & Technology
Institute and DevTec, an independent industrial partner, to
develop and commercialize an automated blood chemistry
analyzer. Another collaborative project is Prof. Bill
Thomson's work with the Washington Technology Center










and Washington Water Power to refine an efficient fuel cell.
West Coast companies such as ARCO, Boise Cascade,
Hanford contractors, SEH America, Kimberly Clarke, and
Reynolds Aluminum offer summer jobs and internships (sum-
mer plus one semester) to students-and often end up offer-
ing them jobs at graduation. They say they are impressed
with WSU's students' practical knowledge and leadership
experiences outside the classroom.
"Chemical engineers need
foundations in science and
math, but also in other disci- -.
lines that will prepare them
to adapt new products into the
culture," says Glenn Butler,
CEO of the ARCO Refinery at
Cherry Point. He and other de-
partment advisers urge devel-
opment of students' skills in
communication, presentation,
and human relations, in addi-
tion to developing a code of
ethics and a knowledge of
overall business and eco-
nomic concepts. "Team think" in action.
sors, researchers, mi
John Wolfe, a 1997 chemi- commercializatio
cal engineering graduate now automated bloc
at ARCO in Anaheim, recently
returned to the College's Ca-
reer Fair-this time as a recruiter! "Basically, my ChemE
degree helped to more than double my salary as a science lab
technician," says Wolfe. "My internship with ARCO during
the summers gave me a foot in the door. And now I have a
good job with them solving technically complex problems in
oil refining." He's now on the lookout for other potential
hires who work well in teams of engineers, scientists, cus-
tomers, plant operators, managers, lawyers, government regu-
lators, and construction workers.

OUTREACH
WSU Pullman's program remains fairly stable in size.
Many families continue sending new generations as lega-
cies. The department also reaches out to the adult, some-
times mid-career, learner through a companion program at
the WSU branch campus in Tri-Cities, with access to Hanford
and the Environmental Molecular Science Lab. It generally
serves full-time employed engineers, offering after-hour
courses so that engineers can upgrade their education or
work toward the MSChE. It is one of few evening-based
MSChE programs in the country. Many of the courses are
taught by PhD chemical engineers who are also employed at
the Pacific Northwest National Laboratory.
Tri-Cities student projects or theses usually are done with
a committee composed of both WSU regular faculty and


SGr<
edici
n ex
od at


"adjunct" faculty from local industry. Three to five students
usually earn the MSChE degree through this program each
year. This year's graduates will be Penny Colton (Dissolu-
tion Kinetics at the Calcite-Water Interface), Dan Schmitt
(10 W Proton Exchange Membrane Fuel Cell Design), Chris
Johnson (Microbial Growth Kinetics Using Colloidal Poly-
mer as Substrate), Scott Estey (Thin Film Poly-
dimethylsiloxane Oil Evaporation), and Brad Knutson (Evalu-
ation of Ion Exchange Perfor-
mance Predictive Tools).

STo develop interest in engi-
S -- neering at earlier ages, for the
past five years chemical engi-
neering faculty has conducted
a NSF-funded summer pro-
gram for secondary school
teachers. The goal is to famil-
iarize them with engineering
and help them develop mod-
ules to teach when they return
to their classrooms. All to-
gether, more than 65 teachers
attended-half from North-
aduate students, profes- west schools, and others from
al professionals, and as far away as Korea, Florida,
perts work on the and Connecticut. Feedback
ialyzer project. from the participants noted
that the experience suc-
ceeded in bringing engineering into their science curricu-
lum. One science teacher developed an engineering-re-
lated module that contributed to a portfolio that won her
state and national teaching awards.

FUTURE TRENDS
Academic destinies will depend a great deal on the eco-
nomic and technological trends transforming higher educa-
tion today. For instance, the Tri-Cities program will be influ-
enced by down-sizing at Hanford. And, as noted earlier,
keeping computer equipment and software current is a major
and constant challenge. By the same token, developing dis-
tance-learning technologies may allow webbing courses
beyond campus sites, which will again revolutionize what
we do at universities.
While state appropriations can no longer meet the
entire departmental needs, outside grants and private fund-
ing is on the rise. The department plans to pursue more
endowments similar to the O.H. Reaugh Lab, to stabilize
funding bases.
Faculty and student numbers are not expected to change
dramatically over the next five years. Expansion of multi-
disciplinary experiences is anticipated, however, particu-
larly in the wake of the Center for Multiphase Environmen-
tal Research's path-finding activities. 0
Chemical Engineering Education











M" book review


Numerical Computation
in Science and Engineering
by C. Pozrikidis
Published by Oxford University Press, Inc., 198 Madison Avenue,
New York, NY 10016; 627 pages including index, $75.00 (1998)

Reviewed by
James N. Petersen
Washington State University

The necessity of obtaining numerical solutions to physical
problems crosses virtually all the discipline boundaries in
engineering and science. Only a limited number of books are
available, however, that provide the fundamentals of scien-
tific numerical computational techniques, together with ap-
plication of those techniques. In general, Pozrikidis has been
able to achieve this result in his book, which is intended to
be used in upper-level undergraduate and beginning gradu-
ate courses and may be suitable for individual study. In so
doing, he has produced a book that strikes a balance between
rigor and practicality.
While not providing lengthy computer codes in the text, he
has communicated the essential aspects of various numerical
methods by relying on code fragments and pseudoCode that
can be translated into any suitable computer language. In so
doing, he has covered most of the topics required in an
introductory course within a manageable number of pages.
To compliment the text, he has also provided access, via the
Internet, to a public domain software library of Fortran 77
programs, organized by book chapter.
The book is organized into eleven chapters. Each of them
is further divided into sections and subsections. Because the
subsections are the intended functional learning entities, the
author provides both theoretical and computational prob-
lems at the end of each. These problems are designed to
complement the theory presented and to provide the student
with an immediate opportunity to practice its implementation.
The chapters are organized in a fashion similar to many
other numerical methods books. In Chapter 1, the author
provides a background of computer hardware, computer arith-
metic including both integer and floating-point representa-
tions, and errors. In the next three chapters, he first lays a
foundation on which to build the solution of linear and
nonlinear sets of simultaneous equations. Thus, in Chapter 2
he covers matrix algebra and matrix calculus, and Chapter 3
is devoted to the solution of sets of linear algebraic equa-
tions. The solutions of sets of simultaneous nonlinear alge-
braic equations is discussed in Chapter 4, and he goes on to
discuss eigenvalues of matrices in Chapter 5.
_Continued on page 65.
Winter 1999


FEATURES
* Graphical Interface
* Integrated Plotting
* Batch Calculations
* On-line Help


gasdynamics chemical equilibrium chemical kinetics


APPLICATIONS
* Education
* Combustion
* Pollution
* Explosions



BI Ad


info@combdyn.com
www.combdyn.com
fax: 403-529-2516


CDLd
Combustion Dynamics Ltd


POSITIONS AVAILABLE
Use CEE's reasonable rates to advertise.
Minimum rate, 1/8 page, $100;
Each additional column inch or portion thereof, $40.

Faculty Position in Chemical Engineering
THE UNIVERSITY OF TEXAS AT AUSTIN

The Department of Chemical Engineering invites applications for a
tenure track faculty position at the Assistant Professor level. A PhD is
required, and applicants must have at least one degree in chemical
engineering, an outstanding record of research accomplishments, and
a strong interest in undergraduate and graduate teaching. Preference
will be given to applicants with skills that will add to the Department's
strengths in bioengineering and applied and computational mathemat-
ics, including fluid dynamics, systems engineering, and statistical
mechanics. The successful candidates are expected to teach under-
graduate and graduate courses, develop a research program, collabo-
rate with other faculty, and be involved in service to the university and
the profession. Applications from women and minorities are encour-
aged. Interested persons should submit a detailed curriculum vitae,
including academic and professional experience, a list of peer-re-
viewed publications and other technical papers, and the names, ad-
dresses, and telephone numbers of three or more references to: Chair-
man, Department of Chemical Engineering, The University of Texas
at Austin, Austin, Texas 78712-1062. The University of Texas is an
Equal Opportunity/Affirmative Action Employer.










Award Lecture ...





DO CHANGES IN

THE CHEMICAL INDUSTRY

IMPLY

CHANGES IN CURRICULUM?



E.L. CUSSLER
Edward L. Cussler, Institute University of Minnesota Minneapolis, MN 55455
Professor at the University of
Minnesota, received his BE de- is p r i i i ri r-
gree from Yale University ins paper is a synopsis of my Union Carbide Lecture-
1961, his MS from the Univer- ship, an award given at the 1998 meeting of the
sity of Wisconsin in 1963, and American Society of Engineering Education. I am flattered
his PhD from the University of to have my research and teaching on diffusion acknowledged. I
Wisconsin in 1965. He rose from
Assistant Professor to Professor know that this lecture can often be a review of the past research,
of Chemical Engineering at centering on a scattering of old slides, like a photograph album of
Carnegie-Mellon University dur- half-remembered vacations.
ing the years from 1967 to 1980, at which time he joined the
faculty at the University of Minnesota as Professor of Chemi- But the lecture and this paper are too good a forum to waste on
cal Engineering. In 1996 he became Institute of Technology my past. Instead of the past, I want to consider the future. In doing
Professor at the University of Minnesota, and is currently at so, I remember a conversation I had thirty years ago with the
Cambridge University in the United Kingdom as Professor historian, L. Pearce Williams. I was visiting him to gush about his
of Chemical Engineering.
Ed has wn nmers biography of Michael Faraday,1' which I had enjoyed enormously.
Ed has won numerous awards during his professional
career, some of which are the AIChE Alan P. Colbum Award I suspect that he found my naive enthusiasm both flattering and
in 1975, seven Minnesota Institute of Technology Teaching embarrassing. To make conversation, Williams asked if I knew the
Awards through the years, the George Taylor Distinguished real difference between science and the arts. I did not. He re-
Teaching Award from the University of Minnesota in 1987, sponded that in the sciences, we wrote papers and books when we
the Donald Katz Lecture Award from the University of
Michigan in 1996, and the Danckwerts Lecture from the felt we knew everything about our topic. In the arts, he asserted,
Institution of Chemical Engineers in London in 1997. authors wrote when they knew little initially and used the writing
Ed serves as Associate Editor of the AIChE Journal and is as a way to focus new questions and to explore possible answers.
on the Editorial Board of the Journal of Membrane Science. Whether this arts-science contrast is true or not, I want to use
He also has served as a Director, Vice President, and Presi-
dent of AIChE, and was Chair of the American Association this paper as a way to learn about possible changes in chemical
of Engineering Societies. engineering curricula. I am not yet sure if these ideas are correct,
He has also been author or co-author of over 160 publica- but I want to see if they make sense. In the next few years, I'll try
tions. He is co-author with Belter and Hu of Bioseparations them out. For now, though, they're best described under three
(John Wiley and Sons, New York, 1988) and with Baker, headings: the changes in the chemical industry, the status in
Eykamp, Koros, Riley, and Strathmann, of Membrane Sepa- academia, and possible curricular changes.
ration Systems, (Noyes Data Corporation, New Jersey 1991).
He is author of the books Diffusion (Cambridge University
Press, London, 1984; second edition, 1997) and Multicom- CHANGES IN THE CHEMICAL INDUSTRY
ponent Diffusion (Elsevier Publishing Company, Amsterdam, Last spring, I taught our introductory chemical engineering
1976). course-the one that covers stoichiometry. Early in the course, I

Copyright ChE Division of ASEE 1999
12 Chemical Engineering Education









showed pictures of chemical plants to the students. I told them that the tall towers
were for distillation and the short, fat ones were often for gas absorption. I pointed
out the reactors, with their preheaters and recycles. I spoke of the excitement of
running a chemical plant and the satisfaction of using chemical technology to
improve our well being.
I did so with hidden reservations that I did not have ten years ago. I know that the
chemical industry has changed and that many of the students will not work in the
commodity chemical plants I was describing. To see why, we need to review the history
of our industry, using as an example the development of synthetic textile fibers.
From 1950 to 1970, the chemical industry produced ever-increasing amounts of syn-
thetic textile fibers, as shown in Table 1. Over the decades, while the production of
natural fibers was about constant, the production of synthetics grew 20% per year. This
growth was comparable to that of the software industry today; indeed, Du Pont in the
1950s was like Microsoft in the 1990s. It was a golden age for chemicals.
But from 1970 to 1990, synthetic textile fibers grew only four percent a year-at about
the same rate as the growth of world population. That's not surprising; after all, any
logarithmic growth can't continue indefinitely. From 1970 to 1990 the industry stayed
profitable by using larger and larger facilities. Bigger profits came from consolidating
production into bigger plants, designed for
greater efficiency in making one particular
product. The interest in computer-optimized TABLE 1
design is a vestige of this consolidation. Such Growth of Synthetic Fibers From
optimization meant small producers were 1950 to 1970
forced out. For example, the number of com- (Source: Spitz, U.S. Department of Commerce)
panies making vinyl chloride shrank from
twelve in 1964 to only six in 1972.[2] 1948 1969 1989
In the last ten years, the industry has used Cotton, Wool 4353 4285 4794
other strategies to stay profitable. These strat- Synthetics 92 3480 8612
egies often centered on restructuring, which
was three times more likely to affect engi-
neers than the general population. Whether called "restructuring," "downsizing," "right-
sizing," or "rationalization," the strategy meant many mid-career engineers were sud-
denly looking for a job. The Engineering Workforce Commission now feels that engi-
neers will average seven different jobs per career, a dramatic change from two per career
when I graduated in 1961.[3] Middle management, that traditional goal of our B-students,
is no longer a safe haven. Starting salaries remain high, the envy of other technical
professions, but they have not increased faster than inflation in thirty years. In this
environment, I applaud the decision of the American Institute of Chemical Engineers
(AIChE) to be a "lifetime home" for members of our profession, providing more help in
job transitions and financial planning. The AIChE can no longer be only a nineteenth
century-style learned society.
Most recently, the chemical industry has become enchanted with the life sciences, often
called "biotechnology." Biotechnology in the industrial sense is most successfully repre-
sented by applied agronomy, i.e., by genetically modified seeds. It is usually different
from the biotechnology represented in academic chemical engineering that often centers
on separations and reactions involving specialty pharmaceuticals. The model for corpo-
rate enchantment is Monsanto, which has spun off its commodity chemical operations to
compete largely in this new biotechnology. Other companies are imitators. In recent five-
year projections, Du Pont has relabeled its chemicals as "materials," is spinning off
Conoco, and plans to double its life-science efforts to one-third of the company's sales.
Hoechst, by some measures the world's largest chemical company, plans to leave chemi-
cals for the life sciences. It's a different world beyond our ivory towers.


I want to

use this

paper as a

way to

learn about

possible

changes in

chemical

engineering

curricula. I

am not

yet sure if

these ideas

are correct,

but I want

to see if

they make

sense. In

the next

few years,

I'll try them

out.


Winter 1999










THE STATUS IN ACADEMIA
While these industrial changes occur, academic chemical
engineering continues along well-established paths. I think
that this is good. Universities are both stable and resilient;
Clark Kerr, the long-time provost of the University of Cali-
fornia, is said to have asserted that universities make up
more than 90% of the social institutions that have lasted over
500 years. Moreover, courses in any field evolve slowly.
Woodrow Wilson, at the time
President of Princeton, said Che
Chemical
that "changing curricula is
like moving graveyards."
Chemical engineering cur-
ricula in the USA are no ex-
ception. To a large extent, Plant
they reflect the scheme first Engineerin
suggested in 1917 by a com-
mission chaired by Arthur D. New Process
Little, founder of the firm that Plants
bears his name. Building on
British precedents, the com- Equipmentern
mission suggested an orga- Design
nization around "unit opera-
tions." This was based on the
assertion that distillation was
based on the same principles Physics
for any chemical system, be Figure 1. Skills in Chemic
it rum or crude oil. This ideas from chemistry, ph
organization was codified jobs use different p
by the book Principles of
Chemical Engineering.[41
L. E. Scriven tells the possibly apocryphal story that the
book was written only because the authors isolated them-
selves at a camp in the Adirondacks, where they could
not be interrupted.
Principles of Chemical Engineering outlines much of what
would be a reasonable, accreditable major today. It begins
with a chapter on stoichiometry and then covers fluid flow
and heat transfer in three chapters. Four chapters on combus-
tion seem the intellectual ancestors of today's reaction engi-
neering. Four chapters on separations center on distillation,
humidification, and drying. Only the two chapters on me-
chanical separations (crushing and grinding) have material
missing from modern chemical engineering curricula. I
don't mean to overemphasize these parallels, because the
contents of these chapters are often qualitative and dated.
Still, I find the parallels vivid.
The curriculum implied by Principles of Chemical Engi-
neering was challenged most successfully by Transport Phe-
nomena, the book by Bird, Stewart, and Lightfoot.'5 This
book, circulated in 1957 and formally published in 1960,
injected more needed science and mathematics into our field.
For a while, our profession was divided into those who


believed in the older Principles testament and those who
converted to the newer Transport gospel. In one recent stimu-
lating article, Astarita and Ottino161 argued that these two
books have supplied the only two organizing ideas that our
profession has had.
In hindsight, I believe that there are two main reasons why
Transport Phenomena was so successful. First, by stressing
parallels between different transport processes, the book
supplies a pedagogical tem-
g plate that helps all to learn
neering
and think about these pro-
Plant cesses. This template is a
Operations mixed blessing. For example,
the fact that there is no par-
Process allel to chemical reactions in
R&D heat transfer means that
chemical reactions are super-
ficially treated. This may
A Chemical
A R&D contribute to our continuing
R& tendency to teach mass trans-
fer without chemical reac-
:t Technical tions, even though much in-
.Chemistry dustrial mass transfer, e.g.,
acid gas treating, takes place
with reaction.
Chemistry The second reason that


al Engineering. These skills are
sics, and engineering. Different
proportions of these ideas.


Transport Phenomena was
so successful is a reflection
of the boom taking place in
the chemical industry when


the book was published. As
outlined above, this boom centered on petrochemicals, which
of course included the monomers used to make synthetic
fibers. When you make petrochemicals, you often deal with
a plethora of compounds characterized by a near continuum
of boiling points. In such a case, continuum mathematics is
appropriate; one can basically ignore the discrete jumps of
the periodic table. Indeed, one can ignore most of chemistry,
with
A+B- C

i.e., argon plus boron goes to carbon. Moreover, as the
petrochemical industry became more competitive, minor im-
provements in existing processes were important to profit-
ability. These minor improvements could often be found
using the mathematical approach in Transport Phenomena.
While Astarita and Ottino argue powerfully that these two
books provide the only two paradigms in our profession, I
feel that Levenspiel's Chemical Reaction Engineering,171 first
published in 1963, is also important, but for a different
reason. The first two books provided a definition of a profes-
sion, which implied a curriculum. Levenspiel, on the other
hand, reorganized what was already acknowledged into a
Chemical Engineering Education


Engil










The changes in the chemical industry are clear-a movement away from commodities, a romance with
biotechnology, and a long-term interest in specialties. ... These changes in the industry do mean
that our students will work much more on chemical products than on chemical processes. As a
result, we will want them to think more about product design in addition to process design.


way that made it easier to learn. This can be
hard for the founder of a discipline to do. For
example, I view T.K. Sherwood as a founder
of mass transfer. I find his 1937 book Absorp-
tion and Extraction, 181 more understandable than
its 1952 successor Absorption and Extraction, [9]
co-written with Pigford. This second edition is Pro
in turn easier for me to understand than the
1975 revision, Mass Transfer,1 o] co-written with
Pigford and Wilkie. Levenspiel built on earlier
reaction engineering books such as Hougen
and Watson's Chemical Kinetics,"'l but he
achieved a new presentation that was much
easier to understand.
These various subjects in the chemical engi- Figure 2
less lil
neering curriculum can be represented on the
triangular diagram redrawn from Gerhard
Froelich, the 1999 AIChE president, and shown
in Figure 1. The three corners of this plot represent training
in the physical sciences, in the chemical sciences, and in the
chemical engineering subjects. Different jobs use these three
elements in different proportions, as shown in the figure.
There is no surprise in this; plant engineering will demand a
greater knowledge of mechanics and a smaller background
in chemistry than research and development. Figure 1 also
suggests national averages. British chemical engineers seem
to have somewhat more chemical engineering and less
chemistry than their US counterparts. Please don't take
this diagram too literally; use it instead as a catalyst for
thought, perhaps for deciding how your department's
curriculum should evolve.

DO INDUSTRY CHANGES
IMPLY ACADEMIC RESPONSES?
So far, I have summarized the revolution in the chemical
industry and the evolution of academic chemical engineer-
ing. I now want to compare the two to see what, if any,
changes are needed in what we teach.
Basically, I don't think many changes are indicated. The
skills we currently teach seem to prepare our students well.
Starting salaries remain high, the envy of most other engi-
neering disciplines. The number of jobs is again high, after
almost a decade of bad years caused by restructuring.1121 In
fact, the job market right now is better than I thought it
would be three years ago. Industrial complaints about our
teaching seem scattered, with about the same number urging
more, say, kinetics as those who urge less kinetics. Most
Winter 1999


Consult


1995


Consult Commodities








Products


Employment in 1975 versus 1995. Current graduates are much
kely to work for commodity chemical producers and more
likely to be involved with products.

industrial complainers who urge us to teach more of a par-
ticular topic are hard pressed to suggest which current topics
they would omit to make room for their favorite.
Thus, I believe our current curriculum is basically in good
shape. One frequent omission does concern me, however. I
want to explore this omission next.
My concern centers on the jobs our graduates now hold
compared with those they held perhaps twenty years ago.
My data for this are fragmentary, so I would be interested in
any other data that are available. My data are probably
biased toward large corporations, about whom our place-
ment office has better records. My data also have a regional
bias towards 3M and food companies such as General Mills
that are based here in Minnesota. Still, the data suggest
major changes in the last twenty years.
The focus of my analysis is the employment in 1975
versus that in 1995. I chose 1995 because the students often
need several years to settle down, to decide which sort of job
they really want to do. As shown in Figure 2, there are
enormous differences between 1975 and 1995. In 1975, three-
quarters of our graduates were working in the commodity
chemicals business. The small number who were not were
split between work on products, either product design or
product development, and work in other areas, which for
convenience I have labeled "consulting." That would in-
clude those working directly for consulting firms as well
as those carrying out specific tasks such as environmen-
tal impact statements.
In 1995, the distribution of jobs is different. The majority


1975










of students (in Minnesota's case, about two-thirds) now
work primarily on products. This includes not only students
who work on materials, but also those who work on pharma-
ceuticals, on specialty coatings, on adhesives, and on spe-
cialty chemicals. The number who work in commodity chemi-
cals has dropped so that it now is less than a quarter of our
graduates. The number who work in consulting has risen
dramatically, as commodity chemical businesses outsource
many of the in-house functions they used to do. For ex-
ample, in one case, a commodity chemical company took its
process engineering group from 1500 to fewer than 50 people.
This is not a business cycle; this is a change in the way they
expect to do business. This is why the number of people
involved in consulting has gone up.
Thus, the nature of the jobs that our students are doing has
changed dramatically. The next question concerns where the
changes are reflected in our curriculum. To explore this, I
have shown a basic generic curriculum in Table 2. It con-
tains the usual stoichiometry, the thermodynamics, and the
transport classes. The three classes in kinetics, process con-
trol, etc., are the place where departments will have unique
offerings. For example, this is the location of courses in
polymers or biochemical engineering or environmental en-
gineering. Such uniqueness is a strength of our departments,
a way in which we add special skills to a common core.
There are a few places in these classes that contain mate-
rial on products, that subject on which our students are most
likely to work. The most logical place to add this type of
material is in the capstone design class. This class usually
focuses on process design, the tradition of our discipline.
The hierarchy suggested by Jim Douglast13] for this process


design seems to me especially strong and appropriate. It is
summarized on the left side of Table 3. After deciding whether
a process is batch or continuous, one then moves on to flow
sheets, which are almost always continuous. The initial flow
sheets center on the stoichiometry. The next level in the
hierarchy, which adds the recycles, often involves a discus-
sion of the chemical reactions. Once these are established,
one moves on to the separation trains and finally to the heat
integration. All of this makes for a good course.
If we want to emphasize product design, we need to go
beyond this hierarchy. We cannot simply substitute a prod-
uct for drug delivery for the existing process and carry out
the same kind of hierarchy. Instead, the hierarchy suggested
by books on product design (e.g., Ulrich and Eppinger[141) is
exemplified by that on the right side of Table 3. After first
identifying a corporate need, one generates ideas to fill this
need. One then decides between these alternatives and fi-
nally decides how to manufacture the chosen product. The
manufacturing step essentially includes all of Jim Douglas'
hierarchy.
Thus the important steps in product design anticipate those
in process design. Product design implies a focus on the
initial decisions around the form of the product and implic-
itly de-emphasizes its manufacture. Such an emphasis shifts
the curriculum away from the common engineering calcula-
tions that have been our bread and butter. Such an emphasis
includes subjects that are normally left to those directly
concerned with the business. I am concerned that if I make
this shift in a design class, I will wind up teaching my
students watered-down business school principles rather
than "real" engineering. I undertake this change because


Chemical Engineering Education


TABLE 2
Generic Chemical Engineering Curriculum

Most universities teach a similar sequence.
f Stoichiometry (1 course)
3 Thermodynamics (3 courses)
3 Transport Phenomena and Unit Operations (3 courses)
3 Reactors, Process Control, etc. (3 courses)
3 Process Design (2 courses)


TABLE 3
Process Design versus Product Design
All of process design is contained in the last step of product design.

Process Desien Product Design
1. Batch vs. Continuous Process 1. Identify Customer Needs
2. Inputs and Outputs 2. Generate Ideas to Meet Needs
3. Reactors and Recycles 3. Select among Ideas
4. Separations and Heat Integration 4. Process Design for Manufacturing


TABLE 4
"Sick House" Ventilation

1. Customer need; ventilate for under $800
2. Ideas: Open window
Controlled vent
Heat exchanger
Heat and humidity exchanger
3. Select heat and humidity exchanger
4. Manufacture follows kidney dialysis










so many more of my students are encountering this shift
in their professional lives. I want them to see how prod-
uct design works.
When I've discussed these ideas with other faculty, I often
get the indignant reaction that the faculty are already doing
this. Some have mailed me syllabi and reports that include
aspects of product design. Without exception, what I have
received represents good education, but almost without ex-
ception, the material seems to skip all steps except the last in
the product design hierarchy in Table 3. These earlier steps
seem to me too important to leave to the MBAs.
As an example of these ideas, consider the so-called "sick
house syndrome" that has developed as houses were built to
be energy efficient. Such houses exchange their air as infre-
quently as twice a day. In contrast, a house built fifty years
ago exchanges its air almost every forty minutes. Thus,
while the modem house does not cost much to heat, it can
concentrate radon from the basement, formaldehyde released
from carpeting and drapery, and carbon dioxide from the
people who live in the house. The modern house needs
more fresh air. Thus the product needed is a device that
allows a house to remain energy-efficient, but which
provides fresh air at the ASHRE standard of 19 cubic
feet/minute/person in the house.
The way in which the product development might proceed
is shown in Table 4. The need is for a device costing less
than about $800 that can provide this degree of ventilation.
Ideas include opening a window, providing automatic con-
trol for opening a window, providing a heat exchanger, and
providing an exchanger for both energy and mass. Opening a
window sacrifices the energy benefits of insulating the house
in the first place. Opening the window with an automatic
controller that might anticipate weather cycles makes sense.
For example, one could open the window only on sunny
winter days and keep the house closed on cold winter nights.
Using a heat exchanger can provide the necessary ventila-
tion at an order of magnitude less heat loss. As anyone who
has bought a house with such a heat exchanger knows,
however, the heat exchanger also exhausts the water va-
por in the house. The heat of evaporation of the water is
about a third of the heating value in the humidified air. If
the heat exchanger runs, the house dries out and becomes
very uncomfortable.
The final alternative is the most complicated, but the most
satisfying. In this case, one uses a heat exchanger in which
the walls are membranes selectively permeable to water
vapor. As a result, one captures 90% of the energy and 90%
of the water vapor, but exhausts the carbon dioxide, formal-
dehyde, and radon in the house. The question is cost. The
students need a more complete design, perhaps using the
manufacturing technology developed for kidney dialysis, to
make the membranes. This is an area of active commercial
development by several heat-exchanger companies.
Winter 1999


CONCLUSIONS
We are now ready to answer the question posed in the title
of this paper: "Do changes in the chemical industry imply
changes in the chemical engineering curriculum?" The
changes in the chemical industry are clear-a movement
away from commodities, a romance with biotechnology, and
a long-term interest in specialties. Major changes in the
curriculum are probably not needed; our students still have
the basic skills necessary not only for the changed chemical
industry but also for the other jobs they now hold.
These changes in the industry do mean that our students
will work much more on chemical products than on chemi-
cal processes. As a result, we will want them to think more
about product design in addition to process design. The work
on product design will follow a different hierarchy than that
which effectively organizes process design.
But I'm not sure of this. You may remember that I began
this article by saying that I was going to follow the lead of
Pearce Williams to write a paper on what I thought might be
done rather than what I had already found effective. With
Geoff Moggridge, I am going to teach product design as a
Zeneca fellow at Cambridge University in the academic year
1998-1999. If we are successful, I will try to move some of
these ideas back into our design courses here at Minnesota. I
am not yet sure they will work. I look forward to discussing
with you what parts do work and what parts do not.

REFERENCES
1. Williams, L. Pearce, Michael Faraday, Simon and Schuster,
New York, NY (1964)
2. Spitz, Peter, Petrochemicals: The Rise of an Industry, Wiley,
New York, NY (1988)
3. Ellis, R.A., "At the Crossroads," Eng. Workforce Bull. No.
130(1994)
4. Walker, W.H., W.K. Lewis, and W.H. McAdams, Principles
of Chemical Engineering, McGraw-Hill, New York, NY (1923)
5. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport
Phenomena, Wiley, New York, NY (1960)
6. Astarita, G., and J.M. Ottino, "Thirty-Five Years of BSL,"
Ind. Eng. Chem. Res., 34, 3177 (1995)
7. Levenspiel, 0., Chemical Reaction Engineering, Wiley, New
York, NY (1963)
8. Sherwood, T.K., Absorption and Extraction, McGraw-Hill,
New York, NY (1937)
9. Sherwood, T.K., and R.L. Pigford, Absorption and Extrac-
tion, McGraw-Hill, New York, NY (1952)
10. Sherwood, T.K., R.L. Pigford, and C.R. Wilke, Mass Trans-
fer, McGraw-Hill, New York, NY (1975)
11. Hougen, O. A., and K.M. Watson, Chemical Process Prin-
ciples III: Kinetics and Catalysis, Wiley, New York, NY
(1947)
12. Weatherall, R.K, "Strongest Market in Years for New Gradu-
ates," Engineers, 4, 2 (1998)
13. Douglas, J.M., Conceptual Design of Chemical Processes,
McGraw-Hill, New York, NY (1988)
14. Ulrich, K.T., and S.D. Eppinger, Product Design and Devel-
opment, McGraw-Hill, New York, NY (1995) 0











g classroom


DISCONTINUITIES IN


ChE EDUCATION


STEPHEN WHITAKER
University of California Davis, CA 95616


From the perspective of the first-year student, the en-
tire four-year chemical engineering program repre-
sents an overwhelming array of courses and subject
matter. One must learn about ionic strength and indefinite
integrals, acoustics and hydrostatics, turbulence and chemi-
cal kinetics, organic chemistry and process dynamics, optics
and quantum mechanics, stoichiometry and process synthe-
sis, radiant energy heat transfer and partial differential equa-
tions, etc., etc. Viewed in its entirety, the typical chemical
engineering program is enough to make a student change
majors; but if taken one step at a time, the overall objective
becomes quite feasible.
In the ideal chemical engineering program, one would like
to develop a seamless passage from ionic strength to process
synthesis. Given the size of the task, it should not be surpris-
ing that the route from A to Z contains a few "leaps of faith."
The failed leap of faith within the confines of the university
leads only to a lurking sense of insecurity and no real physi-
cal damage. Outside the university, however, a failed leap of
faith may be a financial disaster, a physical disaster, or both.
For this reason, we should avoid or minimize the leaps of
faith in our educational programs or we should clearly iden-
tify them as such. A discussion of the so-called principle of
lost work represents an interesting example of the latter.111

THE MECHANICAL DISCONTINUITY
In the first physics course, students encounter Newton's
second law for a particle written in the form

d(mv)=F (1)
dt


Copyright ChE Division ofASEE 1999


Here, m is the mass of the particle, v is the velocity of the
particle relative to an inertial frame, and F is the force acting
on the particle. Given the success of Eq. (1) in predicting the
motion of the planets around the sun and in predicting the
motion of projectiles in a physics lecture hall, students ac-
quire a certain degree of confidence in Newton's second
law. This confidence may begin to weaken when they move
on to a chemical engineering study of fluid flow where they
are often confronted with a dictum of the form


sum of forces
acting on
the control
volume


rate of
momentum out
of the control
volume


rate of
momentum into
the control
volume


rate of
accumulation of (2)
momentum in the
control volume


It is true that the concept of a control volume has already
been presented in a course on material balances; but the
distance between Eq. (1) and Eq. (2) is so great that most
students view the latter with some distrust. The student's
skepticism is quite justified, but the repeated use of Eq. (2)
to solve real problems eventually leads to its acceptance.
Such a leap of faith in the design of an oil pipeline passing
underneath the city of Los Angeles would never be consid-
ered, but Eq. (2) is something that everyone knows is true,
and many students move successfully through their pro-
grams of study with it securely locked in their tool boxes.

THE THERMODYNAMIC DISCONTINUITY
When students have made sufficient progress in their stud-
ies of fluid mechanics, they will often solve a variety of
incompressible flow problems (it is important to keep in
mind that there are no incompressible fluids, but there are
flows that can be approximated as incompressible) using the
Navier-Stokes equations and the continuity equation. These
equations can be expressed as

P +v. Vv Vp + g + gV2v (3)
at )
V-v=0 (4)

Equation (3) represents the governing differential equation


Chemical Engineering Education


Stephen Whitaker received his undergradu-
ate degree in chemical engineering from the
University of California, Berkeley, and his PhD
from the University of Delaware. He is the au-
thor of two undergraduate texts, Introduction to
Fluid Mechanics and Fundamental Principles
of Heat Transfer, and a monograph, The Method
of Volume Averaging. He has received a num-
ber of awards for his contributions to both un-
dergraduate and graduate teaching.










for the fluid velocity, v, and Eq. (4) is the constraining
equation for the vector field represented by Vp. That is to
say that the momentum source, Vp, must be distributed in
such a manner that the velocity determined by Eq. (3) will be
solenoidal. If the pressure is specified at some point, one can
use the vector field, Vp, to calculate the pressure every-
where. The actual determination of the pressure field is
discussed in some detail in Reference 2.
At the same time that students are using Eqs. (3) and (4) to
determine the pressure in a course on fluid mechanics, they
are also calculating the pressure in a course on thermody-
namics using an equation of state. The simplest equation of
state is the ideal gas law given by


p= RT


This expression for the pressure would appear to have no
connection with the pressure that one would determine from
Eqs. (3) and (4), thus suggesting that there is a mechanical
pressure used in the solution of certain fluid-flow problems
and a thermodynamic pressure used in the solution of ther-
modynamics problems. It would be best to think of the
pressure, as determined by an equation of state, as the pres-
sure, and to think of the pressure determined by Eqs. (3) and
(4) as a good approximation of the pressure. If it is not a
good approximation, Eqs. (3) and (4) should not be used to
solve the flow problem under consideration. The resolution
of the conflict between Eqs. (3) and (4) and an equation of
state, such as Eq. (5), relies on Birkhoff's plausible intuitive
hypothesis that small causes give rise to small effects.

THE MULTIPHASE DISCONTINUITY
After having completed courses in fluid mechanics, ther-
modynamics, heat transfer, and mass transfer, chemical en-
gineering students are often confronted with a course on
mass transfer operations or unit operations. Since virtually
all chemical engineering processes involve multiphase sys-
tems, a study of the gas-liquid contacting device illustrated
in Figure 1 is a harbinger of things to come, and students
approach this problem with a great deal of interest. Often
they are equipped with Eqs. (3) and (4) from a course on
fluid mechanics, the thermal energy equation from a course
on heat transfer141

pCp +v-VT= kV2T (6)

and the species continuity equation from a course on mass
transfer'51
ac
A+V(CAVA)=RA A=1,2,...,N (7)

Most students are somewhat dismayed when the process
illustrated in Figure 1 takes on the form shown in Figure 2,
and the rigor represented by Eq. (7) is replaced by the
Winter 1999


suggestion that
mass of A entering mass of A leaving mass of A transferred
in the gas phase in the gas phase + to the liquid phase (

After the struggle to reach Eq. (7) via a series of challenging
courses, it is disappointing to be asked to return to the
concepts encountered in the course on material balances.
What is worse is that the analysis of the process illustrated in
Figure 1 will be heavily based on the intuition suggested by
Figure 2 and most students will have no idea how reliable
the final result will be. The resolution of this problem can be
achieved using the method of volume averaging.161

RESOLUTION OF THE
MECHANICAL DISCONTINUITY
Rather than leap from Eq. (1) to Eq. (2), one can follow a
sequence of steps that begins in the eighteenth century'17 and
leads to our current understanding of continuum mechanics.
A central idea in continuum mechanics is that the laws of
physics can be applied to any body that one imagines as
being cut out of a distinct body. Truesdellv7 attributes this
idea to Euler and Cauchy and refers to it as the cut principle.
Engineering students encounter this idea in a course on
statics where it leads them to the concept of a free-body
diagram. If we accept this idea, we can cut an arbitrary body
from a moving, deforming fluid and state the axioms for
mass and mechanics as follows:
Mass
d f
dt pdV= 0 (9)
',, (t)
Linear Momentum: Euler's First Law
dt pvdV= jpbdV+ tndA (10)
'n "" (',,, ( ,(t)


Figure 1. Gas-liquid contacting
device.


_.._ _

I







Figure 2.
Model of a gas-
liquid contact-
ing device.









Angular Momentum: Euler's Second Law
d JrxpvdV= rx pbdV+ frxt(n)dA (11)
Vm(t) V (t) Am(t)
Here, Vm(t) represents the time-dependent region occupied
by a body, p is the mass density, v is the fluid velocity, b is
the body force per unit mass, t(,) is the stress vector, and r is
the position vector. Both v and r are measured relative to
some inertial frame. The representation of the angular mo-
mentum principle given by Eq. (11) assumes that all torques
are the moments of forces and this ignores the existence of
body torques and couple stresses that have been observed in
polar fluids.18] The forms of these three axiomatic statements
suggest the need for a study of the kinematics of volume
integrals and this leads to the general transport theorem9']
given by
d VdV= dV+ f< w-ndV (12)
dt J J at I
Va(t) Va(t) Aa(t)
Here, Va(t) represents the region occupied by an arbitrary
moving volume, Aa(t) is the bounding surface of this vol-
ume, and w n is the speed of displacement of the bounding
surface. When the arbitrary velocity, w, is set equal to the
fluid velocity, v, we obtain the Reynolds transport theorem
given by

-dt ~ dV= fItdV+ v rndV (13)
V.(t) Vm(t) Am(t)
When applied to Eq. (9), this theorem provides

dt pdV= FapdV+ fpvndV=0 (14)
dt f J at J
V m(t) Vm(t) ^m(t)
and use of the divergence theorem leads us to

f J[+V-(pv) dV=O (15)
[J at
Vm(t)
Assuming that the integrand is continuous and noting that
the limits of integration are arbitrary leads to the continuity
equation
+ +V.(pv)=0 (16)
at
In order to extract the governing differential equations
associated with the linear and angular momentum principles,
we first need to follow the work of Cauchy and prove[71
Cauchy's Lemma:
t() = -tn) (17)
Cauchy's Fundamental Theorem:
t(n) = n T (18)
The first of these is introduced as intuitively obvious in
every statics course where it is applied to the shear stresses
20


acting on opposing surfaces of a beam that has been sub-
jected to an Eulerian cut. The second result is generally
avoided because of its complexity even though most stu-
dents have completed a course on matrix algebra prior to
their study of fluid mechanics.
The use of Eqs. (17) and (18), along with the Reynolds
transport theorem, allows us to extract the following differ-
ential equations from Eqs. (10) and (11):
Cauchy's First Equation
S(pv)+ V(pvv)=pg + VT (19)
at
Cauchy's Second Equation
T=TT (20)
At this point, we are in a position to derive the macroscopic
momentum balance that was described in words by Eq. (2).
We begin by integrating Eq. (19) over an arbitrary, moving
control volume to obtain

ft(pv)dV+ fV-.(pvv)dV= fpbdV+ V -TdV (21)
ga(t) f(t) a(t) f,(t)
We now use the general transport theorem and the diver-
gence theorem to arrange this result in a useful form given
by
dt pvdV+ Jpv(v-w)-ndA= JpbdV+ t(n)dA (22)
V (t) a,(t) V' (t) A.(t)
This represents a precise mathematical description of the
words contained in Eq. (2), and it clearly indicates that the
source of these words is Euler's first law. While the route
from the axiom given by Eq. (10) to the proved theorem
given by Eq. (22) consists of only a few steps, one must
invest a significant amount of time in the study of kinematics
and stress in order to derive this result. (At UC Davis we
have two ten-week courses in fluid mechanics and time is
less of a problem than in most programs.) While kinematics
and stress may be confusing, we should heed the words of
Pucciani and Hamel,110o who provide the following advice to
students:
There is no learning without confusion. It is by the organi-
zation of this confusion that you will progress.
Said another way, it is better to be confused and frustrated by
the concepts of kinematics and stress than to be baffled by
the leap of faith from Eq. (1) to Eq. (2).
In order to follow the development from Eq. (10) to Eq.
(22) in a successful manner, the faculty must be aware of
what students know, what they don't know, and what they
are supposed to know. For example, in the typical statics
course, students make use of Euler's laws to solve problems,
but the laws are never identified in a clear and concise
manner. The concept of an Eulerian cut is presented as being
obvious en route to the development of a free-body diagram,
Chemical Engineering Education









but no mention is made of the fact that it was not obvious in
the eighteenth century. Cauchy's lemma is used in the same
way, i.e., the shear is up on this side of the cut and down on
that side of the cut. In order to unravel this mass of intuition,
the faculty must be aware of the content of previous courses
and must be prepared to extract some order and logic from
the student's previous studies. In addition to statics, these
previous studies include calculus where the definition of a
derivative is presented and the projected area theorem is
given."" These are all that one needs to derive the general
transport theorem. Previous studies also include a course on
matrix algebra where the students learn that a three-by-three
array can be used to transform one set of three numbers to
another set of three numbers. This is the essential feature of
Cauchy's fundamental theorem.

RESOLUTION OF THE
THERMODYNAMIC DISCONTINUITY
In order to resolve the thermodynamic discontinuity, we
begin with a reasonable description of a compressible flow
process. This consists of the governing equations for the
density, velocity, temperature, and pressure that can be ex-
pressed as
Governing equation for p

p+V.(pv)=0 (23)
at
Governing equation for v

p( +v-Vv =-Vp+pg+pV2v (24)
(t a

Governing equation for T

pc T +v-VT kV2T (25)

Governing equation for p
p=p(p,T) (26)

The first of these equations represents a completely general
form of the continuity equation, while the last represents a
completely arbitrary equation of state. Equations (24) and
(25) represent special forms of the equations of motions and
the thermal energy equation, but they are general enough for
our purposes.
The usual concept associated with an incompressible flow
is that the variation of the density is small enough so that the
dependent variable in Eq. (23) can be replaced with a con-
stant, po. This means that one of our four dependent vari-
ables is determined by some means other than a law of
physics and this, in turn, means that we must discard one of
our laws of physics. Our new description of the physical
process is given by
V.vm =0 (27)


Po +vm VVm =-VPm +po+ V2Vm (28)

PC v- + VTm )=kV2Tm (29)

in which we have used Vm, Pm, and Tm to represent the
velocity, pressure, and temperature determined by Eqs. (27)
through (29). These quantities differ from p, v, p, and T that
are determined by Eqs. (23) through (26), and we would like
to understand the asymptotic conditions that lead to

P-*Po V- Vm P- Pm T-Tm (30)
When Eqs. (27) through (29) produce velocity, pressure, and
temperature fields (Vm, Pm, Tm) that are good approximations
of the fields (v, p, T) determined by Eqs. (23) through (26),
we say that the flow can be approximated as incompressible.
Under these circumstances, the pressure can be calculated
by purely mechanical means; but it would seem best not to
refer to pm as a "mechanical pressure," but simply to say that
pm is a "good approximation" of the pressure determined by
an equation of state.
The general asymptotic conditions associated with Eq.
(30) are difficult to develop;[l21 we can, however, explore the
first of these conditions for steady flow in the absence of any
temperature effects without a great deal of effort. This re-
quires that we consider an isothermal process described by
the steady forms of Eqs. (23) through (26), and then search
for conditions that lead to
P-Po (31)
It will be in the nature of a plausible intuitive hypothesis'31 to
assume that v vm and p pm when the condition repre-
sented by Eq. (31) is satisfied.
For the case in which temperature effects are negligible,
we can invert Eq. (26) to obtain

p=p(p) (32)
and a Taylor series expansion about Po leads to

p=Po+(-pP) LP +(P-Po)2 \-- +.. (33)
PPT 2 +p2 T
Here, po is the density determined by Eq. (32) at the refer-
ence pressure po. As an estimate of the density change that
occurs for the process under consideration, we use the first
term of the expansion to obtain

P-P=0o[ Ap (34)

in which Ap is representative of the maximum pressure
change that occurs in the system. From a thermodynamic
analysis, we know that the speed of sound is related to the
derivative of the density with respect to the pressure at
constant entropy. This relation is given by


Winter 1999










lap 1 (35)
-mi =(35)
ap, c
in which c is the speed of sound. As an approximation, we
use


)T C (36)
so Eq. (34) takes the form

p-p0o=02 (37)
SACp
In order to obtain an estimate of the pressure change, we first
make use of the steady form of Eq. (24) to estimate the
gradient of the pressure as
Vp= O(pg) + O(V2v)+ O(pv. Vv) (38)

The idea associated with this estimate is that Vp may be as
large as any of the other terms in Eq. (24) but not signifi-
cantly larger. In addition, it is possible that Vp may be much
smaller than any of those terms, and thus Eq. (38) should be
thought of as an overestimate of the pressure gradient. For
example, in a laminar boundary layer created by a uniform
flow past a flat plate, the pressure gradient is essentially
hydrostatic
Laminar boundary layer flow
Vp pg (39)
while the viscous and inertial terms are essentially equal and
much larger than Vp pg, i.e.,
Laminar boundary layer flow
pv Vv =O(V2)>> Vp-pg (40)

In this case, only the first estimate given by Eq. (38) is valid
and it becomes clear that one must have some idea about the
nature of the flow under consideration in order to use Eq.
(38) successfully.
If uo represents the characteristic velocity for the process,
we can use order-of-magnitude analysis131 to obtain the esti-
mates

pg = O[(pg)Xg] (41a)
gV2v = O[(uo /L2),] (41b)

pv Vv = O(pu / Lp)p] (41c)

Here, L. represents the viscous length, and Lp represents
the inertial length,"141 while Xg,.9X, p are unit vectors that
are parallel to the gravitational term, the viscous term, and
the inertial term, respectively. For the laminar boundary
layer example discussed above, X. and Xp are parallel, and
Xg is an arbitrary unit vector.
In order to make use of Eq. (38) to estimate the pressure
difference that appears in Eq. (37), we express the pressure
gradient as


Vp = O(Ap / L) (42)
Here one must keep in mind that the pressure gradient will
be influenced by all three terms represented by Eqs. (41),
thus Eq. (42) represents three separate estimates and it is left
to the reader to keep this fact in mind. Use of Eqs. (41) and
(42) in Eq. (38) leads to the following estimate for the
pressure change owing to gravitational, viscous, and inertial
effects:
Ap= O[(pg)L] + O[(pu, / L )L] + O[(pu2 / Lp)L] (43)

Substitution of this result into Eq. (37) provides the follow-
ing estimate for the change in density that occurs for the
process under consideration:
P-P = O(gL/c2)+ o( uoL/pL c2)+O(u2L/Lpc2) (44)

If we define a Reynolds number and a Mach number as
Re pu M -- (45)
M c
the estimates given by Eq. (44) take the form
P-Po -=O(gL/c2)+O[Re- M2(L/Lg)]+O[M2(L/Lp)] (46)

From this result we conclude that


P Po << 1,
p
p


or P Po


when the following constraints are satisfied:


gL/c2 <<


M2<

Here, one must remember that L has a different meaning in
each one of these constraints, which are often replaced with
the single condition that the Mach number squared is small
compared to one. This has considerable appeal for the last
constraint in Eq. (48) since Lp is often large compared to L;
the simplification given by M2<<1 has less appeal as a
substitution for the second constraint, however, since L. is
generally small compared to L. A little thought will indicate
that the first constraint given by Eq. (48) is difficult to
violate, and thus it is the constraints involving the Mach
number that must be considered with care.
It seems plausible that when Po is a reasonable approxi-
mation for p, we can assume that Vm and pm are reasonable
approximations for v and p; but a rigorous proof would
require that we identify the asymptotic conditions that allow
us to simplify Eqs. (23) through (26) to the incompressible
approximation represented by Eqs. (27) through (29). One
should keep in mind that the analysis leading to the con-
straints given by Eqs. (48) was based on a steady-flow
process, and that there are unsteady, low-Mach-number pro-
cesses for which the flow cannot be treated as incompress-
ible under any circumstances. An example is given in Figure
3, and for that type of process the approximation represented


p P = O(Vp)L (49)
Chemical Engineering Education










is not at all applicable.
The partial resolution of the thermodynamic discontinuity
required that we clearly identify the general case indicated
by Eqs. (23) through (26) and that we discard a law of
physics in favor of the approximation given by p = po. The
justification of this approxi-
mation was based on the
equation of state that pro-
vided Eq. (37) and the order-
of-magnitude analysis given
by Eqs. (38) through (46).
The use of order-of-magni-
tude analysis allows students
to go beyond the assumptions
compressed gas based on the title of a course,
the title of a text, or the title
of a chapter in a text. Since
these titles are not available
to our students when they
Figure 3. Compression leave the university, we
process, should encourage them to
formulate their own assump-
tions and then follow those assumptions with restrictions
and constraints. 131
It is important to understand that the thermodynamic dis-
continuity cannot be resolved only by discussion in fluid
mechanics courses. Faculty members who teach thermody-
namics must be aware of the problem and speak to the issue.
A righteous attitude about the correctness of Eq. (5), or a
more general equation of state, provides no help to the
student who must deal with the reality of incompressible
flows. One must remember that the students take every
course in the program and they do not have the luxury to
choose their battleground.

RESOLUTION OF THE
MULTIPHASE DISCONTINUITY

Studies of multicomponent mass transport usually include
a derivation of the species continuity equation51, and the
molar form of this result is given by
A +V(CAVA)= RA A = 1,2,3,... N (50)
at
Knowledge of the molar concentration, cA, is a central issue
in chemical engineering since it forms the basis for all sepa-
ration and purification processes, for all reactor design cal-
culations, and for all studies of contaminant transport in the
land, air, and water. The macroscopic mole balance associ-
ated with Eq. (50) can be derived by following the steps that
link Eq. (19) to Eq. (22), and the result is given by


d CAdV+ JcA(VA-w).ndA= jRAdV
va(t) a(t) 'V(t)
Winter 1999


A=1,2,...,N


Both Eq. (50) and (51) represent powerful problem-solving
tools, and most chemical engineering students acquire a
certain degree of skill in the application of these results for a
variety of single-phase transport problems; their application
to multiphase systems, however, is problematic.
Most multiphase transport processes cannot be solved di-
rectly in terms of either Eq. (50) or Eq. (51), but require
instead the local volume-averaged form of Eq. (50).5s-'81 The
development of this form begins by associating an averaging
volume with every point in the region under consideration.
This allows one to define a volume-averaged concentration
everywhere and to generate a spatially smoothed concentra-
tion field. In Figure 4 we have illustrated a two-phase system
and a spherical averaging volume having the centroid lo-
cated at the point identified by the position vector x. For this
system, we identify the point concentration of species A in
the y- phase as CAY and we define the superficial average
concentration by

(cAY)lx= JCAydV (52)
Vy(xt)

Here, Vy(x,t) represents the volume of the y- phase con-
tained within the averaging volume, V, and we have clearly


indicated that the
superficial aver-
age concentra-
tion is associated
with the point lo-
cated by the po-
sition vector x.
For the particu-
lar case illus-
trated in Figure
4, the position
vector x locates
a point in the
- phase where
the point con-
centration of
species A may
be zero.


Figure 4. Two-phase system.


In general, the intrinsic average concentration is preferred
for the analysis of multiphase transport processes, and this is
defined by

cA) V(x, t) JcAdV (53)
v7 (x.t)
The superficial and intrinsic average concentrations are re-
lated according to

CAy)=E y(CAy) (54)
in which e. is the volume fraction of the y-phase defined
explicitly by


S' .- "- '




I ,/ y -phase I





*/' ^Cr










Ey = V(x,t)/V (55)
In many systems, the superficial and intrinsic averages dif-
fer by a factor of three or more, and thus it is important to
make use of a nomenclature that clearly identifies these two
concentrations.
In Figure 5 we have shown a two-phase system for which we
would like to develop the design equation for the concentration
of species A. We think of the flowing fluid as the y phase,
while the o phase could represent porous catalyst pellets, or
droplets of a more dense fluid that is descending through the
y phase. The governing differential equation for the concen-
tration of species A in the y phase is given by

S+V- (CAyVAVy)=RAy (56)

and we begin our analysis of this point equation by forming
the superficial average to obtain

I A dV+- + V-(CAyVAy)dV= RAydV (57)
V7(x,t) V (x,t) Vy(x,t)
In order to transform this result to something useful, we
make use of two theorems that are essentially extensions of
the classic one-dimensional Leibniz rule[9 Prob. 3-5] for differ-
entiating an integral. The first of these is the general trans-
port theorem, which we used earlier in our treatment of the
mechanical discontinuity. This theorem allows us to express
the first term in Eq. (57) in the form


1 CA7 dV
V J tdV
Vy(x,t)


d fv CAydV JAYcwA" dA (58)
S V(xt) Ap(xt)


Figure 5. Mass transfer and reaction process in a
two-phase system.


Here, n, represents the unit normal vector directed from
the y- phase toward the o- phase, and w- n represents
the speed of displacement of the y -o interface. This is zero
if the system under consideration is a packed-bed catalytic
reactor, but would be non-zero for a fluidized-bed reactor or
any fluid-fluid system. The second theorem needed for the
analysis of Eq. (57) is the spatial averaging theorem, and the
derivation of this theorem16'19 is analogous to the derivation
of the general transport theorem. Application of the spatial
averaging theorem provides

I fv.(CcAy VA)dV
v,(x,t)

=V-- JCAYVAydV +- + fCAVA "ndA (59)
Vy (x,t) A (x,t)

and use of this result, along with Eq. (58), in Eq. (57) leads to

d1
d [ vjYdVI + V cvAVAydV
V,(x,t) V,(x,t)

+ IC Ay(V Ay -W)-dA=- fRAdV (60)
A (x,t) V,(x,t)

One should note that this result is a superficial average
transport equation, and thus each term has units of moles of
species A per unit time per unit volume of the y- system.
Use of the nomenclature for the superficial average indi-
cated by Eq. 52 allows us to express Eq. (60) in the more
compact form given by

-(cAY)+V-(CAyVAy)+- CAy(VAy- w)- ndA = (RA) (61)
accumulation y phase A (x,t) homogeneous
transport interfacial transport reaction
Here it is understood that the averaged quantities are associ-
ated with the centroid of the averaging volume identified by
the position vector x in Figure 4. In addition, we have ex-
pressed the accumulation in terms of the partial time derivative
since (cA.y) is associated with a point that is fixed in space.
The first and last terms in Eq. (61) can be expressed in
terms of intrinsic averages by using the relation given by Eq.
(54), and this leads to

atE ( ) + (cAyVAy)

= fcAy Ay7 w). ntdA+e (RAyY (62)
Ap(x,t)
In order to simplify the convective transport term, one can
use the averaging theorem and the divergence theorem to
show that

v.(CAyAy) =1 JAyVAy -ndA (63)
A, (x,t)
Chemical Engineering Education


volume, Vt









in which A, (x,t)represents the area of entrances and exits
associated with the volume, Vy(x,t). For many practical
applications, diffusive transport is negligible compared to
convective transport at entrances and exits, and this encour-
ages the simplification
.(CAyVAy)=V (CAyy) (64)

in which vy is the mass average velocity. At this point we
make use of Gray' s1201 spatial decomposition for the concen-
tration and velocity

CA =CAy) +CAy Vy =(v) +iv (65)
and follow the work of Carbonell and Whitaker121] to express
the convective transport in the form


(CAyVy) = E7(CA)(Vy )Y
volume averaged
convective transport


+ (CAyy)
dispersive
transport


While the intrinsic average concentration is the preferred con-
centration for the description of multiphase mass transport
processes, most workers favor the use of the superficial aver-
age velocity

v}y)=r y)y (67)
and thus we normally express Eq. (66) as

(cAyVy) = (CAy) (vy ) + (Ayy) (68)
volume averaged dispersive
convective transport transport

Use of this result (and the approximation given by Eq. 64)
with Eq. (62) leads to

atV ( CA)+v (CAy)(v

=-V(cAyVy)- CA VAy, w)-ndA+eRAy,) (69)
A,(x,t)
Here we are confronted with the development of appropriate
representations for the dispersive transport, the interfacial
transport, and the homogeneous reaction rate.
The diffusion model of dispersion is an approximation for
active systems;122"231 but it is likely to be quite acceptable for
most design applications. This encourages us to express the
dispersive flux as

(eAyV = -D-.VcA)' (70)
our local volume-averaged transport equation takes the form


S(i y ))) + V.
accumulation volume averaged
convective transport


V D.V(cAY)
dispersive
transport


f cAy(VAy w).nydA + Ey(RAY)
A (x,t) homogeneous
interfacial transport reaction
Winter 1999


(71)


The treatment of the interfacial flux depends on the type of
process that takes place in the a-phase; Equation (71),
however, provides the basic framework for a design equa-
tion that has been derived directly from Eq. (56). This pro-
vides a much more rigorous formulation than one can obtain
on the basis of the statement
moles of A moles of A moles of A moles of A)
entering ) produced by reaction )( leaving J (accumulated)

and it clearly sends a message to our students that their
efforts to understand the details of diffusive and convective
mass transport have not been wasted.
When I teach the second of our two mass transfer courses
at UC Davis, I use the general transport theorem to obtain
Eq. (58) because I know that the students have proved this
theorem in the fluid mechanics course. In addition, I derive
the spatial averaging theorem by a straightforward applica-
tion of the projected area theorem that is presented in the
fluid mechanics course. It also helps to know that the pro-
jected area theorem was encountered originally in one of the
calculus courses.'] 1Chap. 17] If I did not know what the students
had covered in previous courses, I would be forced to retreat
to the approach suggested by Figure 2.

CONCLUSIONS
In this paper we have shown how some of the traditional
discontinuities associated with chemical engineering educa-
tion can be eliminated or minimized. Time and effort are
required to accomplish this, but the effort will convince
students that leaps of faith can be avoided, and this is surely
worthwhile.

ACKNOWLEDGMENTS
This paper is based on a lecture presented at the VII
Encontro Brasileiro sobre Ensino de Engenharia Quimica in
Caxambu, Brazil, 1997. The enthusiastic response by my
Brazilian colleagues provided the motivation to prepare a
written version, and their encouragement is greatly appreci-
ated.

NOMENCLATURE


Aa (t) area of the surface of an arbitrary moving volume, Va (t), m2
m (t) area of the bounding surface of a moving, deforming
body, m2
A (x, t)area of the y- interface contained in the averaging
volume, i, m2
y,e (x, t) area of the entrances and exits of the y phase contained
in the averaging volume, V, m2
b body force per unit mass, N/kg
cA molar concentration of species A, mole/m3
CAy molar concentration of species A in the y phase, mole/
m3
(CAy) superficial volume averaged concentration of species A,
mole/m3
Continued on page 61.










Re' class and home problems


The object of this column is to enhance our readers' collections of interesting and novel
problems in chemical engineering. Problems of the type that can be used to motivate the student
by presenting a particular principle in class, or in a new light, or that can be assigned as a novel
home problem, are requested, as well as those that are more traditional in nature and that
elucidate difficult concepts. Manuscripts should not exceed ten double-spaced pages if possible
and should be accompanied by the originals of any figures or photographs. Please submit them to
Professor James O. Wilkes (e-mail: wilkes@engin.umich.edu), Chemical Engineering Depart-
ment, University of Michigan, Ann Arbor, MI 48109-2136.




NON-ADIABATIC CONTAINER

FILLING AND EMPTYING


NOEL DE NEVER
University of Utah Salt Lake City, UT 84112


A ll thermodynamics textbooks present container fill-
ing and container emptying (often called bottle fill-
ing and bottle emptying) as the simplest examples
of unsteady-state, varying-inventory processes. If the pro-
cess is adiabatic and the contents of the container are well
mixed, then the differential mass and energy balances can be
combined and integrated, leading to closed-form solutions.
The classic problems are of the form:
Problem 1 An evacuated, rigid, adiabatic container
is connected to a compressed air line at a pressure
of 738 kPa and a temperature of 220C. The
connecting valve is opened, and the air flows in
until the container pressure is 738 kPa. What is the
final temperature in the container?
Problem 2 A rigid, adiabatic container contains air
at 641 kPa and 290C. Its valve is opened and it
exhausts to the atmosphere at 86 kPa (at Salt Lake
City, 1460 m above sea level). When the pressure
in the container is the same as atmospheric
pressure, what is the temperature in the container?

GENERAL THEORY
FOR ADIABATIC FILLING AND EMPTYING
For a rigid container with flow of matter in or out, the
energy balance on the contents, ignoring kinetic, gravita-
tional, electrostatic, and magnetic energies,"1 is

d(mu)system hindmin -houtdmout +dQ (1)


and the corresponding mass balance is
dmsystem =dm in -dout (2)
For a container filling from some reservoir (e.g., the atmo-
sphere into an evacuated container, or a large compressed-
air or steam line into a container), we may assume perfect
internal mixing and that hin is a constant, combine Eqs. (1)
and (2), and integrate to

(mu)systemfinal (mu)system,initial = hin finall initial )+ AQ (3)
If the mixing is not perfect, then the specific properties
shown in Eq. (3) and throughout this paper should be inter-
preted as mass-average values. For emptying (discharge,
blowdown), the simple integration leading to Eq. (3) is not
correct because the expansion work done by the fluid during
the emptying process causes the temperature, and hence hout,
to decrease during the process. If one uses an average value
of hout, one can then use this integration.'2' If the material
originally present in the system and flowing in or out is a


@ Copyright ChE Division of ASEE 1999


Chemical Engineering Education


Noel de Nevers is Professor of Chemical En-
gineering at the University of Utah, where he
has been on the faculty since 1963. He has
written texts on Fluid Mechanics for Chemical
Engineers and Air Pollution Control Engineer-
ing. In addition, he works in thermodynamics
and accident investigations.










perfect gas, we can substitute and change Eqs. (1) and (3) to

d(mCvT)system = CpTindmin -CpToutdmout +dQ (4)

For filling, we can integrate Eq. (4) to

(mCvT)system.final -(mCvT)ystem.initial
=CpTin(mfinal- minitial)+AQ (5)

If m,intia and dQ are both zero (initially empty container
and adiabatic process), then for filling, Eq. (5) becomes

Tsystem.final = kTn (6)
If the container is not originally empty, but contains a gas
with the same value of k as the gas that enters, AQ = 0 and
Initial = Tin, then the solution is

Tsystem,final = kTin Tinitia(k 1) initial (7)
final
For adiabatic emptying, the perfect mixing assumption
allows us to set the system temperature equal to the outflow
temperature and integrate Eq. (4) to


Tfinal final
initial m initial )


Replacement of the m terms by their ideal-gas-law values
produces

(k-1
Final Ptinal k~ (9)
Initial \Pinitial )
which is the relation for an isentropic expansion of an ideal
gas.

PROBLEM SOLUTION


Using Eqs. (6) and (9), with the adiabatic assumption and
the further assumption that air is an ideal gas with k = 1.40,
we can solve Problems 1 and 2, finding

Tsystem,final = kTin =(1.40)(295 K)= 413K = 140C
and

k-l (0.4)
Tfinal Pfinal k ) 86 kpa ) =0563
=0.563
Tintial Pinitial 641kpa
Tfinal = 0.563 Timntia = (0.563)(302 K)= 170 K = -103C
But experimental results131 do not agree, even approxi-
mately, with these simple theories. This disagreement is
explained[3' as being due to significant heat transfer. This
appears startling, because the filling and emptying experi-
ments are normally finished in a few seconds. But, as shown


below, it is correct.
General Theory for Non-Adiabatic filling If we now
allow for heat transfer, and replace dQ by hA(T-T,,na,,dings)dt
and keep the ideal gas assumption, then Eq. (4) becomes

d(mCvT)system
CpTindmin -CpToutdmout +hA(Tsystem -Tsurroundings)dt (10)

For filling, we drop the outflow term and rearrange to

d(mT)system hA
d(mT) inkTin h (Tsystem Tsurroundings) (11)

To save writing, we define ca=hA/Cv and drop the sub-
script on Tystem. If we assume a constant filling rate, we can
replace the instantaneous system mass with

m = minimal + mint (12)
and rearrange to

initiall + tmint) = fin(kTin -T) (T- T oundings) (13)

Separating variables, integrating from start to finish, and
assuming that the Tsurroundngs is constant and equal to the
initial temperature of the gas in the container, we find


-T+ ri in kTin + Tsurroundings

mm i kTin + 1,-I Tsurroundings
min+ 0C thin +a )

initial +mint min (14)
m initial J

If, instead of assuming a constant inflow rate, we assume
an inflow rate that declines linearly with time, we can write

system = initial + f (a bt)dt = initial + at 0.5 bt2 (15)

where a and b are data-fitting constants. Substituting this for
system in Eq. (11) and rearranging, we find

dT (a- bt)(kTin -T)- C(T-Ts, urodings)
dt initial +at-0.5bt

which can be integrated numerically.

General Theory for Non-Adiabatic Emptying For con-
tainer emptying, Eq. (11) becomes

d(mT)system
d -sst -mi out kT (T Turroundings) (17)

For an assumed constant flow rate, the equivalent of Eq. (13)
is


Winter 1999











initiall houtt)- = -rmoutT(k- 1)- o(T Tsurroundings) (18) _
dt
and the equivalent of Eq. (14) is
70
[-hout (k- I)-a ]T + Tsrroundings 60
[-rout (k -1)- a]Tinitial + aTsurroundings 50

S (k -)+ 1 40
m initial til Mout Q
minitial-moutt ou (19)
initial ) 30
E
20
For the assumed linearly decreasing mass flow rate out, 20
the equivalent of Eq. (16) is 10

dT -(a bt)(k 1)T (T Tsurroundings) 0 20 40 60 80 00
(20) 0 20 40 60 80 100
dt initial -at+0.5bt2
dt at +0.5bt2 Time from opening of valve, s
which is also suitable for numerical integration.
Figure 2. Temperature-time plot for filling the container
EXPERIMENTAL TESTS from the laboratory compressed air main at 738 kPa. The
The experimental apparatus,[4] sketched in Figure 1, con- pressure reached that value in = 5.3 s. The temperature
rose from 22 to 74.50C, reaching the peak at 6 s. The air
sisted of an ordinary 0.027 m3 propane storage container, a inlet valve was closed at 5 s, and the pressure allowed to
pressure transducer, thermocouples, and a data logger. The decrease as the air in the container cooled. The metal
container was evacuated, filled from a compressed air main, walls of the tank rose in temperature from 26 to 29.30C.
and then emptied to the atmosphere several times, with After the peak temperature, the gas cooled slowly toward
several sizes of thermocouples (see discussion of thermo- room temperature. For the adiabatic assumption, the peak
couple measuring lag below), and with inlet and outlet flow temperature, calculated in the text, was 140C.
restrictors, in some cases, to slow the flow. Figures 2 and 3
show the temperature measurements for typical filling and
emptying experiments. The measured maximum and mini- 30
mum temperatures are far from those computed above.
25 -
We can estimate the heat transfer coefficient between the
20

To atmosphere 15
Pressure and 10
temperature TPI
indicators a
connected to Quarter turn E
data logger ball valves 0
I-
-5
-10
0 20 40 60 80 100 120

~ Time from opening of valve, s
Surface mount From compressed
thermocouple 5.7 Gallon (0.027 m3) air supply line
wth le 5.7 Gallon (0.027 m3) air supply lne Figure 3. Temperature-time plot for emptying the con-
insulation propane container trainer to the atmosphere (at 86 kPa) from an initial pres-
sure of 641 kPa. The pressure reached atmospheric in =
1.7 s. The temperature fell from 29 to -4 C, reaching its
minimum at 9.7 s. The metal walls of the tank fell in
temperature from 28.8 to 27.50C. After the minimum tem-
Figure 1. Flow and instrumentation diagram of the experi- perature, the gas warmed slowly toward room temperature
mental apparatus. The data logger records temperatures with the exit valve open. For the adiabatic assumption, the
and pressures at 1/3-second intervals, minimum temperature, calculated in the text, was -103 C.


Chemical Engineering Education









air in the containers and the container walls if we assume
that the thermocouple lag was small compared to 100 s and
that in each process the gas underwent a step change in
temperature and was then cooled or heated by simple con-
vective heat transfer with a constant-temperature container
wall. With these assumptions, the gas temperature is given
by
TTwall (exp- hA (2
ep tl (21)
After step Twall '
suggesting that a plot of the In of the temperature ratio at the
left of Eq. (21) vs. t should form a straight line, from which h
could be estimated. Figures 4 and 5 show such plots; from
the slopes, one may infer the values of the heat-transfer
coefficients. The choice of Tater tep is arbitrary, made to force
the straight lines through 1.0 on the ordinate. Changing these
values moves the curves up and down without changing
their slopes.






5o.i


Straight line slope = 0.0344/s


0.01 I I
0 20 40 60 80 100
Time from opening of valve, s

Figure 4. Replot of temperature-time data from Figure 2,
in the form suggested by Eq. (21).


From these tests, one can estimate the heat-transfer coeffi-
cients. For example, for the filling test (Figures 2 and 4),
with C = Cv =2.5R

(166g)(2.5)(8.314) J )r 0.0344
h 2mol K s Jmo10.49 W
0.390m2 29g m2K

and for emptying (Figures 3 and 5), with C = Cp = 3.5 R
W
h=0.91
m2K
The surprisingly large difference is largely due to the
difference in air densities (due to differing pressures) for the
two cases. The heat-transfer coefficients, estimated from a
flat-wall natural convection correlation151

Nu = 0.0210(GrPr)04 (22)
using average values of the gas density and the temperature
differences, are 11.9 and 1.7 W/(m2K). The first is close to
the value calculated from the cooling curve and the second
about twice the value calculated from the warming curve.
If we assume that the processes were practically two-step,
with a quick adiabatic process followed by a slow transfer of
heat to or from the walls of the container, then by energy
balance, we can compute that


ATcontainer walls


[mC(Tadiabatic Tinal )]gas in container
(mC)container walls


For the filling experiment, with C = Cv, the value are


ATcontainer walls


{182 g[(25)314) g (140C 29C)}
29[(7.03 gK)(046)

(7.03 kg)(0.46) kJ
kgK


=4.48C


and the corresponding value for emptying, with C = C,, is
-0.940C. The measured values are 3.3 and -1.30C.

THERMOCOUPLE LAG

A major part of the difference between the steep parts of
the temperature curves on Figures 2 and 4 and the values
calculated from adiabatic behavior, or those computed by
Eqs. (14), (16), (19), and (20), is due to thermocouple lag.
This is normally characterized161 in terms of the first-order
time constant of the thermocouple. If we assume that the gas
in the container undergoes a step increase or decrease in
temperature, followed by a first-order decay toward the sur-
rounding temperature, and that the thermocouple responds
as described in Reference 6, then the equation for the ther-
mocouple reading will be


Winter 1999


Figure 5. Replot of temperature-time data from Figure 3 in
the form suggested by Eq. (21).











dTth people b.Twail + (Tend of step -Twall) exp(-at) T (24)
dt
where a and b are the time constants of the cooling or
heating air in the container and of the thermocouple. The
integrated form is

Tthermocouple Twall a -t- e-t]
Tedofstep T = b-a [exp(-at)-exp(-bt)j (25)
Tendofstep -Twall b-a

with the peak value of Tthemocouple occurring at

t=n() (26)
b-a
with maximum value


Tthermocouple Twall
Tendofstep- wall Jmaximum


( b a
Y lb-a)
-IbJ


Figure 6 shows a comparison of the reported temperatures
for two sizes of thermocouple for identical filling experi-
ments. As expected, the smaller thermocouple reports a higher
peak temperature and reaches it sooner. Table 1 shows the
comparison of the time-to-peak reported T and the esti-
mated value of Tend of step calculated from Eqs. (26) and
(27). The computed and observed times-to-peak reported
temperature are in good agreement, but the computed
maximum temperatures are far too high, indicating that
after a few seconds, the two-first-order-processes-in-se-
ries model works well, but its extrapolation to t=0 does
not. If the contained gas temperatures were not changing
rapidly due to heat transfer, this thermocouple lag would
pose no problem.

ESTIMATING MAXIMUM TEMPERATURES
To estimate the maximum temperatures from Eqs. (14),
(16), (19), and (20), the mass flow rates were computed by


0 20 40 60 80 100
Time from opening of valve, s

Figure 6. Reported temperatures for two identical filling
experiments with different size thermocouples. The 0.51 -
mm diameter thermocouple reached its peak reading at
3.3 s; the 1.59-mm diameter thermocouple reached its at
7.67 s.
30


Figure 7. Computed departure of the tank temperature for
tank filling from the calculated adiabatic temperature for
the same conditions. Here, miti,1=3.5 g.


0.5 1 1.5
Time from opening of valve, s


Figure 8. Computed departure of the tank temperature for
tank emptying from the calculated adiabatic temperature
for the same conditions.


Chemical Engineering Education


TABLE 1
Applications of Eqs. (26) and (27) with the Time Constant
of Cooling the Container Assumed = 0.034/s
(See Figure 4)

0.51-mm diameter 1.59-mm diameter
Thermocouple Thermocouple

Reported time constant,16] s 1/4 = 0.25 1/9= 1.11
Observed time to peak reported T,s 3.3 7.67
Calculated time to peak T, by Eq. (26), s 3.2 9.3
Right-hand side of Eq. (27) 0.027 0.099
Calculated Td of sp from Eq. (27), C 2000 500
end of step


20
0




10
. 5
a
U



- 5
-o
I-


0 0.5 1 1.5 2 2.5 3
Time from opening of valve, s


3.5 4










differencing the calculated mass in the container at each 1/3-
second measuring interval. This has the drawback that it
relies on the thermocouple reading, which is known to lag
the true temperature. With this caveat, the flow rates corre-
sponding to Figures 2 and 3 were estimated as

m= 100g 28.6 t; 0 s s
and

mi = 239 --171- t; 0_ S s-
For both filling and emptying, the heat-transfer coefficient
was estimated from Eq. (22) and assumed constant. At the
average density and temperature difference between wall
and gas, the estimated values were 8.9 and 9.0 W/(m2K) for
filling and emptying. Figures 7 and 8 show the calculated
departures from the corresponding adiabatic solutions (Eqs.
7 and 8). From them we see that in both cases the major
departure occurs at the boundary of the process at which the
mass of air in the container is least (the start for filling, the
finish for emptying); this is the natural consequence of di-
viding a calculated heat flow that is assumed independent of
the mass by a small mass rather than a large one. Because of
the strength of the assumptions and the thermocouple lag
problem, these figures should be seen as order-of-magni-
tude. Nonetheless, they make clear that even for these fast
processes, with plausible heat-transfer coefficients, the cal-
culated temperatures are substantially different from the com-
puted adiabatic temperatures.

APPLICATION TO LARGER TANKS
Equations (14) and (19) show that for the constant mass-
flow-rate in or out simplification, and for Tn=Tiniia=Tsurroundings,
the T-t behavior depends only on the two dimensionless
groups m/m, and a/rh. Thus the experimental results shown
here should also be observed in any container for which
these parameters have the same values. The first can take on
any value, but the second is a function of container geom-
etry. If the initial conditions in the container and the reser-
voir conditions for filling and emptying are the same for two
tanks, then ih should be proportional to the cross-sectional
area of the inlet pipe. If the heat-transfer coefficient does not
change, then a is proportional to the surface area of the
container. If the ratio of the inlet pipe cross-sectional area to
the surface area of the container does not change, then the
second of these dimensionless groups should also not change
(or change much with changes in the heat-transfer coeffi-
cient). Thus, while the experiments reported here were all
performed in a 0.027 m3 container, they should be directly
applicable to larger tanks with the same dimension ratios.

CONCLUSIONS
The adiabatic, ideal-gas container filling and emptying
solutions have a traditional place in thermodynamics text-
Winter 1999


books because they are the simplest unsteady-state, varying-
inventory problems that can be solved in closed form. In
practice, it is impossible to conduct these processes without
heat-transfer-producing gas temperatures far different from
the adiabatic flow solutions, mostly because while the amount
of heat transferred is small, the mass of gas into which it is
transferred is also small. The effects of such heat transfer on
the temperature-time behavior of such processes can be esti-
mated with at least order-of-magnitude accuracy.
Thermocouple lag adds to the effect of the heat transfer,
further increasing the difference between the observed tem-
perature extremes and the values calculated for adiabatic
filling and emptying.
A different version of this problem and experiment ap-
peared while this paper was in press.171

NOMENCLATURE
A area
a time constant of cooling or heating air in the container
a,b constants in data-fitting equations
b time constant of thermocouple
C heat capacity
C heat capacity at constant pressure
Cv heat capacity at constant volume
Gr Grashof Number
h heat-transfer coefficient
h specific enthalpy
k Cp/C
m mass
m mass flow rate
Nu Nusselt number
P pressure
Pr Prandtl number
Q heat quantity
R universal gas constant
T temperature
t time
u specific internal energy
a hA/Cv

REFERENCES

1. de Nevers, N., Fluid Mechanics for Chemical Engineering,
2nd ed., McGraw-Hill, New York, NY, 109 (1991)
2. Wisniak, J. "Discharge of Vessels: Thermodynamic Analy-
sis," J. Chem. Ed., 74, 301 (1997)
3. Ryan, J.T., R.K. Wood, and P.J. Crickmore, "An Inexpensive
and Quick Fluid Mechanics Experiment," Chem. Eng. Ed.,
27, 140(1993)
4. Cutler, S., J. Feichko, and R. Waldron, "Bottle Filling and
Emptying Experiments," Senior Process Engineering Labo-
ratory Reports, Department of Chemical and Fuels Engi-
neering, University of Utah (1997)
5. Kreith, F., Principles of Heat Transfer, 3rd ed., IEP, New
York, 395 (1973)
6. Omega Engineering Co., Catalog 29, Page Z-43, Stamford,
CT (1995)
7. Forrester, S.E., and G.M. Evans, "The Importance of Sys-
tem Selection on Compressible Flow Analysis: Filling Ves-
sels," Chem. Eng. Ed., 32, 308 (1998) J











Random Thoughts...





FAQS





RICHARD M. FIELDER, REBECCA BRENT
North Carolina State University Raleigh, NC 27695


A t the teaching workshops we give, we propose a
variety of instructional methods that deviate from
traditional teaching practice. We recommend, for
example, that instructors break up their lectures at frequent
intervals with brief individual or small group exercises. We
suggest using formal cooperative learning, in which students
work on assignments in instructor-formed teams under con-
ditions structured to assure individual accountability for all
of the assigned material. We caution against giving tests that
only the best students in the class have time to finish, and we
argue strongly against curving grades.
Predictably, critical questions are raised about these rec-
ommendations and others we offer. In a series of columns
beginning with this one, we want to review some of the most
frequently asked questions (FAQs) and our responses. We
have two reasons for doing this. First, the suggestions we
offer at the workshops are far from unique with us: they are
being made with increasing frequency by educational re-
searchers, national study commissions, employers of engi-
neering graduates, and accrediting bodies like ABET. If you
have not already been exposed to them, you almost certainly

Richard M. Felder is Hoechst Celanese Profes-
sor of Chemical Engineering at North Carolina
State University. He received his BChE from City
College of CUNY and his PhD from Princeton. He
has presented courses on chemical engineering
principles, reactor design, process optimization,
and effective teaching to various American and
foreign industries and institutions. He is coauthor
of the text Elementary Principles of Chemical Pro-
cesses (Wiley, 1986).
Rebecca Brent is an education consultant spe-
cializing in faculty development for effective uni-
versity teaching, classroom and computer-based
simulations in teacher education, and K-12 staff
development in language arts and classroom
management. She co-directs the SUCCEED Coa-
lition faculty development program and has pub-
lished articles on a variety of topics including
writing in undergraduate courses, cooperative
learning, public school reform, and effective uni-
versity teaching.


will be before long, and some of our responses may be
helpful as you consider the ideas being advanced. Our sec-
ond objective is to offer those of you who are already using
the new methods some answers to give your colleagues,
administrators, and students, who are certain to raise the
same questions with you.
Here, then, is our top ten list of questions frequently asked
at teaching workshops.


1. Is there any real evidence that these methods work?

2. I have a lot of material to get through in a semester.
Can I use these methods and still have time to
cover my syllabus?

3. I teach a class of 175 students in a fixed-seat
auditorium.
Will these methods work in large classes?

4. I'm teaching a course by distance education.
How can I get students active when I'm not in the
same room with them?

5. I tried putting students to work in groups, but some
of them hated it and one complained to my depart-
ment head.
Why are some students so hostile to cooperative
learning and what am I supposed to do about the
hostility?

6. Many of my students are (a) unmotivated, (b) self-
centered, (c) apathetic, (d) lazy, (e) materialistic, (f)
unprepared, (g) unable to do high school math, (h)
unable to write, (i) unable to read, (j) spoiled rotten.
(Pick any subset.)
How can you teach people who don't have the right
background or the willingness to work or even the


Copyright ChE Division of ASEE 1999


Chemical Engineering Education










desire to learn?

7. Engineers constantly have to face deadlines.
What's wrong with giving tests that only the best
students have time to finish?

8. What difference does it make if my test averages are
in the 50's, since I'm going to curve in the end?

9. My department head says that we can't count
teaching too much in promotion and tenure deci-
sions because we don't know how to evaluate
teaching.
Is there a meaningful way to evaluate teaching?

10. The people who go to teaching workshops are
mostly excellent teachers-the ones who most need
to change wouldn't go to a teaching workshop at
gunpoint.
How can I persuade my traditional colleagues to
do some of the nontraditional things you're
recommending?


The workshop participants who ask these questions are
doing what they have been trained to do as scientists and
engineers and educated people, which is to ask for hard
evidence before changing the way they've always done things.
We applaud them for asking. In this column we'll offer an
answer to the first question, and subsequent columns will
deal with the others.
Q: Is there any REAL evidence that these nontradi-
tional methods work?
A: Tons of it.
Cognitive and educational scientists have learned a great
deal about learning in recent years. The near-unanimous
consensus is that we learn mainly by doing things and re-
flecting on the outcomes, taking in relatively little of what
we just see and hear (e.g., in lectures) and retaining even
less. Countless studies have compared the academic perfor-
mance and attitudes of students taught using active and
cooperative methods with the performance and attitudes of
students taught more traditionally. The evidence for the ef-
fectiveness of the nontraditional methods is overwhelm-
ing. (Specific references will be cited shortly.)
Unfortunately, most professors have never seen a mono-
graph, paper, or seminar on research into teaching and learn-
ing and would be hard pressed to name a journal or confer-
ence where such research might show up. When the "Prove
it!" card is played at our workshops (and even if it isn't), we


therefore urge our questioners not to take our word for
anything we say but to approach the matter scientifically and
check the literature. We point them to a series of three papers
in Chemical Engineering Education written by Jim Haile,111
which collectively provide the best summary we've ever
seen of what cognitive science has discovered about the
learning process and the implications of this knowledge for
teaching. We introduce them to the classic Teaching Tips,[2
in which Wilbert McKeachie offers an abundance of practi-
cal suggestions about every aspect of college teaching along
with citations of the research that backs up the suggestions.
We tell them about What Matters in College, 341 Alexander
Astin's monumental study of nearly 25,000 students at over
300 institutions that powerfully demonstrates the deficien-
cies of the traditional instructional model. We cite refer-
ences on cooperative learning (e.g., Johnson, Johnson, and
Smith[51) that in turn cite hundreds of research studies attest-
ing to the effectiveness of this approach, and we discuss the
results of a longitudinal study one of us carried out of the
effectiveness of cooperative learning in chemical engineer-
ing education.[671 "Browse these references," we urge. "Then
decide whether the research and the methods we're advocat-
ing are worthy of serious consideration."
More to come.

REFERENCES
1. Haile, J.M., "Toward Technical Understanding": (a) "Part 1.
Brain Structure and Function," Chem. Engr. Education,
31(3), 152-157 (1997); (b) "Part 2. Elementary Levels,"
Chem. Engr. Education, 31(4), 214-219 (1997); (c) "Part 3.
Advanced Levels," Chem. Engr. Education, 32(1), 30-39
(1998)
2. McKeachie, W.J.,Teaching Tips: Strategies, Research, and
Theory for College and University Teachers, 10th Edition.
Boston, Houghton Mifflin Co. (1999)
3. Astin, A.W., What Matters in College, San Francisco, Jossey-
Bass (1993)
4. Felder, R.M., "What Matters in College," Chemical Engi-
neering Education, 27(4), 194-195 (1993)
(View at < http:/ / www2.ncsu.edu /unity / lockers /users/fl
fielder /public /Columns/Astin. html >.)
5. Johnson, D.W., R.T. Johnson, and K.A. Smith, Active Learn-
ing: Cooperation in the College Classroom, 2nd Edition.
Edina, MN, Interaction Book Co. (1998)
6. Felder, R.M., "A Longitudinal Study of Engineering Stu-
dent Performance and Retention. IV. Instructional Methods
and Student Responses to Them," J. Engr. Education, 84
(4), 361-367 (1995)
(View at < http:/ /www2.ncsu.edu/unity/lockers/users/f/
fielder public Papers /long4.html >.)
7. Felder, R.M., G.N. Felder, and E.J. Dietz, "A Longitudinal
Study of Engineering Student Performance and Retention:
V. Comparisons with Traditionally-Taught Students," J.
Engr. Education, 87(4), 469-480 (1998)
(View at < http://www2.ncsu.edu/unity/lockers/users/f/
fielder public Papers /long5.html >.) D


All of the Random Thoughts columns are now available on the World Wide Web at
http://www2.ncsu.edu/effectiveteaching/ and at http://che.ufl.edu/-cee/

Winter 1999










,f lIaboratory


EVALUATION OF

COMPUTER-SIMULATION EXPERIMENTS

IN A SENIOR-LEVEL

CAPSTONE ChE COURSE



SCOTT R. WHITE, GEORGE M. BODNER
Purdue University West Lafayette, IN 47907-1393


In 1986, the School of Chemical Engineering at Purdue
University began a revision of its senior-level capstone
laboratory courses, including the development of a se-
ries of computer-simulation experiments described else-
where.[171 For each computer simulation, the students are
given a budget (i.e., $35,000) that is the amount they can
spend on experimental runs, wages, and consultation fees.
The computer also keeps track of the "virtual" time the
students use for each run and charges extra for work that has
to be done on weekends.
This paper describes the results of an evaluation of the
effect of using these simulations. It is based on three as-
sumptions:
When we change what we teach, or how we teach, we change
what the students learn.
A systematic evaluation should be done whenever major
changes are made in an established curriculum.
Systematic evaluations should look behind the facade of
answers to the questions, "Do the students like it?" toward
deeper questions such as "What will students learn that they
were not learning before?" and "If we could provide
students with a voice to express their opinions and concerns,
what changes would they recommend?"

The basic research question behind this study was: "How
do the students' experiences with computer simulations com-
pare with their experiences with traditional laboratory ex-
periments?" Corollary research questions included: "What
did the students perceive as a valuable experience in both
laboratory formats?"; "How did the students' decision-mak-
ing processes and other group-related interactions differ be-
tween the two formats?"; "What do the students believe
makes the computer-simulation experiment a legitimate ex-


ercise to include in the chemical engineering curriculum?"

DEVELOPING EVALUATION METHODS
The study was based on a collaboration between members
of a chemical education research groupl18 and faculty and
staff from the School of Chemical Engineering who had
developed and implemented the computer simulations. We
began by scrutinizing a list of questions generated by Profes-
sor R. G. Squires and Dr. S. Jayakumar for use in a quantita-
tive study of student attitudes toward the simulations. Some
of the questions were retained and others were modified to
make them either less complex or less "leading." The result
of this review was a 15-item five-point Likert-scale ques-
tionnaire that included space for students to write additional
comments and/or suggestions. The questionnaire was given
to the students after they had completed both a traditional


Scott R. White is a PhD graduate student in
Science Education at Purdue University. He re-
ceived his BS in Chemistry and Secondary Edu-
cation Certification in 1992 from Harding Univer-
sity. He received his MS in Chemistry from Purdue
University in 1996 with G.M. Bodner. His research
interests are in teaching and learning in science
and curriculum reform.


George M. Bodner is Professor of Chemistry
and Education at Purdue. He received his BS in
Chemistry from the State University at New York,
Buffalo (1969) and his PhD in inorganic and or-
ganic chemistry from Indiana University (1972).
His research interests are learning theory, over-
coming barriers to curriculum reform, and under-
standing the conditions for appropriate use of
technology in teaching and learning chemistry.


Copyright ChE Division of ASEE 1999


Chemical Engineering Education












TABLE 1
Survey Percentage Responses*


Agree Neural Disa g


1. 1 like using computer simulations. 91 3 6 4.2
2. When using the computer simulation, I worried that my 22 16 62 2.4
data would be lost, or that the program would fail.
3. Time and budget constraints made the computer 86 3 ft 4.2
experiments more realistic.
4. The conventional lab experiments worked better that the 5 11 84 1.9
computer experiments.
5. The video tour of the plant added little to the value of the 30 30 40 3.0
computer experiment.
6. It was easy to learn and operate the computer simulanon 95 0 -5 4.3
". Computer-smulation experiments intimidate me. 6 8 86 1.8
8. The speed of data acquisition in the computer experiments 6 5 89 1.7
makes me uneasy.


9. Computer experiments allowed me to focus on the prin- 82
ciples to be learned rather than on the details of
operating a particular piece of equipment.
10. Computer experiments are more interesting that conven- 40
tional experiments.
11. One disadvantageof computer experiments is that I do 60
not gain experinece with the real plant equipment.
12. T would like to see more computer-simulated experiments 73
in the chermcal engineering curriculum.
13 I would rather work on a computer simulation because it 27
is less hazardous than a conventional experiment
14 Conventional experiments give me a beer sense of the 54
kinds of problems likely to be encountered in industry.
15 My group cooperated better dunng the conventional lab 13
experiments.
16. The design problem imposed by the computer simulation 30
is not as challenging as those encountered during conven-
tonal experiments.
17 A higher percentage of our time was spent planning the 86
design of computer experiments.
18. The computer simulaton allowed me to study problems 78
that are more complex and realistic than the conventional
experiments.
19. Computer simulations allow me to make more effective 95
use of time by reducing the amount of time needed to
run experiments.
20. The conventional lab experiments were easier to learn 6
and operate than the computer experiment.
21 Computer simulations are a good way to learn new 79
processes and concepts.
22 Computer simulations work better than the conventional 60
experiments.
23. Computer simulations are more likely to "work" than 81
conventional experiments.
24. Overall, I think the presentcombination of computer 57
simulations and conventional experiments is appropriate.


10 8 4.2


38 22 3.3


24 16

14 13

27 46

22 24

43 44

22 49


3 II 4.1

19 3 4.0


5 0 4.4


8 86 1.9

22 0 4.1


5 3.8


14 5 4.1

14 29 3.4


"This table summarizes the results of two semesters. We combined the "Strongly
Agree" and 'Agree' responses into one category-"Agree. "Strongly Disagree" and
"Disagree" have been combined into the "Disagree" category. The "Undecided"
responses are indicated as "Neutral (N)."

Winter 1999


etI m Starement


experiment and a computer-simulation ex-
periment. Results of this survey for stu-
dents from two semesters are summarized
in Table 1.
The authors developed a qualitative com-
ponent of the evaluation191 based on struc-
tured interviews with individual students or
with groups of students, observations and
field notes collected in the laboratories, writ-
ten comments from the surveys described
above, and interactions with the students
in the labs. As those familiar with quali-
tative techniques might expect, the quali-
tative component provided the "richest"
source of data for this study.
Collection of qualitative data began with
the researcher sitting in a corer of the tra-
ditional lab, taking field notes as he ob-
served what was happening. The students
would frequently start conversations with
the researcher, asking what he was doing
there and relating what they thought about
the experiment they were doing or what
they thought or had heard about the com-
puter-simulation experiments. Frequently the
students would physically point out things
that were working or not working with their
traditional experiments, which helped the
researcher gain an understanding of the ex-
periments the students were performing.
As these interactions continued, the re-
searcher found it useful to switch from the
role of an objective observer sitting in a
corer of the room taking notes to that of a
participant-observer, listening to and talk-
ing with students while they worked. The
students also seemed more comfortable with
this approach. The result was an environ-
ment in which a good rapport was devel-
oped between the researcher and the stu-
dents prior to the structured interviews. This
approach also provided the researcher with
a set of experiences that allowed him to
prod the students' memories during the
subsequent interviews when they were
asked to compare the two different labo-
ratory formats.
Observations collected while students
were working in the computer lab did not
prove useful because most of the decision-
making process had already been accom-
plished during group meetings before the
students came to the lab and the students









were less likely to involve the researcher in their activities
while they worked with the computer. Insight into these
group meetings and the interactions between members of the
group was provided by the structured interviews, however.
The structured interviews were the core of the qualitative
evaluation methods. The researcher developed a list of ques-
tions that he wished to cover during the interviews, covering
many of the same topics as the Likert-scale surveys so that
the researcher could triangulate his conclusions from differ-
ent data sources.10] Using the structured topic list pro-
duced interviews that followed a similar pattern, but the
students had ample opportunity to bring up any subject
they felt appropriate.
The interviews were recorded, transcribed, and then ana-
lyzed using the method of inductive analysis."11 The analysis
consisted of reading the transcripts multiple times and con-
densing the students' comments to common and uncommon
categories by literally cutting and pasting together similar
comments obtained in different interviews.

QUANTITATIVE RESULTS AND DISCUSSION
The results of the Likert-scale survey indicated that the
students liked using the simulations (91%; Q1); found the
simulations easy to learn and operate (95%; Q6); reported
that the computer simulations did not intimidate them (86%;
Q7); would like to see more of them (73%; Q12); believed
that the computer simulations allowed them to study more
complex and realistic problems (78%; Q18); valued the bud-
getary constraints included with the simulations, which made
the simulations more realistic (86%; Q3); and believed that
they spent a higher percentage of their time planning the
design of the computer simulation (86%; Q17), which sug-
gests that the simulations provide the students with an expe-
rience that is different from the traditional lab. The students
liked the simulations for a variety of reasons, including the
fact that they were more likely to work than the traditional
experiments (81%; Q23), thus giving the students reason-
able and workable data.
The computer simulations were very different from tradi-
tional labs because of the speed with which data could be
acquired. This did not bother the students or make them feel
uneasy about the computer experiment (89%; Q8). In fact,
they felt that this made more efficient use of their time (95%;
Q19). The students felt the simulations allowed them to
focus on the principles involved in an experiment (82%; Q9)
and therefore were a good way to learn new processes and
concepts (79%; Q21). But a majority (54%; Q14) of the
students felt that the traditional experiments gave them a
better sense of the problems likely to be encountered in
industry. Thus it is not surprising that a majority (57%; Q24)
felt that the present combination of computer and simulation
experiments was appropriate.
The quantitative results produced a sense of conflict, or


dualism, in the students' opinions. They simultaneously be-
lieved the computer simulations are a good instructional
technique that helped them better focus on the principles
they were expected to apply, and at the same time that the
traditional experiments gave them a better sense of the
problems they might encounter in industry. The source of
this dualism cannot be extracted from the results of a
Likert-scale survey, but they can be obtained by triangu-
lating this data source with the results of qualitative
research techniques.
As we will see, the students simultaneously regarded the
computer simulations as both "good" and "bad." They are
good because they allowed students to tackle more complex
problems in which they were compelled to proceed with
realistic budgetary and time constraints, and because these
experiments were more likely to "work," providing the stu-
dents with data that allowed them to complete a realistic
scale-up. The simulations are "bad" because they are not
real; they cannot fail in the same way a traditional experi-
ment would fail. Even though the students tended to value
the ability to focus on important conceptual engineering
issues in the simulation experiments, they recognized
that this "ability" has little to do with the world in which
they actually live.

QUALITATIVE RESULTS AND DISCUSSION
Twelve students were interviewed after they had com-
pleted one experiment of each type. The theoretical frame-
work for this portion of the study falls within the domain of
hermeneutics1121 in the sense that we are trying to give stu-
dents the opportunity to be heard, to have a "voice," through
interpretations of the meanings of their statements and ac-
tions. The interviews were used to probe more deeply into
the students' experiences, opinions, and beliefs about tradi-
tional versus computer-simulation experiments; to probe how
students constructed the knowledge they gained from doing
the lab experiments; to examine how they perceived com-
puter-simulation experiments (e.g., as just one long equation
to be worked out with data generated by the computer or as a
chance to do meaningful engineering work similar to that
done in industry); to explore their opinions on whether the
computer simulations were more (or less) realistic than tra-
ditional experiments; to discern whether the simulations re-
quire a particular teaching style from the instructor; and to
determine the aspects of the computer simulation that make
it more (or less) difficult than the traditional experiments.
In some ways, the students felt the computer simulation
was more realistic, and perhaps more difficult, than their
other experiments. (In the following vignettes, "I" stands for
the interviewer and the names are nicknames given to pro-
tect the students' identities.)
I: You were talking about the computer simulation being
more "in-depth." What did you mean by that?
Chemical Engineering Education










Andy: Instead of dealing with the unit, you dealt with more of
what you'd deal with in the real plant...the computer inter-
faced you to multiple types of equipment and more "real"
equipment than you would use in industry rather than just the
small glass tube that we used for the cation exchange. And I
thought that was better because you get more of a full view of
the operation rather than just one small aspect of it.
Jody: In addition to that, too, we had a budget that we had to
follow. Which is gonna be true in real life once we graduate
and do what we need to do to get data and stuff like that.
The time and budgetary constraints imposed on the com-
puter-simulation experiments had the tendency to change
the students' decision-making process by forcing them to
reflect on their decisions before taking actions, as illus-
trated by comments made by Adam and Don, who were
in separate groups.
Adam: It made it more "real-world" I guess. Before, on the
other experiments, if you wanted to ask the professor a ques-
tion, we'd just go up and ask; even if it was just a stupid
question. Now if we wanted to talk to the professor it would
cost us $500 for a consulting fee. It made you stop and think
about it instead of just running up and asking the professor
when you could have figured it out yourself if you'djust have
thought about it.
Don: It was good to have a budget. If there was no actual
planning involved, with no budget, we would just have run it
for hours and hours and had stacks of paper for results. We
wouldn't have thought about what we were doing.
These comments are echoed by the results of the survey,
which showed that the majority of the students felt that use
of budget and time constraints made the simulation more
realistic than the traditional laboratory experiments. Darrin
and Laura found the realism introduced by the budget/time
constraints intimidating.
I: Let me ask you about the computer simulation. What did
you think of it when you first saw it?
Darrin: Heh! Intimidating.
I: How?
Darrin: Well, even though we sat through a whole lecture, I
felt that I really didn't know where to begin, and really I was
ready to get another apparatus experimental problem. With
this Compp. simulation] I had no idea how to start. I was
afraid that I was going to make a mistake.... And plus there's
this thing that if you ask a question it would cost you like $500
or something. [consultation fee] So you're kinda tentative.
Laura provided insight into why her group felt intimidated
by the computer experiment when she responded to a ques-
tion that asked for her impression of the computer simula-
tion.
Laura: I was scared because it wasn't like any of our other
labs were, even if you, like, totally get bad data you don't
have anything to lose. You can still write up your report and
say that your results are no good. But on this lab [computer
simulation] you have to find your constants.


Winter 1999


The problem was simple-there was no place for the stu-
dents to "hide." They could not gloss over or "fudge" poor
data collected during the computer experiment the way they
said they could when discussing traditional experiments.
Darrin and Laura's comments are not representative of the
perceptions of the group of students who completed the
computer experiment during the evaluation, but their com-
ments raise an important issue in evaluation. Historically,
evaluations of curriculum-reform projects have been based
on what we have called a "sports-mentality" approach.J3'
Statistical techniques, such as a t-test on the mean scores
of some measure of performance of students in experi-
mental versus control sections of the course, are used to
answer the questions "Is the new curriculum better or
worse than the old curriculum?"
Darrin and Laura's comments remind us that any substan-
tive change in curriculum will have both positive and nega-
tive effects. Some students will benefit, but others will be
hurt. Evaluative studies, such as this one, allow one to search
for both effects and then probe what additional changes
could be made to maximize the positive effect and minimize
the negative effect.
Darrin and Laura's interviews identified another source of
differences in students' perceptions of the computer simula-
tions-the amount of success the students felt they had en-
joyed. In general, the students who were interviewed felt
that they had enjoyed success with the computer simulation.
Darrin and Laura's group did not share this perspective,
however, as illustrated by the following comments:
I: What was the computer simulation supposed to do and
what did it really do as you look back on it now? What was it
supposed to represent?
Darrin: I think that it was supposed to represent a better way
of solving a large problem that we could never have
solved on a laboratory scale. With the amount of trials we
ran. .. it was supposed to demonstrate how much work you
could get done, ... how many trials you could get done on the
computer. But what it turned out to be was just trial and
error.
Laura: I think that what the ... simulation was doing was to
show us how we can use a computer to simulate something
and then to optimize conditions. And then apply them to an
actual plant or whatever. And, ... I guess it did it, ... I don't
know! I don't really know because I still don't really under-
stand how our values correlated to the actual running of the
simulation and the running of pilot plant. I don't really
think that I learned anything from it. I just learned to manipu-
late what we were trying to do ... I'm still a little unclear on
some things.
The interviews provided useful information about the stu-
dents' perception of the role of the computer in their re-
sponse to the question "Which experiment gave you the best
experience?"
Dallas: [The computer experiment] was the best. Granted,
37










that they wanted to try to give us some sort of real-life simula-
tion .... what it's like in real industry .... We were actually
able to get our numbers and do our scale-ups, and do our
actual engineering work without trying to mess around trying
to get something to work. Or trying to get some data or
making up data ... we were actually able to do engineering.
Don: [In the computer experiment] you could kind of estimate
sort of where what would happen under certain circum-
stances. Whereas in the experimental part of the regular lab,
you would get such confusing results. It was so difficult to try
to extrapolate that onto any large scale. We basically said,
"We'll just have to throw this out and we'll see what some-
body else did or make up something."
These comments reflect the common perception among
students that the simulations were more likely to work
than the traditional experiments; that they could acquire
feasible data from the computer simulation. But other
students questioned whether the computer simulation gave
them the best experience.
Adam: The best experience was probably with the water
cation exchange, just because we did a lot more research with
that to learn how to get the right data and stuff like that. The
computer simulation was interesting but pretty much all the
data was right there in front of you. The computer
simulation was pretty neat. But it was a lot of wasted time for
three people to sit there and do it, because only one person
could get on the computer and run it.
Ruth and Tina provided further insight into the computer
experiment.
Ruth: With the computer experiment all we did was calcula-
tions, ... with the spreadsheet and stuff That's all we did. The
whole lab was one big long calculation.
Tina: The whole lab was just finding numbers you had to put
in the simulation. You had to work through a bunch of equa-
tions.

EFFECT OF THE PROFESSOR'S TEACHING STYLE

Unlike surveys, which can only provide answers to ques-
tions that are explicitly stated, interviews often provide data
on topics or questions one might not have anticipated. Con-
sider the role of the professor's teaching style, for example.
This topic was not covered in the survey, but the interviews
showed that it had a significant impact on the students'
experiences. It was clear from the interview data that the
professor's "hands-on" teaching style during the planning ses-
sions had a direct impact on the students' perception of the
computer simulations and the success of these simulations.

CONCLUSIONS
The results of this study suggest that it would be a mistake
to ask which laboratory format is "better" for students. They
indicate that computer simulations and traditional experi-
ments have different roles in the curriculum because they
emphasize different aspects of engineering and require both


different levels and types of expertise.
Students who were frustrated with traditional lab equip-
ment seemed to enjoy "actually doing" the engineering re-
quired to tackle the complex problems provided by the com-
puter simulation. They did not have to worry about "making
up data or seeing what someone else did" when the
traditional lab failed. For these students, the computer simu-
lations were more "realistic" than the traditional lab that
gave results students described as ". I'd turn it in for a
grade but I certainly wouldn't buy it!" For other students, the
simulations were less "realistic" because they cannot fail,
the way a traditional lab fails.
This study provided insight into the role of the environ-
ment in which computer simulations are implemented. Our
results clearly indicated that budgetary and time constraints
played an important role in making the computer simula-
tions seem "realistic"-so realistic that a few students felt
intimidated by this aspect of the simulations.
This study also suggests that computer simulations, by
themselves, are not magic bullets that provide instructional
and pedagogical benefits for the students in the absence of a
human interaction between the students and the instructor.
They are best thought of in terms of being a tool for instruc-
tion rather than a replacement for the instructor.
The authors hope that this study leads others to recognize
the importance of asking the correct questions when evaluat-
ing curriculum reform projects, as required by ABET and
NSF, and the importance of collecting qualitative inter-
view data to both reinforce quantitative data collected in
anonymous surveys and to provide a deeper understand-
ing of the effect of curriculum changes on students' atti-
tudes and opinions.

ACKNOWLEDGMENTS
The authors wish to thank Professor Robert Squires and R.
S. Jayakumar of the School of Chemical Engineering for
their support throughout this evaluation study.

REFERENCES
1. Squires, R.G., G.V. Reklaitis, N.C. Yeh, J.F. Mosby, I.A.
Karimi, and P.K. Anderson, "Purdue-Industry Computer
Simulation Modules: The Amoco Resid Hydrotreater Pro-
cess," Chem. Eng. Ed., 25(2), 98 (1991)
2. Squires, R.G., P.K. Anderson, G.V. Reklaitis, S. Jayakumar,
and D.S. Carmichael, "Multimedia-Based Educational Ap-
plications of Computer Simulations of Chemical Engineer-
ing Processes," Comp. Appls. in Eng. Ed., 1(1), 25 (1992)
3. Jayakumar, S. R.G. Squires, G.V. Reklaitis, P.K. Anderson,
K.R. Graziani, and B.C. Choi, "The Use of Computer Simu-
lations in Engineering Capstone Courses: A Chemical Engi-
neering Example-The Mobil Catalytic Reforming Process
Simulation," Internat. J. of Eng. Ed., 19(3), 243 (1993)
4. Jayakumar, S., R.G. Squires, G.V. Reklaitis, P.K. Anderson,
and L.R. Partin, "Purdue-Industry Computer Simulation
Modules 2: The Eastman Chemical Reactive Distillation
Process," Chem. Eng. Ed., 27(2), 136 (1993)
Chemical Engineering Education










5. Jayakumar, S., R.G. Squires, G.V. Reklaitis, P.K. Anderson,
and B.K. Dietrich, "The Purdue-Dow Styrene-Butadiene
Polymerization Simulation," J. ofEng. Ed., 84(3), 271 (1995)
6. Jayakumar, S., R.G. Squires, G.V. Reklaitis, and K.S. Grassi,
"Simulating the Air Products Cryogenic Hydrogen Reactive
Cooling Process," Chem. Eng. Ed., 29(1), 26 (1995)
7. Squires, R.G., K. Kuriyan, S. Jayakumar, G.V. Reklaitis, M.
Evans, B. Morrato, and R. Gutwein, "The Procter and Gamble
Decaffeination Project: A Multimedia Instruction Module,"
Comp. Appls. in Eng. Ed., 4(4), 269 (1996)
8. Bodner, G.M., and J.D. Herron, "Completing the Program
with a Division of Chemical Education," J. of College Sci.
Teaching, 14(3), 179 (1984)


9. Patton, M.Q., Qualitative Evaluation and Research Meth-
ods, 2nd ed., Sage, Newbury Park (1990)
10. Ibid, pp. 187-189
11. Goetz, J.P., and M.D. LeCompte, Ethnography and Qualita-
tive Design in Educational Research, Academic Press, San
Diego, CA (1984)
12. Gadamer, H.G., Philosophical Hermeneutics, University of
California Press, Berkeley, CA (1976)
13. Bodner, G.M., "Overcoming the Sports Mentality Approach
to Program Evaluation: Action Research as a Metaphor for
Curriculum Evaluation," paper presented at the National
Association for Research in Science Teaching Meeting, At-
lanta, GA, April (1993) 0


M book review


Alternative Fuels
by S. Lee
Taylor & Francis, Bristol, PA; 485 pages, $83.95 (1996)

Reviewed by
Thomas R. Marrero
University of Missouri-Columbia

Knowledge of chemical processes is important in the de-
velopment of more environmentally friendly fuels because
of the implementation of stricter constraints on energy utili-
zation by almost all nations. The main objective of Alterna-
tive Fuels is to comprehensively describe the science and
technology of various process treatments for the clean use of
coal and coal products, synthesis gas, alcohols, shale oil
crude, biomass, and solid wastes. This ambitious objective is
presented in eleven topical chapters that include current
references to the state-of-the-art for each type of fuel pro-
cessing. Dr. Lee has successfully compiled a comprehensive
collection of pertinent data and information that were scat-
tered throughout the literature. Alternative Fuels is necessar-
ily lengthy, but neat, clear, and consistent. It can well serve
as a chemical engineering text and as a reference book for
practicing engineers and researchers.
Alternative Fuels, a book in the Applied Energy Technol-
ogy Series, has 485 pages, 172 one-line process diagrams,
graphs, and sketches, and 96 tables of data. The index lists
470 subject terms, excluding numerous sub-terms. All these
features succinctly provide a wealth of informative data that
is easily accessed by the reader. In addition, each chapter has
a set of problems (useful for students), and a solution manual
is available. It has 586 references, with 250 of them pub-
lished since 1990. References to relatively inactive clean-
coal technologies, such as oil shale, shale oil, and tar sands,
are primarily taken from studies published prior to 1980.
The first chapter presents a global overview of energy
production, consumption, and reserves for coal, gas, and oil.
Additional data are presented for electric power generation
Winter 1999


from renewable energy sources: biomass, geothermal, hy-
droelectric, solar, and wind. This chapter summarizes the
global energy situation with 18 graphs and 13 tables.
Chapter 2, in 60 pages, focuses on three major topics that
could produce environmentally clean solid and liquid fuels
from processed coal. First, the basic properties of coal are
presented along with safety issues related to coal mining and
environmental issues related to coal combustion. In the sec-
ond part, many developments in coal technology are de-
scribed for use as a means to clean fuel. The third part of
Chapter 2 presents environmental issues and regulations,
particularly related to coal mining.
Chapter 3 deals with coal gasification, which includes a
series of processes that convert coal containing C, H, and O
as well as impurities such as S and N into fuel and/or synthe-
sis gas. A total of 10 gasification processes are summarized
in about 30 pages. Then the equations are presented for
stoichiometry, thermodynamics, and reaction kinetics rela-
tive to coal gasification.
Chapter 4 presents more than two dozen processes to
develop alternative liquid fuels from coal by pyrolysis, di-
rect and indirect liquefaction, and several other known es-
tablished chemical-process techniques. This material does
not include process economics.
The next topic, Chapter 5, is the development of gas fuels
from coal. This material summarizes pertinent advances in
the DOE (multibillion dollar) Clean Coal Technology Pro-
grams and an extensive discussion of Integrated Gasification
Combined Cycle (IGCC) systems. The IGCC technology eco-
nomics are discussed. Advantages and disadvantages of com-
bined-cycle systems are delineated as potential sources of fuel.
Chapters 6, 7, and 8 are presentations of more established
technologies (coal slurry, oil shale, and tar sands) as poten-
tial sources of fuel. The coal slurry focuses on transportation
and handleability, but no economics. Descriptions of oil
shale and tar sand are focused around process diagrams and
pertinent chemical reactions.
Continued on page 83.










M1 curriculum


PROCESS ANALYSIS

An Electronic Version


GEORGE B. DELANCEY
Stevens Institute of Technology Hoboken, NJ 07030


Here at Stevens we adhere to the widespread proposi-
tion that using the Internet and the Intranet can be
substantially beneficial to undergraduate chemical
engineering education, both administratively and pedagogi-
cally.['12] There is also evidence that using software such as
spreadsheets and equation solvers13-J not only develops the
skills and flexibility necessary for ready adoption of differ-
ent software packages161 for professional activities in indus-
try, but also substantially supports learning. We believe that
effective use of these tools requires a "cultural" change, or
enhancement, on the part of both faculty and students. Full
course integration is desirable1[6 and requires broad faculty
participation, which comes slowly in many cases."71 The skill
level and motivation of the students, however, can signifi-
cantly stimulate the faculty culture.
At Stevens, our initial focus has been on the students. We
have attempted to forge a paradigm for a departmental
electronic culture in our first chemical engineering course,
Process Analysis, that comes in the second term of the
sophomore year. Our efforts to do this are summarized in
this article.

COURSE SUMMARY
The objectives of the course are to introduce students to
chemical engineering, to chemical processing equipment and
chemical processes, and to apply material balances and basic
phase equilibria to processing systems and the design of
equipment from the equilibrium-stage point of view. At the
same time, the course is being developed in an electronic
environment in order to prepare students for this emerging
characteristic of the workplace, to enhance their learning,
and to establish a basis for distance learning and asynchro-
nous delivery. By "electronic environment" we mean the
ubiquitous use of software for problem solving and trans-
mission, communication between constituencies, presenta-
tions, introduction to process simulators, graphics, and com-
puter-aided instruction and learning.
The course, which uses the Luyben and Wenzel text,s18 is


summarized in Table 1. Throughout the course the physical
and chemical bases of the process or equipment being dis-
cussed are emphasized and class discussions are often based
on the problem assignments and examples in the text. The
students are required to complete a project involving con-
struction or enhancement of a website devoted to a chemical
process or to a class of chemical processing equipment, and
each group presents its project later in the semester. Through-
out the semester, students are encouraged to use the com-
puter-aided instruction modules developed at the University
of Michigan and distributed by CACHE (Computer Aids in
Chemical Engineering).
It should be noted that the integrity of individual and
group work is subject to the Stevens Honor System. This
system is managed by the students through an Honor Board.
Individual cases are investigated and tried on the basis of
well-defined procedures approved by both faculty and stu-
dents. Penalties range from warnings through grade loss to
expulsion from the Institute. All homework and examina-
tions in the course are signed by the students and attest to the
fact that they have adhered to the Honor Code.

SOFTWARE
Students are required to purchase computers when enter-
ing Stevens. The software suite that comes with the comput-
ers includes MS Office Pro 97, Mathcad, Scientific Note-
book, X-win32, WinQVT, Netscape, Java, and Matlab.
Software used in the course that is not included in this
package is available for purchase on campus. An exten-

George B. DeLancey is Professor of Chemical
Engineering at Stevens Institute of Technology.
He received his BS, MS, and PhD degrees in
chemical engineering from the University of Pitts-
burgh. He has presented undergraduate courses
in process analysis, transport phenomena, reac-
tor design, process control, separation processes,
and thermodynamics. His current research inter-
ests are in biotechnology.


Copyright ChE Division of ASEE 1999


Chemical Engineering Education











... the course is being developed in an electronic environment in order to prepare students for
this emerging characteristic of the workplace, to enhance their learning, and to
establish a basis for distance learning and asynchronous delivery.


sive application of Ethernet supports communications
from the academic buildings, computational facilities,
and residence halls, and a gateway from the Ethernet is
provided for access to the Internet.
WebCT, described at (http://homebrew.cs.ubc.ca/webct)
is used to organize the sites summarized in Figure 1, to make
grades, examples, and solutions available to the students,
to provide E-mail and Bulletin Board communication
tools as well as a Calendar. The underlying html files

TABLE 1
Course Summary

Subject Weeks
Overview of Chemical Industry and Processes -
Process Equipment: Construction and Operation 3
Process Analyses: Flow Sheets and Process Conditions
Mass Balances, Degrees of Freedom, and Chemical Reaction 4
Equilibrium Separations: McCabe-Thiele Analyses 4
Multi-Component Phase Equilibrium: Isothermal Flash 2
Class Presentations I



S--HHI


Welcome to Chemical Engineering 210


Boad Chet


Po

Course Infoanaton Schodule. ExamOes dAss gnenm ts


Choec SlPro ca p E
Procssr Eadammet


Click on one ofthe icons above




References I Group Memberships I Grading Policy I Schedule and Assignmentsl
Preparation and Submission of Solutions I Process Equpmentl Chemical Processes I
Project I Micromentor I Scientific Notebook I User Notes for Software I Scanning I
Maling Files I Zipping Files I Downloading/Viewing Solutions and Examples I McLean
Computer Room I Virtual Libraiy I Chemical Engineering News Group I American
Institute of Chemical Engineers I Course Homepage

Figure 1. Course homepage and information hyperlinks.


C2snLcPss-vo d

4k
MyvRecord





Solutions


were created with Microsoft Word, except possibly for
the student project pages.
The Calendar is used on a class basis to post important
events such as examination and assignment due-dates, class
activities, vacations, etc. The Calendar can also be used on a
personal basis by individual students or the instructor with-
out others being able to view their entries.
Examinations are posted on the Bulletin Board and the
students can download/view the examination file. All ques-
tions concerning the examination contents are posted on the
Bulletin Board along with the responses of the instructor.
Postings from both parties may take place at any convenient
time and from any convenient location and may be viewed
by the entire class. This function is very useful, especially
when the exams are not due for several days or over a
weekend. Questions involving problem assignments are
handled in a similar fashion except that viewing may be
limited to the group in question, and the grader for the class
may respond as well. The Bulletin Board is also used on a
group basis to manage group discussions on project and
problem assignments and to transfer files associated with
these activities within the group.
E-mail serves the common communication needs as well
as being the vehicle for transferring homework, examina-
tions, and project files between two parties. The subject
entry on the e-mail message is important for quick over-
views of previous messages and for the search function
supplied by WebCT.
Scientific Notebook (SNB), found at
http://scinotebook.tcisoft.com
is used primarily for solving systems of nonlinear equations
arising from material balances and in phase-equilibrium cal-
culations. SNB has a user-friendly front end with a Maple
kernel and supports a word processing format. No special
code is required, and SNB can interface with the web and
serve as a browser for Tex files. Users can quickly perform
symbolic computations, integration, differentiation, matrix
and vector operations, and many other more complex com-
putations involved in calculus, linear algebra, differential
equations, and statistics. For these reasons, we chose to
adopt SNB rather than Mathcad"151 or Mathematica.131
The solution to simultaneous nonlinear equations associ-
ated with material balances is illustrated in Example 1,
while Example 2 illustrates how SNB is used to solve
equilibrium flash calculations.
Microsoft Excel is used for graphical (McCabe-Thiele)
solutions of equilibrium-stage problems: stripper with a


Winter 1999










reboiler, gas absorption and stripping, distillation, and liq-
uid-liquid extraction. We have elected to preserve the
graphical construction modality available in Excel for
this introductory course rather than the spreadsheet solu-
tions to the material balance equations used at some
other institutions.[4] The stage steps are constructed with
the drawing tool in Excel.
Distillation is illustrated in Example 3 and extraction in
Example 4. Calculations can be done on the spreadsheet or
SNB may be used, saved with a screen capturing software
(e.g., Snagit) and pasted into Excel. Equilibrium data may be
generated from a function (Example 3) or plotted directly as
an input data series (Example 4). Stream points are entered
as a new data series for convenience when zooming in for
stage stepping (Example 3) or zooming out for the operating
point in extraction (Example 4).
The course also makes use of the MicroMENTOR system
for delivery of the educational software modules developed


at the University of Michigan and distributed by CACHE.
The modules suggested for use during the class are: UM-
Units, UM-Beer, MATBAL, UM-Hawaii, UM-POP, and
UM-McCabe. The modules have been required in some
cases and left to the discretion of the students in others.
Distill is a program, available from CACHE, that does
multicomponent flash and distillation calculations. It consid-
ers both the liquid and gas to be nonideal and it includes a
database of 98 compounds. The vapor-phase fugacity is
calculated with the Redlich-Kwong equation of state. The
liquid-phase activity coefficients are based on the Hildebrand
solubility parameter. This program is used in several assign-
ments to illustrate the effect of assuming ideal conditions in
flash and bubble/dew point calculations for hydrocarbons.
ChemWindows is made available to the students for draw-
ing schematics, flow sheets (see the figure in Example 1, for
example) and chemical formulae. They have not yet been
required to use this software and may use others if they wish.


- EXAMPLE 1


Figure 1
The flow diagram in Figure 1 illustrates a simplified version of
the main steps in the production of vinyl chloride (C2H3CI) from
ethylene (C2H4). The reactions, which occur separately in the dif-
ferent reactors, are:
Chlorination: CH4 + Cl, -- CH4CI,
Oxyhydrochlorination: CH4 + 2 HC1 + 1/2 02 C2H4C12 + HO
Pyrolysis: C2H4C1 CH3Cl + HCI
The ethylene feed, F,, is 90 mole % ethylene and the remainder is
inerts. The chlorine and oxygen feeds, F, and F3, respectively, are
pure. All of the ethylene, oxygen, and chlorine react and all of the
hydrochloric acid (HC1) fed to the oxyhydrochlorination unit re-
acts.
Only 50% of the total dichloroethane (C2H4C12) fed to the pyroly-
sis reactor is converted, with the remainder being separated and
recycled with inerts in stream F|2. The inert concentration in the
recycle stream is 50 mole %. Pure hydrochloric acid (HCI) is
recycled in stream F,3. The final product stream, F9, consists only
of vinyl chloride and water.


Determine all of the unknown flow rates, Fj, and mole fractions,
x,. (mole fraction i in stream j). Set F,=l mole/hr as a basis. The
species are labeled as shown in the following Table 1.
TABLE 1
Labeling of Components
Species Index Species Index
CH, I CH4C1 5
Cl1 2 CH3C1 6
HCI 3 HO 7
0, 4 Inerts 8

SNB solves the material balance equations given in Figure 2 in
SNB format (single column) in less than one minute. The solution
is shown in Figure 3 after some reordering for convenience.


F,=0.90F4
F3=(0.90/2)F,
F,+F,=F,
xs6F6=0.90F4
xs,6F6=0.10F4
xS.6+x8.6=
x,7F,=0.90F5
Xs7F7=0.90F5
F,3=2(0.90Fs)
x,,F,7=0.10F5
X5.7+x7.7+x,8=l
0.50(xs.F6+x5,7F7+0.50F,2)=x-


0.50(x,,6F6+x ,7F7+0.50F 2)=x6,.F,
0.50(x,,6F6+x5.7F7+0.50F12)=x3.F,8
X7.7F7=X7.8F
x,,6F6+x,7,F7+0.50F 2=X,8F,
X3.8+X5.8+X6.8+X7.8+X8 8=
X3.8F,=Fi,
x, 8F,=0.50Fo,
X6.8F=x6.9F9
X7.Fs8=x7,9F9
x.s,F,=0.50F,
X6,9+X7.9=
5.8sF F]=FI+FI2


Figure 2. Material Balances

F2= .5,F, = .2,F = .55556,F5 = .44444,
F6 = .55556,F7 = .84444,F8 = 3.6,F, = 1.2,
Fo= 1.6,F,, =.2,F2 = 1.4,F,3=.8,
x,, = .22222,x5., = .22222,x6, = .22222,x,. =.1111,x,88 = .22222,
x6 = .9,x6 =. 1,
x.7 = .47368,x7 = .47368,xg = 5.2632 x 102
x9 = .66667,x7 = .33333.F = 1.0
Figure 3. Solutions


2 Chemical Engineering Education


HCI Separator


Oxyhydrochlorination
Reactor











-EXAMPLE 2

A petrochemical stream consisting of 30 mole % propane, 10 mole % n-
butane, 15 mole % n-pentane, and 45 mole % n-hexane is to be flashed to
200 kPa. (a) Determine strict bounds on the operating temperature. (b)
Find the vapor flow rate per unit of feed and the product compositions for
a temperature mid-way between these limits. The equilibrium data are
given in Table 1.


TABLE 1
Phase Equilibrium Data
Equilibrium data: y = Kx, T = R. and p = psia:
In K(T,p) = (a,/T2) + a, + b, In p + (b,/p)
Species a, a, b, b,
C,H, -970,688.5625 7.15059 -0.76984 6.90224
n-C Hm, -1,280,557 7.94986 -0.96455 0
n-CH,, -1,524,891 7.33129 -0.89143 0
n-C H,4 -1,778,901 6.96783 -0.84634 0


The solution procedure in SNB is as follows:
The functions in Figure 1 are defined in SNB format. Bounds for the
bubble and dew points of the feed are at unity K values for propane (j=1)
and n-hexane (j=4). These values are obtained from the roots of the K
function at the prevailing values of p (200 kPa) and j. The solutions are
shown in SNB format in Figure 2. The bubble and dew points of the feed
are roots of the associated functions at the prevailing p. The results in SNB
format are shown in Figure 3. The vapor flow at the midpoint temperature,
561.4R, is obtained similarly and is shown in Figure 4.


Equilibrium: K(i,T,p)= exp +a + b. n p+-b.3

4
Bubble point: f(T.p)=I- K(i,T,p)z,..
'=l
-1
4
Dewpoint: g(T,p)=l- Zi.I
I K(i,T,p)
i=l


Isothermal flash: h(n,T,p)= +--K(i.T.p)


Figure 1


+a3 +b1 ,np+b3 =0
T Solution is {T=657.27},{T=449.88}
Tc(O,-)

Figure 2


f(T,p)=0 g(T,p)=0
( Solution is {T=509.27 } Solution is T=613.54)
T e(449,658) Te(449,658

Figure 3

h(O,T,p)=0
le(0,1) Solution is {0=.37704}

Figure 4

Winter 1999


- EXAMPLE 3

Find the number of stages, the best feed location, and the minimum
reflux ratio for a distillation column that separates ethanol and propanol
at 101.3 kPa. The ratio of the vapor pressure of ethanol to the vapor
pressure of propanol is approximately constant at 2.10. The feed is 48
mole % ethanol and 40 mole % liquid. The distillate and bottoms compo-
sitions are to be 96 mole % and 4 mole % ethanol respectively. There is
to be a total condenser overhead with no sub-cooling. The reflux ratio is
3.0. The graphical construction is shown in the figure in Excel format.

Elfh.nolProDpnol Syslm 1t 101.3 kPa


-EXAMPLE 4


Acetic acid (species 1) is to be extracted from water (species 2) using
isopropyl ether (species 3) as the solvent at 200C and 1.0 atm. The feed
rate is 1000 kg/hr and contains 35 wt % acid in water. The solvent flow
rate is 1475 kg/hr and is essentially pure ether. The raffinate is to contain
no more than 10 wt % acid. Find (a) the minimum solvent flow rate, (b)
the number of equilibrium stages required for the separation, and (c) the
outlet concentrations and flow rates. Equilibrium data at the operating
conditions are given in the Table and the graphical construction in Excel
is shown in the figure.
Water Layer Ether Layer TABLE
TABLE
X3 x 3 x3P e
1.2 0.69 99.3 0.18 Phase Equilibrium
1.5 1.41 98.9 0.37 Data
1.6 2.89 98.4 0.79
1.9 6.42 97.1 1.93
2.3 13.3 93.3 4.82
3.4 25.5 84.7 11.4
4.4 36.7 71.5 21.6 V FIGURE
10.6 44.3 58.1 31.1 Excel Constuction for
16.5 46.4 48.7 36.2 Example 4





S---- .










WL % isopmropyl thr, 100x,










FILE HANDLING
An extraordinary amount of time is required to open files
to learn the source and/or the contents. This is especially true
if more than one course is underway. Consequently, labeling
of files is very important. Points are deducted from assign-
ments and examinations (which may not even be accepted in
some cases) if the established procedures are not followed.
No files are accepted unless adherence to the Honor Code is
pledged by the students at the end of the file.
Homework files are prepared weekly by students working
in groups. The assignments are numbered and are generally
specific problems in specific chapters of the text. Solutions
to individual problems are labeled as gxpy-zc210.(tex, rap,
doc, xls,...) where x is the group number, y is the chapter,
and z is the problem number. The individual solution files
are then assembled in a single archive, with WinZip for
example. The archive is labeled gxayc210.zip where x is the
group number and y is the assignment number. The archive
is attached to an e-mail message with an informative subject
entry and mailed to the instructor. The archives are placed in
a directory named with the group numbers. Solution files for
each problem are posted for downloading after the due date
for the assignment.
Examination files include a reference to the individual
student: exmc210qnabz.(tex, rap, doc, xls,...), where m is the
examination number, n is the question number, ab are the
student initials, and z is the group number. We have not
encountered students with the same initials in the same group
yet, but many modifications are possible to handle such an
event. The examination files are then grouped in an archive
labeled exmc210abz.zip and attached to a relevantly labeled
e-mail message to the instructor. The archives are then placed
in a directory named with the examination number.
Copies of files received by the instructor are forwarded by
e-mail to the grader. The grader is not permitted to accept
files directly from the students. The grader has a set of
directories similar to those used by the instructor, described
above. The grader uses several additional subdirectories as-
sociated with the grading process. One directory is for graded
files, one for ungraded files, and one for the archive in
process of being graded. After the files are graded and the
comments and points included, the files are returned directly
to the students, the grader retains a copy, and a copy of the
file is sent to the instructor. The same file name is retained so
that the instructor can replace the previously ungraded file
with the graded one in his directory. It has always been
possible to track down "missing" files with this procedure.
The grader is responsible for posting grades in WebCT,
but a separate grade file is maintained by the instructor and
updated with each new assignment or examination. The
grade file maintained by the instructor is the official one so
that grades, once recorded, cannot be changed without the


instructor updating this file. Grade access is limited to the
student, the grader, and the instructor, but statistical infor-
mation is generally available.

PROJECT
The objectives of the project are for the student to gain a
special familiarity with a specific chemical process or a
category of chemical process equipment and a significant
experience with web-based presentation of technical materi-
als. The students are also required to present their projects to
the class, electronically if desired.
The bases on which the sites are initiated or enhanced are
summarized below.
A) Process equipment (nine sites at present covering major
equipment categories):
1) Purpose and operating principles)
2) Historical background
3) Construction-schematic/illustrations
4) Pictures of actual equipment
5) Range of duties-sizes
6) Maintenance required
7) Utilities required
8) (Some) design equations
9) References
B) Chemical Processes (nine sites at present spanning range of
top 50 chemicals):
1) Chemical formula and form/state of product
2) Uses and market price
3) Historical industry development
4) Common methods of production with raw material sources
and side products
5) Major companies, production methods, and production lev-
els
6) Details for a common production method: process chemistry
with implications for most favorable process conditions;
flow sheet(s); operating conditions and problems; environ-
mental considerations; production costs; utility require-
ments; hazards; handling of waste and side products
7) References
C) Grading (Engineers from EXXON Research and Engineer-
ing and experienced faculty outside the department contribute
to site assessment):
1) Content (see above)
2) Layout/format
3) User friendliness
4) Use of colors and graphics
5) Use of relevant hyperlinks and other resources
Html files are prepared or edited with software that is
generally available to the students, such as Microsoft Word,
or that is the individual preference of the students. Scanning
hardware is used for some picture files, which can be edited
with Photopaint. The completed project files, including im-
ages (gif, jpeg,...) are collected in an archive and mailed to
the instructor as described above.


Chemical Engineering Education










DISCUSSION
The students very quickly become accustomed to the elec-
tronic communication features of the course and the file-
handling procedures. It is, however, helpful to deduct points
for improper file labeling as well as to not answer exam
questions online unless posted for viewing by the entire
class. Off-campus students find the electronic version help-
ful and students generally like the option of transferring files
and posing questions at their convenience. Such items ap-
pear at all hours of the day. For the solutions to assignments,
the availability of examples for downloading by the students
is crucial for each new application.
The use of SNB for material balances has greatly im-
proved the sophistication and ability of the students to write
an independent set of material-balance equations and to be
acutely aware of the degrees of freedom. They are also able
to spot-check their solutions with quick calculations using
time components, inerts, etc. Most students begin with doing
the solution on paper and transferring it to SNB. They soon
progress, some to doing the solution completely online. A
series of user sessions would be helpful at the beginning
for learning the software. Some students have prior expe-
rience in mathematics classes, which is beneficial. Stu-
dents have voiced a preference for SNB to similar soft-
ware packages, but these alternatives have not been ex-
plored in a formal way.
The use of Excel for graphical multistage constructions
has been very successful and well accepted by students.
Errors can be quickly corrected. The students cannot imag-
ine doing liquid-liquid extraction constructions by hand.
When not required, the computer-aided instruction mod-
ules in MicroMENTOR are not extensively used. Use of the
modules is greatly increased if the students are held account-
able, for example, in class discussions on the Bulletin Board
of WebCT. The modules, once used, are considered helpful.
Distill is viewed as inconvenient by the students, although
the results are very elucidating for comparison purposes. For
this reason and for broader objectives, it may be worthwhile
to introduce Aspen in a limited way at this stage.
ChemWindows is viewed primarily as a cosmetic add-on
for problem assignments when not required. It is used for all
examination drawings and for some project work. As it
significantly enhances the quality of the presentation, we are
introducing some limited requirements for its use.
The class projects have been most successful and well
received by the students. The project and equipment sites
they have developed are being made an integral part of the
course. The students are extremely creative and enthusiastic
about finding relevant materials on the Web and using other
resources to develop or enhance their site. The technical
content of the class presentations attests to the learning com-
ponent that is present in the projects. The students very


easily adopt a full range of software to suit their personal
needs for developing the required html files and exploit the
extensive range of picture files available on the Web.
The electronic approach used in Process Analysis has been
repeated with much less effort by the author in Transport
Phenomena, a course offered in the following semester. The
approach has not yet been extended further across the chemi-
cal engineering curriculum. Software is, however, used in all
of the courses to some extent and coding is required in
several classes. The students have been observed using
the specific skills (especially the graphical solution to
multistage equilibrium separations) and exhibiting the
electronic dexterity they gained in Process Analysis in
subsequent courses.
The time commitment and the background that must be
acquired on the part of faculty in order to develop an elec-
tronic version of a course as defined here are important
related issues and tied very much to the available support
and the electronic environment. It must be generally recog-
nized, however, that significant effort is required which must
be explicitly recognized and supported within the local edu-
cational context and the electronic capabilities of its con-
stituencies. The benefits include presentation of a course in
an environment that anticipates the workplace, that is more
adaptable to the different life styles of the students, and that
can be conveniently modified and maintained with no more
effort than a conventional course. The electronic version is
more supportive of the learning process and allows more
spontaneity and interaction, especially when the students are
networked with their own computers. A greater focus on
problem structure and formulation can be realized as well as
a pronounced increase in the specific and general electronic
skills of the students.

REFERENCES
1. Bungay, H., and W. Kuchinski, "The World Wide Web for
Teaching Chemical Engineering," Chem. Eng. Ed., 29(3),
162(1995)
2. Marr, D.W.M., and J.D. Way, "Using the Intranet in ChE
Instruction," Chem. Eng. Ed., 31(2), 110 (1997)
3. Al-Dahhan, M.H., "Computing in the Undergraduate ChE
Curriculum," Chem. Eng. Ed., 29(3), 198 (1995)
4. Burns, MA., and J.C. Sung, "Design of Separation Units
Using Spreadsheets," Chem. Eng. Ed., 30(1), 62 (1996)
5. Sandler, S.I., "Spreadsheets for Thermodynamics Instruc-
tion: Another Point of View," Chem. Eng. Ed., 31(1), 18
(1997)
6. Davis, J.F., G.E. Blau, and G.V. Reklaitis, "Computers in
Undergraduate Chemical Engineering Education," Chem.
Eng. Ed., 29(1), 50 (1995)
7. Harb, J.N., A. Jones, R.L. Rowley, and W.W. Wilding, "Use
of Computational Tools in Engineering Education," Chem.
Eng. Ed., 31(3), 180 (1997)
8. Luyben, W.L., and L.A. Wenzel, Chemical Process Analysis:
Mass and Energy Balances, Prentice Hall, Englewood Cliffs,
NJ (1988) 1


Winter 1999










rem.= laboratory


DEMONSTRATING SIMULTANEOUS

HEAT AND MASS TRANSFER

WITH MICROWAVE DRYING



CHERI C. STEIDLE, KEVIN J. MYERS
University of Dayton Dayton, OH 45469-0246


In a recent article, Nirdosh and Baird"'1 emphasized the
critical role that the laboratory plays in the undergradu-
ate engineering experience. It is crucial that laboratory
experiments provide practical reinforcement of the theoreti-
cal chemical engineering concepts developed in lecture
courses, but due to budget constraints, it is often also neces-
sary to develop inexpensive experiments or to make use of
existing equipment. This paper describes an effective and
inexpensive microwave drying experiment that can be used
on a variety of levels. On the introductory level, the data
analysis associated with this experiment illustrates the nu-
merical approximation of derivatives from discrete data,
while on the advanced level, the experiment develops an
understanding of simultaneous heat and mass transfer.

BACKGROUND
Of all the chemical engineering unit operations, drying is
one of the most widely used, with applications in various
processing industries such as food processing, pulp and pa-
per, pharmaceuticals, etc. Because of its widespread use and
the fact that it may account for up to ten percent of industrial
energy consumption,[2] it is therefore essential that the fun-
damentals of the drying process be encountered and under-
stood by undergraduate chemical engineering students.
There are a variety of drying techniques, such as vacuum
drying and spray drying,131 each with its own operational
characteristics. Conventional dryers include both direct and
indirect methods of heat transfer.[4] Direct dryers (also called
convection dryers) use contact between the wet solid and a
hot gas to accomplish heat transfer, with the vaporized liquid
being carried away by the drying gas. In indirect dryers, heat
for drying is transferred through a wall that separates the wet
solid and the heating medium, with the vaporized liquid
being removed independently of the heating medium. Indi-


rect dryers are also called conduction, or contact, dryers.
The driving force for heat transfer in both direct and
indirect dryers is the temperature difference between the
drying medium and the wet solid. In the case of microwave
drying, a magnetron produces a pulsing electromagnetic field.
Polar molecules such as water align with this field, and as
the field direction changes, the molecules are forced to re-
align. These molecular oscillations create friction that gener-
ates heat, raising the temperature and causing liquid evapo-
ration."5 Thus, although supplied from an external source,
the energy for microwave drying is often thought of as being
generated within the wet solid.
Although the mechanism of delivering energy to the wet
solid in microwave drying is different than in conventional
drying, since microwave energy does not penetrate very far
below the surface of the exposed material,16t the transport
processes occurring during microwave drying are very simi-
lar to those that occur during conventional drying. As liquid
evaporates from the surface of the solid-liquid mixture, liq-

Cheri C. Steidle received her Bachelor of Chemi-
cal Engineering degree in 1995 and her MS in
Chemical Engineering in 1998, both from the
University of Dayton. She is currently working
as a process engineer in the glass industry and
is responsible for implementation of new pro-
cess control equipment for the manufacture of
glass tumblers and tableware.



Kevin Myers is a professor in the Department of
Chemical and Materials Engineering at the Uni-
versity of Dayton. He received his BChE degree
from the University of Dayton and his DSChE
from Washington University in St. Louis. His re-
search interests are in multiphase agitation and
chemical reactors.


Copyright ChE Division of ASEE 1999


Chemical Engineering Education











This experiment demonstrates the drying process effectively. It is an extremely flexible, safe, and
inexpensive experiment that can be incorporated into the undergraduate laboratory curriculum.
The experiment is easy to set up and run. Typically, meaningful experimental data for
higher power settings can be collected in about thirty minutes.


uid from within the solid migrates to the surface because o
concentration gradients. Thus, drying, either by convention
or microwave devices, inherently involves simultaneous hea
and mass transfer.
The primary data obtained during a drying experiment ar
the moisture content as a function of time (on a dry basis the
moisture content equals the liquid mass divided by the soli
mass). Curve 1 in Figure 1 presents a typical moisture con
tent curve. The moisture content data can be differentiated t(
yield the drying rate curve, Curve 2 in Figure 1.

Drying Rate= -1 dm (1)
m, )t dt )
Typically, drying rate curves for materials with a thor
oughly wetted surface exhibit three periods.171 The first stag
is the warming-up period. This stage is characterized b
increasing the material temperature to that of the evapora
tion temperature of the wetting liquid. Also, the drying ratf
increases as the liquid begins to vaporize. This is followed
by a period of constant-rate drying where the surface mois
ture evaporates and moisture is steadily brought to the sur
face to maintain a continuous liquid film over the surface
The third, and last, period in the drying-rate curve, called thi
falling-rate period, is characterized by a nonlinear decrease
of the drying rate due to the increasingly uneven moisture
distribution over the surface. The constant-rate and falling
rate stages of the drying-rate curve are separated by the poin
of critical moisture content. This point marks the instan


I Time
Figure 1. Typical moisture content and drying-rate
curves for a solid with a thoroughly wetted surface.
Winter 1999


f
I
t

e
e
d

0


when the liquid no longer forms a continuous film over the
entire surface because the rate of moisture transport to the
surface is less than the rate of evaporation from the surface.
The critical moisture content is not a material property. It
varies with drying rate, thickness of the material, particle
size, and other factors that affect moisture movement. Criti-
cal moisture content is best determined by experiment.

APPARATUS


The equipment required for this experiment is inexpensive
and easy to operate. Specifically, an off-the-shelf Frigidare
MC-1100M microwave with ten power settings ranging from
72 W to 720 W was used. The power settings of the oven are
in terms of percentage of the maximum power. This micro-
wave oven is rather old, but a comparable current model
could be purchased for less than $200. Other required items
include a digital balance accurate to the nearest gram, a
stopwatch (or the microwave timer), microwaveable trays or
d Pyrex glass beakers, and a thermocouple (specifically, a type
T was used). Sand wetted with water was studied, with the
nominal diameter of the sand particles being 600 microns.

S METHOD
e
e Wet sand in the ratio of approximately 1 kg of sand to 0.2
e kg of water was used. This ratio provided the condition that
all of the sand was completely wetted, but no standing pools
t of water were present. The sand and water were thoroughly
t mixed together. The mixture was weighed, spread into an
even layer approximately 0.025 m deep in a 0.15x0.15x0.05
m tray that was placed in the center of the microwave oven.
The microwave was started at a desired power level. The
weight of the mixture was recorded every minute until the
sample was dry. In addition, the surface temperature of the
sand could also be monitored with a thermocouple when the
sample was removed from the oven for weighing. The effect
of removing the sample from the oven for weighing was
examined by using various drying intervals, such as 30
seconds, 1 minute, 5 minutes, and 10 minutes. Only mi-
nor changes in the sample weight were observed for
these different drying intervals. Appropriate safety pre-
cautions include wearing safety glasses and oven mitts
when handling the hot tray.

RESULTS
Figure 2 presents experimental moisture content and dry-
ing rate data for the highest power setting of the microwave
oven (720 W), while Figure 3 compares the drying rate










curves obtained at three power settings (720 W, 540 W, and
360 W). All of the experimental data are very similar to the
idealized data of Figure 1, clearly illustrating the three stages
of drying. Table 1 demonstrates the influence of power
setting on the critical moisture content. Recall that the criti-
cal moisture content is not a material property; rather, it is
dependent on operating conditions. The trend of increasing
critical moisture content with increasing power setting is
typical.[4] As the drying rate is increased (by increasing the
power setting), it becomes progressively more difficult for
the rate of moisture transport to the surface to remain as high
as the rate of surface evaporation. Thus, the falling-rate
period, which begins when the critical moisture content is
reached, occurs at a higher moisture content.
During the constant-drying-rate period, a pseudo-steady
state is achieved, with the power input going to the latent
heat of evaporation of the liquid and energy losses. Table 2
presents the efficiency of the constant drying-rate period
assuming a latent heat of evaporation of 2300 kJ/kg (corre-
sponding to approximately 80'C). The efficiency is calcu-
lated as the energy required for water evaporation (equal to
the evaporation rate, kg/s, times the latent heat of evapora-
tion, kJ/kg) divided by the rate of energy input (kW = kJ/s).
The observed efficiencies are rather high, ranging from 74%
to 87%. Given the advanced age of our microwave oven,
its power output is probably less than the nominal value
(we accepted the nominal power outputs and did not
actually measure the oven's power output). If this is true,
the actual efficiencies are higher than those calculated
here. The high efficiencies of microwave drying are par-
tially due to the energy-transfer mechanism. Since en-
ergy is supplied directly to the wet solid, there is no large
energy requirement for heating the drying medium as in
conventional drying devices.
The efficiency decreases with increasing power level. Ini-
tially this was thought to be due to increased energy losses to
the lower portions of the wet solid by conduction. At low
power inputs, the temperature of the wet solid might be


relatively uniform at the start of the constant-drying-rate
period as conduction has had sufficient time to transport
energy from the surface to the interior of the wet solid. At
high power inputs, conduction may be unable to transport
energy from the surface to the interior regions of the wet
solid rapidly enough to achieve temperature uniformity at
the start of the constant-drying-rate period. If this were the


,0.20 I ,, 0.020
0 E

00.15 Moisture Content 0.015
SDrying Rate
Z0.10 -0.010 .
0 A A -
0 A
20.05 0.005 W
0 A 0)

S0.00 +-*--- ',--A -o 0.000
0 5 10 15 20 25
Time (min)

Figure 2. Moisture content and drying-rate curves for a
power level of 720 W.


0.020

E

S0.015
A A 720 W
5 ...*
M0.010 0 540W
0M A XzI
Sx A360 W
0.005 = A x
x x z z
6X A *0
0.000 A ,
0 5 10 15 20 25 30 35 40 45
Time (min)
Figure 3. Comparison of drying-rate curves for power
levels of 720 W, 540 W, and 360 W.


TABLE 2
Influence of Power Level on
Constant Drying-Rate Period Efficiency
(from Figure 3 drying-rate curves)

Power Level Evaporation Rate Evaporation Energy Efficiency
(kW = kJ/s) (kg/s) (kJ/s)
0.360 1.36 x 10- 0.313 87%
0.540 1.90 x 10-4 0.437 81%
0.720 2.33 x 10-4 0.536 74%

Chemical Engineering Education


TABLE 1
Influence of Power Level
on Critical Moisture
Content for Drying Sand

Critical
Moisture
Power Level Content
720 W 7.4%
540 W 7.1%
360 W 6.2%
144W 5.7%
Literature value'4) 5.9%
(unknown power
level)









case, conduction would still be removing some of the energy
input from the surface to the interior during the constant-
drying-rate period. This energy "loss" would lead to de-
creased efficiencies at high power levels. This explanation,
however, loses its appeal when the energy inputs during the
warming-up period are compared. These energy inputs are
about 240 kJ, independent of power level. This constancy of
energy input during the warming-up period indicates that
energy losses due to conduction are likely to be the same for
all power levels. The reason for decreasing efficiency with
increasing power level requires another explanation that re-
quires further investigation.
The drying-rate curves of Figures 2 and 3 were generated
from the moisture-content curves through numerical differ-
entiation using a central-difference method. The central-
difference method is best used in cases involving large time
intervals1[8 as is the case with this experiment. Assuming
evenly spaced data, at some time, t,, the central difference

0.0040
E
2 0.0032
o nStepsize= 1 min

a 0.0024 Stepsize= 2 min

." 0.0016

o0.0008

0.0000
0 20 40 60 80 100 120
Time (min)

Figure 4. Effect of step size on numerical differentia-
tion at a power level of 144W.


0.020
E
o 0.015 m-"=.
SEqn. 2

S0.010


= 0.005

0.000 U
0.000 -,,----,-1- -,"I-
0 5 10 15 20 25
Time (min)
Figure 5. Effect of truncation on numerical differen-
tiation at a power level of 720 W.
Winter 1999


approximation to the drying rate is

Drying Rate = =L-1 i (, m i+ -mi-1) (2)
msldt MJ I 2(ti,+-ti)

The drying rates at the first and last points were generated
using related forward and backward difference approxima-
tions as shown in the following equations:

First Point:

-' (dmi 1 (-m2 +4mi -3mo)
Drying Rate= -I ---1 (-2+4ml-3mo)] (3a)
mSdt ms 2(t, -to)

Last Point:

Drying Rate t (n (,M-2t -4mt +3mn)] (3b)
-m-dJ ms 2(tn -,tn-)

Figure 4 illustrates the importance of sampling interval on
the numerical differentiation process. These data were taken
at a low power setting (144 W), and the drying-rate curve
was generated using step sizes of both one and two minutes.
The drying-rate curve obtained using a step size of two
minutes is relatively smooth, while the drying-rate curve
obtained using a step size of one minute fluctuates, particu-
larly during the constant-rate drying period and as the drying
rate falls to zero. At this low power setting, the amount of
water evaporated in a one-minute interval was approaching
the accuracy of the scale, introducing significant round-off
error into the calculations.
Figure 5 examines the effect of truncation error on nu-
merical differentiation of the data. This figure compares the
central difference approximation of Eq. (2) to the following,
more accurate, central-difference approximation:

Drying Rate =

[ (dm]. 1 [(-mi+2 + 8 mi 8 mi- +mi-2)]
_Ms dtJ =-m ms 12(ti+, t)

The minimal difference between the derivatives generated
by the two differentiation formulas is indicative that trunca-
tion error is not significant in this instance and the numerical
differentiation formula of Eq. (2) is sufficiently accurate.

EXTENSIONS
This experiment offers great flexibility and can be ex-
tended to examine other concepts, including temperature
profiles, particle size and geometry effects, and drying char-
acteristics of different materials.
E Temperature profiles of the solid surface can be produced
by recording the surface temperature with a thermocouple
when the sample is removed for weighing. Surface
Continued on page 71.










Blm laboratory


MEDICAL SURVEILLANCE

AND THE

UNDERGRADUATE THESIS


IAN A. FURZER
University of Sydney Sydney, New South Wales 2006, Australia


he Department of Chemical Engineering at the Uni-
versity of Sydney has a thesis as a core unit of study
in the final year of the chemical engineering curricu-
lum. Students are required to complete fifty credit units in
their final year, and the undergraduate thesis is worth
eight of them. It normally takes place in Semester 1 and
may overflow into the vacation break before the second
semester begins.
The thesis discussed in this article is an experimental
thesis concerned with the operation of a distillation column
to collect composition data on the trays for a ternary mixture
of ethyl acetate-ethanol-water. The same thesis, however,
can incorporate a significant new component that is de-
signed to make students more aware of occupational expo-
sure to hazardous substances.
This insight can also be valuable in other chemical engi-
neering courses, including risk engineering, hazards and
hazops, environmental pollution, and chemical engineering
design. The awareness of the dangers of human exposure to
hazardous substances is becoming evermore important due
to the long-term health effects on workers in the oil, chemi-
cal, and biotechnology industries, as well as on the general
public. Chemical engineering students need to be aware of
alternative process flowsheets that avoid hazardous substances
and a general chemical reduction-use program. Several popu-
lar books have brought the effect of chemicals on human
health into the public consciousness.111

MEDICAL SURVEILLANCE

The World Health Organization has a medical surveil-
lance program for the early detection of occupational dis-
eases. It is a prevention program that should be brought to


the attention of chemical engineering students. The informa-
tion from such program can help students reduce the risk of
exposure to hazardous substances. While medical treatment
procedures are still not well developed for exposure that
leads to cancer several decades ahead, numerous medical
treatment procedures are outlined on Internet web sites, in-
cluding osha.gov in the USA and worksafe.gov.au in Aus-
tralia. There is also a CD-ROM[2' that contains outlines of
medical treatment and surveillance programs for a wide
range of hazardous substances.
Another CD-ROM search for ethyl acetate provides a
wide range of information, including an eight-hour time-
weighted average (TWA) exposure of 200 ppm in many
countries and an outline of the proper medical treatment for
it. This information confirms the low risk to students of
exposure to ethyl acetate during a series of distillation ex-
periments conducted for their undergraduate thesis. But
chemical engineering students should be made aware of
important and relevant sections of the subject's toxicology
and epidemiology. Benzene is a substance that has under-
gone a reduced TWA over the years and is now recognized
as being carcinogenic to humans. Details on background
levels of benzene, a series of epidemiological studies, and
cancer mortalities can be found in reference 3. The environ-


Copyright ChE Division of ASEE 1999


Chemical Engineering Education


lan Furzer has been a faculty member in the
Department of Chemical Enginering at the Uni-
versity of Sydney for over twenty-five years. He
has extensive teaching and research interests
that include computing, process simulation, and
chemical engineering plant design. He is the
author of over eighty research publications and
the textbook Distillation for University Students.












mental effects of benzene[41 in wastewater and aircraft en-
gine exhausts and their connective pathways to humans need
to be carefully considered by chemical engineering students
in the design of plants that produce benzene or fuels which
produce benzene upon combustion. Can we expect a reduc-
tion in the TWA value for ethyl acetate in the future if there
is medical evidence of an adverse health effect?
In New South Wales, 1966 legislation on Occupational
Health and Safety (Hazardous Substances) Regulation has a
separate division on an employer's duties relating to health
surveillance. The employer must provide workplace health
surveillance if any exposure to a hazardous substance results
in a reasonable likelihood of a disease or other effect on
health. This health surveillance must be under the supervi-
sion of an approved medical practitioner. The type of sur-
veillance is listed in some detail for eleven hazardous sub-
stances, including acrylonitrile, asbestos, isocyanates, orga-
nophosphate pesticides, polycyclic aromatics, and vinyl chlo-
ride. The medical tests include the standard respiratory func-
tion tests such as FEV1 and FVC.

OCCUPATIONAL HEALTH
AND EXPOSURE IN THE LABORATORY

The distillation column for the undergraduate thesis ex-
periment involves medium quantities of solvent. The choice
of solvents to be distilled by students should be based on
occupational health, safety, ease of analysis, shape of the x-y
diagram for binary systems, and the ability to extract knowl-
edge on separation systems. Alcohol-and-water is a com-
monly used system. The occupational health risks to stu-
dents of exposure to ethanol, by inhalation, need to be con-
sidered. How often are laboratory demonstrators advising
students of the need to minimize the inhalation of ethanol?
For an undergraduate thesis or special project, it is useful
to consider the ternary mixture of ethyl acetate, ethanol, and
water. This mixture provides a good working environment
to extract knowledge on ternary and binary azeotropes, dis-
tillation paths, and the appearance of a two-liquid-phase
region in the distillation column. The following method has
been used to actively reduce the exposure to ethyl acetate
and ethanol over the students' extended period of work for
the undergraduate thesis.
The most important aspect of occupational health is to
introduce the subject to the students who will be involved, to
notify them of the low hazard of ethyl acetate and ethanol,
and to discuss with them the methods of reducing the mass
inhaled. This introduces to them the psychological compo-
nent of occupational health, leading each student to develop
his or her own concern about the toxic nature of the inhaled
substances. This component will play an important part in
Winter 1999


The awareness of the dangers of
human exposure to hazardous substances
is becoming evermore important due to the
long-term health effects on workers in
the oil, chemical, and biotechnology
industries, as well as on
the general public.


assessing their own exposure estimates. An approach to
primary prevention of exposure is to increase ventilation in
the laboratory, thus diluting the inhaled air and reducing the
composition of the solvent in the air. The mass of inhaled
solvent per breath can be reduced and the risk factor in
occupational health could also be expected to be reduced.
This involves starting exhaust fans and opening doors and
windows to increase air circulation.
The next aspect of occupational health is associated with
safety and is concerned with the release of solvent vapors
due to a laboratory fire. Students should be advised on the
safety procedures to be used in the event of a fire. They
should be told to evacuate the laboratory immediately and
not to attempt putting out the fire, but to notify the safety
officer instead. This is extremely important for minimizing
the risk of exposure.
Good laboratory practice is important for occupational
health. There is a direct association between hazard identifi-
cation, hazard analysis, and good occupational health. When
release of solvent vapors is the hazard to be minimized, a
hazard identification program must be carefully conducted
on the distillation equipment, identifying all possibilities
of vapor release and listing all actions needed to prevent
such a release.
For example, vapor could be released by a failure of the
cooling water, a mechanical failure of a glass column piece,
a gasket failure, or a low level in the reboiler. Special atten-
tion should also be given to the cooling water supply. Safety
labels on the cooling water supply switches are essential to
prevent an insufficient supply of cooling water and to mini-
mize the release of solvent vapor.
Also associated with good laboratory practice is the need
for the student to be present in the laboratory while the
distillation equipment is operating. Students should be well
trained in emergency procedures such as turning off the
steam supply to the reboiler in the event of a solvent release.

STUDENT EXPOSURE TO ETHYL ACETATE

Students can be exposed to ethyl acetate in a number of the










laboratory experiment phases. The first is concerned
with calibration of the gas-chromatograph (GC)
equipment for analysis of the ethanol-water-ethyl
acetate mixture. This can involve exposure while
transferring ethyl acetate from a 20-L drum to smaller
glass vessels, the preparation of standards, and dur-
ing the running of the GC. Student exposure to ethyl
acetate during this phase in generally low. An inter-
nal standard is used in the GC calibration. In this
case it was 1-propanol, and students should be aware
of its occupational health characteristics.
Filling the distillation column with the ternary
mixture involves transferring and measuring about
20 L of the mixture and often involves mild expo-
sure to ethyl acetate. The distillation column, which
had been tested for leaks with water, will now be
found to have a smell of ethyl acetate, but no visible
liquid leaks. The laboratory is equipped with a gas
alarm system with a sensor adjacent to the distilla-
tion column. This alarm is continuously on but is
not activated by this mild smell of ethyl acetate.
Running the distillation column with good venti-
lation until steady state is reached may require one
hour and result in mild exposure to ethyl acetate.
Additional exposure could occur when liquid
samples are withdrawn from the nine column
trays. Further mild exposure would take place for
each distillation run.


BLOOD TESTS AND STUDENT PRIVACY

Students may be exposed to ethyl acetate before
entering the unit operations laboratory. Ethyl ac-
etate is a well-known solvent and is often used in the
cosmetic industry as a nail polish remover, so fe-
male students who use this substance might be ex-
pected to have a higher ethyl acetate content in their
blood. The body may also generate ethyl acetate
from the complex biochemical pathways in the body.
The initial ethyl acetate content of students' blood
is important before they enter the solvent environ-
ment of the unit operations laboratory, but this can
pose some problems with privacy. Students have a
right to privacy concerning analysis of their blood.
Voluntary agreement must be obtained from the
student for a blood test for ethyl acetate. Students
are advised to contact the student medical center
for this blood test.
The background level of ethyl acetate in the
blood should be around 0.5 mg/L. One female
student, however, had an initial level of 1.4 mg/L
of ethyl acetate. This highlights the importance


0 5 10 15
Day Number


20 25 30


Figure 1. Estimated impulse exposure to ethyl acetate (male student).


OCCUPATIONAL HEALTH: EXPOSURE
40


35


30


E 25


20



0

10 -


5


0
0 5 10 15 20 25
Day Number

Figure 2. Estimated impulse exposure to ethyl acetate (female student).


Chemical Engineering Education


OCCUPATIONAL HEALTH: EXPOSURE
S,. .


25


20


I I l I ,










of a blood test before conducting the distilla-
tion experiment.
Students were then asked to estimate their
exposure to ethyl acetate in the qualitative terms
of low, medium, and high and to estimate the
number of hours at these exposure levels on
each exposure day. This raw data represents a
series of impulses of exposure to ethyl acetate,
the height of the impulse being a measure of the
perceived intensity of exposure.
The body reacts to these impulses by exhal-
ing ethyl acetate, by generating enzymes to con-
vert ethyl acetate to other substances for rejec-
tion, or by storing ethyl acetate in fatty tissue.
The complex response is similar to the simple
impulse response taught in chemical engineer-
ing mathematics and process control.
The process could be made more quantitative
if the inhaled dose of ethyl acetate per impulse
could be measured. This would require per-
sonal monitoring devices attached to the stu-
dent and subsequent analysis of the sample tubes.
These devices are useful in providing an inte-
gration of the impulse of ethyl acetate. Other
methods, which are not practical in a unit op-
erations laboratory, would involve a composi-
tion measurement of ethyl acetate in the air and
a profile of the rate of inhalation of air.


The wuit
operations
laboratory
envi:eotannt
treated as a local
workplace can be
valuable in
introducing
students to
exposure,
occupational
health, and
perceived
impressions of
exposure.



TABLE 1
Blood Analyses
Ethyl Acetate (mg/L)

Before After
Male 0.5 0.5
Female 1.4 0.5


Medical surveillance is introduced to the students through
blood tests both before and after the full laboratory period
covering the undergraduate thesis. Students conducted a cali-
bration, distillation runs, and analyses over a period of four
months, with the number of exposure days limited to about
thirty days. The period of zero exposure to ethyl acetate
between exposure days may have provided time for the body
dynamics to remove excess ethyl acetate from the blood.
Figures 1 and 2 show the estimated impulse exposure to
ethyl acetate on the exposure days for a male and a female
student, respectively. The data for these figures were ob-
tained from the student's own impression of the level of
exposure as low, medium, or high, and the hours of expo-
sure. The differences in the two figures is due mainly to the
different perceived exposure by each student. The two fig-
ures provide important qualitative information on the prob-
lem of estimating exposure without a personal monitor.
The chemical engineering student will meet this type of
occupational health problem when employed in the work-
place. The unit operations laboratory environment treated
as a local workplace can be valuable in introducing stu-
dents to exposure, occupational health, and perceived
impressions of exposure.
The end of the laboratory experiments, with the column


drained and the solvents returned to the solvent
store, marks the end of the occupational health
exposure period. Students are then advised to
have a second blood test for ethyl acetate. Table
1 shows the ethyl acetate content of the blood
before and after the distillation experiments.
The blood results are most encouraging, with
both students having background levels of ethyl
acetate at 0.5 mg/L. One might conclude that
student awareness of the occupational health
study reduced their exposure, or that the ven-
tilation of the laboratory was adequate, or
that the interval between exposures was suf-
ficient for the body to remove any excess
ethyl acetate.

OCCUPATIONAL HEALTH GOALS

3 To use the undergraduate thesis
experiment on distillation to introduce
students to the concept of medical surveil-
lance.
3 To introduce students to the
psychological response of exposure limits
to inhaled chemicals when introduced to
the Material Safety Data Sheets (MSDS).
3 To obtain essential information on
the previous history of exposure to a chemical
through a voluntary blood test before the
experiment begins.


3 To monitor the daily exposure to a chemical in a
qualitative manner, thus introducing perceived
exposures to a chemical and an improved
awareness of occupational health.
3 To introduce ventilation as a key measure in
reducing exposure to a chemical.
3 To have a repeat blood test after the laboratory
experiment to ensure that preventive methods
for reducing occupation health exposure have
been successful.
3 To expect students to perform better in the
occupational health area of chemical engineer-
ing design and that they have an improved
concept of occupational health, both in the
workplace and in general.

REFERENCES
1. Crumpler, D., Chemical Crisis, Scribe Publication (1994)
2. CD-ROM Micromedia, "Environmental Health and Safety"
3. Furzer, R.I., and I.A. Furzer, CHEMECA 96, Sydney, Sept
30 Oct 2, 4, 69 (1996)
4. Furzer, I.A., Chem. Eng. Aust., ChE 21, 22-25 (1996) J


Winter 1999











Ie W1 laboratory


LABORATORY EXPERIMENT


IN BIOCHEMICAL ENGINEERING

Ethanol Fermentation




ALBERTO COLLI BADINO, JR., CARLOS OSAMU HOKKA
Universidade Federal de Sao Carlos Sao Carlos, Brazil


he need for didactic experiments that will prepare
our students for their professional futures, together
with the importance of the ethanol industry in Brazil,
led us to design and construct an experimental bench-scale
kit for determining the kinetic parameters related to the
ethanol fermentation process.
The experiment's design was based on the principles that
guided construction of the Didactic Laboratory of Chemical
and Biochemical Reactions at our university, which included
a laboratory composed of didactic kits for short experiments
prepared by the students that would complement classroom
knowledge.
A set of three bench-scale fermentors was designed and
constructed that would support groups of five students per
fermentor. To standardize the experiments, one of them
was used in two types of cultivation: the first at low
substrate concentration (= 16.7 gL-' of glucose) and the

Alberto Colli Badino, Jr., is Associate Profes-
sor of Chemical Engineering at the Universidade
Federal de S. Carlos. He received his Master's
from the Federal University of S. Carlos in 1991
and his Doctorate in Biochemical Engineering
from the State University of S. Paulo in 1997. His
research interests are in mass transfer in con-
ventional fermentation processes, rheology of fer-
mentation broths, and power requirements in con-
ventional fermentors.

Carlos Osamu Hokka is Associate Professor of
Chemical Engineering at the Universidade Fed-
eral de S. Carlos. He received his Master's from
Osaka University in 1976, and his Doctorate in
Biochemical Engineering from the State Univer-
sity of Campinas in 1983. His research interests
i i are in ethanol production in non-conventional con-
tinuous processes and beta-lactam antibiotic pro-
duction processes in non-conventional reactors.


second at a concentration of 60 gL-' to allow fitting of
traditional kinetic models without and with inhibition by
the product, respectively.

THEORY
We know that in a favorable environment, simple sugars
monosaccharidess) are transformed in ethanol and carbon
dioxide (CO2) by the action of yeasts as follows:


C6H1206 2 C02
(hexose) yeast (30C) (carbon dioxide)


+ 2C2H50H
(ethanol)


It can be seen that for each mole of hexose, equal molar
amounts of ethanol and carbon dioxide are produced.
In fermentation processes that synthesize primary metabo-
lites such as ethanol, cell growth and product generation take
place simultaneously. Here the cellular growth and the prod-
uct synthesis are directly related. Therefore, the ethanol pro-
duction can be predicted from the cellular growth kinetics.
From the hypothesis that the concentration of cells is a good
measure of the enzymatic system responsible for the trans-
formation of the substrate into product, it is convenient to
define the specific growth rate (4) as
1 dX (
I- (1)
X dt
where dX/dt is the variation of the cellular concentration (X)
with time (t).
Several models have been proposed to relate the specific
growth rate (g) to the limiting nutrient and inhibitor concen-
trations. A classic kinetic model relating the specific growth
rate to the limiting substrate was proposed by Monod.{1'21
Working in a continuous process, he obtained the following
relationship, also applied to batch processes:


Copyright ChE Division of ASEE 1999


Chemical Engineering Education











The equipment allowed us to perform short-duration didactic experiments
(4.5 and 3.0 hours) using low-cost raw materials.


= lmax Ks +S (2)

where umax is the maximum specific growth rate achievable
when S >> Ks, S is the concentration of growth-limiting
nutrient, and Ks is the saturation constant or the value of the
limiting nutrient concentration at which the specific growth
rate is half of Umax. It is known that this model is valid for
cultivations using low initial limiting substrate concentra-
tions (So).
In ethanol fermentation, high initial substrate concentra-
tions (So) generate high product concentrations (P) that in-
hibit cellular growth and, consequently, the production of
ethanol. Various relationships relating the effect of ethanol
concentration (P) to the specific growth rate (p) of the
organism have been reported.13-51
One type of relationship that presents similarity to the
non-competitive inhibition in enzyme kinetics has been pro-
posedI31 for modeling ethanol inhibition of Saccharamyces
cerevisiae. A term accounting for ethanol inhibition is added
to the simple Monod kinetic model, giving

S K (3)
tlmax SK p (3)
maxKs+S K,+P

where P is the ethanol concentration and Kp is the product
inhibition constant. From Eq. (3), we see that the higher the
ethanol concentration, the higher its negative effect will be
on the specific growth rate (pi).
Still, in a fermentation process, the concentration of cells
(X) and product (P) can be related to the limiting substrate
concentration (S) by the yield coefficients


dX X-X0
1Yx/s^ --- ----
x dS So-S
dP P-Po
Y dS S -S


MATERIAL AND METHODS:

0 Microorganism Saccharomyces
cerevisiae (commercial Fleischmann's
baker's yeast) was grown in two different
media for ethanol production.
3 Media Two different media have been
used in the experiments: media 1 and 2 in
ASSAYS 1 and 2, respectively.
0 Medium 1 (low substrate concentra-
tion) in gL ': commercial corn glucose
(90% w/w in glucose), 20.0; KH2PO4,
Winter 1999


5.0; MgSO4.7 H0, 0.4; yeast extract, 3.0; (NH4)2SO4, 1.8;
commercial antifoam (dilution 1:10), 5 drops; pH=4.6; sol-
vent, distilled water.
3 Medium 2 (high substrate concentration) in gL-': com-
mercial corn glucose, 66.0; KH2PO4, 5.0; MgSO4.7 H20,
0.4; yeast extract, 3.0; NH4C1, 2.5; commercial antifoan
(dilution 1:10), 5 drops; pH=4.6; solvent, distilled water.
3 Experimental Equipment The fermentors were made of
glass, adapting 1000-mL "kettle"-type (Pyrex) recipients with
flat bottoms. Fermentor lids were made of "technyl" (nylon)
and adapted to the opening of the recipients. Agitation was
accomplished using magnetic stirrers adapted to the base of
the fermentors. The temperature was controlled through so-
lenoid valves activated by bulb and capillary controllers.
Wells for controlling and monitoring temperature, and tubes
for heat transfer ("U" tubes), sampling, inoculation, and gas
exit, were made of stainless steel and connected to the lids of
the fermentors.
Devices for determining the volume of CO2 liberated by
the fermentation were constructed of PVC (poly-
vinylchloride) pipes with internal diameters of 10 cm and
heights of 100 cm and connected to the gas exits of the
fermentors, based on equipment proposed by Nilsson, et
al.16] A scheme of the experimental apparatus is shown in
Figure 1.
3 Analytical Methods Cell concentration was evaluated as
dry weight. Broth samples were centrifuged, washed twice,
and resuspended with distilled water. Aliquots were diluted
and the absorbance of the suspension was measured at 650
nm with a spectrophotometer (Micronal B-395). The concen-


Figure 1. Schematic of the experimental apparatus.










tration of the yeast suspension (X) in gL-' was related to the
absorbance according to the following calibration curve:
X=0.817 Abs, valid for X < 0,45 gL-'. Glucose concentration
(S) was determined as total reduction sugars by the colori-
metric Somogyi methodt71 using the same spectrophotom-
eter. Ethanol concentration (P) was determined by the method
of oxidation by potassium dichromate after distillation and
indirectly relating to the volume of CO2 formed.
3 Measurement of CO, Volume Produced by Fermenta-
tion According to the stoichiometry of the microbial reac-
tion, for each mole of consumed glucose, two moles of
ethanol and CO2 are produced. Therefore, if it is possible to
measure the amount of CO2 produced by the reaction, this
can be related to the concentration of ethanol present in the
fermentation broth. Such a procedure facilitates the didactic
experimental routine as ethanol analysis by the method of
oxidation by potassium dichromate or other conventional
method is time consuming, requiring the supernatant of the
centrifuged sample to be distilled and diluted for titration.
Measurements of the CO2 volume produced by the fer-
mentation process were accomplished in the following man-
ner. Initially, the pipe was filled with water from the bottom
(valve 1) up to the level yo measured from the top of the pipe,
maintaining valve 2 open. Before cultivation was started,
valves 1 and 2 were closed. The CO2 produced left the
fermentor through the gas exit tube and entered from the top
of the pipe, pushing the water level down. The water level
was monitored by a transparent tube connected vertically to
the PVC pipe. According to hydrostatic principles, in order
to maintain constant atmospheric pressure in the head of the
fermentor, the water level in a thin tube connected to the
base of the PVC pipe was controlled manually by draining
the water.
The number of moles of CO2 produced during fermenta-
tion can be calculated at any time by the ideal gas law

PatmVi PatmiD2(yi -Yo)
ni = (6)
RT 4RT
where
n, number of moles of CO2 evolved at time t=i
V, gas volume at time t=i
R gas constant [R=82.04 atm cm3/(g-mole K)]
T gas temperature [K]
D internal diameter of the PVC pipe [D=10 cm]
y, distance from the top of the PVC tube to the water level at
time t=i
yo Distance from the top of the PVC tube to the water level at
t=0

3 Experimental Procedure Two fermentation assays were
carried out at 300C using the culture media previously pre-
sented. First, the 1000-mL fermentors containing 700 mL of
culture media were sterilized in an autoclave at 121C for
thirty minutes. After sterilization, the temperature was main-


trained at 300C. Then, 100 mL of inoculum activated in a
shaker with different cell concentrations (Xo) was added,
completing 800 mL of initial working volume. The inocu-
lum is activated prior to its addition to the culture medium in
order to eliminate the adaptation stage of the microorganism
to the culture medium (the "lag" phase of the process).
Samples of 10 mL were withdrawn approximately every half
hour, right after inoculation up to complete depletion of the
glucose (end of fermentation). At the moment of sample
withdrawal, the distance between the top of the PVC pipe
and the water level (yi) was measured. Samples were divided
into two aliquots of 5 mL for analysis of the concentrations
of substrate (S), ethanol (P), and cellular mass (X). The
assays carried out at low and high initial substrate concentra-
tions (So) were designated as ASSAY 1 and ASSAY2, re-
spectively. Ethanol concentrations were determined analyti-
cally only in ASSAY 2 to be related to the CO2 volume
produced. In ASSAY 1, the ethanol concentrations were
determined indirectly by the CO2 volume produced.

RESULTS AND DISCUSSION
The results obtained in ASSAYS 1 and 2 are shown in
Table 1. Usually, fermentation processes are relatively too
time-consuming to be used in didactic experiments, but stan-
dardization of the assays provided reasonably short experi-
ments of 4.5 and 3.0 hours, respectively.
From the experimental data of ASSAY 2, it was possible
to relate the numbers of moles of ethanol and CO2 generated
by the ethanol fermentation. The number of moles of CO2
evolved (nco2) was calculated by Eq. (6) and the number of
ethanol moles (nethanol) formed was estimated as

ethanol = PVMethanol (7)
where P is the ethanol concentration in the broth (in gL '), V
is the broth's volume (in L), and Metanol is the ethanol mo-
lecular weight (Me,,,o-=46).


TABLE 1
Experimental Results Obtained in ASSAYS 1 and 2


ASSAY 1
time (h) S(gL-') x(gL ') P(gL')


ASSAY2
time(h) S(gL') x(gL') P(gL')


60.00
48.25
35.73
26.60
15.38
5.86


18.38
19.80
22.51
23.78
24.43
25.83


0.00
2.79
7.18
10.37
15.16
18.35


Chemical Engineering Education










Figure 2 illustrates the good linear relationship between
values of nethanoi and nco,. The linear regression of the ex-
perimental data resulted in a slope of 1.006 and a regression
coefficient (R2) of 0.994, showing that the theoretical-ex-
perimental methods, proposed to determine the ethanol con-


0 10 20 30 40
nethanol 102 (moles)
Figure 2. Relationship between the number of moles of
ethanol formed (netha.no and CO, evolved (nco,) during
ASSAY 2.


time (h)
Figure 3. Fit of Monod's model to the experimental
data ofASSAY 1.


Figure 4. Fit of the Aiba, et al., model3' to experimental
data ofASSAY 2.
Winter 1999


centration indirectly in the broth by the volume of CO2
produced, generated very good results.
Yield coefficients Yx/s and Y,, were determined by ex-
perimental data (Table 1) as being: Yxs=0.11--Ys=0.46
(ASSAY 1) and Yxs,=0.14--Yps=0.35 (ASSAY2), respec-
tively.
Experimental results were analyzed based on two classical
kinetic models. Initially, Monod's model (Eq. 2) was fitted
to the experimental data of ASSAY 1. The specific growth
rate (4) was estimated by Eq. (1) where dX/dt was calcu-
lated from the polynomial equation fitted to the curve X(t).
The kinetic parameters, lmx=0.32 h-' and Ks=0.63 gL-',
were obtained by nonlinear regression of g and S values
according to Eq. (2) using Marquardt's algorithm.
In the same way, the kinetic model proposed by Aiba, et
al.,131 was fitted to the experimental values of ASSAY 2,
considering the kinetic parameters, max and Ks, obtained
previously (ASSAY 1). The product inhibition constant (Kp)
was estimated as being 6.29.
Figures 3 and 4 illustrate the good fits of the kinetic
models of Monod and Aiba, et al., to the experimental values
of the assays carried out at low and high initial substrate
concentration, ASSAY 1 and ASSAY 2, respectively. A
fourth-order Runge-Kutta technique was applied to simulate
the predicted curves of S, X, and P as a function of time.
For comparison, Table 2 shows the values of the kinetic
parameters evaluated from the data of other workers."s] It can
be seen that the values of the kinetic parameters obtained by
the present work are within the range of values encountered
in the literature showing that the proposed methodology can
be of great value for educational and research purposes.

CONCLUSIONS
The equipment allowed us to perform short-duration di-
dactic experiments (4.5 and 3.0 hours) using low-cost raw
materials. The relationship of ethanol/CO2 produced by the
fermentation was very close to unity (1.006), showing the
precision of the equipment. Use of the CO2 meter in didactic
Continued on page 70.

TABLE 2
Kinetic Parameters

T max Ks K YP
(C) (h') (gL') (gL-') ()

Egamberdiev and lerusalimskiil81 28 0.31 20.6 0.39
Aiba and Shodal81 30 0.43 55 0.35
Pironti181 30 0.26 15.5 13.7 0.47
Cysewskit81 35 0.58 4.9 5.0 0.44
Bazua and Wilke'8' 35 0.64 0.24 40 0.52
Hoppe and Hansfords81 30 0.64 3.3 5.2 0.43
This work 30 0.32 0.63 6.29 0.35-0.46










S1classroom


PERMEATION OF GASES IN

ASYMMETRIC

CERAMIC MEMBRANES




CARLOS FINOL, JOAQUIN CORONAS
University of Zaragoza 50009 Zaragoza, Spain


interest in and applications of inorganic membranes have
been growing exponentially for the last fifteen years.'"
Inorganic membranes can be classified into two catego-
ries: dense and porous. Dense membranes are dominated by
palladium and its alloys (metal membranes) and solid elec-
trolyte membranes (commonly, simple or complex oxides or
oxide-solid solutions). Porous membranes are made of ox-
ides, carbon, glass, metals, zeolites, etc. The most unusual
commercial porous ceramic membranes have an asymmetric
structure consisting of a support layer (generally, alpha-
alumina) with large pores and a separation layer made of a
different material (gamma-alumina, zirconia, silica, etc.) that
controls the permeation flux.
The industrial development of inorganic membranes started
in the 1940s with the purification of nuclear fuels. This
process is based on the separation of U235Fj/U238F6 by Knudsen
diffusion. Knudsen diffusion occurs when the mean-free-
path of the gas molecules is much larger than the average
pore dimensions of a porous material through which the
molecules diffuse, and permeate will be enriched in the
molecule of the lower molecular weight.
In this short article, we will describe an easy experiment
that helps to understand the transport through ceramic mem-
branes when Knudsen diffusion occurs.

THEORY
In general, transport through porous ceramic membranes
can be related to the pore diameter, dp, according to IUPAC
definitions:121 macropores with dp > 50 nm, where basically
viscous flow (in this case, no separation of mixtures is pos-
sible) and Knudsen diffusion occur; mesopores with dp be-
tween 2 and 50 nm, where Knudsen diffusion and multilayer
diffusion/capillary condensation take place; and micropores


with dp < 2 nm, where molecular sieving effects can be
expected. Surface-multilayer diffusion and capillary con-
densation are achieved when the permeating molecule pref-
erentially absorbs on the membrane pores or condenses within
the pores due to capillary forces, respectively. Both trans-
port mechanisms, which allow the separation of mixtures
with very high selectivities, are especially important at
relatively low temperatures and with small pores,
mesopores, and even micropores.
When permeances through mesoporous membranes are
studied for permanent gas or vapors at low relative pres-
sures, the transport mechanism is controlled by the Knudsen
diffusion. Sometimes laminar flow (viscous flow at laminar
regime) can appear if the membrane has defects of macropore
size, decreasing the separation power of the membrane. Con-
sidering both the laminar and Knudsen flow contributions,
the transport equation can be written as


Carlos FInol is a PhD student in the Depart-
ment of Chemical and Environmental Engi-
neering at the University of Zaragoza. His
research interests are in using special reac-
tors for partial oxidation of hydrocarbons.


Joaquin Coronas is Assistant Professor in En-
vironmental Engineering at the University of
Zaragoza. He received his PhD in chemistry in
.. 1995 from Zaragoza University. His research
S interests are in developing membranes forsepa-
ration and ceramic membrane reactors.


@ Copyright ChE Division of ASEE 1999


Chemical Engineering Education










4 /2 r Tr2
FT = FK +FOVP = -- + --- P (1)
3 it Lir RTM 8 LTrRT
where
FT total permeance [mol/(m2 s Pa)]
FK is the Knudsen contribution
Fov factor that multiplied by P gives the laminar contribution
to the total permeance
P average pressure across the membrane [Pa]
e porosity of the membrane
r pore radius [m]
L membrane thickness [m]
T tortuosity of the medium
R gas constant [8.314 m3Pa/(mol K)]
T absolute temperature [K]
l viscosity of the gas [Pa s]

A plot of FT vs. P allows one to estimate the relative contri-
butions of both transport mechanisms: the higher the slope
of such a plot, the higher the laminar contribution and vice
versa. On the other hand, for a certain membrane material, if
the laminar contribution is not important, or P=0, it is pos-
sible to calculate the Knudsen separation factor for two
gases, A and B, by using the first term of Eq. (1)

FK (2)B
aA/B = MA (2)
FKB MA
Since Knudsen diffusion occurs when the mean-free-path
of the molecules (X) is much larger than the mean pore
radius (r) of the membrane, the Knudsen number, defined by
the equation

16 b 4 RT
Kn =- (3)
r 5 tP 2M

is a convenient dimensionless group to quickly estimate if
Knudsen diffusion will likely be the dominant transport
mechanism. When Kn>l, the molecules collide with the
pore wall much more often than with each other.

EXPERIMENTAL
The membrane used was a commercial ceramic tube 10
cm long and 0.7 cm i.d. purchased at SCT (Soci6t6 des
C6ramiques Techniques, a subsidiary of US Filter). The
ceramic tube had an asymmetric structure consisting of a
support layer of a-alumina and an inner layer of 7-
alumina with 5-nm diameter pores. The ends of the mem-
brane were non-permeable, restricting the permeation area
to a length of 5 cm.
Figure 1 shows the permeation system. The membrane
was sealed in a stainless-steel module by silicone o-rings,
and the pressure at the feed site (or retentate) was measured
by a pressure transducer. At the permeate side, the perme-
ation flux was measured using a bubble flowmeter and,
finally, an electric furnace allowed modification of the mem-


A very simple system can be used
with a commercial ceramic membrane of
y -alumina to measure permeances of single
gases (e.g., N2 and He) at several total
pressures and temperatures.

brane temperature. The permeate side was open to the atmo-
sphere, and permeate pressure was always the atmospheric
one. The permeance F in mol/(m2 s Pa) was calculated as

F= (4)
AAP
where
Q molar flux [mol/s]
A permeation area referred to the internal side of the ceramic
tube (1.lx10-m)
AP pressure drop [bar] between retentate and permeate sides


RESULTS AND DISCUSSION
The variation of He and N2 single-gas permeances vs. the
total pressure (calculated as [P, + P2]/2, where P, and P2 are
the pressures at the permeate and retentate sides, respec-
tively) through an alumina membrane was studied at 298
and 473K. Figure 2 shows that the He permeances at 298 and
473K did not change with total pressure, and the same can
be stated for the N, permeances (Figure 3), at least in the
pressure range tested, which means that the viscous con-
tribution is not important, and in the equation (Eq. 1) that
governs transport through the membrane, the second term


Pressure
measurement


Needle
valve 9


Temperature
control




Membrane
module










SBubble
flowmeter


Figure 1. Experimental system.


Winter 1999










(Fov) can be neglected.
The Knudsen number calculated using Eq. (3) for He and
N2 in the temperature and pressure ranges tested is 44-124
and 14-30, respectively (it decreases with increasing pres-
sure and decreasing temperature), i.e., in this case, this num-
ber is always higher than one, which means (as seen above)
that Knudsen diffusion will dominate over viscous flow.
Moreover, permeances decreased with increasing tempera-

40-

35 298 K

S30-

S25-

20
A20-
2 '- -u B -- -- -- --p- a
S15-
: 473 K
10-

5-

0O ..


1.00


1.25 1.50
Total Pressure [bar]


1.75 2.00


Figure 2. He permeance as a function of total pressure
at 298 and 473 K.


18
16- 298 K
14-
12-
10-




48-
2-
0
1.00 1.25 1.50 1.75 2.00
Total Pressure [bar]

Figure 3. N2 permeance as a function of total pressure
at 298 and 473K.


TABLE 1
He and N2 permeances and aHe / N, at 298 and 473 K

Permeance
Temperature [mol/(m2sPa)] cHe / N,
[K] He N, Theoretical Experimental

298 37.2x10-6 15.3x10-6 2.65 2.43
473 17.2x10-' 6.81x10-6 2.65 2.53


ture, as expected after analyzing Eq. (1). Table 1 lists the
averages of N2 and He permeances. From these permeances,
the experimental separation factor He/ N2 can be calculated
as the ratio of He and N, permeances. The XHe/N, is 2.43 at
298K and 2.53 at 473, converging with increasing tempera-
ture to the theoretical value of 2.65 calculated by using Eq.
(2). Probably, N, absorbs weakly, but stronger than He at
low temperature, contributing to the total N2 flux. This con-
tribution will decrease at higher temperature, increasing the
separation factor slightly.
Finally, the ratio of pure gas permeances is also known as
the ideal separation factor, and very often ideal separation
factors do not allow extrapolation to mixtures. To carry out
separation of mixtures of compounds with selectivities higher
than those of Knudsen, microporous membranes (silica, zeo-
lite, carbon) can be employed. Zeolite membranes have re-
cently received increasing attention (because of their great
potential for use in industrial applications, especially under
high temperature conditions) and have been used for gas or
liquid-phase separations of non-adsorbing compounds, 31 or-
ganic/organic,[41 permanent gas/vapor,'[5 and water or polar
molecules/organic161 mixtures. Although molecular sieving
is mainly invoked to justify the separation of permanent gas
mixtures with significant differences in their molecular size,
the majority of separations using zeolite membranes can be
explained in terms of surface diffusion and, sometimes,
capillary condensation.

CONCLUSIONS

A very simple system can be used with a commercial
ceramic membrane of Y -alumina to measure permeances of
single gases (e.g., N2 and He) at several total pressures and
temperatures. The experiment gives the student a complete
and easy analysis of the permeation measurements in terms
of Knudsen transport.

REFERENCES
1. Hsieh, H.P., Inorganic Membranes for Separation and Re-
action, Elsevier Science B.V., Amsterdam (1996)
2. Keizer, K., R.J.R. Uhlhorn, and T.J. Burggraaf, "Gas Sepa-
ration Using Inorganic Membranes," in R.D. Noble and S.A.
Stern (eds) Membrane Separation Technology: Principles
and Applications, Elsevier Science B.V., Amsterdam (1995)
3. Bakker, W.J.W., F. Kapteijn, J. Poppe, and J.A. Moulijn,
"Permeation Characteristics of a Metal-Supported Silicalite-
1 Zeolite Membrane," J. Membrane. Sci., 117, 57 (1996)
4. Coronas, J., J.L. Falconer, and R.D. Noble, "Characteriza-
tion and Permeation Properties of ZSM-5 Membranes,"
AIChE J., 43, 1797, (1997)
5. Piera, E., A. Giroir-Fendler, H. Moueddeb, J.A. Dalmon, J.
Coronas, M. Men6ndez, and J. Santamaria, "Separation of
Alcohols and Alcohols/O, Mixtures Using Zeolite MFI Mem-
branes," J. Membrane Sci., 142, 97 (1998)
6. Kita, H., K. Horii, Y. Ohtoshi, and K. Okamoto, "Synthesis
of a Zeolite NaA Membrane for Pervaporation of Water/
Organic Liquid Mixtures," J. Mater. Sci. Lett., 14, 206 (1995)

Chemical Engineering Education











Discontinuities
Continued from page 25.

(cA/Y intrinsic volume averaged concentration of species A,
mole/m3
cAy spatial deviation concentration, mole/m3
c constant pressure heat capacity, J/kgK
c speed of sound, m/s
D dispersion tensor, m2/s
F force acting on a particle, N
g gravity vector, m/s2
k thermal conductivity, J/msK
LP viscous length, m
Lp inertial length, m
L length associated with the pressure change, Ap, m
M u /c, Mach number
m mass of a particle, kg
n number of moles
n unit normal vector
ny unit normal vector directed from the y phase toward
the a phase
p pressure, N/m2
p., pressure for an incompressible flow, N/m2
r position vector, m
R gas constant, Nm/mole K
Re puoL,/p, Reynolds number
RA molar rate of production of species A owing to homoge-
neous reaction, mole/m3s
RAY molar rate of production of species A in the y phase
owing to homogeneous reaction, mole/m3s
T temperature, K
Tm temperature for an incompressible flow, K
t time, s
t(W stress vector, N/m2
T stress tensor, N/m2
u characteristic velocity, m/s
v mass average velocity vector, m/s
vm velocity for an incompressible flow, m/s
vA velocity of species A, m/s
VAy velocity of species A in the y phase, m/s
vy mass average velocity in the y phase, m/s

(vy) superficial mass average velocity in the y phase, m/s

(v7Y intrinsic mass average velocity in the y -phase, m/s
vy spatial deviation velocity, m/s
V volume, m3
*/ averaging volume, m3
Vy (x, t) volume of the y phase contained in the averaging
volume, m3
w arbitrary velocity, m/s
x position vector locating the centroid of an averaging
volume, m


Greek Letters
Ey V,(x,t)/V, volume fraction of the y phase
k unit vector
p fluid viscosity, Ns/m'
p mass density, kg/mi

REFERENCES
1. Whitaker, S., "The Development of Fluid Mechanics in
Chemical Engineering," p. 47 in One Hundred Years of
Chemical Engineering, edited by N. Peppas, Kluwer Aca-
demic Publishers, Dordrecht (1989)
2. Gresho, P.M., and R.L. Sani, Incompressible Flow and the
Finite Element Method, John Wiley & Sons Inc., New York,
NY (1998)
3. Birkhoff, G., Hydrodynamics: A Study in Logic, Fact, and
Similitude, Princeton University Press, Princeton, NJ (1960)
4. Whitaker, S., Fundamental Principles of Heat Transfer,
R.E. Krieger Publishing Company, Malabar, FL (1983)
5. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport
Phenomena, John Wiley & Sons, Inc., New York, NY (1960)
6. Whitaker, S., "A Simple Geometrical Derivation of the Spa-
tial Averaging Theorem," Chem. Eng. Ed., 19, 18 (1985)
7. Truesdell, C., Essays in the History of Mechanics, Springer-
Verlag, New York, NY (1968)
8. Aris, R., Vectors, Tensors, and the Basic Equations of Fluid
Mechanics, Prentice-Hall, Englewood Cliffs, NJ (1962)
9. Whitaker, S., Introduction to Fluid Mechanics, R.E. Krieger
Publishing Company, Malabar, FL (1981)
10. Pucciani, O.F., and J. Hamel, Langue et Langage, 4th ed.,
Holt, Rinehart, and Winston, New York, NY (1983)
11. Stein, S.K., and A. Barcellos, Calculus and Analytic Geom-
etry, McGraw-Hill, Inc., New York, NY (1992)
12. Majda, A., Compressible Fluid Flow and Systems of Conser-
vation Laws in Several Space Variables, Springer-Verlag,
New York, NY (1984)
13. Whitaker, S., "Levels of Simplification: The Use of Assump-
tions, Restrictions, and Constraints in Engineering Analy-
sis," Chem. Eng. Ed., 22, 104 (1988)
14. Whitaker, S., "Laws of Continuum Physics for Single-Phase,
Single-Component Systems," in Handbook of Multiphase
Systems, edited by G. Hetsroni, Hemisphere Publishing Cor-
poration, New York, NY (1982)
15. Anderson, T.B., and R. Jackson, "A Fluid Mechanical De-
scription of Fluidized Beds," Ind. Eng. Chem. Fund., 6, 527
(1967)
16. Marle, C.M., "Ecoulements monophasique en Milieu Poreux,"
Rev. Inst. Franqais du Petrole, 22(10), 1471 (1967)
17. Slattery, J.C., "Flow ofViscoelastic Fluids Through Porous
Media," AIChE J., 13, 1066 (1967)
18. Whitaker, S., "Diffusion and Dispersion in Porous Media,"
AIChE J., 13,420 (1967)
19. Whitaker, S., The Method of Volume Averaging, Kluwer
Academic Publishers, Dordrecht (1999)
20. Gray, W.G., "A Derivation of the Equations for Multiphase
Transport," Chem. Eng. Sci., 30, 229 (1975)
21. Carbonell, R.G., and S. Whitaker, "Dispersion in Pulsed
Systems II: Theoretical Developments for Passive Disper-
sion in Porous Media," Chem. Eng. Sci., 38, 1795 (1983)
22. Paine, M.A., R.G. Carbonell, and S. Whitaker, "Dispersion
in Pulsed Systems I: Heterogeneous Reaction and Revers-
ible Adsorption in Capillary Tubes," Chem. Eng. Sci., 38,
1781 (1983)
23. Quintard, M., and S. Whitaker, "Convection, Dispersion,
and Interfacial Transport of Contaminants: Homogeneous
Porous Media," Adv. Water Resour., 17, 221 (1994) O


Winter 1999










MR, -!curriculum


A JOINT

CHEMICAL/ELECTRICAL

ENGINEERING COURSE IN

ADVANCED DIGITAL PROCESS CONTROL


JOSEPH J. FEELEY, Louis L. EDWARDS
University of Idaho Moscow, ID 83843


his article describes a course in the digital control of
industrial processes jointly offered by the Depart-
ments of Chemical Engineering and Electrical Engi-
neering at the University of Idaho. The course grew out of a
perceived need for engineers with the ability to design multi-
input multi-output (MIMO) digital controllers for industrial
processes, especially in the pulp and paper industry. Evi-
dence for this need lies in the numerous opportunities for
improved process performance offered by advanced multi-
variable digital control techniques.
The primary goals of the course are to
Teach students how to design MIMO controllers for
industrial processes
Give students laboratory experience with represen-
tative industrial processes and control systems
Introduce students to the use of state-of-the-art
computer-aided design software for control-system
design
Provide an opportunity for chemical and electrical
engineering students to work together in interdisci-
plinary teams.

The following sections describe how the course attempts to
achieve these goals, highlights some of the distinctive fea-
tures of the course, and discusses some plans for the future.

COURSE DESCRIPTION
Digital Process Control is a three semester-credit hour
course jointly offered by the chemical and electrical engi-
neering departments. It is a required course for chemical
engineering (ChemE) seniors and an elective course that


satisfies a breadth requirement in the control-systems area
for electrical engineering (EE) seniors. ChemE students have
previously had a course in chemical process control using
classical control-system design methods. The prerequisite
for EE students is a junior-level course in signals and sys-
tems. Some EE students, however, have had a prior course in
classical control-system design as applied, primarily, to
electro-mechanical systems. All of the EE students and many
of the ChemE students have had a prior course in linear
algebra or have had some exposure to vectors and matrices
in previous courses.
Computer-aided design is an essential part of the course,
and all students have had prior experience with Matlab,'" the
main computer program used in the course. The course is
taken by about twenty-five ChemE students and fifteen EE
students each year and is jointly taught by a faculty member
from each department.

Joseph J. Feeley is Associate Professor of
Electrical Engineering at the University of Idaho.
He received his BS from the New Jersey Insti-
tute of Technology, and his MS and PhD de-
grees from the University of Idaho, all in electri-
cal engineering. He is currently writing a text-
book on digital process control.



Louis L. Edwards has been Professor of
Chemical Engineering at the University of
Idaho since 1971. He teaches process con-
trol and his research focuses on simulation of
pulp and paper processes. He has published
over fifty papers and has consulted with more
than thirty pulp- and paper-related compa-
nies worldwide.


Copyright ChE Division of ASEE 1999


Chemical Engineering Education










The course is organized around three major laboratory
experiments of progressive difficulty. Each of the experi-
ments is a laboratory-scale simulation representative of a
typical industrial process. The lecture material and the
laboratories are coordinated so that when a lecture se-
quence describing a new design technique is completed,
it is followed immediately by a laboratory experiment
using that technique.
The choice of an appropriate design method, or methods,
for this class is not easy. Current practice in the process
control community is focused
on some form of internal
model control (IMC),'1231 of-
ten model predictive control
(MPC) or dynamic matrix Stabilization
control (DMC). On the other tank
hand, methods based on the
linear quadratic regulator
(LQR)[41 and its extensions to
incorporate state estimation Bii
via linear quadratic estimation Vent
(LQE), recover stability mar-
gins via loop-transfer recov- dp cell
ery (LTR), and design for ro-
bustness directly through HI Measured Digit
and p-synthesis[5' have been level Contro
more popular in the electro-
mechanical control commu- Figure 1. Experiment 2:
nity. The distinctions are consisting of
not, of course, that clear,
and there are many successful LQR applications in the
process industry[61 and many PID applications in the
electro-mechanical area.I71
A compromise, of sorts, has been reached by using the
discrete-time version of the linear quadratic regulator (DLQR)
as the main design tool used and pointing out the relation-
ships among DLQR, DLQG, PID, and MPC as the occasions
arise. DLQR is used because of its generality, the availabil-
ity of Matlab design tools and the familiarity of most of the
students with Matlab, and the simplicity of the resulting
control algorithm. Once a suitable process model is devel-
oped, the design can be efficiently completed in an interac-
tive computer session using Matlab, Simulink,l8' and the
Control Systems Toolbox.[9] Students quickly learn the rela-
tionship between performance index weightings on the one
hand and state performance and control effort on the other
hand, and are able to find control gains that yield satis-
factory system performance.
After an introduction to sampling and development of a
discrete-time model, the DLQR method is introduced through
a scalar single-input single-output (SISO) example. The ex-
ample concerns a simple nonlinear process model associated
with the liquid-level control problem addressed in the first
Winter 1999


laboratory experiment. The model is linearized at an operat-
ing point and discretized at a selected sampling rate to form
the basis for the controller design. The control problem is
presented as an optimization problem, trading off control
effort for state performance. The dynamic programming ap-
proach is used to develop the Riccati equation, and its steady-
state solution is used to find the optimal DLQR gain. By
beginning with this scalar example, students get a good
understanding of the basic DLQR approach before moving
to the multivariable problem, with the additional complexity
of vector-matrix notation,
studied later. The effect of
state variable feedback in
changing the system time
constant is also clearly seen
in this simple example. Stu-
dents solve this control prob-
lem by hand and then use
Matlab to verify their calcu-
ctor lations. Simulink is then
used to simulate the continu-
Outlet flow
Outt fw ous-time nonlinear process
l model under DLQR control.

Actuator I Experiment 1

Control signal The first experiment is a
simple water-level control


a simulated bioreactor system
two cascaded tanks.


problem that illustrates the
application of DLQR state


variable feedback control
methods to a first-order system. It allows the chemical engi-
neering students to become familiar with the notion of sam-
pling and digital-control-system hardware and familiarizes
the electrical engineering students with the relatively long
time scale and nonlinearity of process systems. The relation-
ship of scalar DLQR control to traditional proportional con-
trol is pointed out, and the need for integral augmentation to
eliminate steady-state error is clearly illustrated. The inte-
grator-augmented DLQR controller is then shown to be
equivalent to a traditional PI controller.
Experiment 2
The second experiment involves a second-order process
consisting of two cascaded tanks simulating a bioreactor
system. The objective of this experiment is to control the
level in the second tank, in spite of disturbances caused by a
fluctuating flow from the first tank into the bottom of the
second tank. Both tank levels are measured so that full state-
feedback can be used. The control input is the position of an
outlet valve on the second tank. A schematic of the experi-
mental setup is shown in Figure 1. The system is modeled as
a two-state discrete-time SISO process and the vector-ma-
trix form of the state and output equations is introduced. The
controller for this process is designed using the integrator-


al
ller










augmented DLQR method"l0 extended to accommodate level
set-point changes and zero steady-state error. The necessity
of using vector-matrix notation and computer-aided design
tools begins to become evident to students when solving this
relatively simple second-order problem. The relationship of
this simple DLQR controller to the classical PID controller
is also examined.
Experiment 3
The third experiment requires the design of a two-input/
two-output controller to simultaneously control the water
level and total head in a laboratory-scale model of a paper
machine head box. A schematic diagram of the process is
shown in Figure 2 and details of its construction are avail-
able from the second author. Zero steady-state error and set-


point control are required
for both controlled vari-
ables. This is the most chal-
lenging of the control prob-
lems and involves four
states, two inputs, and two
outputs. Nevertheless, most
students are now familiar
with the DLQR design-it-
eration procedure and are
able to complete the con-
troller design with little dif-
ficulty. In the laboratory,
students are encouraged to
attempt this MIMO control
manually to fully appreci-
ate the difficulty of the task
and the power of the DLQR
design method. The control
state feedback basis.


Figure 2. Experiment 3: a la
paper- machine


design is carried out on a full


COURSE STRUCTURE
Daily homework assignments are focused on the upcom-
ing laboratory experiment. Each assignment addresses a dif-
ferent phase of the design effort, with the last assignment
before a given lab producing the controller algorithm re-
quired for that experiment.
Student performance is assessed through the homework,
two one-hour tests, a two-hour final exam, and reports on
each of the laboratory experiments. The lab reports are a
crucial part of the course requirements and account for about
one-third of the course grade. Lab groups generally consist
of three students, two from chemical engineering and one
from electrical engineering. Students pick their own lab
partners, but "bonus points" are awarded to interdisciplinary
lab groups to encourage interaction between the electrical
and chemical engineering students. Individual lab reports in
the form of a technical memo with supporting appendices
are required of each student.


STUDENT REACTION
Student reaction to the course has been quite favorable.
The ability to test controller designs in the laboratory is men-
tioned by most students as the most positive feature of the
course. Many students comment on the power of the com-
puter-aided design tools to facilitate control system design. A
smaller number view the computer approach as "magic."

DISTINCTIVE FEATURES OF THE COURSE
This course includes a number of distinctive features. The
most obvious is the combination of chemical and electrical
engineering students in a joint class. This helps bridge the
disciplinary gap and teaches the students how to work with
specialists in another field on challenging interdisciplinary
problems. The student-
to-student learning pro-
.... cess generally involves
Air outlet flow the ChemE students
teaching the EE students
about process dynamics
and modeling, while the
Water EE students take the lead
level
on computer and digital
controller topics. The fact
Water outlet flow that the class is jointly
taught by a faculty mem-
dp Total head ber from each department
measurement also gives the students
complementary views of
boratory-scale model of a a given control or mod-
e headbox. eling issue. Both instruc-


tors attend each class,
with the EE instructor primarily addressing the control theory
and design topics and the ChemE instructor focusing on
process modeling and the relationship of the labs to actual
industrial processes.
A second key feature of the course is that it approaches the
digital control-design problem from the "direct digital de-
sign" perspective and uses the DLQR design method as the
basic design approach. This approach is taken for a number
of reasons. First, the DLQR method approaches controller
design as an optimization problem, trading off control effort
for state performance. This is a crucial concept that is often
overlooked, or addressed after the fact, in other design meth-
ods. Second, currently available CAD tools, such as Matlab,
Simulink, and associated toolboxes, simplify computations
so that students can focus on design objectives and not
become bogged down in mathematical complexities. Stu-
dents are taught the advantages of testing controller designs
by nonlinear simulations on a desktop computer before try-
ing them out on the actual process. Third, the DLQR ap-
proach seems more direct in that a good design can be
accomplished almost entirely in the discrete-time domain,
Chemical Engineering Education


Air inlet flow



Controller


| Water
inlet flow

Level measurement


Air







dp


I
r/










avoiding, for the most part, the complexities introduced by
transform mathematics. Finally, we realize that students com-
pleting this course will not be experts in DLQR design
methods. They will, however, have had a practical introduc-
tion to the topic that will alert them to the power and poten-
tial of advanced control techniques. They will also be aware
of some of the relationships among three control methods
representative of the electro-mechanical, process, and tradi-
tional control communities, i.e., DLQR, MPC, and PID.
The final important aspect of the course is the laboratories
and their integration with the lecture material. In a sense, the
course is driven by the laboratory experiments and the stu-
dents are motivated to master the lecture material so that
they can perform the laboratory experiments satisfactorily.
Homework assignments are carefully chosen so that satisfac-
tory completion of the last assignment before a lab should fully
prepare the student to carry out the experiment properly.

PLANS FOR FUTURE DEVELOPMENT
A number of plans are underway for further development
of the course. New laboratory experiments are always under
consideration. We plan to include an experiment involving
system identification to develop an empirical model of the
headbox process next year. Longer-range plans include add-
ing a third input and a third output to the headbox experi-
ment and developing a new experiment using a distillation
column as the process to be controlled. The lecture material
is also being continually refined and updated to find more
logical and consistent ways to introduce this relatively ad-
vanced material to students at this stage in their development.
More emphasis will be given to IMC methods such as MPC
and DMC in future versions of the course. The course notes
currently being used as the course text are being readied for
publication and should be submitted to a publisher this year.

ACKNOWLEDGMENT
The authors are very grateful to Barry King for construct-
ing and maintaining the laboratory experiments and for his
assistance with the students in the laboratory.

REFERENCES
1. Matlab: High Performance Numeric Computation and Visu-
alization Software, The Mathworks, Inc., Natick, MA (1993)
2. Garcia, C.E., D.M. Prett, and M. Morari, "Model Predictive
Control: Theory and Practice, A Survey," Automatica, 25,
335 (1989)
3. Morari, M., and E. Zafiriou, Robust Process Control, PTR
Prentice Hall, Englewood Cliffs, NJ (1989)
4. Kuo, B.C., Digital Control Systems, Holt, Rinehart, and
Winston, Inc., New York, NY (1980)
5. Zhou, K., J.C. Doyle, and K. Glover, Robust and Optimal
Control, Prentice Hall, Upper Saddle River, NJ (1996)
6. Ramirez, W.F., Process Identification and Control, Academic
Press, Boston, MA (1994)
7. Franklin, G.F., J.D. Powell, and A. Emami-Naeini, Feed-
back Control of Dynamic Systems, Addison-Wesley, Read-
Winter 1999


ing, MA (1994)
8. Simulink: Dynamic System Simulation Software, The
Mathworks, Inc., Natick, MA (1993)
9. Control Systems Toolbox, The Mathworks, Inc., Natick, MA
(1993)
10. Kwakernaak, H. and R. Sivan, Linear Optimal Control Sys-
tems, Wiley-Interscience, New York, NY (1972). D



BOOK REVIEW: Numeric Computation
Continued from page II.
The next three chapters are devoted to function interpola-
tion and differentiation (Chapter 6), numerical integration
(Chapter 7), and approximation of functions, lines, and sur-
faces (Chapter 8). In the final chapters, the author discusses
the solution of ordinary differential equations and initial
value problems (Chapter 9), boundary-value, eigenvalue,
and free-boundary problems (Chapter 10), and finite-differ-
ence methods for partial differential equations (Chapter 11).
Finally, the author provides four appendices. In Appendix
A, he provides a brief calculus refresher, with a table of
various series approximations to a variety of functions. In
Appendix B, he discusses orthogonal polynomials. Again,
he includes a number of tables of various orthogonal polyno-
mials, including Legendre, Chebyshev, Jacobi, Radau,
Lobatto Laguerre, Hermite, and Gram, and their properties.
Developing the ability to proficiently use the computer
and interface with the operating system is one of the main
hurdles often encountered by students in a numerical meth-
ods course. Appendix C provides this essential background
information, including an introduction to the c-shell, to the
unix file system, to the vi editor, and to the compilation and
linking of typical programs. Finally, Appendix D provides a
Fortran primer, while an index to the publicly available
programs discussed above is provided in Appendix E.
Although they are covered briefly in Chapter 10, I would
have liked additional coverage of weighted residual tech-
niques. These methods provide a powerful tool for the solu-
tion of ordinary differential equations and provide the basis
of finite element methods for the solution of partial differen-
tial equations. In devoting only seventeen pages to this topic,
the author has given the students only a brief glimpse of the
power of such methods and may leave students thinking that
they are not as useful as others that are covered in much
more detail. Further, as is often the case, the coverage of
techniques for solving partial differential equations is quite
limited. I assume that this is because the author expects this
materials to be covered in a later course.
In general, Pozrikidis has met his goals and has produced a
usable text in which he covers the fundamentals of numeri-
cal methods, while at the same time enabling the reader to
understand how to use the various techniques to solve physi-
cal problems in science and engineering. 0
65










[^] 1 laboratory


INTRODUCING

PROCESS-DESIGN ELEMENTS

IN THE UNIT OPERATIONS LAB


CHRISTINE L. MCCALLUM, L. ANTONIO ESTiVEZ*
Cornell University Ithaca, NY 14853-5201


ne of the main problems of the classical chemical
engineering or unit operations laboratory is the dan-
ger of its becoming repetitive and boring. This pa-
per will address that problem and show how a simple labora-
tory experiment can be turned into an interesting experience
for both the students and the instructor by giving it a slight
flavor of process design. This was accomplished by adding
to the requirements of the final lab report the solution of a
design problem related to the experiment done in the lab.
The problem combined scale-up elements and changes to the
material's chemical nature, or to the physical characteristics
of the laboratory setup.
This paper describes our approach to a classical stripping
experiment where air was used to remove toluene from
water in a packed column. Although the approach is de-
scribed in the context of this specific experiment, it can be
applied to any classical experiment.

THE UNIT OPERATIONS LAB AT CORNELL
To better understand the approach presented in this paper,
we will first give a brief description of how our lab is
organized. The senior class is divided into five groups (Mon-
day, Tuesday, Wednesday, Thursday, and Friday), with four
teams in each group. Six experiments are carried out in four
rotations. During the first week, all groups work on a short,
introductory experiment. Then the first rotation starts on the
second week and lasts four weeks, as do all rotations: a
rotation involves preparation, experiments (2 weeks), and a
report. (The report on one rotation and the preparation in
next rotation may overlap.) Two shorter experiments are
done in one rotation. The stripping experiment (described
in this paper) is done in one full rotation. The Unit Op-
erations Laboratory at Cornell has a page on the world

* Present Address: University of Puerto Rico, Mayagiiez, Puerto
Rico 00681-9046


wide web, currently at
http://speedrcr.cheme.comell.edu/Uo/
that can be consulted for additional details on the course
organization.

ABOUT THE EXPERIMENT
The essence of this experiment is to remove a trace amount
of pollutant (a volatile organic compound) from water by
stripping it with air. The experimental apparatus and basic
operating procedure are as follows.
Apparatus In this experiment, air and a toluene-con-
taminated water stream are brought into counter-current con-
tact in a stripping column. The experimental apparatus is
shown in Figure 1. The column is 15 cm (0.5') in diameter,
packed with 5/8" plastic Pall rings. The approximate pack-
ing height is 1.15 m (3.75'). A 0.5-m3 (135-gal) tank with an
electric mixer is used to homogenize the feed supply. Rota-
meters are used to read liquid and air-flow rates, and the
thermocouples are used to read tank and column tempera-
tures. The liquid feed flows by gravity at low flow rates and
is driven by a centrifugal pump at higher flow rates. Samples
are collected at the water-stream inlet and outlet and ana-
lyzed using a Gow-Mac gas chromatograph equip-ped with
a flame ionization detector and Spectra-Physics integrator.

Christine L. McCallum received her MS in chemical engineering at Cornell
University, where she conducted research in the area of molecular simula-
tion and thermodynamics. She received her BS from Bucknell University,
also in chemical engineering. She currently works for Intel Corporation in
Phoenix, Arizona.
L. Antonio Estevez received his PhD from the University of California,
Davis. He holds a BS degree from the University of Santiago, Chile, and a
MS from the Central University of Venezuela. He has been on the faculty of
the University of Puerto Rico since 1987, having previously taught at the
University of Santiago, Chile, and the Simon Bolivar University in Caracas.
He was on sabbatical leave at Cornell University during the academic year
1996-97. His research interests are supercritical fluids, separation pro-
cesses, and multiphase reactors.


Copyright ChEDivision ofASEE 1999


Chemical Engineering Education









Procedure The tolu-
ene/water feed is prepared
at least one hour before the
lab begins by filling the
feed tank with water and
adding an amount of tolu-
ene to produce the desired
concentration. The mixer
runs until the lab period
begins. After verifying
proper valve settings, air
is allowed to flow through
the system by opening
valve V4, and its flow rate
is regulated with valve V3.
The water is then intro-
duced by opening valve
V2. A water seal is main-
tained by keeping the wa-
ter depth at the bottom of
the column within 3 to 5
cm. Steady-state condi-
tions are reached after
about five minutes. Water
samples are collected at
ports S1 and S2 (drain
pipe) in glass vials. Corks
wrapped with aluminum
foil are used to seal the
samples, which are then la-
beled and stored on ice.
The samples are analyzed
using the gas chromato-
graph. Pressure drop,
gauge pressure at the bot-
tom of the column, and
column temperature are
also measured.

THE
NEW APPROACH
As explained previously,
the students have four
weeks to complete the ex-
periment and submit a fi-
nal report. In the first
week, the TA meets with
the group for a brief ex-
planation of the experi-
ment's objectives and op-
eration of the apparatus. A
memo with the require-
ments is distributed at this


This paper presents a simple way to
enhance the learning experience in the unit
operations laboratory by adding a
related design problem to the
lab requirements.


Figure 1. Apparatus for the stripping experiment.



TABLE 1
Design Problem

Clearwaters and Associates is a company that produces bottled
drinking water on a large scale. In their process they chlorinate
their source using standard processes, and therefore the level of
chloroform (and of other trihalomethanes, THMs) goes beyond
specifications. They are considering reducing the THM level by
stripping with air in a packed column. The water after chlorina-
tion may contain up to 400 ppb (by mole) of THM, 30% of
which is chloroform, 70% bromodichloromethane, and negli-
gible amounts of bromoform and dibromochloromethane. Their
typical daily production is about 60,000 5-gal bottles. As the
engineering team of Clearwaters and Associates your assign-
ment is to design the THM-removal system using stripping.
They plan to use a column packed with 5/8" plastic Pall rings.
The final THM level must be less than 50 ppb (by mole).
Typically, the operation will take place at 77F.


meeting. The students are
also referred to the
experiment's reference
manual, which is available
on the world wide web at
http://cheme. cornell.edu/
-lae/432/stripping.htm.
Using the traditional ap-
proach, the students would
figure out what variables to
measure to determine the
overall mass-transfer coef-
ficient and would probably
present some plots showing
the effect of, say, flow rates
on mass-transfer coefficient
and on the height of a trans-
fer unit. With this new ap-
proach, the students must
also solve a design problem
and include the solution in
their final lab report. In this
particular design problem,
the students use the infor-
mation collected in the lab
to compute the height of the
column. In addition to scale-
up, they must also consider
a difference in the chemical
composition/nature of the
inlet stream or the size/type
of the column packing. The
students need to determine
whether or not that differ-
ence is important, and if so,
how to account for it.

A SAMPLE
DESIGN PROBLEM
Table 1 shows a sample
design problem given to the
students in one of the fall
1996 rotations. Each team
was responsible for provid-
ing a final design for the
company. This task required
determination of the pack-
ing height and the column
diameter. The students were
given guidelines to help
achieve this goal, and in this
particular case they were re-
ferred to Veldzquez and


Winter 1999










Estdvez"'l to find the needed THM's properties. Addition-
ally, they were asked to submit three pieces of documenta-
tion: a pre-experimentation memo, a preliminary-calcula-
tions memo, and a final report. Each of these parts had a
specific objective and helped break up the student workload
during the four-week lab period. The three parts are dis-
cussed in detail in the following section.

REPORTING REQUIREMENTS

Pre-Experimentation Memo
To help meet the final objectives of the design problem,
the students were asked to first prepare an experimental
plan, including a factorial design of experiments, to measure
the mass-transfer coefficients and the height of a transfer
unit in the laboratory. The plan was to be submitted to the
TA in a short memo at least two days before the day the
experiment was to begin. The next day, a copy of the memo
would be returned to the team with comments.
Some of the questions to be answered in this memo were:
What is the problem being addressed? What variables will
be measured? Which flow rates will be run? How much
toluene is needed to prepare the feed solution? What proce-
dure will be followed? What calculations will be performed?
How will the results be presented?
The pre-experimentation memo was an important part of
the lab, as it required the team to become familiar with the
apparatus and to carefully plan the experiment. The group
also needed to figure out how to do the necessary calcula-
tions ahead of time so they could record all the relevant data.

Preliminary Calculations
On the second lab day, each team had to submit calcula-
tions of the height of a transfer unit and the mass-transfer
coefficient for each experimental point taken on the first lab
day. Results had to be presented in a short memo (2 or 3
pages) along with a table showing L and V in lab units, and
G, Gy, m, S, xx/Xb, Nox, Ho,, and Ka in SI units. A brief
description of the procedure that would be used to solve the
design problem also had to be included in the memo.
In the preliminary-calculations memo, the students had to
use the actual data they collected on the first lab day. At this
point, they will realize if they are doing something wrong,
or if they have failed to measure a variable they need
(e.g., atmospheric pressure, column temperature). This
also takes part of the load of writing the lab report from
the last week and spreads out report writing over two
weeks rather than just one.
Final Report
In addition to the traditional final lab report components,
the students were asked to use a correlation for their data of
the form


Kxa= GWGG (1)

where W, a, and are constants. To find these constants, they
used the linear-regression option of any standard spread-
sheet, with three contiguous columns containing log Kxa, log
Gx, and log Gy. Each team also had to prepare a plot of the
height of an overall transfer unit and the overall mass-trans-
fer coefficient versus the liquid mass velocity (liquid mass
flow rate per unit cross-sectional area) at various gas mass
velocities. Experimental results, and the values predicted by
Eq. (1) must also be included. Finally, the students had to
prepare a parity plot with their results. This is a plot of the
experimental Kxa (or Hox) versus the values computed by
Eq. (1). Included as a reference were the diagonal line and
constant-error lines (e.g., +5% and -5% lines).
The solution to the design problem was to be discussed in
a separate section of the laboratory report. Using the correla-
tions proposed in the earlier section of the report, duly cor-
rected for chloroform and bromodichloromethane, the re-
quired packing height and column diameter (for a gas-flow
rate equivalent to 1.6 times the minimum gas-flow rate) had
to be computed. In the calculation of the diameter, a gas-
flow rate equivalent to 50% of the flooding value should be
used. After performing all the calculations, the team had to
report on the recommended design and the criterion or crite-
ria they used.
A company-type format of about ten pages was suggested
for the final report. This required the students to evaluate the
process by which they used actual lab data to solve an
industrial design problem and to justify their final design
proposal.


RESULTS OF THE EXPERIMENTS

We found it interesting to present the results of this experi-
ment. Although this is not directly related to the main objec-
tive of the paper, we realized that correlations for these types
of systems are less abundant than originally thought. To
present the results in a compact manner, all the data col-
lected during the course of the semester (four rotations, 219
points) were consolidated in a single spreadsheet. Then, a
correlation of the form suggested above in Eq. (1) was
obtained. The result was

Ka=8.7G0778G -0.05 (2)

The correlation coefficient (R2) for this fit was 0.908. With
Eq. (2), a parity plot was prepared to show the goodness of
fit for the correlation obtained (shown in Figure 2). In a
parity plot, the points should align along the diagonal, which
represents the perfect fit. Although the exponent for Gy is
small, its standard error determined that it is statistically
significant.


Chemical Engineering Education










ON THE SOLUTION
OF THE DESIGN
PROBLEM
The design problem
presented here requires
the computation of mass-
transfer coefficients for
trihalomethanes in a sys-
tem similar to the one
used in the lab. The cal-
culation of the packing
height and column diam-
eter is relatively straight-
forward. If the data ob-
tained for toluene are
used, however, the stu-
dents have to determine
how to correct for the fact
that the solutes in the lab
and in the problem are
different. To stimulate
creativity, no guidelines


10,000 i _--

_




*

t 1,000


I-






100
100 1,000 10,000
Experimental Kxa, [mol/m3*s]
Figure 2. Parity plot for the mass-transfer coefficient.


were given on how to ap-
proach this aspect of the
problem. Perhaps the most elegant way to do this is to obtain
a dimensionless correlation to fit the experimental data, i.e.,
Shox =pReaRe ScY (3)

The exponent for Sc cannot be obtained from the experimen-
tal data since only one solute is used. A value taken from the
literature can be used instead. The students can determine
the exponents a and p from the experimental data and then
use the correlation to obtain the mass-transfer coefficient for
the design problem. The five groups in this rotation used a
similar, but somewhat lengthier, approach. They took a di-
mensionless correlation from the literature121 and used it to
figure out a "corrected" value of y in Eq. (1) (one for
chloroform and one for dichlorobromomethane). Then they
used the "corrected" Eq. (1) for the corresponding
trihalomethane. The important thing here is that all the
groups could find their way in solving the problem cor-
rectly. This contributes more to the learning process than
guiding them through a solution that could be more to the
instructor's liking.

OTHER DESIGN PROBLEMS
During the fall 1996 semester, we prepared four design
problems, one for each rotation. The goal was to encourage
independent thinking about how to use a routine lab experi-
ment to solve a real-life problem. To provide some ideas of
how this approach can be implemented, we highlight here
the main features of the other three problems.
Problem 1. Here the students were asked to design a


stripping system to remove
toluene from a wastewa-
ter stream. The packing
available was 1.5" Pall
rings. Gas flow rate to be
used was 1.5 times the
minimum value. The stu-
dents had to figure out how
to correct for the different
packing size.
* Problem 2. Addition-
ally, the students had to
find the value gas-flow
rate (relative to the mini-
mum value) that mini-
mized the amount of pack-
ing in the column.
* Problem 3. Similar to
the problem presented in
extenso, but with just one
trihalomethane (chloro-
form).


STUDENT FEEDBACK


No formal specific survey to explore the students' impres-
sion of the approach was conducted because the idea of
writing this paper came after the term was over. But a three-
page course-evaluation questionnaire was distributed to give
the students an opportunity to comment generally on the
course. One of the questions was, "Please comment on each
experiment in terms of its value as a chemical engineering
learning experience." We also informally asked some
students what they thought of the new approach. The
feedback was highly positive. The most praised fact was
the opportunity to connect an isolated lab experiment
with a real-world problem.

Some of the actual responses were:

"This was the most valuable learning experience. I
got a lot out of it in terms of learning what it takes
to design a real column..."
"I really like this lab. It was complicated, but not
impossible. Overall, I learned a lot."
"I liked this experiment. The lab itself is simple, but
understanding and utilizing the data is very
complex."

"Of all the experiments, Ifeel that this one was the
one I learned the most in."
"Good learning experience in terms of scale-up
design."


Winter 1999










CONCLUSIONS
This paper presents a simple way to enhance the learning
experience in the unit operations laboratory by adding a
related design problem to the lab requirements. This idea can
be implemented in any classical unit operations experiment
such as distillation, absorption, liquid-liquid extraction, or
even in a heat-exchanger experiment.
By implementing this approach, the students connect lab
work to real-life problems, their creativity is stimulated, and
the concepts are learned enduringly. The feedback from the
students confirms this. It is apparent that they felt more
comfortable the following term when they had to apply these
concepts to design absorption, stripping, and even distilla-
tion columns as part of the task in the senior design course.

NOMENCLATURE
a gas-liquid interfacial area per unit total volume [m2/m3]
Gx liquid-phase flow rate per unit cross-sectional surface area
[m]
Gy gas-phase flow rate per unit cross-sectional surface area
[m]
HoL height of an overall transfer unit based on liquid-phase
concentrations [m]


K overall mass-transfer coefficient [m/s]
L liquid-phase flow rate [kg/s]
m Henry's constant expressed as a ratio of mole fractions
Nox number of overall transfer units based on liquid-phase
concentrations
Rex liquid-phase Reynolds number
Re gas-phase Reynolds number
S stripping (desorption) factor
Scx liquid-phase Schmidt number
Sho0 liquid-phase Sherwood number based on KL
V gas-phase flow rate [kg/s]
xa mole fraction of solute in the liquid phase at the inlet
xb mole fraction of solute in the liquid phase at the outlet
Greek Symbols
a exponent in Eqs. (1) and (3)
p exponent in Eqs. (1) and (3)
y exponent in Eq. (3)
cp coefficient in Eq. (3)
y coefficient in Eq. (1)

REFERENCES
1. Velazquez, C., and L.A. Estdvez, "Stripping of
Trihalomethanes from Drinking Water in a Bubble-Column
Aerator," AIChE J., 38(2), 211 (1992)
2. Treybal, R.E., Mass-Transfer Operations, 3rd ed., McGraw-
Hill, New York, NY (1980) O


Ethanol Fermentation
Continued from page 57.


experiments of ethanol fermentation dispensed with the dis-
tillation of the sample and the chemical analysis of ethanol.
The experimental data obtained in the assays allowed fits of
a simple kinetic model (Monod) and of one that takes into
account the inhibition by the product.'3' Comparison of the
kinetic parameters obtained in the present work with others
in the literature showed that the experimental device has
attained the expected aim.

NOMENCLATURE
D internal diameter of the PVC pipe [D=10 cm]
K product inhibition constant [gL-']
Ks saturation constant [gL-']
Methano ethanol molecular weight [Methanol=46]
nco, number of moles of ethanol formed
n number of moles of CO, evolved at time t=i
P product concentration [gL-']
Pam local atmospheric pressure or barometric pressure [atm]
R gas constant [R=82.04 atm cm3/(g-mole K)]
R2 regression coefficient [-]
S limiting substrate concentration [gL ']
T gas temperature [K]
t time [h]
V broth's volume [L]
Vi gas volume at time t=i [mL or cm3]
X cellular concentration [gL-']
yo distance from top of PVC tube to water level at t=0 [cm]


y, distance from top of PVC tube to water level at t=i [cm]
Yxs cell yield coefficient [gg ']
YP/ product yield coefficient [gg ']
Greek
specific growth rate [h' ]
Pmax maximum specific growth rate [h-']
Subscripts
0 refers to the time t=0
i refers to the time t=i
REFERENCES
1. Aiba, S., A.E. Humphrey, and N.F. Millis, Biochemical En-
gineering, 2nd ed., Academic Press, New York, NY (1973)
2. Bailey, J.E., and D.F. Ollis, Biochemical Engineering, 2nd
ed., McGraw-Hill, New York, NY (1986)
3. Aiba, S., M. Shoda, and M. Nagatani, "Kinetics of Product
Inhibition in Alcohol Fermentation," Biotechnol. Bioeng.,
10,845(1968)
4. Levenspiel, O., "The Monod Equation: A Revisit and a Gen-
eralization to Product Inhibition Situations," Biotechnol.
Bioeng., 22, 1671 (1980)
5. Luong, J.H.T., "Kinetics of Ethanol Inhibition in Alcohol
Fermentation," Biotechnol. Bioeng., 27, 280 (1985)
6. Nilsson, B.K., I. Bjerle, and H.T. Karlsson, "A Simple Meter
with Zero Pressure Drop for Gas Flows," Ind. Eng. chem.
Res., 27, 1553 (1988)
7. Somogyi, M., "Notes on Sugar Determination," J. Biol. Chem.,
195,19(1952)
8. Hoppe, G.K., and G.S. Hansford, "Ethanol Inhibition of
Continuous Anaerobic Yeast Growth," Biotechnol. Lett., 4(1),
39(1982) 0


Chemical Engineering Education











Microwave Drying
Continued from page 49.
temperatures show trends similar to the drying-rate curves
(see Figure 6). The temperature profiles demonstrate the
warming-up period, a constant- temperature period, as well
as a rapid falling-off period. Due to the difficult nature of
temperature measurement in microwave drying, these data
are applicable only for representing trends in the tempera-
ture profile. It should be noted that the surface-temperature
values of Figure 6 did not reach 1000C, the boiling point of
water at atmospheric pressure. This can be accounted for by
the effect of surface evaporation "1 and by the measurement
technique that allowed for some cooling of the sample upon
removal from the microwave. In general, during microwave
drying, the surface temperature falls between the wet-bulb
temperature in the oven and the boiling temperature.41 Our
microwave oven was not equipped with a carousel, and
given the uneven nature of the energy field in a microwave
oven, the temperature probably varied with position. We
did not study this phenomenon, however, and only
measured the surface temperature near the middle of the
sample.
U Students can be asked to design an experiment that
illustrates that mass transport effects are responsible for the
falling-rate period. Evaporation of water with no solid
present can be used to do this. In this instance, the falling-
rate period is eliminated, with the constant-drying-rate
period ending abruptly as the last of the water is evapo-
rated.
E Particle-size effects can be examined by using sand that has
been sifted into different particle sizes. Critical moisture
content increases with decreasing particle size since smaller
particles pack closer together, slowing the movement of
moisture to the surface.[4'
Datta151 states that geometry plays a role in microwave
processing. This effect can be explored by changing the
shapes of the containers and the thickness of the bed height
while maintaining a constant mass.
B The analysis presented here does not yield detailed
information about the fundamental mechanisms of heat and
mass transfer during the drying process, It is possible,
however, to estimate interphase heat and mass transfer
coefficients and effective liquid diffusivities using methods
presented in the literature.[4] Also, complex mathematical
models of the drying process have been developed.191
Drying characteristics of different materials can be
compared. Various materials have been used in our
laboratory, including calcium carbonate, sponges, bread,
and fruit.

CONCLUSIONS

This experiment demonstrates the drying process effec-
tively. It is an extremely flexible, safe, and inexpensive
experiment that can be incorporated into the undergraduate
laboratory curriculum. The experiment is easy to set up and
run. Typically, meaningful experimental data for higher power
Winter 1999


100


80


60


a. 40
E
1--
20


0 I- I ---I---
0 5 10 15 20 25
Time (min)
Figure 6. Surface temperature profile for
a power level of 720 W.

settings can be collected in about thirty minutes. This experi-
ment also involves data analysis that introduces students to
the various methods of treating data and the errors associ-
ated with each method.

ACKNOWLEDGMENTS
We would like to thank the reviewers of this paper for
their positive comments and pertinent suggestions that im-
proved it.

NOMENCLATURE
i index for numerical differentiation (unitless)
m mass of liquid (kilograms)
m mass of solid (kilograms)
t time (minutes)
to time of first data point (minutes)
t time of last data point (minutes)

REFERENCES
1. Nirdosh, I., and M.H.I. Baird, "Low-Cost Experiments in
Mass Transfer," Chem. Eng. Ed., 30(1), 50 (1996)
2. Mujamdar, A.S., "Drying Neglected," Letter to Chem. Eng.
Prog., 93(10), 9 (1997)
3. Oakley, D.E., "Produce Uniform Particles by Spray Drying,"
Chem. Eng. Prog., 93(10), 48 (1997)
4. Porter, H.F., P.Y. McCormick, R.L. Lucas, and D.F. Wells,
"Gas-Solid Systems," Chap. 20 of Perry's Chemical Engi-
neers' Handbook, Fifth Ed., McGraw-Hill, New York, NY
(1973)
5. Datta, A.K., "Heat and Mass Transfer in the Microwave
Processing of Food," Chem. Eng. Prog., 86(6), 47 (1990)
6. Keey, R.B., Drying: Principles and Practice, Pergamon Press,
Oxford, p. 3 (1972)
7. Williams-Gardner, A., Industrial Drying, Leonard Hill, Lon-
don, England, p. 45 (1971)
8. Ray, M. Engineering Experimentation, McGraw-Hill, New
York, NY, p. 169 (1989)
9. King, C.J., and J.P. Clark, eds., Water Removal Processes:
Drying and Concentration of Foods and Other Materials,"
AIChE Symposium Series, 73(163) (1977) 0











= assessment


RANKING GRADUATE PROGRAMS

Alternative Measures of Quality


JOHN C. ANGUS, ROBERT V. EDWARDS, BRIAN D. SCHULTZ
Case Western Reserve University Cleveland, OH44106-7217


A assessing the relative quality of graduate programs is
of great interest to policy makers, academic admin
istrators, prospective students, employers of gradu-
ates, alumni, and the general public. Rankings by federal
advisory panelst1 and the popular press[21 are widely quoted,
but despite the importance assigned to the rankings, there
has been little critical, detailed analysis of their relevance
and accuracy. In this paper we will present an analysis of the
most prominent of these reports,l11 especially as it relates to
chemical engineering programs. Although our discussion is
confined to chemical engineering, we believe that the gen-
eral conclusions and methods are also relevant to the other
engineering disciplines.
The National Research Council (NRC) in 1995 released a
massive study"' of research doctorate programs in the U.S. It
was the product of a committee of eighteen academics from


Editorial Comment...
A strength of the engineering education system in the United States is its diver-
sity. It is evident in such characteristics as student demographics, college missions
and sources of funding, enrollment levels, research strengths, collaborations, and
curricula. A positive outcome of our system is the diverse pool of graduates pro-
duced to meet the varied and dynamic workforce needs of the world. Choices,
however, are being made that can significantly impact programs. Industry selects
schools for recruiting visits; students commit to pursue graduate or undergraduate
degrees at specific schools, and private and public foundations and agencies award
grants, contracts and gifts to selected institutions. The perceived quality of an
institution is often an important factor in these decisions, and rankings by
institution, college, or degree program contribute to defining perceptions.
In recent years, chemical engineering departments have been asked to assess the
quality of their programs to direct improvement strategies. This movement is being
driven internally as well as externally by regional and national accreditation entities
and by funding agencies (e.g., state governments). Ideally, rankings would directly
assess the quality of graduates and the improvement in students while they were
enrolled. Since no method to do this has been devised, ranking schemes typically
use a combination of numerical program data and peer ranking to determine a score
instead of doing a direct assessment. Although the efficacy of this approach is still
being debated, it has increased the importance of peer comparison and the availabil-
ity of program data. This issue of CEE presents a paper by Angus, Edwards, and
Schultz that proposes alternative measures of graduate program quality. Not sur-
prisingly, an extensive review process revealed that the subject of rankings is a


various disciplines and is a follow-on study of a similar
report issued in 1982. The report contains reputational
rankings, based on a survey of graduate faculty, as well as an
impressive amount of factual data from several independent
sources. Unfortunately, most attention has been focused on
the survey results-in particular the reputational ranking
based on perceived faculty quality. This is apparently due, at
least in part, to the method by which the data were displayed
in the report and the normal tendency to simplify complex
data sets into a single, easily understood, quality index.
Several aspects of the NRC report have caused concern.
(See, for example, the summary article by Mervis.'31) One
striking feature of the results, noted by the authors, was the
"remarkable stability among programs rated in the top and
bottom quarter" between the 1982 and 1995 reports. An-
other striking feature was the heavy reliance on the survey


contentious issue. The reviewers as well as the authors identified many of the
shortfalls in assessing the quality of something as complex, multidimensional, and
subjective as graduate programs. Identified issues included the establishment of false
goals, for example publishing papers in smaller segments or of lesser quality, or
hiring faculty in publication-intensive research areas simply to increase the publica-
tion count, all of which would increase ranking but could likely decrease program
quality. Another issue was that inaccurate or inappropriate data would undermine the
conclusions. Examples include the counting of non-competitive research funding
(e.g., state funding) and the use of a limited set of journals for citation searching or a
limited set of societies in award counting that would not recognize programs with an
emphasis on emerging research areas.
This paper presents a sound analysis of the recent NRC ranking, and many of the
conclusions as well as the analytical approach can be extended to other ranking
schemes. Although the proposed alternative measures of quality may be open to
criticism, the authors clearly show the sensitivity of rank order to the selected set of
measures and to weighting algorithms (e.g., intensive vs. extensive). We hope that
this paper will increase awareness of the shortfalls of any approach to ranking and
help one gauge what can and cannot be ascertained from a ranking. We do not
believe that a composite index can be developed to accurately measure the relative
quality of chemical engineering degree programs in our complex graduate education
system, nor is it desirable to drive programs to conformity. We recognize, however,
that ranking schemes have increased their presence in our profession and to ignore
their impact would be a mistake. We hope that this article will stimulate serious
discussion in the community. O


Copyright ChE Division of ASEE 1999
'2 Chemical Engineering Education











results, which were used to generate reputational rankings
rather than "quantitative" measures of quality. Another
concern, quite apparent to engineers and scientists, was
the minimal distinction drawn between intensive (size
independent) and extensive (size dependent) measures of
quality. Although the report included a number of statisti-
cal tests of the data, no detailed analysis of sources of
error in the data sets was provided. Finally, there was no
assessment of program quality based on student outcomes
in their subsequent professional life. The committee was
TABLE 1
Survey Questions Used for the NRC Report"'

Bl Familiarity with work of Program Faculty
1. Considerable familiarity
2. Some familiarity
3. Little or no familiarity
B2 Scholarly Quality of Program Faculty
1. Distinguished
2. Strong
3. Good
4. Adequate
5. Marginal
6. Not sufficient for doctoral education

9. Don't know well enough to evaluate
B3. Familarity with Graduates of this Program
1. Considerable familiarity
2. Some familiarity
3. Little or no familiarity
B4 Effectiveness of Program in Educating Research Scholars/Scientists
1. Extremely effective
2. Reasonable effective
3. Minimally effective
4. Not effective

9. Don't know well enough to evaluate
B5 Change in Program Quality in Last Five Years
1. Better than five years ago
2. Little or no change in the last five years
3. Poorer than five years ago

9. Don't know well enough to evaluate


John C. Angus is Professor of Chemical Engineering at Case West-
ern Reserve University. He received his BS, MS, and PhD degrees
from the University of Michigan. He worked on thermoelectric materials
at the 3M Company for three years before joining the faculty at Case.
He has worked on the growth of diamond by chemical vapor deposition
and various electrochemical problems for almost forty years.
Robert V. Edwards received his PhD from Johns Hopkins University
in 1968 and took a post-doctoral position at Case Western Reserve
University to work on the then-new field of laser light scattering for
transport measurements. He joined the Case faculty in 1970 and has
subsequently made numerous contributions to the theory and practice
of laser light scattering with collaborators, both here and abroad.
Brian D. Schultz obtained his BS in chemical engineering from Case
Western Reserve University in 1977. He won a National Science
Foundation Fellowship, which he used to pursue his Master's degree
at Case, and in the fall of 1998, he began working toward his PhD at
the University of Minnesota. Research interests include ternary phase
diagrams for low-pressure crystal growth as well as the thermochemi-
cal behavior of group III nitrides.

Winter 1999


aware of many of these concerns and, in fact, was unable to
address some of them for lack of time and resources. The commit-
tee was also aware that the report might be used in superficial ways
that were not intended.
The NRC report is being used by deans, legislators, and founda-
tions in the allocation of resources and in other critical decisions. It
is therefore useful to understand the report and to critically exam-
ine its conclusions. In this paper we will give an analysis of the
data for the chemical engineering programs covered in the report.
We will also give alternative rankings using data from the NRC
report and other sources. We emphasize that the rankings pre-
sented here are meant only to illustrate the methods employed and
to reach general conclusions. Because of limitations in the data
available to us, the position of a particular individual program in
the rankings should be treated with caution.


HE NATNL RESEARCH COWCIL REPORT

Methods The most discussed part of the NRC report is the
faculty survey conducted in 1993. Questionnaires were sent to
randomly selected faculty and each participant was asked to rank
approximately fifty programs. Other than a list of faculty, the
participants were provided no other information about the pro-
grams. The survey questions are shown in Table 1.
Forty-one graduate fields of study were covered in the NRC
report, one of which was chemical engineering. For chemical
engineering programs, 206 usable responses were obtained from
361 questionnaires, a 57% response rate. Within chemical engi-
neering, 93 of the 121 engineering departments awarding PhD
degrees during the 1986-92 time period were included. These
93 departments produced 96% of the chemical engineering
PhDs awarded during that period.
The results of the survey were tabulated in the Appendices to the
NRC report. The programs were listed in order in the tables ac-
cording to the results of the first survey question (the average
ranking of faculty quality). This procedure was used in response to
complaints that data in the 1982 report were difficult to interpret
because programs were listed alphabetically. The result, however,
has been to focus on this one single measure of quality, despite the
fact that rankings in the other categories (e.g., program effective-
ness and visibility) are also provided in the report.
One of the purposes of the committee that compiled the 1995
NRC Report was to expand the "objective" measures developed
by prior committees. Some program statistics were provided: num-
ber of faculty, number of PhDs granted, number of PhDs awarded
to female and minority students and non-citizens, and the average
length of time to receive a PhD. Quantitative measures of quality
were also provided: 1) percentage of faculty with research support
(%SUPP), 2) percentage of faculty publishing during 1986-92
(%PUB), 3) total publications during 1986-92 (PUB), and 4) total
citations to published work during 1986-92 (TC). The latter two
were also reported on an intensive (normalized per faculty)
basis, i.e., PUB/TF and TC/TF (see Table 2 for a description of
73











the terms in the NRC report).

Survey Results In Table 3 we give the acronyms used
in subsequent tables for identification of universities, and
in Table 4 we list the graduate programs in chemical
engineering as they were rank-ordered by perceived
faculty quality (93Q) in the NRC report. This is the
order in which the programs are listed in Appendix P
of the NRC report.

A striking, but not widely appreciated, feature of the
NRC report is shown in Figure 1,* which is a plot of the
survey results for faculty quality (93Q) versus program
effectiveness (93E). A very strong correlation is evident.
For example, R2=0.97 when the data are fit with the
equality (93Q)=(93E). This strong correlation can arise
simply because high-quality faculty will produce effec-
tive graduate programs. We believe it is far more likely
that the respondents did not discriminate between faculty
quality and program effectiveness and treated both ques-
tions the same. This strong correlation was noted in the
NRC report in Appendix 0-8 where a Pearson product-
moment correlation coefficient of 0.98 between 93Q and
93E was given for chemical engineering. Similar strong
correlations between 93Q and 93E were observed for the

'In Figure 1 and subsequent figures we indicate the square of
the degree of correlation by R2, the coefficient of determina-
tion. The magnitude of R2 is simply described as the fraction
of the raw variance in the data set accounted for by using the
fitted equation. The plots and values of R2 were obtained
using an Excel spreadsheet.


TABLE 3
Acronyms Used to Identify Universities


Arizona State University
Auburn University
Brigham Young University
California Inst. of Technology
Clarkson University
Clemson University
Camegie Mellon University
Columbia University
Cornell University
Colorado School of Mines
CUNY-Grad Sch & Univ Center
Case Western Reserve Univ.
Duke University
Georgia Institute of Technology
Illinois Institute of Technology
Iowa State University
Johns Hopkins University
Kansas State University
Lehigh University
Louisiana St. U & A&M College
Massachusetts Inst. of Tech.
Michigan State University
North Carolina State University
University of Notre Dame


Northeastern University
New Jersey Inst. of Technology
Northwestern University
Ohio University
Oklahoma State University
Oregon State University
Ohio State University
University of Pittsburgh
Polytechnic University
Princeton University
Pennsylvania State University
Purdue University
Rice University
University of Rochester
Rensselaer Polytechnic Institute
Rutgers St. Univ-New Brunswick
Stevens Institute of Technology
Stanford University
State Univ of New York-Buffalo
Syracuse University
Texas A&M University
Tulane University
University of Akron


University of Arizona
University of Califomia-Berkeley
University of California-Davis
University of Cincinnati
Univ. of Califomia-Los Angeles
University of Colorado
Univ of California-Santa Barbara
University of Connecticut
University of Delaware
University of Florida
University of Houston
University of Iowa
University of Illinois-Chicago
University of Idaho
Univ of Illinois,Urbana-Champaign
University of Kansas
University of Kentucky
University of Louisville
Univ. of Massachusetts-Amherst
Univ. of Maryland-College Park
University of Maine
University of Michigan
University of Minnesota


UMOC
UMOR
UMS
UOK
UPA
URI
USC
UTA
UTN
UTUL
UUT
UVA
UWA
UWI
UWY
VAND
VPI
WASU
WPI
WSU
WU
WVU
YALE


University of Missouri-Columbia
University of Missouri-Rolla
University of Mississippi
University of Oklahoma
University of Pennsylvania
University of Rhode Island
Univ. of Southern California
University of Texas-Austin
Univ. of Tennessee-Knoxville
University of Tulsa
University of Utah
University of Virginia
University of Washington
Univ. of Wisconsin-Madison
University of Wyoming
Vanderbilt University
Virginia Polytech Inst & State Univ
Washington State University
Worcester Polytechnic Institute
Wayne State University
Washington University
West Virginia University
Yale University


4 Chemical Engineering Education


TABLE 2
Definition of Terms Used in this Paper

NOTE: In the NRC report, the symbols for the variables referred to both the rank order,
and, where applicable, to the average score of the ratings on the scale of I to 5. We give
both definitions in the list below. The definitions are taken from Appendix P, page 469,
and Table 2-4, page 25, of the NRC report. Not all of the categories used in the NRC
report were used in this paper. The terms below the dashed line were not used in the NRC
report.

93Q Rank order of "scholarly quality of program faculty." (Average score on a scale
of 0 to 5, with 5 representing "Distinguished.")
93E Rank order of "program effectiveness in educating research scholars and
scientists." (Average score on a scale of 0 to 5, with 5 representing "Extremely
Effective.")
VIS Rank order of visibility of the doctoral program (Percentage of the question-
naires that reported some knowledge of the program by an answer other than
"Don't know well enough to evaluate" or "Little or no familiarity" to one or
more of the five questions.)
TC Rank order of the total number of citations attributed to program faculty in the
period 1988-92. (Total number of citations attributed to program faculty.)
Source: Institute of Scientific Information.
C/F Rank order of the citation density for the program faculty (Total number of
citations (TC) divided by the number of program faculty (TF)). Source:
Institute of Scientific Information.
PUB Total number of publications attributed to program faculty for the period 1988-
92.
TF Number of program faculty in fall 1992. NRC Report for calculating PUB/TF
and TC/TF; AIChE1'1 for calculating HON/TF and SUPP/TF.


HON Number of honors received by faculty."16 See text for details
SUPP Total research support from all sources. Source: National Science Foundation.4


ASU
AUB
BYU
CIT
CLAR
CLMN
CMU
COL
CORN
CSM
CUNY
CWRU
DUKE
GIT
IIT
ISU
JHU
KSU
LEH
LSU
MIT
MSU
NCSU
ND











other engineering programs. This level of
correlation strongly suggests that these two
of the five survey questions gave indistin-
guishable results.

In Figure 2a, we show a plot of perceived


TABLE 4
Rank Order of ChE Faculty
Quality Survey Results (93Q) Given
in the NRC Report


Rank
Order
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21.5
21.5
23
24
25
26
27
28
29
30.5
30.5
32.5
32.5
34
35.5
35.5
37
38
39
40
41
42
43
44
45
46
47.5
47.5
49
50


University
University of Minnesota
Massachusetts Institute of Technology
University of California-Berkeley
University of Wisconsin-Madison
Univ of Illinois, Urbana-Champaign
California Institute of Technology
Stanford University
University of Delaware
Princeton University
University of Texas at Austin
University of Pennsylvania
Carnegie Mellon University
Cornell University
Univ of California-Santa Barbara
Northwestern University
Purdue University
University of Houston
University of Michigan
CUNY-Grad Sch & Univ Center
University of Washington
Univ of Massachusetts-Amherst
Rice University
Pennsylvania State University
University of Notre Dame
North Carolina State University
University of Colorado
Lehigh University
University of California-Davis
State University of New York-Buffalo
University of Virginia
Georgia Institute of Technology
Yale University
Iowa State University
University of Florida
Rensselaer Polytechnic Institute
Johns Hopkins University
Texas A&M University
Washington University
University of Califoria-Los Angeles
University of Rochester
Ohio State University
Virginia Polytech Inst & State Univ
Rutgers State Univ.-New Brunswick
University of Pittsburgh
Michigan State University
Case Western Reserve University
Syracuse University
Illinois Institute of Technology
Clarkson University
Brigham Young University


faculty quality (93Q) versus faculty size (TF). The value of R2 is 0.40, suggest-
ing that the survey results for faculty quality are influenced to some extent by
program size, but that other factors are also important.

The program visibility (VIS) was defined as the percentage of respondents
who reported some knowledge of the program. In Figure 2b we plot the faculty
quality (93Q) versus the visibility (VIS). A strong correlation, R2=0.84, is
observed. One cannot prove cause-and-effect relationships through correlation
alone; however, these results suggest that the perceived faculty quality (93Q)
scores arise, at least in part, because respondents rate highly those faculty with

5.00 -
2oo

R =0.97
4.00


S3.00


2.00


1.00


0.00
0.00 1.00 2.00 3.00 400 5.00
Program Effectivenesa (93E)


Figure 1. Survey results of program effectiveness (93E) versus faculty
quality score (93Q). R2=0.97 when fit with (93E)=(93Q).


a -
5 00-

R' =0.40 0
4.00 aQ

08 o
Figure 2. 3.o00 B

(a) Survey I .
results of faculty 2.00
quality (93Q)
versus total .oo
faculty (TF). 0.00
R = 0.40 for 0 5 10 15 20 25 30 35
linear fit, Total Faclty (TF)
y=mx+b.


(b) Survey 1 s.
results of facultyR284
quality (93Q) 4.00 -
versus visibility
(VIS). R2=0.84 03.0o "
for linear fit,
y=mx+b. 2.00 o ea


40





,


0
t oo o

0.00 -
30 40 50 60 70 80 90 100
Visibility (VIS)


Winter 1999











whom they are familiar. If this familiarity arises
because of the true quality of the faculty, this re
benign; otherwise it is not.

A relatively strong correlation is observed between r
der of visibility and rank order of total citations (R2=0.6
a weaker correlation between rank order of visibility ar
number of faculty (R2=0.39). These results are encou
because they show that smaller departments can h;
impact by virtue of their research output. Significant
lations are also found between perceived faculty (
(93Q) and the number of publications (R2=0.73), nunr
publications per faculty (R2=0.64), to-
tal citations (R2=0.71), and citations
per faculty (R2=0.56).

In summary, it appears that the respon-
dents made no distinction between the
survey questions on faculty quality (93Q) A
and program effectiveness (93E). To University
some extent, sheer size influenced the UMN 10(
MIT 7!
quality rankings and respondents gave UTA 7
high ranks to programs with which they UCB 5(
were familiar. Strong positive correla- UDE 3'
tions exist between the survey results of UWI 3,
faculty quality and the publication and TAM 2:
SPSU 2:
citation rates ..


W- --1 ;.' -




Alternative Measures of Quality Four
extensive measures of program quality
are used: 1) Number of publications, 2)
number of citations to publications, 3)
research funding, and 4) faculty honors.
In addition, each of these extensive mea-
sures is normalized by the number of
faculty to provide four intensive mea-
sures of quality. We use these data to
develop alternative rankings of programs
based on both the extensive and inten-
sive criteria. We also provide a final com-
posite ranking based on the extensive
and intensive rankings. We are quite
aware that these so-called "objective"
measures are imperfect, and we will at-
tempt to point out potential problems with
each of the measures we use.

Quantitative measures of quality are
not new. Some were used in the NRC
Report as mentioned above. Also, the
often-maligned U.S. News and World
Report'21 used a lumped score in which


UCLA
UCSB
PRU
NCSU
NWU
PUR
STAN
UMI
CIT
UIL
CORN
LEH
UPA
CMU
GIT
SUNY
CWRU
UMA
LSU
UCO
UTN
UWA
UCD
JHU
UUT
UOK
UH
OSU
RPI
PITT
ISU
UVA


4.
2
2
2:
2'
3'
2
21
2'
2:
3(
3(
2t
11
21
2:
1:
1
2'
1c

2'
2:
1:
2:

1:
1I
1
1:
1'
1:


TABLE 5
Programs not Considered in Final Rankings
Because of Lack of Data on Research Support


Brigham Young University
City University of New York
Duke University
Illinois Institute of Technology
Northeastern University
Rice University
University of Akron
University of Idaho
University of Kansas


University of Louisville
University of Maine
University of Mississippi
University of Notre Dame
University of Rhode Island
University of Tulsa
University of Wyoming
Washington University
Worcester Polytechnic University


TABLE 6
Rank Order and Scaled Scores of ChE Graduate Programs
Using Extensive Criteria


publications
ore Rank
3.0 1
9.2 2
5.6 3
3.3 4
4.2 6
1.1 7
2.8 24
5.7 17
3.4 5
7.6 13
1.0 27
2.6 25
4.7 20
4.0 8
7.6 13
8.4 11
4.0 21
5.7 17
3.2 9
3.2 9
5.1 16
8.9 30
8.2 12
3.8 22
2.0 50
9.1 29
4.9 19
6.5 33
3.7 87
7.6 13
3.5 23
5.1 36
2.5 26
1.8 51
2.2 48
3.5 31
7.3 32
6.4 34
2.6 44
4.3 38


Citations
Score Rank
100.0 1
65.0 3
76.6 2
45.2 4
23.3 8
16.2 19
6.4 45
13.0 24
39.6 5
24.9 7
17.5 16
16.8 17
13.7 23
21.4 9
27.7 6
18.8 13
17.9 15
18.3 14
19.0 12
20.3 10
19.3 11
9.6 38
10.3 35
14.8 22
15.8 21
12.1 26
11.1 31
9.4 39
0.5 91
16.5 18
15.8 20
11.1 32
11.1 32
7.9 44
5.3 52
9.2 40
9.9 37
5.8 48
12.2 25
10.0 36


Suwoart
Score Rank
76.6 3
100.0 1
55.6 5
13.4 41
30.7 12
41.6 7
77.8 2
23.0 19
12.2 44
28.2 13
25.7 16
35.9 9
13.1 42
26.0 15
25.4 17
21.2 24
21.7 23
15.4 34
21.9 22
23.9 18
8.4 56
10.7 48
26.7 14
10.0 50
32.2 11
11.9 45
22.1 21
16.9 31
61.6 4
14.4 37
8.8 53
34.4 10
21.2 25
51.6 6
23.0 20
11.2 47
15.3 35
18.9 29
8.7 55
10.4 49


Average of Extensive Composite Scores for All Unive


Honors
Score Rank
77.9 2
100.0 1
72.6 3
56.1 6
68.3 4
58.6 5
21.8 27
54.0 7
18.7 35
32.5 16
46.9 10
35.0 12
52.8 8
20.3 32
20.3 31
32.0 17
29.8 20
34.0 15
21.8 27
17.2 37
35.0 12
47.6 9
20.8 29
34.1 14
22.4 25
35.3 11
16.5 39
31.9 18
8.5 53
13.8 43
24.0 24
11.1 48
16.5 39
0.0 83
27.2 21
26.7 22
20.2 34
18.1 36
22.4 25
20.3 32
'rsities- 71.85


Extensive Composite
Score Rank
354.5 1
344.2 2
281.4 3
165.0 4
156.6 5
150.6 6
128.7 7
115.7 8
114.0 9
113.2 10
111.1 11
110.4 12
104.3 13
101.6 14
101.0 15
100.4 16
93.4 17
93.4 18
92.9 19
91.5 20
88.7 21
86.8 22
86.0 23
82.6 24
82.3 25
78.3 26
74.8 27
74.7 28
74.3 29
72.3 30
72.1 31
71.6 32
71.3 33
71.3 34
67.7 35
65.6 36
62.7 37
59.2 38
55.9 39
55.0 40


Chemical Engineering Education










40% was based on a reputational survey of deans and mem-
bers of the National Academy of Engineering. The remain-
ing 60% was derived from quantitative measures of research
support, faculty honors, and student selectivity. The U.S.
News and World report also included both extensive and
intensive measures of quality.
We have included only the top forty programs in the
extensive, intensive, and composite rankings. Our purpose
is to focus on alternative methods of ranking rather than
the rank order itself. We have no wish to identify any
program as being of low quality.
Publications and Citations We use both the total number
of faculty publications (PUB) and citations to published
papers (TC) from the NRC report as extensive measures of
research quality. The same variables normalized by the total
number of faculty (TF) are used as intensive measures of
research quality. We recognize these are imperfect mea-
sures. For example, research with the longest range and most
profound impact may go unnoticed for decades. Also, it is
difficult to agree on what constitutes a publication, and there
is a proliferation and duplication of research papers of mar-
ginal merit. The number of times a research paper has been
cited is a summary judgment, albeit imperfect, of its relevance
and importance. But papers with classic errors (for example,
cold fusion) may attract numerous citations. More signifi-
cantly, a single review paper or a paper describing a widely
used test or procedure can generate an inordinate number of
citations not closely related to research quality. Finally, certain
sub-fields within chemical engineering may more easily pro-
duce publishable results than others.
Research Support The NRC report contained data on the
percentage of faculty that received research support (%SUPP)
and the percentage of faculty that published (%PUB). We
found these variables provided little discrimination, especially
between high-ranked programs, and we did not use them in our
analysis. Instead, we elected to use total research support from
all sources (SUPP) collected by the National Science Founda-
tion[41 as an extensive quality measure. These figures were used
without modification.* We emphasize that the compilation
reported by the NSF is meant to be complete; it includes
state support and support from other federal agencies, indus-
try, and foundations. We also note that total research support
is one of the primary measures used in recent scholarly
studies of the relative quality of research universities.[51
The data in the NSF report are reported by the individual
institutions and may not be reported on a similar basis;
research support from ancillary research institutes or unrelated
programs may be included in some cases. Also, the amount of

We made one exception to this generalization. For our own uni-
versity, we removed the expenditures of the Macromolecular Sci-
ence Department from the NSF figures. This lowered the CWRU
extensive ranking and left the intensive ranking unchanged. We
were unable to make a similar correction for other programs.
Winter 1999


state support for research may not be uniformly reported.
Eighteen chemical engineering programs in the NRC re-
port were not listed in the National Science Foundation
report. There is no indication whether this is because no data
were submitted by these programs or whether they had too
little research income to appear on the table (only the top
100 engineering programs were included in the table). Rather
than estimate the research support from other sources, we
excluded these programs from our rankings. The programs
that were excluded are listed in Table 5.
Faculty Honors Inexplicably, faculty honors were not
used as a quality index for engineering and science programs
in the NRC report; they were used, however, in the NRC
rankings of programs in the arts and humanities. There are
certain categories of honors and awards for chemical engi-
neering faculty that can easily be tabulated. For junior fac-
ulty we used the number of recipients of NSF Career Devel-
opment Awards, NSF Young Investigator awards, and Presi-
dential Investigator awards over the period 1988 to 1996;[61
for mid-career faculty we used winners of the principal
AIChE awards from 1987 to 1996;71] and for senior faculty
we use the sum of the current number of National Academy
of Engineering members181 plus one-half of the number of
Fellows of the AIChE.171 This arbitrary choice is based on the
observation that there are approximately twice as many
AIChE fellows as NAE members in the departments sur-
veyed. Retired and emeritus faculty were excluded. The
three categories (junior, mid-career, senior) were scaled to
give each equal weight. We believe that including only
AIChE honors and awards over-emphasizes the traditional
areas of chemical engineering. In future rankings we suggest
including honors and awards from other professional organi-
zations (e.g., the American Chemical Society, the Electro-
chemical Society, and the Materials Research Society).
Alternative Rankings The numerical data in each exten-
sive category (Publications, Citations, Support, and Honors)
were scaled so the maximum value in each category was
100. The total extensive score for each program was ob-
tained by summing the four scores for each extensive cat-
egory. The overall extensive rank order was determined
from these summed scores (see Table 6). The programs are
listed in Table 6 in the order of their total extensive score.
The intensive scores in each category for each program
were obtained by dividing the extensive scores by the appro-
priate number of program faculty. For calculation of PUB/
TF and TC/TF, all data were taken from the NRC report. For
calculation of HON/TF and SUPP/TF, the data were taken
from references 4,6,7,8, and 9. The intensive scores were
also scaled so that the maximum value in each category was
100. A total intensive score was obtained for each program
by summing the scaled intensive scores of the four catego-
ries. The programs are listed in Table 7 in the order of their
total intensive score.











It is tempting to use the intensive rankings in Table 7 as a
measure of the average, individual quality of the program
faculty. But one should be cautious in doing so, especially
for the smaller programs. In some cases the average inten-
sive score is heavily influenced by the activities of one or
two particularly strong individuals. This effect was mea-
sured in the NRC report by the Gini coefficient, which is a
measure of the non-uniformity of the distribution of scores
among the individuals. Since we did not have access to the
raw data, we could not make this estimate.


value of A In (TC) is the fractional change in number of total
citations required to change one place in the rank order. In
the middle range, the average fractional change required to
move one place in the rank order of citations is approxi-
mately 0.03; however, greater fractional changes (over 0.30)
are required to move one place in the rank ordering at either
extreme. Similar behavior is observed for the other exten-
sive variables, also shown in Figure 3. These results show
that while it is relatively easy to move in the middle range of
rank orders, it will be more difficult for programs to move


A composite extensive plus intensive rank-
ing was also calculated. A simple summation
of the total extensive plus intensive scores
gave undue weight to the intensive scores.
We rescaled the intensive total scores to give
the same average score as the extensive
scores. A composite extensive/intensive score
was calculated for each program using the
total extensive score and the rescaled inten-
sive total score. For example, for MIT the
composite extensive/intensive score was ob-
tained from

71.85
344.2+ (312.1)= 515.8
130.66

This procedure gave the same overall weight-
ing to the extensive and intensive scores.
The composite extensive/intensive scores and
rankings are given in Table 8.

It is most appropriate to compare programs
using the separate extensive and intensive
measures in Tables 6 and 7. Nevertheless,
the composite extensive/intensive ranking in
Table 8 has value. For example, when mak-
ing a choice of a graduate program, a pro-
spective student will make an integrated as-
sessment of both extensive and intensive mea-
sures. Small programs that are rated very
highly on a per-faculty basis may have a
limited range of course work and research
options; large programs with high extensive
scores may not have the desired level of indi-
vidual faculty quality. A composite score also
permits comparisons with other lumped
scores-for example, the U.S. News and
World report rankings.

Sensitivity and Error Analysis The sen-
sitivity of the rank ordering to changes in the
extensive data sets (Publications, Citations,
Support, and Honors) can be calculated by
calculating A In X( A X/X) for each of the
rank ordered data sets. We show plots of
these results in Figure 3. For example, the


TABLE 7
Rank Order and Scaled Scores of ChE Graduate Programs
Using Intensive Criteria

Publications/Faculty Citations/Faculty Support/Faculty Honors/Faculty Intensive Composite
University Score Rank Score Rank Score Rank Score Rank Score Rank


UMN 100.0 1
UTA 81.7 4
MIT 81.7 4
STAN 88.3 2
UCB 84.8 3
CIT 77.0 6
UWI 57.4 10
JHU 60.4 8
CWRU 38.3 33
UIL 54.8 11
PRU 44.8 25
UDE 54.8 11
SUNY 54.3 13
UCLA 60.4 8
UCSB 49.1 20

NWU 46.5 24
TAM 36.5 37
SYR 61.3 7
UPA 52.2 14
CORN 50.9 17
PSU 30.4 48
UMA 47.0 22
UOK 34.3 43
NCSU 42.6 27
UMI 50.4 19

YALE 51.3 16
UCO 35.2 41
CMU 35.7 40
LEH 50.9 17
OSU 42.2 28

OKSU 20.9 70
UVA 38.3 33
RPI 39.6 30
UWA 49.1 20
UH 21.7 68
PUR 43.5 26
COL 31.7 45
UCD 47.0 22
UTN 7.8 92
LSU 29.6 52


100.0 1
81.7 3
67.1 5
88.7 2
76.2 4
57.3 6
27.4 27
44.3 10
50.4 8
39.0 12
37.3 14
37.3 14
33.8 19
55.1 7
44.2 11

25.9 31
10.2 62
36.6 16
38.6 13
32.1 21

15.4 45
29.9 24
22.9 35
31.7 22
33.5 20

48.0 9
20.0 40
18.0 41
34.1 18
21.1 39

7.7 71
26.7 30
22.5 36
29.4 25
9.4 67
27.4 27
9.8 65
31.7 23
1.1 92
13.2 52


56.7 11
58.3 9
74.1 6
63.9 8
19.8 47
54.7 12
58.2 10
96.2 4
73.7 7
32.3 24
40.5 15
30.9 26
22.9 40
30.8 27
41.7 14

18.3 53
97.9 3
39.2 16
13.2 61
36.8 19

30.5 29
18.7 51
100.0 1
37.7 18
25.4 34

19.6 49
30.5 28
15.9 56
33.4 21
21.7 45

98.0 2
23.7 38
27.5 30
22.6 43
38.6 17
26.2 32
43.1 13
9.2 69
86.2 5
32.8 23


Average of Intensive Composite Scores for all Univ


69.5 13 326.2 1
91.8 4 313.5 2
89.2 6 312.1 3
61.5 19 302.4 4
100.0 1 280.8 5
90.2 5 279.1 6
98.8 2 241.8 7
37.4 32 238.3 8
61.6 18 224.0 9
86.0 10 212.0 10
88.9 8 211.5 11
82.9 12 205.9 12
93.9 3 205.0 13
56.7 22 203.1 14
58.0 21 193.0 15
89.0 7 179.6 16
33.0 38 177.7 17
40.6 31 177.6 18
66.4 16 170.3 19
44.0 29 163.8 20

86.2 9 162.5 21
66.8 15 162.3 22
0.0 83 157.2 23
44.2 28 156.3 24
46.1 27 155.6 25

36.1 34 155.1 26
69.2 14 154.8 27
84.8 11 154.4 28
28.9 42 147.3 29
62.3 17 147.3 30

19.5 53 146.0 31
55.9 23 144.6 32
43.8 30 133.4 33
26.1 44 127.2 34
55.0 24 124.7 35
24.6 47 121.6 36
36.3 33 120.9 37
30.3 39 118.2 38
14.4 60 109.5 39
29.5 41 105.1 40

'ersities 130.66


Chemical Engineering Education












TABLE 8
Example of Rank Order of ChE Graduate Programs using a Single, Composite Extensive/Intensive Criterion


Composite Score
University Rescaledf Normalized
University of Minnesota 533.9 100.0
Massachusetts Inst. of Technology 515.8 96.6
University of Texas at Austin 453.8 85.0
University of California-Berkeley 319.4 59.8
University of Wisconsin-Madison 283.5 53.1
University of Delaware 269.8 50.5
Stanford University 267.3 50.1
California Institute of Technology 246.9 46.2
Princeton University 227.4 42.6
Texas A&M University 226.4 42.4
Univ. of California-Los Angeles 225.6 42.3
Univ. of California-Santa Barbara 219.3 41.1
Univ. of Illinois, Urbana-Champaign 210.0 39.3
Case Western Reserve University 205.5 38.5
Pennsylvania State University 205.1 38.4
Northwestern University 203.1 38.0
Johns Hopkins University 202.7 38.0
North Carolina State University 196.4 36.8
State Univ. of New York-Buffalo 195.3 36.6
University of Michigan 185.9 34.8
Cornell University 183.0 34.3
University of Pennsylvania 182.4 34.2
Lehigh University 172.5 32.3
Carnegie Mellon University 171.7 32.2


Deviation"Rank
4.9 1
2.1 2
2.2 3
18.8 4 ('
8.2 5
6.7 6
9.9 7
7.3 8
6.6 9
20.4 10 d,
15.2 11
4.6 12
8.1 13
14.5 14
15.2 15
15.8 16
16.4 17
7.0 18
15.2 19
7.5 20
6.8 21
20.9 22"'
11.9 23
16.7 24


University
Purdue University
Univ. of Massachusetts, Amherst
University of Colorado
University of Oklahoma
Syracuse University
Ohio State University
University of Washington
University of California-Davis
University of Houston
Rensselaer Polytechnic Institute
University of Tennessee-Knoxville
University of Virginia
Georgia Institute of Technology
Louisiana State University
Oklahoma State University
Yale University


Composite Score
Rescalef' Normalied DeviationlRank
168.5 31.6 13.3 25
167.6 31.4 13.8 26
159.9 29.9 10.1 27
157.7 29.5 30.4 28
151.4 28.4 14.2 29
146.6 27.5 10.9 30
142.2 26.6 12.8 31
137.1 25.7 18.1 32
136.2 25.5 21.6 33
136.0 25.5 2.9 34
134.6 25.2 37.7 35'"
134.5 25.2 7.5 36
134.1 25.1 16.4 37
132.5 24.8 13.3 38
131.0 24.5 31.0 39d"
125.8 23.6 18.1 40


'"Extensive score plus rescaled intensive score as described in text.
"'Standard deviation of the rank order numbers of the eight quality
measures as described in text.
"'Ranking may be low because of different basis or error in Research
Support category.
'"'Ranking may be high because of different basis or error in Research
Support category.


Figure 3. (a) --
Fractional D 0.60
T. 0.60 0---------------- 60----------------
change in 0.60 0.60
citations, 0.50 0 50
A(TC)/(TC), _4 0.40

order; top- o 0.30- 0.30 -


on left. (b) o.o 0.10
Fractional .....0.
change in0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90
publications, Rank Order of Publications Rank Order of Citations
A (PUB)/
(PUB), versus
rank order; c
top-ranked 0.6 0- 0.60
programs are 0.50 0.50
on left. (c) I
Fractional 0.40 0.40
change in 0.30 0.30
research nl
support, 0.20 0.20
(SUPP), 0.,,.10 0.10
(SUPP)/
versus rank 000 o.oo00
order; top- 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
ranked Rank Order of Research Support Rank Order of Honors
ranked __ I
programs are
on left. (d) Fractional change in honors, A (HON)/(HON), versus rank order; top-ranked programs are on left.


Winter 1999










into the first decile or out of the tenth decile.
Without knowing details of the data collection, it is not
possible to make a rigorous assessment of the uncertainty in
the rank ordering. However, a heuristic assessment can be
made. We assume, based on experience and for purposes of
argument, that there are independent errors of 10% in each
of the four extensive numerical data sets: Publications, Cita-
tions, Support, and Honors. In the middle range of each of
the extensive data sets, the fractional change, AX/X, re-
quired to move one place in the rank order is approximately
0.03 (see Figure 3). A fractional error of 0.10 therefore
corresponds to approximately 0.10/0.03 = 3 places in the
rank ordering. Also, if the errors are independent, one would
expect an error in the composite rank ordering of extensive
criteria to be approximately r4(3) = 6 places. If similar argu-
ments are used for the four intensive data sets, we also find
an approximate error of six places. This error is not indepen-
dent of the error in the extensive rank ordering. We conclude
that, in the middle range, programs within 5 to 10 places on
the composite extensive/intensive rank ordering are essen-
tially indistinguishable from each other. This estimate is
consistent with our common-sense interpretation of the rank
ordering, e.g., programs in the second decile are probably
superior to programs in the third decile, and so on.
The eight separate (though not completely independent)
measures of quality give the rankings a certain degree of
robustness that a single criterion would not have. The inter-
nal consistency of the eight measures of quality is estimated
by computing the standard deviation of the rank order number
of the eight separate quality categories for each program (see
Table 8). For example, for MIT the average of the eight rank
orders is 3.5 and the standard deviation of rank orders is just

{(2-3.5)2 +(3 3.5)2 +} 207

Large values of the deviations indicate programs where di-
Large values of the deviations indicate programs where indi-


70 I
S R2 = 0.72 g
60 q---- 0

50 e 0

40
30 A
10


S* 6


0 10 20 30 40 50 60 70
Rank Order of Extensive Measure


= 0
0o

*
*) 1


* 0


o .


0 10 20 30 40 50 60 70
Rank Order of Intensive Measure


Figure 4. (a) Composite extensive rank order versus rank
order of faculty quality (93Q); top-ranked programs are
near origin. R2=0.72 for linear fit, y=mx+b. (b) Composite
intensive rank order versus rank order of faculty quality
(93Q); top-ranked programs are near origin. R2=0.65 for
linear fit, y=mx+b. (c) Composite intensive/extensive rank
order versus rank order of faculty quality (93Q); top-ranked
programs are near origin. R2=0. 72 for linear fit, y=mx+b


TABLE 9
Summary of R2 Values for Linear
Fits Between Rank Order of Faculty
Quality (93Q) and Rank Order in the
Individual Rankings

Ranking
Category Fit R2
PUB y =mx + b 0.7469
PUB/FAC y = mx + b 0.6005
CIT y = mx + b 0.7482
CIT/FAC y = mx + b 0.6028
SUPP y = mx +b 0.2484
SUPP/FAC y = mx+b 0.1351
HON y = mx + b 0.6993
HON/FAC y = mx + b 0.6370


Chemical Engineering Education











a


20




10-

5

0
0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100
Normalized Extensive Score Range


25






2 -0 i0--0--- -303--0-40-0---6----70-- -80-0-9--0-10
20


is


5

0
I l l
0-10 10-20 20-30 30-40 40-50 5060 60-70 70-80 80-90 90-100
Normalized Intensive Score Range

C
25

20



10
l-
tZ t=


0 1111


M


0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100
Normalized Composite Score Range


Figure 5. (a) Histogram showing number of programs ver-
sus scaled total extensive score; top-ranked programs are at
the right. (b) Histogram showing number of programs ver-
sus scaled total intensive score; top-ranked programs are at
the right. (c) Histogram showing number of programs ver-
sus scaled total composite extensive/intensive score; top-
ranked programs are at the right.
Winter 1999


F


Interpretation of Rankings There is no calibration stan-
dard for quality against which any methodology can be
tested. Nevertheless, we find it very suggestive that our
composite extensive/intensive ranking and the NRC
reputational survey identify the same few top programs. For
example, comparison of Tables 4 and 8 shows that the same
top two programs, MIT and Minnesota, and nine out of the
ten top-ranked programs are the same in both the NRC
reputational ranking and our numerical ranking. But only
three out of ten programs in the second decile and two of ten
in the third decile are the same. The NRC reputational


vidual quality rankings are the least internally consistent. In
many cases, large deviations are associated with small pro-
grams that rank higher in intensive than in extensive catego-
ries. However, in some cases, large deviations may indicate
problems in the data. For example, four programs have much
higher rankings in research support than in the other quality
categories: Texas A&M, Oklahoma State University, the
University of Oklahoma, and the University of Tennessee.
On the other hand, the University of California at Berkeley
and the University of Pennsylvania have much lower rankings
in research support than they have in the other categories.
We believe that these disparities likely arise from different
reporting bases and may not reflect true differences in re-
search support. Three programs have a much higher ranking
in Honors than in the other quality categories: Pennsylvania
State University, Carnegie Mellon University, and North-
western University.
The rank order of perceived faculty quality (93Q) was
correlated with the rank order of each of the separate quality
categories (see Table 9). The weakest correlations were found
with the two support categories, SUPP and SUPP/TF, con-
sistent with our belief that some of these data are not re-
ported on a consistent basis. Nevertheless, we are reluctant
to exclude research support from the quality measures. Re-
search support is probably a better current and leading indi-
cator of quality than the other categories. Also, total research
support is a primary criterion used for assessing quality of
research universities.151 Rather than re-ranking the programs
excluding SUPP and SUPP/TF, we believe it is more reason-
able to identify programs where different bases for reporting
support may have strongly influenced the rankings.
Plots of the rank order from the faculty quality survey
(93Q) from the NRC report versus the overall extensive,
overall intensive, and composite rankings are given in Fig-
ure 4. In the figure, the high-quality programs are near the
origin. The figure clearly shows how the quality survey
and our methods identify the same several programs as
the highest quality.


t'" -' -- Ph- .- --. IO '-










rankings, which rely heavily on anecdotal, word-of-mouth
information, will be most accurate for the few, high-profile,
extremely good programs, and will be less accurate for
smaller, lower-profile, and second-tier institutions. But the
numerical measures of quality should remain useful in as-
sessing the relative quality of all institutions. We conclude,
subject to the caveats given about the accuracy of the data
itself that our simple numerical measures do correlate with
program quality as it is normally understood.
Further comparison of Tables 4 and 8 leads to additional
insights. One may divide programs into three broad catego-
ries. First are the programs that are highly rated on both the
NRC reputational survey and the numerical ranking. Prime
examples are the University of Minnesota, Massachusetts
Institute of Technology, the University of Wisconsin, the
University of California at Berkeley, and Stanford Univer-
sity. Second are programs that rank significantly higher in
the numerical ranking than in the survey. These programs
are often (but not always) associated with smaller, research-
intensive programs. Examples are the University of Califor-
nia at Los Angeles, Case Western Reserve University, and
Johns Hopkins University. Finally, there are well-known
programs, which do well in the reputational ranking, that do
not do as well in the numerical measures. These programs
may be relying on past, rather than current, performance.
Further insight can be obtained from histograms of the
final scores, shown in Figure 5. For ease of interpretation, in
the figure the scores from Tables 6, 7, and 8 were scaled to
give maximum values of 100. Figure 5a and Table 6 clearly
show that three programs (MIT, Minnesota, and Texas) have
extensive scores well above all other departments. This dis-
parity is lessened somewhat when the intensive scores are
compared (Figure 5b and Table 7). This same uneven distri-
bution of scores is found in Figure 5c, which shows the
distribution of composite extensive/intensive scores. The top
half of the composite score range contains only seven pro-
grams; the remaining programs fall in the lower half. The
summary shown in Table 8 and Figure 5 indicates that the
highest quality chemical engineering programs are relatively
few in number and significantly higher in quality than the
rest. Below the top five or six programs there is a wide range
of programs with relatively similar quality.
Finally, while we believe that for most programs the
rankings given here are an accurate reflection of quality, we
emphasize once again that one should be cautious in draw-
ing conclusions from the absolute position in the rankings of
a single program.
Limitations of Ranking Systems Respondents to the
NRC questionnaire were asked to rate fifty separate pro-
grams. An individual respondent will only have personal,
detailed knowledge about a small fraction of these. The
resulting reputational rankings will inevitably be influenced
by the network of informal contacts and acquaintances of the


respondents. This will lead to a bias against smaller pro-
grams and will also make the reputational rankings a lagging
indicator of program quality.
Another major problem with the NRC report, recognized
by the committee, was the lack of data on the performance of
graduates from the programs. We were unable to find any
direct quantitative measure for assessing the performance of
graduates of chemical engineering graduate programs. Since
one of the principal goals of a graduate program is the
education of the next generation of researchers, this is a
serious omission indeed. Personnel departments of major
corporate employers of PhD chemical engineers often main-
tain internal ratings of programs based on the performance
of their employees. Perhaps this information can be pro-
vided in some suitable blind format to future NRC com-
mittees. This is a project that could be addressed by the
AIChE and the other engineering societies. Another pos-
sible measure of performance is the number of graduates
that obtain tenure-track appointments at research univer-
sities other than their own.
We suggest that future ranking systems also include some
measures of the effectiveness of technology transfer. To
partially accomplish this, the Publication category could be
expanded to include patents issued to faculty and graduate
students. Similarly, Citations could include papers or patents
cited within patents. More difficult to count, but very useful,
would be the number of new businesses formed as a result of
activities within the program.
The quality measures used in the NRC report and in this
paper are appropriate for doctoral-level, research-based gradu-
ate programs. However, master's-level programs, especially
practice-oriented programs, are of growing importance. Fu-
ture ranking systems should attempt to separately measure
the quality of these programs.
The difficulty in accounting for the rapidly changing, in-
terdisciplinary nature of modern engineering is another prob-
lem encountered when developing ranking systems. Tradi-
tional academic boundaries do not always reflect the reali-
ties of engineering practice. The NRC report addressed this
problem by ranking "programs" rather than "departments."
For chemical engineering, these two categories are usually
commensurate, but this may not be the case for chemical
engineering programs with strong efforts in biotechnology
or advanced materials. Ranking programs with major com-
mitments in these fields can be difficult when the academic
administrative units do not correspond to the categories used
in the ranking scheme. Very strong, interdisciplinary efforts
may not appear in the data set, or conversely, remote extra-
neous efforts can be included. Obtaining a reliable data set,
based on uniform criteria, is a formidable task. The NRC
committee had great difficulty in defining program bound-
aries in modern biology and molecular biology, where the
pace of change is particularly great.
Chemical Engineering Education










Neither the NRC report nor this paper uses any mea-
sures of the quality of graduate teaching. The lack of
quantitative measures of teaching performance is a con-
tinuing, long-term problem.
The very long time between the NRC reports (1982 to
1995) is yet another problem. Waiting more than a decade
for an assessment is slow, even by the standards of academia.
Some form of continuing assessment, for example on a
triennial basis, would be more useful. This would give more
timely results and would also permit running averages of
several years to average out fluctuations in the data.
Concluding Remarks With all of these difficulties, one
can legitimately ask why bother with rankings at all? We
believe that universities will be under ever-increasing pres-
sure to justify tuition rates and the cost of performing re-
search. Whether we like it or not, ranking of academic pro-
grams will continue and will likely increase. It is in the
profession's interest to see that the rankings are based on
rational, measurable criteria. But there is little reason to
continue relying on surveys. Reputational rankings only con-
firm the obvious about the top few programs, permit declin-
ing programs to remain complacent, and fail to recognize
increasing quality where it occurs.


Conclusions and Recommendations

I) Alternative, measurable quality indices exist that correlate
well with graduate program quality as it is normally under-
stood.
C The professional societies, the National Academy of Engi-
neering, and the National Science Foundation should take
the lead in developing these quantitative measures of pro-
gram quality and appropriate data bases to support these
measures.
C Special attention should be paid to developing methods for
assessing the performance of students after they receive their
graduate degrees; this should include using information from
employers of graduates.
C) Methods of assessing the effectiveness of technology trans-
fer and impact on industry should be developed.
C Assessments should be made on a more frequent schedule,
perhaps triennially.

REFERENCES
1. Goldberger, M.L., B.A. Maher, and P.E. Flattau, Eds., Re-
search-Doctorate Programs in the United States: Continuity
and Change, National Academy Press, Washington, DC
(1995)
2. "America's Best Graduate Schools," U.S. News and World
Report, 2400 N St. NW, Washington, DC 20037-1196 (1996)
3. Mervis, J., Science, 168, 1693 (1995)
4. National Science Foundation, Academic Science and Engi-
neering R&D Expenditures: Fiscal Year 1993, NSF 95-332
(Arlington, VA, 1995) Table B-48; and National Science
Foundation, Academic Science and Engineering R&D Ex-
Winter 1999


penditures: Fiscal Year 1994, NSF 95-332 (Arlington, VA,
1996) Table B-48.
5. Graham, H.D., and N. Diamond, The Rise of American Re-
search Universities, Johns Hopkins University Press, Balti-
more, MD (1997)
6. National Science Foundation web site, April, 1997
7. Activities Directory '96, American Institute of Chemical En-
gineers, AIChE, New York, NY (1996)
8. Directory of Members and Foreign Associates, National Acad-
emy of Engineering, Washington, DC, September (1996)
9. Chemical Engineering Faculty Directory: 1996-1997, 45,
AIChE, New York, NY (1996) O


BOOK REVIEW: Alternative Fuels
Continued from page 39.

The use of geothermal energy is presented in Chapter 9.
This topical discussion notes that at depths of about six miles
from the earth's surface, the temperature is greater than
100'C. This equates to more energy storage than the total
thermal energy in all the nuclear and fossil fuel resources-
only solar energy is comparable. Along with scientific and
technological updates, the advantages and disadvantages of
geothermal energy utilization are outlined; this alternative
source of energy will potentially become a larger part of the
world's energy consumption in the near future because geo-
thermal energy is both available and economical. In the
United States, approximately 3 GW of electric power is
produced in 20 power plants from geothermal reservoirs.
Geothermal energy also has great potential as a practical
provider of heat to local areas.
The overall conversion routes of biomass are described in
Chapter 10. They include thermal (combustion, gasification,
liquefaction, and pyrolysis), anaerobic digestion, and fer-
mentation to liquid ethanol fuel. The descriptions in-
clude 15 process diagrams and several tables of data. A
selected amount of cost data is provided for ethanol pro-
duction from lignocellulose.
Chapter 11 presents a comprehensive overview of rela-
tively recent developments in the generation of energy from
municipal solid wastes, including spent tires and polymeric
materials. Processes include incineration, anaerobic diges-
tion and landfill gas recovery, pyrolysis, thermal cracking,
and partial oxidation via supercritical fluids.
In summary, Alternative Fuels superbly achieves its pur-
pose by bringing together a wealth of practical information
required for a thorough understanding of those chemical
process technologies urgently needed for the development
of fuels for future use. Dr. Lee is to be commended for his
extraordinary efforts in synthesizing all these facts and sys-
tems in a clear and consistent manner. Possibly his next
book could focus more on the use of biomass, geothermal,
and solid waste resources-three areas that are undergoing
rapid development. O










u learning in industry


This column provides examples of cases in which students have gained knowledge, insight, and experience in the
practice of chemical engineering while in an industrial setting. Summer internships and co-op assignments typify such
experiences; however, reports of more unusual cases are also welcome. Description of the analytical tools used and the
skills developed during the project should be emphasized. These examples should stimulate innovative approaches to
bring real world tools and experiences back to campus for integration into the curriculum. Please submit manuscripts to
Professor W. J. Koros, Chemical Engineering Department, University of Texas, Austin, Texas 78712.



FROM THE CLASSROOM TO

THE WORKPLACE

Motivating Students to Learn in Industry

A. CHRISTIAN FRICKE
Rensselaer Polytechnic Institute Troy, NY 12180


What makes a successful engineer? No one would
deny that technical expertise is critical to master-
ing real-world engineering problems. Yet techni-
cal mastery is only half the battle; there are also many com-
plex social skills that must be learned in order to make success-
ful use of technical knowledge in a workplace setting.
The purpose of this article is twofold. First, it will note a
few examples of specific nontechnical skills that can be
useful in managing the day-to-day workplace realities of a
BS-level engineer. These skills and strategies are taken from
the author's personal experience in working as a process
engineer for three years at a mid-sized manufacturing con-
sulting firm, as well as from conversations with and observa-
tions of dozens of colleagues working in varied chemical
and mechanical product design and manufacturing settings.
The majority of these engineers were within six years of
graduation and were in the process of learning the social
skills necessary for moving up the corporate ladder from
technically oriented process positions to more business-ori-
ented managerial functions. It is hoped that these observa-
tions will prove useful to engineering professors who have
not worked for an appreciable amount of time at the BS level
and who therefore have experienced the industrial setting in
a much different context.
Second, this article will outline several ways in which the
need for acquiring these informal skills can be communi-
cated to the vast majority of students who will end their
education at the BS level. The conveyance of technical con-
cepts, skills, and information is undoubtedly what the under-
graduate experience is all about, but by suggesting some of
the social contexts within which these skills will be mobi-


lized, engineering educators can increase students' effec-
tiveness in putting this technical material to use in the work-
place.
There exists a real need to alert undergraduates to the fact
that excelling in the classroom, although critical, is only half
the equation in preparing to be an effective professional.
Otherwise naive students need to be explicitly made aware
of the seemingly commonsense notion that one must indeed
"learn in industry" in order to be a successful corporate
engineer.

STUDENT PERCEPTIONS
This need is illustrated by a recent survey of seventy-six
undergraduate engineers at Rensselaer Polytechnic Institute
(RPI) in which 95% indicated they had a "very firm" or at
least a "somewhat firm" idea of what the daily work experi-
ence of an average engineer is like (see Figure 1). A close
look at the numbers indicates that students likely do not have
the firm grasp of engineering workplace realities that they
profess.
For instance, there was no correlation between students'

A. Christian Fricke graduated from North Caro-
lina State University with BS degrees in chemi-
cal engineering and biochemistry. He has
worked in a Merck Pharmaceuticals production
facility, as a research assistant in a molecular
genetics laboratory, and as a process engineer
for CTC, a manufacturing consulting firm with
headquarters in Johnstown, Pennsylvania. He
is currently a doctoral candidate in the Science
and Technology Studies program at RPI.

Copyright ChE Division of ASEE 1999
Chemical Engineering Education











There exists a real need to alert undergraduates to the fact that excelling in the classroom,
although critical, is only half the equation in preparing to be an effective professional... Success on
the job ... entails learning many complex social behaviors in addition to those necessary for classroom
success. It also entails developing an entirely new perspective on what constitutes "engineering."


reported knowledge of "what it is that engi- TA
neers do on a daily basis" and their personal Overall
relationships. Students who had no close rela- (76 e
tives or acquaintances with engineering back-
grounds (more than 60% of those surveyed) Resonseto
were just as likely to indicate a firm knowl- 'Do you have
edge of daily working realities as those with it is that engine
engineers "in the family." In the absence of basis?"
actual engineers to talk with and observe, Very
student conceptions of "workplace realities" Somewhat
NoF
are vague and simplistic at best. No Idea W
This simplistic view can undermine an
engineer's effectiveness in accomplishing personal, profes-
sional, and societal goals in the workplace. In addition, the
apparently prevalent student attitude of believing they al-
ready know what professional working realities are all about
can seriously limit the benefits to be gained from intern and
co-op experiences. These experiences provide the ideal set-
ting for observing the practical day-to-day social skills and
strategies necessary for BS-level success. In order to realize
this benefit, however, students must be actively looking for
these potential lessons in the first place.

UNDERGRADUATE SOCIALIZATION
The majority of undergraduate students form their first
concrete conception of "engineering" through survey courses
and introductory seminars that are structured to help fresh-
man and first-semester sophomore students choose a par-
ticular discipline. At RPI, for example, second-semester fresh-
man students take a course, titled "Engineering Seminar,"
that is designed to "provide the student with information
relative to the various engineering fields and curricular
areas."'1 These types of survey courses generally focus
on the end products of engineering work. In other words,
they emphasize what it is that the various disciplines
accomplish. They leave undergraduates with a feeling
that they understand what it is that engineers do, but
without an appreciation for the social realities of how
these tasks are accomplished.
This distinction is significant. Undergraduate coursework
fosters the perception that the engineering working experi-
ence is one of solving highly conceptual, well-defined, sci-
ence-based problems in a largely individualized setting, with
emphasis on arriving at a single, objectively "correct" solu-
tion. But the reality that working engineers encounter is one
of solving highly practical, undefined, procedural-based prob-
lems in an extremely tight-knit social setting, resulting in


BL
iurve
spon

eque
a firn
eers

'irmI
'irmI
irm I
latsoe


Winter 1999


,E 1 multiple potentially "correct" solutions. In
y Ress his book Designing Engineers,12) Louis
dents) Bucciarelli characterizes these fundamentally
social aspects of engineering design and prac-
*stion: tice in the following manner:
n idea of hat
do on a daily [P]articipants in design work within a rich,
multidimensional environment that reaches
dea- 30% well beyond the narrow confines of their
dea: 65% own object worlds. A customer's needs are
dea: 5% not given or discovered, but must be cre-
ver. 0% ated; an operator's capabilities must be de-
fined; building codes need interpretation;
costs must be tried out; budget limits must be agreed
upon. The task must be organized into subtasks; suppli-
ers must be coaxed to commit to a price and delivery
date; the dropout problem at Photoquik must be con-
structed. All of this is designing. In all of this, choices
are being made, decisions foreshadowed, and possibili-
ties discounted.

In other words, working engineers must create and manage
formal and informal social structures in order to generate
built products.
Success on the job therefore entails learning many com-
plex social behaviors in addition to those necessary for class-
room success. It also entails developing an entirely new
perspective on what constitutes "engineering." Without re-
structuring the entire undergraduate experience to incorpo-
rate these workplace lessons, engineering educators can nev-
ertheless prepare students for this impending paradigm shift
by at least bringing it to their attention. In addition, there are
many specific exercises that can be easily incorporated into
the existing undergraduate curriculum to reinforce some of
the nontechnical social skills necessary for success in the
corporate workplace.

WORKPLACE REALITIES
According to one early '80s study, "technical profession-
als typically spend over a third of their work week writing,
editing, or preparing reports."'3' If you also include compos-
ing letters, proposals, drafting schedules and procedures,
taking field notes, and generating other more informal modes
of written communication, then "writing" easily occupies
more than half of the typical engineer's work experience.
Oral communication also occupies a major portion of the
engineer's time. This can include time spent in meetings or on
the phone with vendors or customers, time spent on the shop
floor interacting with technicians and workers, etc.










Taken together, these two activities comprise by far the
bulk of an engineer's work week. In the workplace setting
that the vast majority of graduates will enter, typical engi-
neers will likely use only 10% of their technical background
10% or so of the time. Of course, the specific 10% will vary
widely for each individual, making technical breadth within
the curriculum essential. But the fact remains that most of a
working BS-level engineer's time will be spent not actively
solving technical problems, but instead communicating po-
tential technical solutions to others. Engineers' effective-
ness, reputation, and career success will be based on techni-
cal expertise, yet determined by how well they manage to
translate this expertise into action through mastering such
nontechnical workplace skills as effective communication,
organization, and persuasion.
This is, of course, not a one-way flow of information. For
every memo that is written or presentation that is given,
someone (presumably) reads and listens. The successful en-
gineer also has to take in and interpret an enormous amount
of written and verbal information. Organizing and making
effective use of this information requires good critical read-
ing and listening skills. Given the enormous amount of in-
formation generated in the typical corporate workplace,
quickly and effectively separating the wheat from the chaff
is an important skill in itself.
The vast majority of a BS engineer's time is taken up with
both taking in and communicating information. Success re-
quires possessing the "nontechnical" skills necessary to first
recognize and then convince and organize others to act on
information that is important.

CONVEYING WORKPLACE REALITIES
So, how can you alert engineering undergraduates to this
reality? One strategy is to suggest that effective communica-
tion is an essential engineering skill-one that can be just as
important as any technical ability. In the words of historian
Henry Petroski,'4' "some of the most accomplished engi-
neers of all time have paid as much attention to their words
as to their numbers, to their sentences as to their equations,
and to their reports as to their designs." Pointing out to
students the vital importance of mastering effective writing,
reading, speaking, and listening reinforces the notion of
engineering practice as a social activity.
Yet there are many other "nontechnical" skills that are
also important to success at the BS level. If presented to
students at all, these skills are most often communicated in
the most general of terms, with successful engineers de-
scribed as possessing "curiosity," "perseverance," "self-con-
fidence," "common sense," and so forth. What undergradu-
ates need is a resource that highlights the importance of
specific skills, motivated by a concrete social context and
picture of the day-to-day realities of corporate engineering
practice. The key is to motivate students to appreciate the


complex social realities of engineering practice by giving
them a tangible feel for the workplace setting that most will
find themselves in.
Using Popular Culture One resource for accomplishing
this is the comic strip "Dilbert." In many respects, "Dilbert"
is an entirely accurate ethnographic account of the typical
BS-level engineering experience. According to one leading
management consultant,'5' "It's not a comic strip, it's a docu-
mentary-it provides the best window into the reality of
corporate life that I've ever seen." It therefore provides an
excellent resource for undergraduates' (or anyone else, for
that matter) interested in the daily interactions of practicing
corporate engineers. "Dilbert" can be read as providing very
specific, contextualized examples of the many workplace
issues and challenges that BS-level engineers must confront
and overcome in the process of applying their technical
knowledge to real-world problems.
Of course, illustration does not imply prescription. "Dilbert"
should certainly not be taken as illustrating a social ideal or
model for how engineering professionals ought to navigate
these issues. It can, however, offer a view of what some of
these issues are and motivate students to contemplate how
they would manage similar circumstances in a more con-
structive manner. "Dilbert" provides an alternative insider's
perspective that, if presented as serious social satire and
critique, can be a valuable learning tool for preparing for the
reality of the engineering workplace.
Taking "Dilbert" as serious social commentary can pre-
pare students for making the leap from viewing corporate
engineering as a purely technical activity to seeing it as a
technically mediated, yet essentially social, endeavor. It can
also prevent the disillusionment commonly generated by the
experience of realizing that daily workplace realities are
quite different from naive undergraduate preconceptions.
Discussing Specific Strategies A truly comprehensive
list of specific social skills useful for managing "Dilbert"-
like situations in a constructive manner would be almost
infinite in length. This section merely presents five strate-
gies that can be particularly critical to workplace success.
"Newly minted" BS-level corporate engineers usually learn
these strategies only after a sometimes painful and poten-
tially damaging period of trial and error. Discussing these
important nontechnical skills within the context of appropri-
ate undergraduate coursework can benefit graduates by ac-
celerating their on-the-job learning curve enormously.
1. Save everything that crosses your desk. Undergradu-
ate education reinforces the notion that when some-
thing is "done," it's over with. With the end of each
semester, textbooks are returned to the bookstore and
class notes are relegated to recycled paper bins. But
in the real world, projects never really come to an
end. You never know when, say, a cost analysis done
for a long-forgotten proposal might come in handy.
Chemical Engineering Education










Saving old work (even draft work) can prevent future
duplication of effort.
2. Document ci r '\ hilnL, in writing. There is no such
thing as an unambiguous verbal order.
3. Learn to use a daily planner. Some corporations
provide a standard dayplanner system free of charge
to technical employees, but even if their use is not
officially encouraged, dayplanners are an essential
tool for maintaining order in personal and project
schedules (planning meetings, scheduling travel,
keeping notes, maintaining contacts, etc.). Under-
graduates are used to having order imposed for
them-tests, project deadlines, class times, course
materials, etc., are all organized in advance. This
leaves novice engineers completely unprepared for
the job of creating their own order from the chaos of
daily events. A good dayplanner system is an
indispensable tool for managing this process.
4. Use the "plus a fifth" rule. One of the most difficult
things to learn in managing complex technical efforts
is how to account for the unexpected. Even the most
detailed, well-researched proposal or project plan can
be subject to unanticipated delays, setbacks, cost
overruns, and instances of Murphy's Law in action.
Planning for the unforeseeable is a management skill
that can only be learned through experience. In the
meantime, beginning engineers can instead simply
assume that all but the most straightforward tasks
will take 20% longer and cost at least 20% more than
expected. Even if nothing goes wrong, coming in
significantly under budget and ahead of schedule can
be much preferable to the alternative for all involved.
5. Learn where and when to compromise. A critical
skill for managing working relationships is knowing
when an issue is important enough to battle over. One
engineer who is employed by a large tool and home-
appliance manufacturer characterizes this as realizing
that "the sun doesn't rise and set on a toaster oven."
Maintaining an uncompromising stance on, say, the
color of a new product can make collaboration
impossible. But no compromise should ever be made
on any aspect of engineering design, production, or
management that infringes on the health or social
welfare of others. Students should recognize that they
will have to make such distinctions for themselves and
that the answers will rarely be clear-cut and obvious.
Practicing Workplace Strategies in the Classroom In
addition to simply discussing the aforementioned strategies,
there are also many relatively simple, straightforward teach-
ing techniques that can be employed to help students de-
velop positive social working skills like good communica-
tion, organization, planning, patience, etc. Once again, pre-
senting a truly comprehensive list would be impossible-
Winter 1999


this is merely a collection of seven specific activities to
illustrate the breadth of possibilities.
1. Performing peer evaluations for student oral presen-
tations. Having undergraduates evaluate one another's
presentations in a structured manner would focus
listening skills and give students practice in recogniz-
ing key points, initiating critical discussion, etc.
2. Practicing giving oral and written equipment
operation and sampling procedure directions. Unit-
op labs provide a plethora of opportunities for
sharpening interpersonal communication skills. For
example, students could actively direct and observe
each other rather than passively following TA
instructions. Also, students could be required to
generate written operation and sampling procedures for
subsequent lab groups to follow. These experiences
would highlight the importance of precision and clarity
in giving both written and verbal direction.
3. Swapping notes for lab reports. Another potential
unit-ops exercise would be to require groups to
exchange notes and generate reports based on each
other's data. This would highlight the importance of
preserving data and other information for unexpected
future uses while also stressing the necessity of
precision and clarity in all forms of engineering
communication.
4. Writing and presenting standard business communi-
cations. Practice in writing and speaking can be
combined with aiding students in their job search.
First, students could be asked to research and
produce a short report and presentation on a particu-
lar industry or market. Then students could generate
a resume and letter of application to an appropriate
company based on their research. This would give
students practice in evaluating what is important to
individuals working within other organizations, while
at the same time reinforcing proper business communi-
cations etiquette, sharpening business research skills,
etc. Students could also gain valuable experience from
attending departmental seminars and producing short
memos or similar communications detailing key
information presented, summarizing discussions, etc.
5. Producing detailed project plans. Senior design
courses also provide a wealth of opportunities for
practicing nontechnical, "real-world," social and
organizational skills. For example, students could be
required to generate detailed project proposals
outlining specifically what is to be done and how it
will be accomplished, complete with a breakdown of
activities, timelines for completion, etc. This would
give students experience in organizing work and
delegating responsibility in a formal and considered
fashion. Coupling this activity with a proposal










presentation would also give students practice in
clearly articulating and advocating a proposed course
of action. Also, requiring periodic project updates
would prompt students to manage delays and
setbacks in an organized fashion.
6. Dealing with vendors. Another good senior design
experience would be to ask students not just to model
a process, but to also locate, spec out, and price the
specific equipment necessary to make the process run.
This could be done by simply giving students access to
a Thomas Register (now available on the Internet) and
a telephone. This would expose students to the realities
of uncertainty and would likely require management of
time and (imaginary) cost overruns.
7. Rotating group members and responsibilities. Per-
forming well on corporate-engineering project teams
means responding constructively to change. This can
be simulated in the classroom by requiring students to
periodically reshuffle project, lab, and homework
groups. Similarly, specific roles such as coordinator,
note taker, etc., can be rotated within groups. Having
to constantly fill new roles and interact with different
individuals would sharpen both leadership and col-
laborative, cooperative interpersonal skills.
In general, implementing these activities would require
some extra TA work and additional time spent in providing
more qualitative feedback on assignments. But all of these
suggestions can be incorporated into existing coursework
with a minimum of curricular disruption.
Outlining the Positive "Qualities of Success" Finally,
simply presenting a few specific contexts of engineering
practice can also motivate students to begin thinking about
the contrasts between their undergraduate educational expe-
rience and their impending BS-level working reality. For
example, successful engineers (whether employed in sales,
processing, design, management, or any other capacity) are
often called upon to
Recognize problems that aren't apparent
(For example, being able to walk through a production
floor and see opportunities for cost-savings, or recognize
subtle ethical questions that others overlook.)
Define problems that are nebulous
(Problems in the real world rarely come numbered for
easy reference.)
Choose solutions that are realistic
(A skill that doesn't necessarily mean limiting the range
of possible solutions-often the most successful engineers
are ones who recognize the practical possibilities of
seemingly impractical approaches.)
Plan how to make solutions work
(A process that includes marshalling resources, motivat-
ing others, keeping people on task, recognizing potential


pitfalls, and a complex combination of many other
discrete social skills.)
Convince others to follow
(This involves recognizing that the importance and
urgency of a "problem" and the feasibility of a "solu-
tion" are directly proportional to the skill and clarity
with which they are presented and defined.)
Cooperate in dealing with contingencies
(This involves realizing that exercising patience and
understanding with one's group members generates the
cooperative solidarity necessary for overcoming crises.)

Undergraduates should view themselves as continually
striving to meet this positive ideal through the constant ac-
quisition of constructive practical social skills.

CONCLUSION
In summary, undergraduate engineering students likely
confuse familiarity with what engineers produce with what
professional engineers actually do on a daily basis. This
confusion is reinforced through introductory engineering sur-
vey courses and an overall curriculum that emphasizes the
built products and technical aspects of engineering over the
social processes through which these products are generated.
Although students are given an opportunity for direct expo-
sure to engineering workplace realities through intern and
co-op experiences, the aforementioned preconceptions are
counterproductive to using these experiences in the context
of developing genuine, conscious insight into the essential
social aspects of engineering practice.
Engineering educators would benefit students by simply
alerting them to the fact that the creative, challenging pro-
cess of learning to "do" engineering will not end, but only
begins, at graduation. Likewise, undergraduates would ben-
efit from being presented with contextualized examples of
the daily workplace realities of corporate engineering life.
Taking a few moments to illustrate the social side of engi-
neering practice, along with devoting some curricular effort
to reinforcing these aspects of engineering work, would help
motivate students to think about their professional futures in
concrete terms and provide undergraduates with a construc-
tive context for developing positive professional social skills.
In the end, this would result in more reflexive, more effec-
tive engineering professionals.

REFERENCES
1. "Engineering Seminar," Rensselaer Catalog 1996-97, RPI
Publications, Troy, NY, 317 (1996)
2. Bucciarelli, Louis L., Designing Engineers, MIT University
Press, Cambridge, MA, 149 (1994)
3. Lampe, D.R., "Engineers' Invisible Activity: Writing," Tech-
nology Rev., 73, April (1983)
4. Petroski, H., "Engineers as Writers," Amer. Scientist, 423,
September-October (1993)
5. Levy, S., "Dilbert's World," Newsweek, 53, August 12 (1996)
Chemical Engineering Education
Chemical Engineering Education












AUTHOR GUIDELINES




This guide is offered to aid authors in preparing manuscripts for Chemical Engineering Education (CEE), a
quarterly journal published by the Chemical Engineering Division of the American Society for Engineering
Education (ASEE).
CEE publishes papers in the broad field of chemical engineering education. Papers generally describe a
course, a laboratory, a ChE department, a ChE educator, a ChE curriculum, research program, machine
computation, special instructional programs, or give views and opinions on various topics of interest to the
profession.


SSpecific suggestions on preparing papers *
TITLE Use specific and informative titles. They should be as brief as possible, consistent with the need for
defining the subject area covered by the paper.

AUTHORSHIP Be consistent in authorship designation. Use first name, second initial, and surname. Give
complete mailing address of place where work was conducted. If current address is different, include it in a
footnote on title page.

ABSTRACT: KEY WORDS Include an abstract of less than seventy-five words and a list (5 or less) of
keywords

TEXT We request that manuscripts not exceed twelve double-spaced typewritten pages in length. Longer
manuscripts may be returned to the authors) for revision/shortening before being reviewed. Assume your
reader is not a novice in the field. Include only as much history as is needed to provide background for the
particular material covered in your paper. Sectionalize the article and insert brief appropriate headings.

TABLES Avoid tables and graphs which involve duplication or superfluous data. If you can use a graph, do
not include a table. If the reader needs the table, omit the graph. Substitute a few typical results for lengthy tables
when practical. Avoid computer printouts.

NOMENCLATURE Follow nomenclature style of Chemical Abstracts; avoid trivial names. If trade names
are used, define at point of first use. Trade names should carry an initial capital only, with no accompanying
footnote. Use consistent units of measurement and give dimensions for all terms. Write all equations and
formulas clearly, and number important equations consecutively.

ACKNOWLEDGMENT Include in acknowledgment only such credits as are essential.

LITERATURE CITED References should be numbered and listed on a separate sheet in the order
occurring in the text.

COPY REQUIREMENTS Send two legible copies of the typed (double-spaced) manuscript on standard
letter-size paper. Submit original drawings (or clear prints) of graphs and diagrams on separate sheets of paper,
and include clear glossy prints of any photographs that will be used. Choose graph papers with blue cross-
sectional lines; other colors interfere with good reproduction. Label ordinates and abscissas of graphs along the
axes and outside the graph proper. Figure captions and legends will be set in type and need not be lettered on the
drawings. Number all illustrations consecutively. Supply all captions and legends typed on a separate page.
State in cover letter if drawings or photographs are to be returned. Authors should also include brief biographi-
cal sketches and recent photographs with the manuscript.


Send your manuscript to
Chemical Engineering Education, c/o Chemical Engineering Department
University of Florida, Gainesville, FL 32611-6005







Full Text

PAGE 1

chemical engineering education .. ... C ... ... ... ::::: '= t.;i 0.() ::: ... lo,. ::: ... 0.() ::: t.;i AWARD LECTURE lo,. ti ... ... C "l ::: feature articles on ... ... ... lo,. :., lo,. Discontinuities in ChE Education (pg. 18) ............................................................................... Whitaker E ::: Permeation of Gases in Asymmetric Ceramic Membranes (pg. 58) ............................... Fino/, Coronos ... ::: 0.() C ::: ... t.;i :., ... Non-Adiabatic Container Filling and Emptying (pg. 26) ........................................................ de Nevers Random Thoughts: FAQS (pg. 32) .................................................................................... Felder, Brent ::.. ... ... ... 0.() E ::: ... -::: lo,. I,) 'c.. ... Computer-Simulation Experiments in a Senior-Level Capstone Course (pg. 34) ........... White. Bodner Demonstrating Simultaneous Heat/Mass Transfer with Microwave Drying (pg. 46) ..... Steidle, Myers Medical Surveillance and the Undergraduate Thesis (pg. 50) ..................................................... Fur-:,er Lab Experiment in Bioc h emical E n ginee r ing: Ethanol Fermentation ( pg. 54) ............... Badino, Hokka ... 0.() ... ::: ::::: Introducing Process-Design Elements in the Unit Operations Lab (pg. 66) .... ....... McCa/111111, Este\'e-:, t.;i ... ... ... "' ::: ... ... .... Process Analysis: An Electronic Version (pg. 40) .................................................................. Delancey ChE/Electrical Engineering Course in Advanced Digital Process Control (pg. 62) .... Feeley, Edwards E ::: -::: ... I,) ... lo,. Ranking Graduate Programs: Alternative Measures of Quality (pg. 72) ........ Angus. Edwards, Schult~ From the Classroom to the Workplace (pg. 84) ........................................................................... Fricke E and Ch at ... ~

PAGE 2

UNITEDSTilTES POSTilLSERVICE Statement of Ownership Management and Circulation 1 PubllcationTin, squ f 39 USC-3685) 2 P\tlllcallcn~r 3 fllngOai. CHE'J1ICAL ENGI NEERIN G EDUCATION 1 1 1 /-1~1 I 9 /8/9 8 4 IHu efrequeney S NlffbwollUU.1PublllhedArn.1111y 6 Anr-u. SublCripliDn Price 4 See Attached (R lred b Rate s h eet 7 Comcilet. AddrtN of Kn:lvmCllleeof Publb.rlon (Nalpml.,J (SITH(, dry; CCIUn')': st.le anc,ZJP .4 } """"'""""' CHEM ICAL ENGINEERING EDUCATION, Room 319, Chemlcai E ng 0 ine~rins Dept. Ca r ole Yocum University of Florida, Gai ne s v ille Alach u a, Flori d a 326 11 -6005 T el1phone 352 392-0S61 II Comp111 Malling Add. IU ol Headqua1ter1 Of a.n.,.i &uu.u Office of Publllhat (Nolpr.tl/91 Chemical Engineering Dhision, American Society for Engin~ering Education 11 DuPont Cir-cle, Washington DC 20030 9 Full Name Ind COmpl,11 Mailing AddtHMI 01 Publllh1r, E dilor, and Manllgi-,o Edi tor (Oo not,._.., OlanJI } Publllhlr ( N .rn e lll'ld ~/11 ,najJjng addrt:U ) A SEE-C h c m icol E n g. Divi sio n, 11 DuPont Circle, Was h ington DC 20030 Edilor{,,,.,,,_atldl.'.iCIIJJN/tlmailngllddtwu J Tim An ders on, C hem, Eng Dept. Univ o f F l orida Gai ne sville FL 32611 6005 Caro l e Yocum chem. Eng. Dept,, University of Florida, Gainesville FL 32611-6005 Official publi cation of Publisher Any mail a ddr essed to owne r should go t o a s listed above Editor o r Hanaging Edito r listed above. 11 Known 8ondholdart. MOl'lgagau and Olhar Sac:urily Hoklars Ownng or Hoidwlg 1 P en:ent Of Mota of Tollll Amounl of Bonds, Mol'loagu o, Ol'letS.:urlCN llrw:ine cneckbox 12.. TuStaiu. ( Fof~IJyrronproltOl(/IIWlklfllwthoriudlOmailal,pJ111fQJ(CMckon. J ~~:!n':.Otg4(W tiot, ltld IN fflPll'-tul lorr.deralhxwna IUpupoMS: 0 Has Changed Cuing P1Kedlng 1 2 Monlhl (~IIIUII .ubtril uplMliltian ol dlatlg. l'l'lfn IIQ s,./9flllll'l(I PS Fotm 3 526 5-iitarri>ar 1 11115 13. Publcdon Tide 1<1 i.-01tebCmalianDetaBalow C II F, MI C AL ENGINEERING EDUCAT ION Summe r 1 998 E.llta nt a nd Na lu r o l Clrculatlon Av11rage No Cop l 111 Each I H U II Actu1 I N o, Cople ol Sl ngl1 l11u 1 During PrlCldlng 1 2 M onlh, Pu bll1 h1d N rn l lo FIUng 0111 a l'OlalNl.fflllrolCoples(N 4 prusnn) 2075 2060 I c11s.1 1ThroughDN11tswldc.mer. stree1vendl;Q 0 0 b Paldaro'or and Countar Sain (Hof maJ l d ) A quNt d C"""'lion 1 {2 )PaldorA qullSledMal Sublcrlpllonl(lnc:AlcM 1978 200J adWta'sllnp,oo/~and~co,HI) c Tow! Pald~or A~Htltd Cirl:Wltion 1978 200J (Sumol 15b(t)Md t 5b(2)) d. FrM 011tnb1.11ion by Mail ( SMr'pllS, ~'-):MJrJallrerfrN) J5 JS F .... Chtributlon Oulaldl the Mail (C.t rrlftra a, o/MrmNIMI) 0 0 I. lolal Free Olstrilutlon (Sum ol 15d and tS.J JS JS g Tolal0istri>uion(5'1mol15candl5f) 2013 20)8 l( I J()IIQUM,L11loven,Sp;,ll d 62 22 h Copillrv:it Distributed I (2) Aatums f rom News Age.nts 0 0 I. lolal (Sumol l Sg. 1511(1 }. nd 1511{2 )) 2075 2060 Psetnl Paid ard'or A quesled Cm.Mtlon 98 26% 98 28% (15t:ll6gx 100 ) I eerily Iha! an ln loon don lurFQhed on !his loon Is true .xi comp ,11 I lrlMl'1 t wld lh.lt anyon who 1\1~ lal11 or mlsleailng inlonnalion on lti lorm o,whoOITT\I rret1rlal Ot lnlomllllion reques t ldonlht!OflTlmaybl,ubjlctlocrinr.el NirlCllons (lnc:tudng ftne andlmprltonm nl) arldJof dYII unctlonl (lncludng ITUtipltl dama9M 9fld civil penalliN) Instructions to Pu b lishers 1 Complet e and file ona copy o l this f onn with your postmas t er annualy on o r be! ore Oclober 1 Kee p a copy o l the comple ted form lor your records 2.. I n cases whera the 1 tockholder or security holder ii a trustaa, lndude In items 10 and 11 the name ol the person or corporation tor whom the trustee la actlng. Also incl ude the names and addles e s ol lndMduals who are stoc:kholdera who own or hold 1 percent or more of the total amount or bonds, mortg;ego a, or other aecllitlea ol lhe l)l.bllsl'ing corporation In ite m 11 II none check the box. Use b l ank a hee l s 11 more s pace Is required 3 Be sure to fumiah al circula t ion in!Offllation c alled tor I n item 15. Free clrculatlon must be shown I n item& 15d. e, and I 4 If lhe publication had seconck:l ass authorization as a 99M~I or requeste r publication, this Statitment ol Ownership Management and Cln:u1allon must be publilhed; it must be printed In any Issue 1n October or ii' the publcation Is no t J)\bllshed during Oc l ober, the flBI issue prin ted after Oc t ober 5. In item 16 indicate the date o l the lsaue In which this Statemen t o f Ownership wil be published 6 llem 17 mual be signed. Fa ilure ro fl/a o r publfsh a s ta tement of ownersh(o may lead lo suspen4ion of /UICOfld.c-lB" authorization PS Form 3526, Sept m;!ff 11195 ( Re'II,.. )

PAGE 3

EDITO R IAL AND BUSINESS ADDRESS: Chemical Engineering Education Department of Chemical Engineering University of Florida Gainesville, FL 32611 PHONE and FAX: 352-392-0861 e-mail: cee @ che.ufl edu Web Page : http:/ /www.c h e .ufl. e d11!ce e/ EDITOR T. J. Anderson ASSOCIATE EDITOR Phillip C. Wankat MANAGING EDITOR Carole Yocum P R OBLEM EDITORS Ja m es 0 Wilkes Univers i ty of Michigan LEARNING IN INDUSTRY EDITOR William J. Koros University of Texas, Austin PUBLICATIONS BOARD CHAIRMAN E. Dendy Sloan, Jr. Colorad o School of Mines PAST CHAIRMEN Gary Poehlein Georgia In s titut e of T ec hnolog y Klaus Timmerhaus University of Colorado MEMBE R S Dianne Do r land University of Minnesota Duluth Thomas F Edgar University of Texas at Austin Richard M. Felder North Carolina State University B r uce A. Finlayson University of Washington H. Scott Fogler University of Michigan David F. Ollis North Carolina State Univ e rsity Angelo J. Pema New J ersey In s titut e of Technology Ro n ald W. Rousseau Georgia In stitute of Te c hnolog y Sta n ley I. Sandler University of Delawar e Richard C. Seagrave I owa State University M. Sami Selim Colorado School of Mines James E. Stice University of T exas at Austin Donald R. Woods McMaster University Winter 1 999 Chemical Engineering Education Volume 33 Number 1 Winter 1999 EDUCATOR 2 Maurice A. Lar so n of Iowa State University DEPARTMENT 6 Washington State University AWARD LECTURE 12 Do Changes in the Chemical Indu stry Imply Changes in Curriculum? E.L. Cuss/ e r CLASS R OOM 18 Discontinuities in ChE Education, Stephen Whitaker 58 Permeation of Gases in Asymmetric Ceramic Membranes Carlos Fin o /, J oa qu{n Corono s CLASS AND HOME PROBLEMS 26 Non-Adiabatic Container Filling and Emptying Noel de Nevers RANDOM THOUGHTS 32 FAQS, Ri chard M. Felder, R ebecca Brent LABORATORY 34 Eva lu ation of Computer-Simulation Experiments in a Senior-Level Capstone ChE Course S co tt R. White Geor ge M Bodner 46 Demonstrating Simultaneous Heat and Mass Transfer with Microwave Drying Cheri C. Steidle, Ke v in J M ye r s 50 Medical Surveillance and the Undergraduate Thesis, Ian A. Fur ze r 54 Laboratory Experiment in Biochemical Engineering: Ethanol Fermentation, Alberto Colli Badino Jr ., Carlos Osamu Hokka 66 Intr od ucin g Process-Design Elements in the Unit Operations Lab, Christine L McCallum, L. Antonio Esteve z CURRICULUM 40 Process Analysis: An Electronic Ver s ion, G eo rge B. Delancey 62 A Joint Chemical/Electrical Engineering Course in Advanced Digita l Proce ss Control, Joseph J. Fe e le y, Louis L. Edwards ASSESSMENT 72 Ranking Graduate Programs : Alternative Measures of Quality, J ohn C. Angus Robert V. Edwards Brian D. Schult z LEARNING IN INDUSTRY 84 From the Classroom to the Workplace : Motivating Students to Learn in Indu stry, A. Christian Fri c k e 11 Positions Available 11, 65 Book Reviews CHEMICAL ENGINEERING EDUCATION (ISSN 0009-2479 ) is published quarterly by tire Chemical E11gineering Division, American Society for Eng ineering Educatio11 and is edited at the U ni ve rsity of Florida. Cor r espo11de11ce regarding editorial maller circulati.on a11d c ha11ge s of address sho uld be se nt to CEE, Chemical Engineering Department U niv ers ity of Florida, Gainesville Fl 3261/.6005. Copyright 1999 by the Chemical Engineering Divisio n American Society for EtJgineering Education. The state m e nt s and opinions ex pr essed in this periodical are those of the writers and ,wt nec essar il y those of the ChE Division ASEE which body assumes no respo11sibility for them Defectiv e copies replaced if notified within 120 days of publicatio11. Write for i11formatio11 on s ubs cri ption costs and for back copy costs and availability. POSTMASTER : Send addre ss c hang es t o CEE, Chemical E11gi11eeri11g Department., University of Florida, Gainesville FL 32611-6005. Periodicals Postage Paid at Gainesville, Florida.

PAGE 4

.t3_..blllliiii._e_d_u_c_a_to_r _________ ) MAURICE A. LARSON of Iowa S tate University "Mr. Crystallization" W idely re spected and recognized the world over for his re searc h contributions, Maurice Larson i s a distinguished and accomplished chemical engi neer who has been a influential figure in chemical engineer ing education and practice for nearly forty years. In the area of crystallization which he was instrumental in changing from an art to a science and into a thriving field of research, he is looked upon as a father figure His contribution s have been both pioneering and original in the context of their time s Mauri was born on a farm in the Loess Hill s near Mi sso uri Valley, Iowa, the second child in a family of two girls and one boy He attended a one-room country school that had twenty students in eight grades. (Interestingly, of the six students in his clas s, three ultimately received PhDs in engi neering! ) When he was nine bis family moved to another farm, this one near Ayrshire, Iowa 150 miles away. The school there consisted of all twelve grades and had about 220 students. Mauri grad uat ed in a class of twenty-one s tudent s in 1944. In 1946 he enlisted in the Army and was later discharged just in time to begin study at Iowa State College in the fall of 1947 He initially planned to st ud y chemistry but influenced by a newspaper article he happened to read he enrolled in chemical engineering in s tead Subsequent to his graduation in 1951, he accepted employment with the Dow Corning Corporation in Midland, Michigan Mauri had always had in mind going on to grad u ate schoo l Mauri as a farm boy in 1936 as an enlistee in the Armed Forces in 1946 and as a distinguished professor in 1998. and, after marrying Ruth Gugeler an ISU chemistry gra du ate and a co-employee at Dow Corning, he decided the time had come to do so He e nrolled in gra duate sc hool at Iowa State and sec ured employment as a teaching assista nt. He found he greatly e njoyed working with s tud e nt s a nd was soon was promoted to in s tructor. Hi s thesi s research was in the area of phosphate fertilizer chemistry and production After receiving hi s PhD in 1958 Mauri was appointed Assistant Profe sso r in the department, and his re searc h broad ened from fertilizer technology to process dynamics and control. He organized and taught one of the first undergradu ate process control courses in the U.S and was also one of the fir s t to organize a course around the seminal t ext Trans port Ph e nom e na by R B Bird W. E Stewart and E. N Lightfoot. He and hi s graduate s tudents published in the areas of fertilizer technology and proces s dynamics and nationally he was elected Chair of the Division of Fertilizer and Soil Chemi st ry of the American Chemical Soci ety. In 1959 a new graduate st udent Alan Randolph arrived at Iowa State with indu s trial experience in the crystallization of ammonium sulfate. A l an was concerned about the limit cycle in crystal size distribution that occurred in lar ge co ntinuou s Copyright ChE Division of ASEE 1999 2 C h e mi ca l Engin ee rin g Ed u ca ti on

PAGE 5

crystallizers. Mauri's interest in both proc ess d y namic s and fertilizer t ec hnol ogy resulted in a perfect interest m a t c h with Alan and a re s ult of their collabo ration was the l a ndmark paper Tran sie nt a nd Steady State Si ze Di s tribution s in Continuous Mix e d Suspension Crystallizers -an import a nt public a tion that se t th e age nd a in indu s trial crystallization r esearc h for the ne x t thirty-fiv e years "Ma uri and I met in 1959 and we imm e diatel y realized we h a d s imilar research inter es t s," says Alan R a ndolph now an Emeritus Profe ssor of Chemi cal Engineering at the University of Arizona. I was fortunate to be one of hi s first PhD s tudent s, a nd he and I ha ve been lifelong co lleague s and friends ever s inc e th at tim e Mauri is clearly responsible for much of the early fund a ment a l work in c r ysta lliz a tion. H e s timul ated the work of many others and th e field ha s ben efite d great l y from hi s work," R a ndolph co ntinue s "A goo d example of hi s s upporti ve approach i s an occasion w h e n h e was visiting our family in Trona California. M a uri and I went on a trip to see the Panamint Mountain s and our old blue Rambler was not able to mak e it up one of the s teep grades. I s houted to Mauri Get out and pu s h! Amazing as it m ay see m he actually pu s hed the car uphill illu s trating that Mauri i s not only imperturbable but he i s al so a true friend who know s the meaning of s upport! Mauri's s ub se quent research at Iowa St ate was co ncerned with m a n y aspects of so lution crys talli zatio n In a ddition to a nal yzi ng co ntinuou s crystallizer sta bilit y, he developed methods to co ll ec t data on and reali s tic a ll y de scr ib e nucl ea tion a nd growt h seco ndar y nucleati o n effec t s of additives and grow th di s per s ion Alan and Mauri were the first to pro v ide a firm mathemati ca l foundation to the analysis of crystallization by introducing and expanding the use of Alan 's population balance model s. Their early studies on population bal a nc es formed the ba s i s for severa l entirely new areas of crystalJization re sea rch and theory Several of Mauri 's twentys ix PhD and twentyseve n MS st udent s hav e been continuing contributors to crystallization funda ment a l s a nd practic e. Many if not all fields of chemical e n g in eeri n g ha ve pioneerin g name s indel ibly assoc iated with their growth B y any acco unt Maurice A. Lar so n i s one s uch nam e No book on crystallization n o review on the s ubject indeed no reaJiy worthwhile paper in the s pecific areas of hi s r ese arch wou ld be complete without reference to his contributions in particular to the book he coau thored with Alan, Th e Th eory of P a r ticulate Pro cesses, Analysis and T ec hniqu es in Continuous Cr ysta lli zation (A. D. Randolph and M. A. Lar so n 1 st ed. 1972; 2nd ed. 1988 ; Academic Pre ss). It ha s become one of th e sta ndard reference s for the field, says Dr. L. K Dorai swa my Anson Marston Di sti n g ui s hed Prof essor and Herbert L. Stile s Prof essor of Chemical Engineering. Mauri s international s tanding in the area of c ry s tallization i s firmly estab lished His activities ha ve included collaboration with experts from the United Kingdom, Australia, Indi a, China The Netherland s, Poland C zec ho s lovakia and many other countries For twent y-five years he ha s also been active with the European Federation of Chemical Engineering Working Group on Cry s tal lization. To s tr e ngthen hi s re searc h he s p e nt a yea r at Stanford on a National Science Foundation F ac ulty Fellow s hip a yea r at University College London as Shell Visiting Profe ssor, a nd a year at th e U ni ve r s ity of Manche s ter In s titute of Science a nd Technology (UM IST ) Hi s intere s t in undergraduate education and opportunities for undergraduate s tudy led him to initiate several chemical engineeri ng student international Wint er 1999 "His vision and example have shaped the culture of the department a balance of teaching, research, and service. ISU undergraduates, alumni, faculty members across campus, and crystallization researchers around the world hold Mauri in the highest regard." 3

PAGE 6

Gordon Youngquist pre sents Mauri with a framed copy of the program prepared for a special AIChE session honoring Mauri's 70th birthda y and his man y contributions to the field of chemical engineering. Mauri's office door has alwa ys been open to students and colleagues alike (1997) T exchange programs, notably with Bradford University, Ham burg University, and University College London. The most successful ha s been the s ummerst ud y program at Univer sity College London which he organized in 1972 during hi s sabbatical there. The program is still in full force and is now shared by several additional American universities including the University of Wisconsin Georgia Institute of Technol ogy Case Western Reserve University and Virginia Poly technic Institute and State University John Garside Pro-Vice-Chancellor and Profe sso r of Chemi cal Engineering UMIST remarks "I have known Mauri now for well over twenty years. He took a faculty-improvement leave at University College London in 1972 when I was on the faculty there and I s ub seq uently s pent a year working with him at ISU in 1976-77. At UMIST, we u se Mauri and Alan 's book, The Theory of Parti c ulate Pr ocesses, for both undergraduate and graduate co ur ses. It is an inspiring book and has had an important impact on teaching in this field throughout the world. Their work has been so influential that use of population balance modeling is now accepted as a matter of course when dealing with crystallization pro cesses. Before their work, the population balance was almost unheard of." Crystallization ha s b ee n a traditional area of chemical engineering education and research, having evolved from its empirical beginning s to the highly sop histicated approaches 4 Mauri is surrounded by well-wishers on the occasion of his retirement. Each of them has at one time or another served as Chair of the ISU ChE Department: (left to right) Terry S. King Charles E. Glatz, Maurice A. Larson, R. C. Seagrave, and George Burnet. practiced today Maurice Larson ha s been associated with all s tage s of recent development of this s ubject and has been a trend se tter and a pioneer in nurturing and s hapin g it. Indeed, it is inconceivable that any account of crystalli zation toda y would be complete without a reference to Mauri s dominating role In 1991, Mauri organized an informal group known as the Association for Crystallization Technolo gy. This organiza tion brings together 70-100 researchers and technologist s from industry and academe for three d ays every year for the so le purpose of discus sing the scie nce and technology of crystallization. Mauri' s extension program s in the form of workshops in crystallization have always s ucc eeded in bring ing together chemical engineers from many parts of th e country, indeed the world, and ha ve resulted in reports and valued s uggestions for future research Mauri ha s been vigorously involved in many national programs and activities of the American Institute of Chemi cal Engineers. He was recognized as an AIChE Fellow in 1982 and se rved as Vice Program Chair for the organization's 1992 Miami Beach meeting. He ha s also been selected to receive the 1998 AIChE Separation Division 's Clarence G Gerhold Award In the American Society for Engineering Education, h e Chemical Engineering Education

PAGE 7

received the Western Electric Fund Award in 1970 served as an officer of the Chemical Engineering Department Heads Group in 1973-74 and was Co-Chair of the 1987 Chemical Engineering Division Summer School. He has served on the editorial boards of the AIChE Journal and the Journal of Separation Pro cess T echnology, and received the U. K. Wilton Park Award in 1978 and the Iowa Governor 's Sci ence Medal in 1989. In addition to his research work, Mauri has always been concerned with the teaching of both graduate and under graduate courses. Over the years he has taken an active interest in practically all matters pertaining to education and student affairs at ISU The high national reputation that undergraduate education in chemical engineering at Iowa State enjoys today may be largely credited to his efforts Evaluated by his students as one of the best teachers they have encountered, they cite his ability to make his classes rewarding and enjoyable The are challenged by his dedica tion and thoroughness E. L. Cussler, ChE Professor at the University of Minnesota, remembers, "Mauri was a marvel ous mentor for young faculty. When I was starting, he found me and encouraged me when I was still surprised to be treated as an adult. I and many like me owe an enormous debt to Mauri for his generosity and time in encouraging and shaping our own careers." Mauri 's combined interests of teaching and research resulted in his promotion to Asso ciate Professor in 1961 and to Professor in 1964. In 1977 he was awarded the highest rank in engineering at ISU-that of Anson Marston Distinguished Professor. Throughout his career, Mauri has also been active in de partmental organization and administrative services. He has served on many university college, and departmental com mittees and has always contributed to curriculum develop ment both at the college and the department levels. He served as Department Chair from 1977 to 1983, and many who have visited the department testify to its increased stat ure as a result of his leadership. As an educator and adminis trator Mauri influenced the growth of engineering at ISU through the breadth of his interests and his ability to fashion a common theme from diverse views. Near the end of his term as Chair Mauri initiated a cam paign to raise funds for a new addition to Sweeney Hall the chemical engineering department building, to provide ex panded research space and an updated teaching laboratory More than $1.5 million was raised from friends and alumni and this, with an appropriation of over $5 million from the State Legislature, resulted in a new wing in 1994. As a result of Mauri s campaign, one of the donors also gave a million dollars to endow the Herbert A Stiles Professor ship in Chemical Engineering. A special session was programmed at the 1997 AIChE Winter 1999 Annual Meeting in Lo s Angeles to recognize Mauri 's distin guished career and to honor him on the occasion of his 70th birthday Dr. Terry King, former head of the ISU Chemical Engineering Department and currently Dean of the College of Engineering at Kansa s State University, declared "Mauri's determination to provide for the building needs of the future for chemical engineering, even when su pport for such a request was not initiall y evident, set a pattern for what is now being achieved b y the !SU College of Engineering as a whole He melded a partnership of alumni, industry, and the state to create enthusiasm and funding for a building project from whose labora tories ha ve come many important results being presented toda y in technical sessions worldwide. That partnership was novel for /SU at the time, and it has been eno rmously successful, resulting in our new construction. His vision and example have shaped the culture of the department-a balance of teaching, res earc h and service. As a result, !SU undergraduates, alumni, faculty members across campus, and crystallization researchers around the world hold Mauri in the highest regard." Larson has served as a tea c her, researcher, ISU supporter, colleague, and an inspiration to his friends for over four decades He and hi s wife Ruth are the parents of three children-Richard, who was lost as a teenager in an acci dent Janet who has a horse ranch in New Mexico, and John who is employed in San Francisco. Mauri and Ruth could not get farming out of their systems and thus have a substantial number of acres of farmland near Ames. Ruth retired in 1996 after working in the ISU Department of Animal Science and the couple has used their free time to travel extensively. Mauri Larson is the quintessential scholar, an outstanding prof essional, and one of the nicest people I have ever known, states friend and collaborator Ronald W Rousseau cur rently Profes sor and Chair of the School of Chemical Engi neering, Georgia Institute of Technology. Dr. James R. Fair, McKetta Chair Emeritus Professor from the University of Texa s-A ustin makes things crystal clear when he states, The achievements of Mauri Larson have been notable One of the foremost has been his rise to a pr e-em inent stature as an expert in crystallization. H e i s considered to be a foremost authority in this field and his many publications bear this out. The record is clear-and well documented-on this issue. H e r ea ll y is Mr. Crystallization." 5

PAGE 8

~S=i department ) lli-1111-1111111---.:.. _____ 6 Washington State University I Celebrating Fifty Years of Chemical Engineering at the Stroke of the New Millennium I Marc VanderSchale and Brian Erickson remove CO 2 from air in packed-column gas absorber S o what can 50 years of chemical engineering education mean in the context of 2 000 years of calendar time, especially when generated from a relatively small program in the rural town of Pullman Washington, located in the middle of wheat fields? Surprisingly, Washington State University s chemical engineering department is a unique examp l e of one of the oldest chemical engi neering programs in the country that is still a thriving concern for the university and related indu s try in today's world Having evolved from being part of the chemistry department in 1917 into a sepa rate department in 1950, ChemE now focuses on education a nd research in advanced gas and chemical processing, hazardou s ma terial cleanup, and bioengineering. The department has graduated more than 1 100 s tudent s, the major ity of whom have taken significant roles in industries s uch as oil and c h emicals, pulp and p a per pharmaceutical s, food, microchip manu facturing, environmental protection, and bioengineerin g Alumni can be found in s u ch diver se industries as petroleum refinerie s, pulp and paper mills nu clear and sy nthetic fuel proce ssing facilities, and food processing plant s They are hired by such companies as Dow Chemi ca l Westinghouse ARCO Boeing Weyerhaeuser, Kai ser Aluminum Intel, B atte ll e Pacific Northwest National Lab s, as well as other Hanford contractors and sma ll er companies in the region "Grad u ates from WSU's chemical engineering program are hi g hly marketable ," reports Richard Zollars, department chair. "S tarting sa la ries in the mid-$40 000s are common, as qualified candidates in these fields are highly so u ght after by industry and agencies." The small size of the program and connections with indu stry work in its favor. With about 100 undergraduate 30 graduate st udent s, and 10 permanent facu lt y, class sizes of 25 35 afford semi nar possibilities Copyright ChE Division of ASEE 1999 Chemical Engineering Education

PAGE 9

Washington State University s chemical engineering department is a unique example of one of the oldest chemical engineering programs in the country that is still a thriving concern for the university and related industry in today's world. and close interaction between faculty and students Be ca use of its select nature the program attracts students with above-average scholastic standing, and as many as one-third of all those certified in the chemical engineering program are in WSU Honors Program. Given the above collegiate atmosphere, the faculty, research indu stry partnerships and educational activities take on a distinctive quality and focus. THE UNDERGRADUATE PROGRAM The undergraduate curriculum allows for broad education in the sciences and sound basics in the discipline, but with a flexibility that allows s tud ents to individualize studies. In the upper-division after st udent s certify, the following courses provide the basics in c h emica l engineering: Sophomore Y ear ChernE Process Principles ; Proces s Simulation Junior Year Transport Phenomena ; F lui d Mechanics/Heat Transfer ; Thermodynamics; Sepa rations; Kinetics Senior Year Chemical E n g in eering Lab I and II; Control; Design I and 11 ; Seminar Another seven or eight e l ectives allow juniors and seniors to customize their focus in bioengineering, environmental, or other a llied fields. They also ha ve opportunities for multidisciplinar y s tudy and research. For example, Beck y Russell a recent graduate who knew s he wanted a career that impacted people explains I decided to make chemical engineering my preparation for medical school. The ski ll s I learned in technical problem solving and biomedical applicatio n s have served me well. Not all is serious study, however. Student s are encouraged to join the College Ambassadors group and to participate in the AIChE student chapter or other engineer ing societies for networking and le a dership growth WSU s s tudent AIChE chapter recently won first prize at its regional conference for proposing a national competition around student -built chemically controlled cars. The gro up will put this to the test at the national conference in Miami in mid-November (after the d ea dlin e for this article). President Wendy Anna and other officers were able to raise funds from industry to defray the cost of several students travel to Florida for this purpose As the nationwide need to recruit and retain engineering st udent s escalates, WSU's chemical engineering department continues to step up its efforts for sc holar s hips and other incentives. There 's a good chance that students will qualify for a sc holar s hip if they achieve a 3.4 GPA or better and show promise while enrolled, reports Zollars Because of the generous support of a lu mni and corpora tion s, about one-third of the Winter 1 999 ABET Accredited E-mail student services thomtd@che .ws u edu Website at http :/ /www .c he .ws u.edu FACULTY Ivory, Cornelius PhD 1980 : Princeton Biopro cess ing, separations, modeling Lee,James PhD 1978: Kentucky Bioproce ssing, mixing Liddell,KNona PhD 1979: Iowa State Hazardous waste s electro d epositio n Mahalingham, R. PhD 1968: Newcastle-Upon Tyne, England Hazardous was tes shock reactions Miller, Reid PhD 1968 California, Berkeley Thermodynamics Petersen, James PhD 1979: Iowa State Bio remediat ion Peyton, Brent PhD 1992 : Montana State Bioremediation, extremo phili c bioprocessing Thomson William PhD 1969, Idaho Materials, kinetics ca tal ysis Van Wie, Bernie PhD 1982 Oklahoma Bioprocessing, biomedical eng ine ering Zollars, Richard PhD 1974 Colorado Colloidallinte,facial phenom ena, reactor engineeri ng 7

PAGE 10

Ethics are taught in one senior class by using the Dilbert Ethics Challenge game. Teams ponder how to handle case studies. undergraduate students receive some form of financial aid. With the exception of athletics chemical engineering's alumni tradition ally give the highest per-annum donations of any other unit at WSU. GRADUATE/RESEARCH PROGRAM While virtually every undergraduate student who wishes to may get involved in faculty research, the research program is the heart of the graduate program. Currently more than $2 million in grant activity is in progr ess, with healthy prospects for more. Emphases in bioengineering, environmental restoration and hydrocar bon processing result in such projects as biotreatment of hazardous con tamination, diagnostic medical devices and converting natural gas to u seful products. The current roster of graduate-student research covers such topic s as electrodeposition, isoelectric focusing, development of biosensors pro tein production in plant cell systems, bioremediation of chlorinated sol vents and heavy metals, oxidative coupling of methane etc. The Center for Multiphase Environmental Research (CM ER), under the direction of Profes sor James Peter se n, conducts interdi sci plinary research addressing important environmental problems for indu s tries and govern ment agencies. It seeks opportunities to transfer this technology to industry and, in the process, to educate the next generation of environ mental professionals. CMER faculty are drawn from the civil, environmental, chemical, me chanical, and biological syste ms engineering departments. The Center and chemical engineering faculty currently play a part in three such projects with large government grants: remediation and recycling of creosote treated piers at Navy ports and two D O.E project s to clean up to xic metal s in soils and aquifers. As a result of her quality work on hazardous waste treatment through CMER doctoral student Juli Sherwood last year won the university 's Harriet Riga s Award from the Association of Faculty Women as WSU' s outstanding doctoral woman student, citing her for her research, teach ing and leadership. Another recent graduate student s uccess story is illustrated by Sherman Xu, who within five years gained both his master' s and doctoral degree valuable research expertise, and a job with the Amoco Research Center in Illinoi s. Hi s research explored a more productive procedure to convert natural gas to ethylene-the feed stock for many of today 's plastics. 8 ABOUT THE UNIVERSITY WSU is a land-grant research university dedicated to excellence in undergraduate education. Founded in Pullman in 1890, it has 21,000 students at four campuses, sev eral Learning Centers and other sites throughout the state. WSU's nine colleges and approximate 150 undergraduate ma jors, along with its distance Extended De gree Programs, can almost ''Take You Any where You Want to Go!" Pullman's resi dential campus designates special honors and math-science-engineering residence halls. WSU is one of the most "wired" campuses in the west, with good access to computers and learning labs Semester classes are scheduled from late August through mid-December, January through mid-May, and at varying summer sessions. ABOUT THE COLLEGE The Chemical Engineering Department is part of WSU's College of Engineering and Architecture, which is committed to quality curricula and innovation in learn ing. Through industry-partnered programs and practical approaches, students prepare for technology-based professions and life long learning. Distinguished faculty and alumni network with industries that offer students scholar ships, mentoring, internships, and project opportunities. Approximately 120 perma nent faculty serve an average of 2,000 un dergraduates, who may join more than 20 student clubs or professional societies for fun and extra challenge. Facilities include numerous labs, classrooms, and other learn ing spaces in seven buildings on the Pull man campus. A modem 100,000 square foot Engineering Teaching and Research Laboratory recently opened and holds state of-the-art testing and analysis facilities. The College assists with tutoring, skills workshops, scholarship applications, ori entations, and other special student needs It also helps recruit and retain individuals from underrepresented groups through the Minority and Women's Engineering Pro grams. Chemical Eng in ee rin g Education

PAGE 11

O.H. Reaugh Lab dedication: Student Carrie O'Rourke, Chair Dick Zollars, lab director Bill Thomson, WSU President Sam Smith, and alumnus O.H. Reaugh. "With my WSU credentials, I found myself very competitive in the job market ," says Xu. 'T m grateful to my superior graduate professor who gave me the big picture, and allowed me to find my own answers His close rapport with industry pro vided me opportunities to see practi cal uses of our research." FA CI LIT I E S Reaugh, Thomson, and student Matt Fountain tweak the new equipment. T The department's premier research lab the O H Reaugh Oil and Gas Processing Laboratory-was dedicated by its namesake thi s fall in a ceremony (attended by his family, university leaders, students, industry partners and friends) that epitomized the momentum that the department's history brings to its future. Orland Harry Reaugh, a 1933 graduate of the program became a petroleum engineer and a leader in independent oil production in Oklahoma, Kansas New Mexico and Illinois. He has since become a major benefactor of WSU' s chemical engineering department which now has the only such alumni endowed and -named laboratory within the university Reaugh provided $250,000, ARCO another $50,000, and matching donation s are being sought to complete the lab 's $500,000 endowment. The income will ensure a lab that is equipped with state-of-the-art instrumentation. Lab director William Thomson says the facility "allows us to conduct research that is fast disappearing from the univer sity sce ne-innovative studies to find les s expensive and ecologically friendly ways to add octane to gas, bring natural gas to remote locations, and to create compact, economical hydrogen fuel cells. More than 9,500 square feet of lab space are located in the Winter /9 99 A Jim Petersen and gr adu ate student Lin Sha look at grimy creosote being eaten by bacteria and turne d into CO 2 in the lab. new Engineering Teaching and Re search facility containing analytical equipment such as spectrometers, chromatographs fitted with auto samplers and various detectors, anaerobic incubators dynamic x-ray diffractometer, laser light scattering, and electrophoresis de vices. The George T Austin endowment for undergraduate lab equipment, augmented by funds from alumnus Gene Voiland help keep learning technologies current. Computing equipment and workstations with parallel pro cessing features are provided for all graduate students and researchers. Keeping such quality computing current is al ways a concern of technical educators, however. Typically, alumni donations help replace one-third of the undergradu ate machines every year in the 15-station computing lab. Full-scale commercial versions of Pro Vision, MatLab, Mathematica MathCAD spreadsheets, word processors, and other programs are used on these machines I NDUSTRY CONNECTIONS A hallmark of WSU 's engineering college is its continuing strong connection to industry. Companies and agencies pro vide internships, scholarships, collaborative research, and even job exchanges. WSU in turn provides on-site or webbed distance learning, qualified potential hires, and R&D. One collaborative project is Prof. Bernie Van Wie's work with the Spokane Interdisciplinary Research & Technology Institute and DevTec, an independent industrial partner to develop and commercialize an automated blood chemistry analyzer. Another collaborative project is Prof Bill Thomson's work with the Washington Technology Center 9

PAGE 12

and Washington Water Power to refine an efficient fuel cell. West Coast companies such as ARCO, Boise Cascade Hanford contractors, SEH America Kimberly Clarke, and Reynolds Aluminum offer summer jobs and internships (sum mer plus one semester) to students-and often end up offer ing them jobs at graduation They say they are impressed with WSU 's students' practical knowledge and leadership experiences outside the classroom. adjunct" faculty from local industry Three to five students usually earn the MSChE degree through this program each year This year's graduates will be Penny Colton (Dissolu tion Kinetics at the Calcite-Water Interface), Dan Schmitt (10 W Proton Exchange Membrane Fuel Cell Design) Chris Johnson (Microbial Growth Kinetics Using Colloidal Poly mer as Substrate), Scott Estey (Thin Film Poly dimethylsiloxane Oil Evaporation) and Brad Knutson (Evaluation of Ion Exchange Perfor mance Predictive Tools ) Chemical engineers need foundations in science and math, but also in other disci plines that will prepare them to adapt new products into the culture," says Glenn Butler CEO of the ARCO Refinery at Cherry Point. He and other de partment advisers urge devel opment of students' skills in communication, presentation and human relations, iri addi tion to developing a code of ethics and a knowledge of overall business and eco nomic concepts. John Wolfe, a 1997 chemi cal engineering graduate now at ARCO in Anaheim, recently returned to the College 's Ca"Team think" in action Graduate students, profes sors, researchers, medical professionals, and commercialization experts work on the automated blood analyzer project. To develop interest in engi neering at earlier ages for the past five years chemical engi neering faculty has conducted a NSF-funded summer pro gram for secondary school teachers. The goal is to famil iarize them with engineering and help them develop mod ule s to teach when they return to their classrooms. All to gether, more than 65 teacher s attended-half from North west schools, and others from as far away as Korea, Florida, and Connecticut. Feedback from the participants noted reer Fair-this time as a recruiter! Ba s ically my ChemE degree helped to more than double my salary as a science lab technician, says Wolfe "My internship with ARCO during the summers gave me a foot in the door. And now I have a good job with them solving technically complex problems in oil refining." He 's now on the lookout for other potential hire s who work well in teams of engineers, scientists, cus tomers, plant operators, managers lawyers, government regu lators, and construction workers. OUTREACH WSU Pullman's program remains fairly stable in size. Many families continue sending new generations as lega cie s. The department also reache s out to the adult, some times mid-career learner through a companion program at the WSU branch campus in Tri-Cities with access to Hanford and the Environmental Molecular Science Lab. It generally serves full-time employed engineers, offering after-hour courses so that engineers can upgrade their education or work toward the MSChE. It is one of few evening-based MSChE programs in the country. Many of the courses are taught by PhD chemical engineers who are also employed at the Pacific Northwest National Laboratory. Tri-Cities student project s or theses usually are done with a committee composed of both WSU regular faculty and JO that the experience suc ceeded in bringing engineering into their science curricu lum One science teacher developed an engineering-re lated module that contributed to a portfolio that won her state and national teaching awards FUTURE TRENDS Academic destinies will depend a great deal on the eco nomic and technological trends transforming higher educa tion today. For instance, the Tri-Cities program will be influ enced by down-sizing at Hanford And as noted earlier, keeping computer equipment and software current is a major and constant challenge. By the same token, developing di tance-learning technologie s may allow webbing courses beyond campus sites, which will again revolutionize what we do at universities While state appropriations can no longer meet the entire departmental needs outside grants and private fund ing is on the rise The department plans to pursue more endowments similar to the O.H Reaugh Lab to stabilize funding bases Faculty and student numbers are not expected to change dramatically over the next five years. Expan s ion of multi disciplinary experiences is anticipated, however particu larly in the wake of the Center for Multiphase Environmen tal Research's path-finding activities. 0 Chemical En g in ee rin g Edu ca tion.

PAGE 13

.,~. 111 ij 11111111 :j...__b_o_o_k_,-,_e_v_ie_w ___ __,) Nume r i cal Computation in Scie nc e and E ng i neerin g by C. Po z rikidis PubUshed by Oxford University Press Inc. I 98 Madison Avenue New York, NY 10016 ; 627 page s including index, $75.00 (1998) Reviewed by James N. P etersen Washington State U n iversity The necessity of obtaining numerical solutions to physical problems crosses virtually all the discipline boundaries in engineering and science Only a limited number of books are available, however, that provide the fundamentals of scien tific numerical computational techniques, together with ap plication of those techniques In general, Pozrikidis has been able to achieve this result in his book which is intended to be u se d in upper-level undergraduate and beginning gradu ate courses and may be suitable for individual study. In so doing, he has produced a book that s trike s a balance between rigor and practicality While not providing lengthy computer codes in the text, he has communicated the essential aspects of various numerical methods by relying on code fragments and pseudoCode that can be translated into any suitable computer language. In so doing, he has covered most of the topics required in an introductory course within a manageable number of pages. To compliment the text he has also provided access, via the Internet to a public domain software library of Fortran 77 programs organized by book chapter. The book is organized into eleven chapters. Each of them is further divided into sections and subsections. Because the subsections are the intended functional learning entities, the author provides both theoretical and computational prob lems at the end of each. These problems are designed to complement the theory presented and to provide the student with an immediate opportunity to practice its implementation. The chapters are organized in a fashion s imilar to many other numerical methods books In Chapter 1, the author provide s a background of computer hardware computer arith metic including both integer and floating-point representa tions, and errors In the next three chapters, he fust lays a foundation on which to build the solution of linear and nonlinear sets of simultaneous equation s. Thus in Chapter 2 he covers matrix algebra and matrix calculus, and Chapter 3 is devoted to the solution of sets of linear algebraic equa tions. The solutions of sets of simultaneous nonlinear alge braic equations i s di sc u sse d in Chapter 4, and he goes on to discu ss eigenvalues of matrices in Chapter 5. ________________ Continued on pag e 65. Winter /999 The next generation analysis software for: gasd yna m i cs ch e mi ca l equ il i br i u m chem ic al k i netics SuperSTATE APPLICATIONS Ed u ca ti o n Combus t io n Pollut i on Explos i on s ~00 1-0-c c=-o( :}---j '1 @@~ El FEATURES G r a p h ic a l I nt e rf ace I nt eg r ated P l ott i ng Batch Calculations O nl i ne Help Free Demo Version CDL C o mb11 s ri o 11 Dynami c s l..Jd. I info@co mbd y n.com www.co mbd y n.c o m fax : 40 3 52 9-2 5 1 6 POSITIONS AVAILABLE Use CEE's reaso n ab l e ra t es to adver ti se. Minimum rate 1/8 page, $100; Each addi tion a l co lumn inch o r portion thereof, $40. Fac u l t y Po s ition in C h e m ica l E n g in ee ring THE UNIVE R SITY O F TEXAS AT AUSTIN The D epart ment of Chemical Engineering invites a pplication s for a tenure track fac ult y position at the As i stant Profe sso r l eve l. A PhD is required and app li cants must have at least one degree in chemical e n g in eering an outstanding record of r esearc h accomplishments and a stro n g int erest in under grad u ate a nd grad uat e teaching. Preferen ce will b e give n to applicants with sk ill s that will add to th e D e partment 's stre n gths in bioengineering a nd app lied and co mputati onal mathemat i cs includin g fluid dynami cs syste m s enginee ring a nd statistical m ec hani cs The s uccessful candidates are ex pect e d to teach under gra duate a nd graduate courses, de ve lop a re sea rch program collabo rate with other faculty and be in vo l ve d in serv ice to the university and th e prof ess i on. Applications from wo men and minoritie s are encour aged. Intere s ted per so ns s hould s ubmit a detailed curriculum vitae, includin g academic and profe ss ion a l experience a li s t of peer-r viewed publications and ot h er technical paper s, and th e names ad dresses, and telephone numb er s of thr ee or more references to : Chair man Department of Chemical Engineering The U ni vers ity of Texas at A u stin A u stin Texas 78712-1062. The University of Texas is an Equal Opportunity/Affirmative Action Employer. II

PAGE 14

Award Lecture ... 12 DO CHANGES IN THE CHEMICAL INDUSTRY IMPLY CHANGES IN CURRICULUM? Edward L. Cussler, In s titut e Profe sso r at the University of Minn eso ta r ece i v ed hi s BE d g re e from Y a le University in 1961, hi s MS from the Univer s it y of Wisconsin in 196 3, a nd hi s PhD from the University of Wi sco n s in in 1965 Hero se fr o m Assistant Profe sso r to Pro fessor of Chemical Engineering at Carnegie-Mellon University durin g the years from 1967 t o 1980 a t w hi c h time h e joined the faculty at the University of Minn eso ta as Profe sso r of Chemi ca l Engineering. ln 1996 h e b eca m e In s titute of Technology Profe sso r at the University of Minnesota an d is currently a t Cambridge U ni ve r s it y in th e U nit ed Kingdom as Profe sso r of Chemical Engineering. Ed ha s won numerou s awards durin g hi s prof essio nal career, some of which are the ATChE A l a n P Colburn Award in 1975 seven Minnesot a In s titut e of Technology Teachin g Awards through the years, the George Ta y l or Distinguished Teaching Award from th e University of Minnesota in 1987 th e D o nald K a tz Lectur e Award from the University of Michigan in I 996 and th e D a n ckwe rt s Lecture from the In s tituti o n of Chemical Engineers in London in 1997 Ed se r ves as A ssocia t e Editor of th e A/Ch J ourna l and i s o n th e Ed it orial Board of th e J ou rnal of Membrane Science H e also ha s se rved as a Dir ec t or, Vice Pr es ident a nd Pr es i d e nt of AlChE, and was Chair of the American Association of Engineering Societi es. H e ha s also been author or co-author of over 160 public tion s. H e i s co -author with B e lt er a nd Hu of Bios epara tion s ( John Wiley and Son s, N ew York 1988 ) and with Baker Eykamp, Koros Riley and Strathmann, of M e mbr ane Sepa r at i on Systems, ( Noyes Data Corporation, New J ersey 1991 ). He i s a uthor of the book s Diffu s i on (Ca mbrid ge University Pr ess, L o nd o n 1984 ; seco nd edit i o n 1997) a nd Multicom ponent Diffu s i o n (Elsevier Publi sh in g Co mp a n y, Amsterdam, 1 976). E.L. CussLER University of Minnesota Minneapolis, MN 55455 T his paper i s a s ynopsi s of my Union Carbide Lecture s hip an award given at the 1998 meeting of the American Society of Engineering Education. I am flattered to have my research and teaching on diffusion acknowledged I know that this lectur e can often be a review of the past research, centering on a scattering of old slides, like a photograph album of half-remembered vacations But the lecture and this paper are too good a forum to waste on my past. Instead of the past, I want to consider the future. In doing s o I remember a conversation I had thirty years ago with the hi s torian L. Pearce William s I was visiting him to gush about hi s biography of Michael Faraday 111 which I had enjoyed enormously. I s u spect that he found my naive enth u siasm both flattering and embarrassing To make conversation Williams asked if I knew the real difference between science and the arts. I did not. He re spon ded that in the sciences, we wrote paper s and book s when we felt we knew everything about our topic. In the art s, he asserted, authors wrote when they knew little initially and used the writing as a way to focus new question s and to explore possible answers. Whether this art s -science contrast is true or not I want to u se this paper as a way to learn about possible changes in chemical engineering curricula. I am not yet s ure if these ideas are co rre c t but I want to see if they make sense. In the next few years, I'll try them out. For now though, they're be s t described under three heading s: the changes in the chemical industry the s tatus in academia, and possible c urricular c h anges. CHANGES IN THE CHEMICAL INDUSTRY La s t spring, I taught our introductory chemical engineering co ur se-the one that covers stoichiometry. Early in the course, I Copyr i gh t ChE Division of ASEE 1999 Chemical Engin ee rin g Education

PAGE 15

showed pictures of chemical plants to the s tudents. I told them that the tall towers were for distillation and the short, fat ones were often for gas absorption. I pointed out the reactors, with their preheaters and recycles. I spoke of the excitement of running a chemical plant and the satisfaction of using chemical technology to improve our well being I did so with hidden reservations that I did not have ten years ago I know that the chemical industry has changed and that many of the students will not work in the commodity chemical plants I was describing To see why, we need to review the history of our industry, using as an example the development of synthetic textile fibers From 1950 to 1970, the chemical industry produced ever-increasing amounts of syn thetic textile fibers as shown in Table 1. Over the decades, while the production of natural fibers was about constant the production of synthetics grew 20 % per year. This growth was comparable to that of the software industry today; indeed, Du Pont in the 1950s was like Microsoft in the 1990s. It was a golden age for chemicals But from 1970 to 1990, synthetic textile fibers grew only four percent a year-at about the same rate as the growth of world population. That's not surprising; after all, any logarithmic growth can t continue indefinitely. From 1970 to 1990 the industry stayed profitable by using larger and larger facilities. Bigger profits came from consolidating production into bigger plants, designed for greater efficiency in making one particular product. The interest in computer-optimized design is a vestige of this consolidation. Such optimization meant small producers were forced out. For example the number of com panies making vinyl chloride shrank from twelve in 1964 to only six in 1972. (2] In the last ten years, the industry has used other strategies to stay profitable. These strat egies often centered on restructuring, which was three times more likely to affect engiTA B LE 1 Grow th of Sy nth etic F iber s F ro m 19 50 to 19 70 ( Sour c e : Spit z U.S. D e partment of Commer c e) Cotton Wool 4 3 5 3 4285 4794 S y nthetic s 92 3480 8612 neers than the general population. Whether called "restructuring," "downsizing," "right sizing," or "rationalization," the strategy meant many mid-career engineers were sud denly looking for a job. The Engineering Workforce Commission now feels that engi neers will average seven different jobs per career, a dramatic change from two per career when I graduated in 1961 Y 1 Middle management, that traditional goal of our B-students, is no longer a safe haven. Starting salaries remain high, the envy of other technical professions, but they have not increased faster than inflation in thirty years. In this environment, I applaud the decision of the American Institute of Chemical Engineers (AIChE) to be a l ifetime home" for members of our profession, providing more help in job transitions a n d financial planning. The AIChE can no longer be only a nineteenth century-style learned society. Most recently, the chemical industry has become enchanted with the life sciences often called "biotechnology." Biotechnology in the industrial sense is most successfully repre sented by applied agronomy i.e. by genetically modified seeds. It is u s ually different from the biotechnology represented in academic chemical engineering that often centers on separations and reactions involving specialty pharmaceuticals. The model for corpo rate enchantment is Monsanto, which has spun off its commodity chemical operations to compete largely in this new biotechnology Other companies are imitators. In recent five year projections, Du Pont has relabeled its chemicals as "materials ," is spinning off Conoco, and plans to double its life-science efforts to one-third of the company s sales. Hoechst, by some measures the world's largest chemical company, plans to leave chemi cals for the life sciences. It s a different world beyond our ivory towers. Wint e r 1999 I want to use this paper as a way to learn about possible changes in chemical engineering curricula. I am not yet sure if these ideas are correct, but I want to see if they make sense. In the next few years I' 11 try them out. 13

PAGE 16

THE STATUS IN ACADEMIA While the se indu s trial changes occur, academic chemical engineering continues along well-established paths. I think that thi s i s good. Univer s itie s are both s t a ble and resilient ; Clark Kerr the long-time provo st of the University of Cali fornia i s sa id to have asserted that univer s itie s make up more than 90 % of the socia l institutions that have la s ted over 500 years. Moreover courses in any field evolve s lowl y. Woodrow Wil so n at the time b e liev e d in th e o ld e r Pri ncip l es te s tament a nd those who converted to the n ewe r Transport gos pel. In one recent st imu lating article, Astarita a nd Ottino [ 6 l argued that these two book s have supplied the only two organizing ideas t h at our profe ss ion ha s had In hind s ight I believe that ther e are two main reasons why Tran spo rt Ph enomena was so s ucce ssfu l. First b y stressi n g parallels between diff e r e nt t ra n s p o rt processes the book s upplie s a ped agogica l tem Pre si dent of Princeton sa id that "c hanging curricula is like moving graveyards." Chemical Engineering plate that h e lp s a ll to l ea rn and think a bout these pro cesses Thi s t e mpl a t e is a mixed bles s in g. For example, the fact that there is no par allel to chemical re ac tion s in heat tran sfer m ea n s that chemical reaction s are s uper ficially treated Thi s m ay contribute to our co ntinuin g tend e ncy to teach ma ss trans fer without chemical reac tion s, even thou g h much in du s trial m ass transfer e.g., acid gas treatin g, t a k es place with reaction. Plant Plant Physics Chemistry Chemical engineering cur ricula in the USA are no ex ception. To a large extent, they reflect the sc heme first suggested in 1917 by a com mi ss ion chaired by Arthur D. Little founder of the finn that bear s hi s name. Building on Briti s h precedent s, th e com mi ssio n suggested an orga nization around unit opera tions. This wa s based on th e assertion that distillation was ba se d on the same principle s for a ny chemical sys tem be it rum or crude oil. Thi s organization was codified b y the book Pr incip l es of Chemical Engineering. [ 4 l Figure 1. Skills in Chemical Engineering Th ese skills are ideas from c h emistry physics and eng in eering Different jobs use diff erent proportions of these ideas The seco nd reason that Transport Ph enomena was so s ucc essfu l is a reflection of the boom taking pla ce in th e chemical indu stry w h e n L. E. Scriven tell s the po ss ibl y apocryphal s tory that the book was written only becau se the authors isolated them se lve s at a camp in the Adirond acks, where they could not be interrupted Prin ci pl es of Chemical Engineering outlines much of what would b e a reasonabl e, accreditable major today It be g in s with a chapter on stoichiometry and then covers fluid flow and heat transfer in three chapters. Four c hapter s on combus tion see m the intellectual ancestors of today 's reaction engi neering. Four chapters on se paration s center on di s tillation humidification and dryin g. Onl y th e two chapters on me chanical se paration s (c ru s hing and gri ndin g) have material mi ss ing from modern chemical e n g ineerin g curricula. I don t mean to overemphasize th ese parallel s, becau se th e contents of these chapters are often qualitativ e and dated Still I find the parallel s vivid. The curriculum implied by Prin cip l es of Chemical Engi neering was challenged mo s t s ucc ess fully b y Transport Ph e nomena th e book b y Bird Stewart, and Li g htfoot Y 1 Thi s book circ ulated in 1957 and formally published in 1960 inject e d more needed sc i e nc e and mathematic s into our field For a while, our profe ssio n was divid e d into tho se who 1 4 the book was publi s h ed. As outlined above, thi s boom centered on petrochemic a l s, which of course included the monomer s used to make sy nth e ti c fi ber s When you make petr oc hemic a l s, you often deal with a plethora of compounds characterized b y a n ea r co ntinuum of boiling point s. In s u c h a case, continuum math e mati cs i s a ppropriate; one ca n ba s i ca ll y ignor e the di scre t e jumps of the periodic table. Indeed one can ignore mo s t of c h e mi s tr y, with A+B~C i.e., argon plu s boron goes to carbon. Moreover as the petrochemical industry b ecame mor e competitive, minor im pro ve m e nt s in exis tin g processes were import a nt t o profit ability. These minor impro ve ment s co uld often be fo und u s ing the mathematical approach in Transport Ph enomena. While Astarita and Ottino argue powerfull y that th ese two book s provide the only two paradigm s in our prof essio n, I feel th a t Leven s piel 's Chemical R eact ion Engineering, l 7 J firs t publi s hed in 196 3, is also import a nt but for a different reason. The first t wo b ooks provided a d efinitio n of a profes sio n which implied a c urri c ulum. Levenspiel on the ot h er hand reorganized what was alrea d y ac knowl e d ged int o a Che mi ca l Engineering Education

PAGE 17

The changes in the chemical industry are clear-a movement away from commodities, a romance with biotechnology, and a long-term interest in specialties ... These changes in the industry do mean that our students will work much more on chemical products than on chemical processes. As a result, we will want them to think more about product design in addition to process design. way that made it easier to learn. Thjs can be hard for the founder of a discipline to do For example, I view T K. Sherwood as a founder of mass transfer. I find his 1937 book Absorp tion and Extraction, l Sl more understandable than its 1952 successor Absorption and Extraction l 9 l co-written with Pigford Trus second edition is in turn easier for me to understand than the 1975 revision, Mass Transfer ci oi co-written with Pigford and Wilkie. Levenspiel built on earlier reaction engineering books such as Hougen and Watson's Chemical KineticsY 1 1 but he achieved a new presentation that was much easier to understand Produc 1975 Consult Commodities 1995 Commodities Products These various subjects in the cherrucal engi neering curriculum can be represented on the triangular diagram redrawn from Gerhard Figure 2. Employment in 1975 versus 1995. Current graduates are much Jess likely to work for commodity chemical producers and more likel y to be involved with products Froelich, the 1999 AIChE president, and shown in Figure 1. The three corners of this plot represent trairung in the physical sciences in the cherrucal sciences and in the cherrucal engineering subjects. Different jobs use these three elements in different proportions, as shown in the figure. There is no s urpri se in this; plant engineering will demand a greater knowledge of mechanics and a smaller background in cherrustry than research and development. Figure 1 also suggests national averages British cherrucal engineers seem to have somewhat more chemical engineering and less chemistry than their US counterparts. Please don't take this diagram too literally; use it instead as a catalyst for thought, perhaps for deciding how your department's curriculum should evolve. DO INDUSTRY CHANGES IMPLY ACADEMIC RESPONSES? So far, I have s ummari zed the revolution in the chemical industry and the evolution of acaderruc cherrucal engineer ing I now want to compare the two to see what, if any, changes are needed in what we teach. Basically, I don't think many changes are indicated. The skills we currently teach seem to prepare our students well. Starting salaries remain high, the envy of most other engi neering disciplines The number of jobs is again rugh after almost a decade of bad years caused by restructuring. 1 1 21 In fact the job market right now is better than I thought it would be three years ago Industrial complaints about our teaching seem scattered, with about the same number urging more, say, kinetics as those who urge less kinetics Most Wint e r 1999 industrial complainers who urge us to teach more of a par ticular topic are hard pressed to suggest which current topics they would orrut to make room for their favorite. Thus, I believe our current curriculum is basically in good shape. One frequent orrussion does concern me, how ever. I want to explore trus omission next. My concern centers on the jobs our graduates now hold compared with those they held perhaps twenty years ago. My data for this are fragmentary, so I would be interested in any other data that are available. My data are probably biased toward large corporations, about whom our place ment office has better records. My data also have a regional bias towards 3M and food companies such as General Mills that are based here in Minnesota. Still the data suggest major changes in the last twenty years The focus of my analysis is the employment in 1975 versus that in 1995 I chose 1995 because the students ofte n need several years to settle down, to decide which sort of job they really want to do. As shown in Figure 2, there are enormous differences between 1975 and 1995. In 1975, three quarters of our graduates were working in the commodity cherrucals business. The small number who were not were split between work on products, either product design or product development, and work in other areas, which for convenience I have labeled consulting." That would in clude those working directly for consulting firms as well as those carrying out specific tasks such as enviro nmen tal impact statements. In 1995, the distribution of jobs is different. The majority 15

PAGE 18

of students (in Minnesota 's case, about two-third s) now work primarily on products. Thi s include s not only students who work on material s, but also those who work on pharma ceuticals, on specialty coatings, on adhesives, and on s pe cialty chemicals. The number who work in commodity chemi cals has dropped so that it now is less than a quarter of our graduates. The number who work in consulting has risen dramatically, as commodity chemical businesses outsource many of the in-house functions they used to do. For ex ample, in one case, a commodity chemical company took its process engineering group from 1500 to fewer than 50 people. This is not a business cycle; this is a change in the way they expect to do business. This is why the number of people involved in consulting has gone up. Thus, the nature of the jobs that our students are doing has changed dramatically. The next question concerns where the changes are reflected in our curriculum. To explore this, I have shown a basic generic curriculum in Table 2. It con tains the usual stoichiometry the thermodynamics and the transport classes. The three classes in kinetics, process con trol, etc. are the place where departments will have unique offerings. For example, this is the location of courses in polymers or biochemical engineering or environmental en gineering. Such uniqueness is a strength of our departments, a way in which we add special skills to a common core. There are a few places in these classes that contain mate rial on products, that subject on which our students are most likely to work. The most logical place to add this type of material is in the capstone design class. This class usually focuses on process design the tradition of our discipline. The hierarchy suggested by Jim Douglas 11 31 for this process TABLE2 Generic Chemical Engineering Curriculum de sig n see m s to me especially strong and appropriate. It is s ummarized on the l eft side of Table 3. After deciding whether a proce ss is batch or continuou s, one then move s on to flow s heet s, which are almost alway s continuo u s The initi a l flow s heet s center on the s toichiometry The next lev e l in the hierarchy which adds the recycles often involve s a discus s ion of the chemical reaction s Once these ar e es t a bli s hed one moves on to the separation train s and finally to the heat integration. All of this makes for a good course. If we want to emphasize product design, we need to go beyond this hierarchy. We cannot s imply substitute a prod uct for drug delivery for the existing process a nd carry out the same kind of hierarchy Instead the hierarchy s u ggested by books on product design (e.g., Ulrich and Eppingerf 1 4 1 ) is exemplified by that on the right s ide of Table 3 After first identifying a corporate need one generates ide as to fill thi s need. One then decides between these alternatives and fi nally decides how to manufacture the chosen product. The manufacturing ste p essentially includes all of Jim Dougla s' hierarchy Thus the important s tep s in product design anticipate those in proce ss de sig n. Product de s i g n implies a focus on the initial decision s around the form of the produ ct and implic itly de-emphasize s its manufacture. Such an emphasis s hift s the curriculum away from the common engineering calcula tions that have been our bread and butter. Such an emphasis includes s ubject s that are normall y left to those directly concerned with the bu si nes s. I am co ncerned that if I make this shift in a de sig n class I will wind up teaching my students watered-down bu s ine ss sc hool prin c iple s rather than real" engineering. I undertake this c han ge because TABLE3 Process Design versus Product Design Most universities teach a similar sequence. 0 Stoichiometry ( l course) All of proc ess d es ign is co ntain e d in th e last step of product d esig n Process Design 0 Thermodynamics (3 courses) I Batch vs. Continuou s Proce ss 0 Transport Phenomena and Unit Operation s (3 courses) 0 Reactor s, Process Control etc. (3 courses) 2. Input s and Outputs 3. Reactor s and Recycl es 0 Proce ss Design (2 courses) 16 4. Separations and H eat Integration TABLE4 "Sick House" Ventilation l. Cu s tomer need; ve ntil ate for und er $8 00 2. Ideas : Open window Controlled vent Heat exc han ger Heat and humidity exc h a nger 3 Select he at and humidity exc h anger 4. Manufacture follows kidney dialysis Product De sign l. Identify Customer Needs 2. Generate Id eas to Meet N eeds 3. Select among Id eas 4 Process Design for Manufacturin g Chemi c al Engineering Edu c ation

PAGE 19

so many more of my students are encounteri n g this shift in their professional lives. I want them to see how prod uct design works When I've discussed these ideas with other faculty, I often get the ind i gnant reaction that the faculty are already doing this. Some have mailed me syllabi and reports that include aspects of product design. Without exception, what I have received represents good education but almost without ex ception the material seems to skip all step s except the last in the product design hierarchy in Table 3 These earlier steps seem to me too important to leave to the MBAs As an example of these ideas consider the so-called "sick house syndrome" that has developed as houses were built to be energy efficient. Such houses exchange their air as infre quently as twice a day. In contrast a house built fifty years ago exchanges its air almost every forty minutes Thus, while the modern house does not cost much to heat it can concentrate radon from the basement formaldehyde released from carpeting and drapery, and carbon dioxide from the people who live in the house The modern house needs more fresh air. Thus the product needed is a device that allows a house to remain energy-efficient but which provides fresh air at the ASHRE standard of 19 cubic feet/minute/person in the house. The way in which the product development might proceed is shown in Table 4. The need is for a device co s ting less than about $800 that can provide this degree of ventilation. Idea s include opening a window providing automatic con trol for opening a window providing a heat exchanger, and providing an exchanger for both energy and mass. Opening a window sacrifices the energy benefits of insulating the house in the first place. Opening the window with an automatic controller that might anticipate weather cycles makes sense. For example one could open the window only on sunny winter days and keep the house closed on cold winter nights. Using a heat exchanger can provide the necessary ventila tion at an order of magnitude less heat loss. As anyone who has bought a house with such a heat exchanger knows, however the heat exchanger also exhausts the water va por in the house. The heat of evaporation of the water is about a third of the heating value in the humidified air. If the heat exchanger runs the house dries out and becomes very uncomfortable. The final alternative is the most complicated but the most satisfying. In this case, one uses a heat exchanger in which the walls are membranes selectively permeable to water vapor. As a result, one captures 90 % of the energy and 90 % of the water vapor, but exhausts the carbon dioxide, formal dehyde and radon in the house. The question is cost. The students need a more complete design, perhaps using the manufacturing technology developed for kidney dialysis to make the membranes. This is an area of active commercial development by several heat exchanger companies. Wint e r 1999 CONCLUSIONS We are now ready to answer the question posed in the title of this paper: "Do changes in the chemical industry imp l y changes in the c h emical engineering curric u lum?" The changes in the chemical industry are clear-a movement away from commodities, a romance with biotec hn o l ogy, and a long-term interest in specialties. Major changes in the curriculum are probably not needed; our students still have the basic s kills necessary not only for the changed chemical industry but also for the other jobs they now hold. These changes in the industry do mean that o u r students will work much more on chemical products than o n chemi cal processes. As a result, we will want them to think more about product design in addition to process design. T h e work on product design will follow a different hierarc h y th an that which effectively organizes process design. But I'm not sure of this You may remember that I began this article by saying that I was going to follow t h e l ead of Pearce Williams to write a paper on what I thought might be done rather than what I had already found effective With Geoff Moggridge, I am going to teach product design as a Zeneca fellow at Cambridge University in the academic year 1998-1999. If we are successful, I will try to move some of these ideas back into our design courses here at Min n esota. I am not yet sure they will work I l ook forward to discussing with you what parts do work and what parts do not. REFERENCES 1. Williams, L Pearce, Michael Faraday, Simon and Schuster, New York, NY ( 1964 ) 2 Spitz Peter Petrochemicals: The Rise of an Industry, Wiley New York, NY ( 1988 ) 3. Ellis R.A., "At the Crossroads Eng. Workforce Bull. No. 130 ( 1994 ) 4. Walker, W.H W K. Lewis, and W.H. McAdams, P r inciples of Chemical Engineering, McGraw Hill, New York, NY (1923) 5. Bird, R.B., W E Stewart, and E N Lightfoot, Transport Phenomena, Wiley, New York, NY ( 1960 ) 6. Astarita, G and J M Ottino, "Thirty-Five Years of B SL," Ind. Eng. Chem. Res 34, 3177 (1995 ) 7 Levenspiel 0., Chemical Reaction Engineering, Wiley, New York, NY ( 1963 ) 8. Sherwood, T.K., Absorption and Extraction, McGraw-Hill, New York NY ( 1937 ) 9. Sherwood, T.K., and R.L Pigford, Absorption and Extrac tion, McGraw-Hill, New York, NY ( 1952 ) 10. Sherwood, T.K., R.L Pigford, and C R. Wilke, Mass Trans fer, McGraw-Hill, New York, NY (1975) 11. Hougen, 0. A., and K.M. Watson, Chemical Process Prin ciples III: Kinetics and Catalysis, Wiley, New York, NY (1947 ) 12. Weatherall, R.K., "Strongest Market in Years for New Gradu ates Engineers, 4 2 ( 1998 ) 13. Douglas J.M., Conceptual Design of Chemical Processes, McGraw-Hill, New York NY ( 1988 ) 14 Ulrich K.T ., and S.D Eppinger, Product Design and Devel opment, McGraw-Hill, New York, NY (1995 ) 0 17

PAGE 20

t.3 ... bllllii._c_l_a_s_s_r_o_o_m _________ ) DISCONTINUITIES IN ChE EDUCATION ST EPHEN WHITAKER University of California D avis, CA 95616 F rom the perspective of the first-year student, the en tire fo uryear chemical engineering program repre se nt s an overwhelrrting array of courses and subject matter. One must learn about ionic s trength and indefinite integral s, acoustics and hydro sta tic s, turbulence and cherrti cal kinetics, organic cherrtistry and proces s dynamics, optics and quantum mechanic s, s toichiometry and process syn th sis, radiant energy heat transfer a nd partial differential equa tions, etc., etc. Viewed in it s entirety, the typical chemical e n gineering program is enough to make a stude nt change majors ; but if taken one step at a time, the overall objective become s quite feasible. In the ideal cherrtical engineering program one would like to develop a sea mless pa ssage from ionic strength to proces s synthesis. Given the size of the ta sk, it should not be s urpri s ing that the route from A to Z contains a few leaps of faith. The failed leap of faith within the confines of the university leads only to a lurking sense of in sec urity and no real physi cal damage. Outside the univer s ity however a failed leap of faith may be a financial di sas ter a phy sica l disaster, or both For this reason, we s hould avoid or rrtinirrtize the leaps of faith in o ur educational programs or we should clearly iden tify them as such A discus s ion of the so-called principle of lost work represents an intere sti ng example of the latter. [1 1 THE MECHANICAL DISCONTINUITY In the first physics course, s tudent s encounter Newton 's second law for a particle written in the form 18 ct dt(mv)=F ( I) Stephen Whitaker received his undergradu ate degree in chemical engineering from the University of California Berkeley, and his PhD from the University of Delaware He is the au thor of two undergraduate texts, Introduction to Fluid Mechanics and Fundamental Principles of Heat Transfer, and a monograph The Method of Volume Averaging He has received a num ber of awards for his contributions to both un dergraduate and graduate teaching Copyright ChE Division of ASEE 1999 Here m is the mas s of the particle v i s the velocity of the particle relative to an inertial frame, and Fis the force acting on the particle. Given the s ucces s of Eq. ( 1 ) in predi ct in g the motion of the planets around the s un and in predictin g the motion of projectile s in a phy s ic s lecture hall students ac quire a certain degree of confidence in Newton's seco nd law This confidence may begin to weaken when they mo ve on to a cherrtical engineering st udy of fluid flow where they are often confronted with a dictum of the form sum of forces rate of rate of rate of ac t ing o n momentum out momentum i nt o accumulation of (2) the contro l = of the contro l the control + momentum in th e volume volume vo lume control volume It is true that the concept of a contro l volume ha s already been presented in a course on material balance s; but the distance between Eq. ( 1 ) a nd Eq. (2) i s so great that mo st s tudents view the latter with so me distrust. The s tudent 's s kepticism is quite justified, but the repeated u se of Eq. (2) to s olve real problem s eventually leads to it s acce ptance Such a l eap of faith in the de s ign of an oil pipeline passing underneath the city of Los Angele s would never be consid ered, but Eq (2) i s s omething that everyone knows is true, and many students move s ucce ssf ully through their pro grams of study with it sec urely locked in their tool boxe s. THE THERMODYNAMIC DISCONTINUITY When students have made s uffici e nt progres s in their st ud ie s of fluid mechanic s, they will often solve a variety of incompressible flow problem s (it is important to keep in rrtind that there are no incompre ss ible fluids but ther e are flows that can be approximated as incompressible) u s ing the Navier-Stokes equations and the continuity equation. The se eq u ations can be expressed as pl~: +v-Vv J =-v'p+pg+V 2 v (3) V-v=O (4) Equation (3) represents the governing differential equation Chemica l Engineering Educat i on

PAGE 21

for the fluid velocity, v and Eq. (4) is the constraining equat ion for the vector field represented by Vp. That is to say that the momentum source, Vp, must be distributed in such a manner that the velocity determined by Eq. (3) will be solenoidal. If the pressure is specified at some point one can use the vector field, Vp, to calculate the pressure every where. The actual determination of the pressure field is discussed in some detail in Reference 2. At the same time that students are using Eqs. (3) and (4) to determine the pressure in a course on fluid mechanics they are also calculating the pressure in a course on thermody namics using an equation of state The simplest equation of state i s the ideal gas law given by p=~RT V (5) This expression for the pressure would appear to have no connection with the pressure that one would determine from Eqs. (3) and (4), thus suggesting that there is a me c hani ca l pressure used in the solution of certain fluid-flow problem s and a thermodynamic pr ess ur e used in the solution of ther modynamics problems. It would be best to think of the pres s ure, as determined by an equation of s tate as the pres sure, and to think of the pressure determined by Eqs. (3) and (4) as a good approximation of the pressure. If it is not a good approximation, Eqs (3) and (4) should not be used to so lve the flow problem under consideration. The resolution of the conflict between Eqs. (3) and (4) and an equation of state, such as Eq. (5), relies on Birkhoff's plausible intuitive hypothesis that small causes give ri se to sma ll effects. THE MUL Tl PHASE DISCONTINUITY After having completed courses in fluid mechanics, ther modynamics heat transfer and mass transfer chemical en gineering students are often confronted with a course on mass transfer operations or unit operations. Since virtually all chemical engineering proces ses involve multiphase sys tem s, a s tudy of the gas-liquid contacting device illustrated in Figure I is a harbinger of things to come, and students approach this problem with a great deal of interest. Often they are equipped with Eqs (3) and (4) from a course on fluid mechanics, the thermal energy equation from a course on heat transfer 141 {6) and the species continuity equation from a course on mas s transfer I 5I A=l 2, .. ,N {7) Most students are somewhat dismayed when the process illu s trated in Figure 1 takes on the form shown in Figure 2, and the rigor represented by Eq (7) is replaced by the Winter 19 99 suggestion that ma ss of A entering in the gas phase = m ass of A lea v ing in th e gas p h ase + ma ss of A tran s ferred to the liquid pha se (8) After the struggle to reach Eq. (7) via a series of challenging courses, it is disappointing to be asked to return to the concepts encountered in the course on material balances. What is worse is that the analysis of the process illustrated in Figure 1 will be heavily ba se d on the intuition suggested by Figure 2 and most students will have no idea how reliable the final result will be The resolution of this problem can be achieved using the method of volume averaging_l 6 J RESOLUTION OF THE MECHANICAL DISCONTINUITY Rather than leap from Eq. (1) to Eq. (2), one can follow a sequence of steps that begins in the eighteenth centuryf7 1 and leads to our current understanding of continuum mechanics A central idea in continuum mechanics is that the laws of physics can be applied to any body that one imagines as being cut out of a distinct body. Truesde!Jl7 1 attributes this idea to Euler and Cauchy and refers to it as the cut prin c iple. Engineering students encounter this idea in a course on statics where it leads them to the concept of a free-body diagram. If we accept this idea, we can cut an arbitrary body from a moving deforming fluid and state the axioms for ma ss and mechanics as follows: Mass _5!_ fpctY=O dt o/m(t) Linear Momentum: Euler's First Law (9) :t J pvdV = J pbdV + J t r n ; dA {10) Vm(t) o/m( t ) Jilm(I) liquid Figure 1. Gas-liquid co ntacting device. t Fig u re 2. Model of a gas liquid contact ing device. /9

PAGE 22

Angular Momentum: Euler's Second lAw :t fr X pvdV = Jr X pbdV + Jr X t ( n ) dA (I I) 'J/m{ t) 'J/m( t ) Ylm{ t ) Here, '1/m(t) represents the time-dependent region occupied by a body p is the ma ss den sity, vis the fluid velocity bis the body force per unit mass t e n > is the stress vector, and r is the position vector. Both v and r are mea s ured relative to some inertial frame The representation of the angular mo mentum principle given by Eq (11) assumes that all torques are the moments of forces and thi s ignores the existence of body torques and couple s tresse s that have been observed in polar fluids csi The forms of the se three axiomatic s tatement s suggest the need for a study of the kinematic s of volume integrals and this leads to the genera l transport theor eml 91 given by J\jldV = I a\jl dV + J\j/W ndV dt at (12) 'Ila (t) 'Ila (t) Yl 3 (t) Here 'v' 3 ( t) represents the region occupied by an arbitrary moving volume, J.t 3 (t) is the bounding surface of thi s vol ume, and w n is the speed of displacement of the bounding surface. When the arbitrary velocity, w, is set equal to the fluid velocity, v, we obtain the R ey nolds transport theorem given by (13) When applied to Eq. (9), this theorem provides JpdV= J ap dV+ Jpv ndV=O ( 14 ) dt at 'J/rn (t) '11 m (t) Yim ( t ) and use of the divergence theorem leads us to J[: +\7 (pv)]dV=O (15) 'llm(t) Assuming that the integrand is continuous and noting that the limits of integration are arbitrary leads to the continuity equation ap + \7 (pv)=O at (I 6) In order to extract the governing differential equations associated with the linear and angular momentum principles, we first need to follow the work of Cauchy and prove C7 1 Cauchy 's Lemma: t ( n ) = t (n ) Cauchy's Fundamental Theor em: t ( n ) = n T (I 7) (18) The first of these is introduced as intuitivel y obvious in every statics course where it is applied to the shear stresses 20 acting on opposing s urf aces of a beam th at h as been s ub jected to an Eulerian c ut. The seco nd result is genera ll y avoided becau se of its complexity eve n though most st dent s have completed a course on matrix algebra prior to their study of fluid mechani cs. The u se of Eqs. ( 17 ) and (18) a long with the Reynold s tran s port theorem allows u s to extrac t the following differ ential equations from Eqs. (10) and ( 11 ): Cauchy 's First Equation ;t (pv)+ \7 (pvv)=pg+ \7 T Cauchy 's Second Equation T = T T ( I 9) (20) At this point we are in a po s ition to derive the macroscopic momentum balance that was de sc ribed in words by Eq. (2) We be gi n by integrating Eq. (19) over an arbitrary, moving control vol ume to obtain f ; t (pv)dV + f \7 (pvv)dV = f pbdV + J \7 Td V (2 1 ) 'II, ( t ) 'II ( t ) 'II ( t ) 'II, (t) We now u se the general transport theorem and the diver gence theorem to arrange this result in a u sefu l form given by :t JpvdV + Jpv (v-w) nd A= Jpbd V+ Jt ( n ) dA (22) '11,( t ) Yl,(t) '11 (t) Yl,(t) Thi s represents a precise mathematical description of the words co ntained in Eq. (2), a nd it clearly indicates that the so urce of the se words is Euler's first law. While th e route from the axiom give n b y Eq. (10) to the pro ved theorem given by Eq. (22) consists of only a few s tep s, one must in vest a s i g nifi ca nt a mount of time in the st udy of kinematics and s tre ss in order to deri ve thi s result. (At UC D avis we have two ten-week courses in fluid mechanics a nd time i s les s of a problem than in mo st pro grams.) While kinematics and s tre ss may be co nfu si n g, we s hould heed the word s of Pucciani and Hamel,C' 01 who provide the following advice to s tudent s: Th e re i s no learning without co nfusion. I t is by the organi zation of this confusion that yo u will p r ogress Said another way, it i s b etter to b e confused a nd frustrated by the concepts of kinematics and stress than to be b affled b y the leap of faith from Eq. (1) to Eq. (2) In order to follow the development from Eq. (10) to Eq (22) in a successful manner the faculty mu s t be aware of what students know what they don t know and what the y are supposed to know For example, in the typical sta tic s course, students make u se of Euler 's law s to so l ve problem s, but the laws are never identified in a clear and concise manner. The concept of an Eulerian cut i s pre sented as being obvious en route to the development of a free-body diagram Chemical Engineering Education

PAGE 23

but no mention is made of the fact that it was not obvious in the eighteenth century Cauchy 's lemma i s used in the same way i. e the shear is up on this side of the cut and down on that side of the cut. In order to unravel this mass of intuition the faculty must be aware of the content of previous courses and must be prepared to extract s ome order and logic from the s tudent 's previous studie s In addition to statics these previou s studies include calculu s where the definition of a derivative is presented and the projected area theorem is given 1 1 1 1 These are all that one needs to derive the general transport theorem. Previous studies also include a course on matrix algebra where the students learn that a three-by-three array can be used to transform one set of three numbers to another set of three numbers. This is the essential feature of Cauchy's fundamental theorem RESOLU T ION OF THE T HERMODYNAMIC DISCONTINUIT Y In order to resolve the thermodynamic discontinuity, we begin with a reasonable description of a compressible flow process. This consists of the governing equations for the density velocity temperature, and pressure that can be ex pressed as Governing equation for p i)p + V (pv) = 0 a1 Go v erning equation for v ( av ) 2 p ar+v Vv = Vp+pg+V V Governing equation for T pep ( ~: +vVT ) =kv' 2 T Governing equation for p p=p(p T) (23) (24) (25) (26) The first of these equations represents a completely general form of the continuity equation, while the last represents a completely arbitrary equation of state Equations (24) and (25) represent special forms of the equations of motions and the thermal energy equation, but they are general enough for our purposes. The usual concept associated with an incompressible flow is that the variation of the den s ity is small enough so that the dependent variable in Eq. (23) can be replaced with a con s tant p 0 This means that one of our four dependent vari ables is determined by some means other than a law of physics and this, in turn, means that we must discard one of our laws of physics. Our new description of the physical process is given by (27) Wint e r 1999 p 0 ( a;; +vm Vv m ) =-Vpm +p 0 g+V 2 vm (28) p 0 cp ( 0 ;; +vm v'Tm )=kv' 2 Tm (29) in which we have used vm Pm and T m to represent the velocity pressure and temperature determined by Eqs (27) through ( 29). These quantities differ from p v, p, and T that are determined by Eqs. (23) through (26) and we would like to understand the asymptotic conditions that lead to (30) When Eqs (27) through (29) produce velocity, pressure, and temperature fields (v m, Pm, Tm) that are good approximations of the fields (v, p, T) determined by Eqs. (23) through (26), we say that the flow can be approximated as incompressible. Under these circumstances, the pressure can be calculated by purely mechanical means; but it would seem best not to refer to Pm as a "mechanical pressure," but simply to say that Pm is a "good approximation" of the pressure determined by an equation of state. The general asymptotic conditions associated with Eq. (30) are difficult to develop; r 1 21 we can, however, explore the first of these conditions for steady flow in the absence of any temperature effects without a great deal of effort. This re quires that we consider an isothermal process described by the steady forms of Eqs. (23) through (26), and then search for conditions that lead to p p 0 (31) It will be in the nature of a plausible intuitive hypothesis 1 3 l to assume that v v m and p Pm when the condition repre sented by Eq. (31) is satisfied. For the case in which temperature effects are negligible, we can invert Eq. (26) to obtain p = p(p) (32) and a Taylor series expansion about p 0 leads to ( ap ) 1 / a 2 p1 P=P o +(P-P o ) op T +2(P-Po) l ap 2 )T + .. (33) Here, p 0 is the density determined by Eq. (32) at the refer ence pressure p 0 As an estimate of the density change that occurs for the process under consideration, we use the first term of the expansion to obtain p-p 0 =0[ ( :~ t ~P] (34) in which ~p is representative of the maximum pressure change that occurs in the system. From a thermodynamic analysis, we know that the speed of sound is related to the derivative of the density with respect to the pressure at constant entropy. This relation is given by 21

PAGE 24

( :~ l (35) in which c is the speed of sound. As an approximation, we use (: ~ l C; (36) so Eq. (34) takes the form p-p o =0 ( :r) (37) In order to obtain an estimate of the pressure change, we first make use of the steady form of Eq. (24) to estimate the gradient of the pressure as v'p = O(pg) + o( v' 2 v) + O(pv v'v) {38) The idea associated with this estimate is that Vp may be as large as any of the other terms in Eq (24) but not signifi cantly larger. In addition, it is possible that Vp may be much smaller than any of those terms and thus Eq. (38) should be thought of as an overestimate of the pressure gradient. For example, in a laminar boundary layer created by a uniform flow past a flat plate the pressure gradient is essentially hydrostatic Laminar boundary layer flow v'p pg (39) while the viscous and inertial terms are essentially equal and much larger than Vppg i.e ., Laminar boundary layer flow pv v'v = o( V 2 v) v'p pg {40) In this case, only the first estimate given by Eq. (38) is valid and it becomes clear that one must have some idea about the nature of the flow under consideration in order to use Eq. (38) successfully If u 0 represents the characteristic velocity for the process, we can use order-of-magnitude analysisr 1 3 1 to obtain the esti mates pg= o[(pg)11. g ] v' 2 v = o[(u o / L~ )"l pv v'v = o[(pu ~ / LP )"-p] ( 4 la) (41b) {41c) Here, L represents the viscous length, and Lp represents the inertial length ,l1 41 while A g, Ap are unit vectors that are parallel to the gravitational term, the viscous term, and the inertial term respectively For the laminar boundary layer example discussed above, A and Ap are parallel and A g is an arbitrary unit vector In order to make use of Eq (38) to estimate the pressure difference that appears in Eq (37), we express the pressure gradient as 22 v'p = 0( Llp / L) {42) Here one must keep in mind that the pres s ure gradient will be influenced by all three terms represented by Eqs. (41), thus Eq. (42) represents thre e separate estimates and it is left to the reader to keep this fact in mind Use of Eq s. ( 41) and (42) in Eq. (38) leads to the following estimate for the pressure change owing to gravitational, viscous and inertial effects: Llp = o[(pg)L]+o[(u 0 /Lt)L]+o[(pu ~ /Lp)L] (43) Substitution of this result into Eq. (37) provide s the follow ing estimate for the change in density that occurs for the process under consideration: p -/ 0 = o(gL / c 2 ) + o(u 0 L/ pLtc 2 ) + o( u ~ L/ Lpc 2 ) {44) If we define a Reynolds number and a Mach number as M=~ {45) C the estimates given by Eq. (44) take the form p-/ 0 = o( gL!c 2 )+ o[R e1 M 2 (LIL)]+o[M 2 (L1Lp)] {46) From this result we conclude that p -po << I p or p P o {47) when the following constraints are satisfied : gL/c 2 l M 2 Re(LIL) M 2 Lp!L) {48) Here one must remember that L has a different meaning in each one of these constraints which are often replaced with the single condition that the Mach number sq uared is sma ll compared to one. This has considerable appeal for the last constraint in Eq. (48) since LP is often large compared to L; the simplification given by M 2 <<1 has les s appeal as a substitution for the second constraint however s ince L is generally small compared to L. A little thought will indicate that the first constraint given by Eq. (48) is difficult to violate, and thus it is the constraints involving the Mach number that must be considered with care. It seems plau si ble that when Po is a reasonable approxi mation for p, we can assume that v m and P m are reasonable approximations for v and p; but a rigorous proof would require that we identify the asymptotic conditions that allow us to simplify Eqs (23) through (26) to the incompressible approximation represented by Eqs. (2 7) through (29). One should keep in mind that the analysis leading to the con straints given by Eqs. (48) was based on a steady-flow process, and that there are unsteady low-Mach-number pro cesses for which the flow cannot be treated as incompress ible under any circumstances An example is given in Figure 3, and for that type of process the approximation represented by pP o= O(Vp)L {49) Ch e mi c al En g in ee rin g Edu cat i on

PAGE 25

is not at all applicable. The partial resolution of the thermodynamic discontinuity required that we clearly identify the general case indicated by Eqs. (23) through (26) and that we discard a law of physics in favor of the approximation given by p = p 0 The compressed gas justification of thi s a pproxi mation was based on the equation of s tate that pro vided Eq. (3 7 ) and the order of-magnitude analysis given by Eqs (38) through (46). The use of order-of-magnitude analysis allows s tudents to go beyond the assumptions based on the title of a course, the title of a text, or the title of a chapter in a text. Since the se titles are not available to our students when the y Fig u re 3 Compression leave the university we pro cess. should encourage them to formulate their own assump tion s and then follow those assumptions with re stric tions and constraints 1131 It is importa n t to understand that the thermodynamic dis continuity cannot be reso l ved only by discussion in fluid mechanics courses. Faculty member s who teach thermody namics must be aware of the problem and speak to the is s ue A righteous attitude about the correctness of Eq (5), or a more general equation of state, provides no help to the student who must deal with the reality of incompressible flows One must remember that the students take every course in the program and they do not have the luxury to choose their battleground RESOLUTION OF THE MUL Tl PHASE DISCONTINUITY Studies of multicomponent mass transport usually include a derivation of the species continuity equation 15 1 and the molar form of this result is given by acA ( ) -+'v CAVA =RA a1 A= 1 ,2,3, ... ,N (50} Knowledge of the molar concentration cM i s a central is s ue in chemical engineering since it forms the basis for all se pa ration and purification processe s, for all reactor design cal culations, and for all studies of contaminant transport in the l and, air, and water. The macroscopic mole balance associ ated with Eq. (50) can be derived by following the ste ps that link Eq (19) to Eq (22), and the re s ult is given by :t fcAdV+ f cA(vA-w) ndA= JR AdV A=l,2, ... N 'll ( 1 ) J'l ( t ) V ( t ) Wint e r 1999 (51) Both Eq. (50) and (5 1 ) repre se nt powerful problem-solving tools, and mo st chemical engineering students acquire a certain degree of ski ll in the application of these results for a variety of single-phase tran s port problems; their application to multiphase systems, however is problematic. Most multipha se transport processes cannot be solved di rectl y in term s of either Eq. (50) or Eq (51), but require in s tead the local vo lume-averaged form ofEq. (50) 11 5 181 The de ve lopment of this form begin s by associating an averaging volume with every point in the region under consideration. Thi s allows one to define a volume-averaged concentration everywhere and to generate a spatially smoothed concentra tion field In Figure 4 we have illustrated a two phase system and a spherical averaging volume having the centroid l cated at the point identified by the position vector x. For this sys tem we identify the point concentration of species A i n the y phase as c Ay and we define the superficial average concentration by (cAr) I, = J cA y dY (52) V y( x 1 ) Here Y y (x t ) represents the volume of they-phase co n tained within the averaging volume 'J/, and we have clear l y indicated that the s uperficial aver age concentra tion is associated with the point lo cated by the po si tion vector x For the particu lar case illus trated in Figure 4 the position vector x locates a point in the x o phase where the point concentration of Figure 4 Two-phase system species A ma y be zero. In general, the intrinsic average co n ce ntration is preferred for the analysis of multiphase tran s port processes and this is defined by ( C )Y I =-'J C dV Ay x Y y (x t) Ay V y( x t ) (53) The superficial and intrinsic average concentrations are r e lated according to (54) in which Ey i s the volume fraction of the yphase defined explicitly by 23

PAGE 26

(55) In many systems, the superficial and intrinsic averages dif fer by a factor of three or more, and thus it is important to make use of a nomenclature that clearly identifies these two concentrations. In Figure 5 we have shown a two-phase system for which we would like to develop the design equation for the concentration of species A. We think of the flowing fluid as the yphase, while the cr phase could represent porous catalyst pellets, or droplets of a more dense fluid that is descending through the y phase The governing differential equation for the concen tration of species A in the yphase is given by acAy ( ) at+ y' CAyV A y =RAy (56) and we begin our analysis of this point equation by forming the superficial average to obtain f a;t ctv + J v (cA y v Ay )ctv = J RA y dY (57) v 1 ( x t ) V 1 ( x t ) V 1 ( x t ) In order to transform this result to something u sefu l we make use of two theorems that are essentially extensions of the classic one-dimensional Leibniz rulec 9 Pr ob 3 5 l for differ entiating an integral. The first of these is the general trans port theorem, which we used earlier in our treatment of the mechanical discontinuity. This theorem allows us to express the first term in Eq. (57) in the form 24 1 f ac A ct [ 1 J ] 1 J __ Y dV=C A ydY -cA y w nyadA (58) 'f/ at dt 'f/ 'f/ V y ( x t ) Vy ( x t ) A l" ( x t ) averaging volume, '11 Figure 5. Mass transfer and reaction process in a two-phase system. Here rcr represents the unit normal vector directed from the yphase toward the cr pha s e and w nyc; repre s ent s the speed of displacement of the y cr interface. Thi s i s zero if the system under consideration i s a packed-bed catalytic reactor, but would be non-zero for a fluidized-bed reactor or any fluid-fluid system. The second theorem needed for the analysis ofEq. (57 ) is the s patial av e ra g in g th e orem a nd the derivation of this theorem r 6 191 i s analogou s to the derivation of the general transport theorem. Application of the spatial averaging theorem provides J v ( c A y v Ay ~v v 1 ( x t ) =v'[~ JcA y V Ay dY]+ J cAy V Ay nyadA {5 9) V Y ( x t) A,., ( x t) and u se of this result, along with Eq. (58 ), in Eq (57 ) lead s to :t[~ Jc Ay dY]+v' [~ JcA y V Ay dY ] V 1 ( x t ) V 1 ( x t ) +~ Jc Ay (v Ayw) n ya dA=~ JR Ay dY (60 ) A,.,( x t ) V 1 ( x t ) One should note that this result is a superficial average transport equation and thus each term has unit s of mole s of species A per unit time per unit volume of the ycr system Use of the nomenclature for the superficial average indi cated by Eq 52 allows us to express Eq (60) in the more compact form given by f(cA y )+v (cAy VA y) + Jc Ay (v Ay -w) n ya dA= ( R Ar) {61) ac c umul a tmn y pha se A ,., ( x 1 ) h o m oge n eo u s tr a n s p o rt in t e r fac i a l tr a n spo rt r eac ti o n Here it is understood that the averaged quantities are as s oci ated with the centroid of the averaging volume identified by the position vector x in Figure 4. In addition we have ex pressed the accumulation in terms of the partial time derivative since ( c A y ) is associated with a point that is fixed in space. The first and last term s in Eq. (61) can be expressed in terms of intrinsic average s by using the relation given by Eq. (54), and this leads to ;t ( e y( c Ay )Y )+ v (cAy v Ay) 'II I CAy (v Ay -w) nyadA +Ey (R Ay r (62) Al"( x t ) In order to simplify the convective transport term, one can use the averaging theorem and the divergence theorem to show that v -(cAy vA y) = -~ f c Ay v Ay n ye dA A.,.,( x t) (6 3) Ch e mi c a l En g in ee rin g Edu c ati o n

PAGE 27

in which Aye(x,t)represents the area of entrances and exits associated with the volume, Yy(x,t) For man y practical application s, diffusive transport i s negligible compared to convective transport at e ntrance s and exit s, and this encour ages the si mplification (64) in which v y is the ma ss average velocity. At this point we make use of Gray' sr 201 s patial decompo sition for the concen tration and velocity (6 5 ) and follow the work of Carbonell and Whitaker r 211 to express the co nvecti ve tran s port in the form (cAyvr) = Ey(cAyr (vr)y + (cAy v y) (66) vo lume averaged convec ti ve transport dispersive transport While the intrin sic average concentration i s the preferred con centration for the de sc ription of multipha s e mass transport processes, most workers favor th e u se of the s uperficial aver age ve locity ( v y ) =E y( v y r and thu s we normally express Eq. (66) as vol um e averaged co n vective transport dispersive transport (67) (68) Use of thi s result (an d the approximation given by Eq 64) with Eq (62) l ea d s to ;t ( Ey ( C Ayr)+ v' -( ( C Ayr (Vy)) =-v'-(cAy"y)f CAy(vAy-w) n ycr dA+E y( R Ayr (69) A-,.,(x,t) Her e we are confronted with the development of appropriate representations for the disper s ive transport the interfacial transport, and the homogeneou s reac tion rate. The diffusion model of dispersion is an approximation for active sys tem s;l2 2 2 3 1 but it is likely to be quite acce ptable for mo st design applications. Thi s encourages u s to express the di s per s ive flux as (70) our local volume-averaged transport equation takes the form ;t ( Ey ( C Ayr) + v' ( ( C Ayr (Vy)) acc umulati o n vo lume averaged convec tive tran spo rt v'-( D v'(cA yr) dispersive tran spo rt 'II f CAy(vAyw) n ycrdA + E y( R Ayr (71) A-,., ( x t) homogeneou s interfacial transport reaction Wint e r 1 999 The treatment of the interfacial flux depends on the type of process that takes place in the cr-ph ase; Equation (71) howe ver, provides the basic framework for a design equa tion that ha s been derived directly from Eq ( 56) This pro vi de s a much more rigorous formulation than one can obtain on the ba sis of the statement ( mole s of A) ( mole s of A ) ( mole s of A) ( moles of A ) entering + produced by reaction = l eav in g + acc umul ated ( 72 ) and it clearly se nds a message to our st udents that their efforts to under s tand the details of diffu s ive and convective ma ss transport have not been wasted. When I teach the second of our two ma ss transfer courses at UC D av i s, I use the general transport theorem to obtain Eq. ( 58 ) because I know that the s tudents have proved t h is theorem in the fluid mechanic s course In addition, I derive the s patial averaging theorem by a s traightforward applica tion of the projected area theorem that i s presented in the fluid mechanics course. It also help s to know that the pro jected area theorem was encountered originally in one of the calculus co ur ses.PI.Chap 171 If I did not know what the students had covered in previous courses, I would be forced to retreat to the approach suggested b y Figure 2 CONCLUSIONS In thi s p a per we have s hown how so me of the traditional discontinuities associated with chemical engineering educa tion can be eliminated or minimized. Time and effort are required to accomplish thi s, but the effort will convince stu dent s that leaps of faith can be avoided, a nd thi s is surely worthwhile. ACKNOWLEDGMENTS This paper is based on a lecture presented at the VII Encontro Brasileiro sobre Ensino de Engenharia Qufmica in Caxambu Brazil, 1997 The enthusiastic response by my Brazilian colleagues provided the motivation to prepare a written versio n and their encouragement is greatly appreci ate d. NOMENCLATURE JI.a ( t) area of the s urface of an arbitrary moving volume, 'Ila ( t ), m 2 Jim ( t) area of the bounding s urface of a movin g, deforming b ody, m 2 J!ycr(x,t)area of the y-cr interface contained in the averaging volume, ,,; m 2 Jty, ( x t) area of the entrances and ex.its of the y phase contained in the averaging volume ,,; m 2 b bod y force per unit ma ss, N/kg cA molar concentration of species A mole/m 3 c Ay molar concentration of s pecie s A in the y phase, mole/ m J ( c Ay) s uperficial volume average d concentration of s pecies A mole/m 3 -------------Continued on page 61. 25

PAGE 28

195 class and home problems ) The object of this column is to enhance our readers collections of intere s ting and novel problems in chemical engineering. Problems of the type that can be used to motivate the student by presenting a particular principle in class, or in a new light or that can be assigned as a novel home problem are requested as well as tho se that are more traditional in nature and that elucidate difficult concepts. Manuscripts s hould not exceed ten doublespaced pages if possible and should be accompa nied by the originals of any figures or photograph s. Ple ase submit them to Professor James 0. Wilkes (e-mail: wilkes@engin.umich.edu), Chemical Engineering Depart ment, Univer s ity of Michigan, Ann Arbor, MI 48109-2136. NON-ADIABATIC CONTAINER FILLING AND EMPTYING NO EL DE NE VERS University of Utah Salt Lake City, UT 84112 A II thermodynamics textbooks present container fill ing and container emptying (often called bottle fill ing and bottle emptying) as the simplest examples of unsteady -s tate, varying-inventory proces ses If the pro cess is adiabatic and the contents of the container are well mixed, then the differential ma ss and energy balances can be combined and integrated leading to closed-form solutions The classic problems are of the form: Problem 1 An evacuated, rigid, adiabatic container is connected to a compressed air line at a pressure of 738 kPa and a temperature of 22 C. The connecting valve is opened and the air flows in until the container pressure i s 738 kPa. What is the final temperature in the container? Problem 2 A rigid adiabatic container contains air at 641 kPa and 29C. It s valve is opened and it exha u sts to the atmosphere at 86 kPa (at Salt Lake City 1460 m above sea l evel). When the pressure in the container is the same as atmospheric pressure what is the temperature in the container? GENERAL THEORY FOR ADIABATIC FILLING AND EMPTYING For a rigid container with flow of matter in or out, the energy balance on the contents, ignoring kinetic, gravita tional electrostatic, and magnetic energies,l' 1 is (I) and the corresponding m ass balance i s dm sys t em =d m in -dmout (2) For a container filling from so me reservoir (e g., the atmo s phere into an evacuated container, or a large compressed air or s team line into a container), we may assume perf ect internal mixing and that h i n is a constant, combine Eqs ( 1 ) and (2), and integrate to ( mu ),ystem final -( mu ),ystem,initial = h in ( mfinal miniti a l )+ L',.Q (3) If the mixing is not perfect then the s pecific propertie s shown in Eq (3) and throughout thi s paper should be inter preted as mass-average values. For emptying (discharge, blowdown) the si mple integration leading to Eq. (3) is not correct because the expansion work done by the fluid during the emptying proce ss causes the temperature, and hence h 0 u,, to decrease during the proce ss. If one u ses an average value of h 0 m, one can then use thi s integration. r 21 If the material originally present in the sys tem and flowing in or out i s a Noel de Nevers is Professor of Chemical En gineering at the University of Utah where he has been on the faculty since 1963 He has written texts on Fluid Mechanics for Chemical Engineers and Air Pollution Control Engineer ing. In addition he works in thermodynamics and accident investigations. Copyr i gh t ChE Division of ASEE 1999 26 Chem i c al Engineering Education

PAGE 29

perfect gas, we ca n s ub s titute a nd c han ge Eqs. (1) and (3) to (4) For filling we can integrate Eq (4) to ( mCv T) fi ( mC v T) ... sys tem rn a l sys tem mllial =Cp Tin (mfinal min itial )+ L'.Q (5) If m in i t i a l and dQ are both zero ( initiall y empty container a nd adiabatic pro cess), th e n for filling, Eq. ( 5 ) become s T sys tem final = kTin (6) If the co ntainer is not originally e mpty but co ntain s a gas with the same value of k a s the gas that enters, t.Q = o and Tinili a l =Tin then the so lution i s T kT T (k I) min itial sys t e m fin a l = in initial -m final (7) For adiabatic emptying the perfect nuxmg assumption a llow s u s to se t th e system temperature e qual to the outflo w temperature and integrate Eq. (4) to T final m final ( J ( k-1 ) Ti.niti al = miniti a l (8) R e pla ce ment of the m term s by th eir ideal-gas-law va lu es produces ( k 1 I Tfin a l ( Pfin al Jl k J T init i al = Pi nitial (9) w hich i s th e relation for an i se ntropic expansion of an id ea l gas ( PROBLEM S OL U TIO N ) Using Eq s. (6) and (9), with the adiabatic assumption and the further assumption that air i s an ideal gas with k = 1 .40 we can solve Problem s l and 2 finding T sys tem final = kTin = ( 1 .40)(295 K ) = 4 1 3 K = l 40 C a nd ( k-1 1 ( 04 1 Tfin a l =( Pfin al J l k) =( 86 kpa ) l ii ) =0 563 Tiniti a l Pin i ti al 64 1 kpa Tfin a l =0.563Tinitial =(0 563)(302K)=170K=-103 C But experimental re s ult sf 3 l do not agree eve n approxi mat e l y, with the se s imple theori es Thi s di sag reement is ex plained [3 1 as bein g du e to sig nifi ca nt h ea t tran sfer. This appears s tartling becau se the filling and e mpt yi ng experi ments are normally finished in a few seconds. But as shown Wint e r 1 999 below it i s correct. General Theory for Non Adiabatic filling If we now allow for heat tran sfer, a nd replace dQ by hA ( T-T urr ou ndin g J dt and keep the ideal gas assumption, then Eq ( 4) become s d(mCvT) = sys t em CP Tindrn i n CP T o utdm o ut + h A(T sys t e m -Tsurro unctin gs )ctt ( I 0) For filling, we drop th e outflow t er m and rearrange to d( mT \ystem hA { ) -~= m ink Tin T ,ystem -Tsurroundi n gs dt Cv ( 1 I) To save writing, we define a= h A / Cv and drop the sub sc ript on T ,ys t e m If we assume a constant filling rate we can replace the instantaneous sys tem ma ss with rn = mi nitial + mi n t ( I 2) and rearr a n ge to ( m init i al +mint)::= m in (kTin -T)-a(T-T s urrounding s ) (13) Separating varia bl es integrating from sta rt to finish, and assuming that the T s urr o undin gs i s constant and equal to the initial t e mperature of the gas in the container, we find ( m in J ( a J -T + -kT in + -T urrounding s mi n +a mi n +a k a T Tin + -I su r ro unding s min +a min +a ( 14 ) If instead of assuming a constant inflow rate we assume an inflo w rate that d ec lin es linearly with time we can write m sys tem = minitial + f (a bt)dt = m initia l + at 0 5 bt 2 ( 15 ) where a and bare dat afittin g constants. Substituting this for m ,yste m in Eq. ( 11 ) and rearranging, we find dT {a bt)(kTin -T)-a{T-T s urroundings) -= 2 dt m initial + a t 0.5 bt { 16 ) which can be integrated numerically General Theory for Non-Adiabatic Emptying For con tainer emptying, Eq. (1 1 ) b ec ome s d( mT) sys t em dt -m o ut kT a( T T urroundings) ( I 7) For a n ass umed constant flow rate, the equivalent ofEq (13) is 27

PAGE 30

( m initial 111 out t) ~: = -mout T (k I)a( T-T s urroundin gs ) and the equivalent ofEq (14) is [-m 0 u 1 (k1 ) a ]T + aT u rroundin gs [-m ou t (k-1)a ]T initia l + aT,urr o unding s ( J[(k-i}+ --,-!:.] = m initial ffiout t mout minitial (18) (19) For the assumed linearly decrea si ng mass flow rate out, the equivalent ofEq (16) is dT = ( a bt )(k I )Ta( T-T,urroundings) dt minitial a t+0 .5 bt 2 (20) which i s also suitable for numerical integration EXPERIMENT AL TESTS The experimental apparatus/ 41 ske tched in Figure 1, co sisted of an ordinary 0 027 m 3 propane storage container, a pressure transducer thermocouples and a data logger. The container was evacuated, filled from a compressed air main and then emptied to the atmosphere several times, with several sizes of thermocouple s (see discu ss ion of thermo couple measuring lag below ), and with inlet and outlet flow restrictors in some cases, to s low the flow Figures 2 and 3 show the temperature measurement s for typical filling and emptying experiments. The measured maximum and mini mum temperature s are far from those computed above. We can estimate the heat transfer coefficient betwe e n the Pressure and temperature indicators connected to data logger 0 Surface mount thennocouple with external insulation 5.7 Gallon (0.027 m3) propane container To atmosphere Quarter turn ball valves From compressed air supply line Figure 1. Flow and instrumentation diagram of the experi mental apparatus. The data Jogg er records t empera tures and pressures at 1 / 3-second intervals. 28 80 70 0 60 0 cu 50 ... .;! 40 co ... C1l a. 30 E 20 10 0 0 20 40 60 80 100 Time from opening of valve, s Figure 2. Temperature-time plot for filling the container from the lab oratory compressed air main at 738 kPa. The pressure reached t hat value in = 5.3 s The temperature rose from 22 to 74 5 G, reaching the peak at 6 s The air inlet va l ve was closed at 5 s and the pressure allowed to decrease as the air in the container cooled. The metal wa ll s of the tank rose in temperature from 26 to 29 3 G. After the peak temperature the gas cooled slow l y toward room temperature. For the adiabatic assumption the peak temperature, calcu lat ed in the text was 140 G 30 25 20 (.) 0 ~15 ::, 10 iii ... C1l 5 a. E C1l 0 I-5 -10 0 20 40 60 80 100 120 Time from opening of valve, s Figure 3. Temperature-time plot for emptying the con tainer to the atmosphere ( at 86 kPa) from an initial pres sure of 641 kPa. The pressure reached atmospheric in = 1. 7 s. The temperature fell from 29 to 4 C, reaching its minimum at 9 7 s. The metal walls of the tank fell in temperature from 28 8 to 27 5 G After the minimum t e perature, the gas wa rm ed slowly toward room temperatur e with the ex it va l ve open For the adiabatic assumption the minimum t emperat ure ca l culated in the text, was -103 C Chemical En g in ee rin g Educati o n

PAGE 31

air in the containers and the container wall s if we as s ume that the thermocouple lag was s mall compared to 100 s and that in each process the ga s underwent a s tep change in t emperat u re and was then coo l ed or h eated by simple con vective h eat tran sfer wit h a co n sta nt-t emperat ur e container wall. With these ass umpti ons the gas temperature is given by T T wal l ( hA 1 = ex p lmC t ) T af t e r s t e p T wa ll (2 1 ) suggesting that a plot of the In of the temperature ratio at the left ofEq (21) vs t s h o uld form a s traight line from which h could b e estimated. Figures 4 and 5 show suc h plots ; from the s l opes, o n e may infer the values of the heat-transfer coefficie nt s The choice ofT ar i ers t ep is arbitrary, made to force the straig ht lines through 1.0 on the ordinate. Changing these values moves the curves up a nd down witho ut changing their slopes. (.) 0 (.) en N 0 en 0.1 N (.) 0 IlO 0 Straight line slope = 0 0344/s 0.01 0 20 40 60 BO 100 Time from open in g of valve, s Figure 4. Replot of temperatur e -tim e data from Figure 2 in the form suggested by Eq (21) (.) 0 II) r--: IN I(.) 0 II) r--: IN (.) 0 II) '7 0.1 0 Straight line slope = 0.0150/s 20 40 60 80 100 120 Time from opening of valve, s Figure 5. Re plot of temperatur e -tim e data from Fi g ure 3 in th e form suggested by Eq. (21) Wint e r /999 From the s e te s t s, one can es timate the heat-transfer coeffi cient s. For example for the filling te s t ( Figures 2 and 4), with C = C v =2.5R ( I66 g )(2 5)( 8.3 14 l 1 J( 0 0 3 44 1 l mol K l s ) mol W h ------~---=10.49-0 390m 2 29 g m 2 K and for emptying (Figure s 3 and 5 ), with C = C P = 3.5 R w h=0 912 m K The surprising l y large difference is largely due to the difference in air densities (due to differing pressures) for the two cases. The heat-transfer coefficie nt s, esti mated from a flat-wall natural convection correlation [ 5 l Nu =0 0210(Gr Pr 4 (22) using average values of the ga s density and the temperature differences are 11.9 and 1.7 W/(m 2 K) The first is close to the value calculated from the cooling curve a nd the seco nd about twice the value calculated from the warming c ur ve If we as s ume that the processes were practically two-step, with a quick adiabatic proce s s followed by a slow transfer of heat to or from the wall s of the container then b y energy balance we can comp ut e that [ mC ( T ad i aba t ic -T fi n a l) ] gas in co nt a in e r ti. T contai n e r wa ll s ( me ) co nt a in er wa ll s (23) For the filling experiment with C = C v the val ue are ti. T con t ainer wa ll s { 1 s 2g [( 2 5 ); ~ 314 ) ~ ] (140 c 29 c) } g =4.48 C [ ( 7 0 3 k g )(0.46) ~ ] kgK and the corresponding val u e for emptying, wit h C = Cp, is -0 94 C. The measured values are 3 3 and -l.3 C. THERMOCOUPLE LAG A major part of the difference between the steep parts of the temperature curves on Figures 2 and 4 a nd the val ue s calculated from adiabatic behavior or those comp ut ed b y Eqs (14) (16) (19) and ( 20), is due to thermocouple la g. Thi s i s normally characterizedl 6 1 in terms of the first-order time con s tant of the thermocouple. If we ass um e that the gas in the container undergoes a step increase or decrease in temperature, followed by a fir s t-order decay toward the s ur rounding temperature, and that the thermocouple responds a s described in Reference 6 then the eq u ation for the ther mocouple reading will be 29

PAGE 32

dTth e rm oco upl e {[ ( ) ( )] } ( ) dt = b Twall + T e nd of s t e p T wa ll exp -at T 24 where a and b are the time constants of the cooling or heating air in the container and of the thermocouple. The integrated form is Tth e rm oco upl e T wall a [ ( ) ( b )] -exp -at -exp t Tend o f s t ep -T wa ll b-a (25) with the peak value of T t h e rm oco upl e occurring at t= ln (b/a) ( 26 ) b-a with maximum value ( \ ( b I l Tthermocoup l e -Tw a ll ) =(~ Jb-a ) (27) Tend ofs t e p -Twall maximum Figure 6 shows a comparison of the reported temperature s for two sizes of thermocouple for identical filling experi m e nts As expected, the smaller thermocouple report s a higher peak temperature and reaches it soo ner. Table 1 shows the comparison of the time -topeak reported T and the esti mated value of T e nd of s t ep calculated from Eqs. (26) and (27). The computed and observed times-to-peak reported temperature are in good agreement, but the computed maximum temperatures are far too high indicating that after a few seconds, the two-fir st-o rder-proce ssesinse ries model works well, but its extrapolation to t=O doe s not. If the contained gas temperatures were not changing rapidly due to heat tran sfer, this thermocouple lag would pose no problem ESTIMATING MAXIMUM TEMPERATURES To estimate the maximum temperatures from Eqs (14), (16), (19), and (20), the mass flow rates were computed by 90 80 (.) 0 7 0 ai ... ::, 6 0 cu 1.59 m m diameter TC cii a. 50 E Q) I40 30 2 0 0 20 40 60 80 100 Time from opening of valve, s Figure 6. Reported temperatures for two identical filling experiments with different size thermocouples The 0.51 mm diameter thermocouple reached its peak reading at 3.3 s; the 1.59-mm diameter thermocouple reached its at 7.67 s 30 TABLE 1 Applications of Eqs. (26) and (27) with the Time Constant of Cooling the Container Assumed = 0.034/s (See Figure 4) 0.51-mm diameter 1.59-mm diameter Thermocouple Thermocouple R e port e d time constant ,1 61 s 1/4 = 0.25 1 /9 = 1.11 Ob se rv e d tim e to peak report ed T ,s 3.3 7 67 Calcu l ated time to peak T by Eq. (26), s 3.2 9.3 Right-hand side of Eq (27) 0 .027 0.099 Calc ul ated T ,ndor,iep from Eq. (2 7 ), C 2000 500 20 (.) 0 "O Equati o n 14 -----------;;;/_ ... / Q) 15 ::, 0 ca 0 t10 ._g ro ..a ro 5 "O ro lo I .J l 0 ..,.,,--,.. ... ......... ... _,.._,;.,,,..,.~' .,,..--Eq. 16, numerical integration 0.5 1 5 2 2.5 3 3.5 Time from opening of valve, s 4 Figure 7. Computed departure of th e tank temperature for tank filling from the calculated adiabatic temperature for the same conditions Here m initia/ =3 5 g. 35 (.) 0 30 0 :;: ro 25 ..a -~ "O 20 ro I15 "O Q) 10 -;;; ::, 0 5 ro (J I0 -5 0 0.5 1 1.5 Time from opening of valve, s Figure 8. Comput e d d e partur e of the tank temperatur e for tank emptying from th e ca l c ulated adiabatic temp e ratur e for the same conditions. Chemi c al Engineering Edu c ation

PAGE 33

differencing the calculated mass in the container at each 1/3seco nd measuring interval. This has the drawback that it relie s on the thermocouple reading, which i s known to l ag the true temperatur e With thi s caveat, the flow rate s corre s ponding to Figures 2 and 3 we re es tim ated as m = I OOi-28.6 t ; 0'.,t'.,3 5s s s a nd rn =2391171 {r; 0 '., t '., 1.5s s s For both filling and emptying, the heat-tr ansfer coefficient was estimated from Eq (22) and assumed co n s tant. At the average density and temperature difference between wa ll and gas, the estimated values were 8.9 and 9 0 W/(m 2 K ) for filling and emptying. Figures 7 and 8 s how the calculated departure s from the corresponding adiabatic so lution s (E q s. 7 and 8). From them we see that in both cases the major departure occur s a t the boundar y of the process at which the ma ss of air in the container i s le as t (the s tart for filling, the finish for emptying); thi s i s the natural co n se quence of di viding a calculated he a t flow that i s assumed independent of the ma ss b y a s mall ma ss rather than a l arge one. Becau se of th e stre ngth of the assumptions a nd th e th er mocouple la g problem the se figures s hould be see n as order-of-magni tude Nonetheless the y mak e clear that eve n for the se fas t proce sses, with plau s ible heat-tran sfer coeff icient s, the cal culated temperature s are substantially different from the com puted adiabatic temperatures. APPLICATION TO LARGER T ANK S Equations ( 14 ) and ( 19 ) s how that for the constant ma ss flow-rate in or out si mplification and for T in = T ini1ia1 = T urr o undin gs the T-t behavior depend s only on th e two dimen s ionle ss groups m/m 0 and a/ m Thu s the experimental results s hown here should also be observed in any container for which the se parameter s have the same values. Th e fir s t can take on a ny value, but the seco nd i s a function of container geom e tr y. If the initial conditions in th e container a nd the reser vo ir conditions for filling and emptying are th e sa me for two tanks then m should be proportional to the cross-sectional area of the inlet pipe. If the heat-transfer coefficient does not change then a is proportional to the s urface area of the container. If the ratio of the inlet pipe cross-sectional area to the s urface area of the container doe s not change, then the seco nd of the se dimen s ionle ss gro up s s hould also not change (or change much with changes in the heat-tran sfe r coeffi cient). Thus while the experiments reported here were all performed in a 0 027 m 3 container, the y s hould be directl y applicable to larger tanks with the same dimen s ion ratios. CONCLUSIONS The adiabatic, ideal-gas container filling and emptying so lutions have a traditional place in thermodynamics text Winter 1 999 book s becau se they are th e s imple s t unsteady-state, varying inventory problem s that can be solved in closed form. In practice, it i s impossible to conduct the se processes without h ea t-tran sfer -produ ci n g gas temperatures far different from the adiabatic flow so lution s, mo s tl y be ca u se while the amount of he at tran sfe rr e d is s m a ll the ma ss of gas into which it i s transferred i s also sma ll. The effects of s uch heat transfer on th e temp era ture tim e b e ha v ior of s uch processes can be esti mated with at lea s t order-of-magnitude accuracy Thermocouple l ag a dd s to the effect of the heat transfer further increasing th e difference between the observed tem perature extre me s a nd the va lue s calculated for adiabatic filling and emptying. A differ e nt version of thi s problem and experiment ap p eare d while thi s p a per was in pre ssY 1 NOMENCLATUR E A area a tim e co n s t a nt of coo lin g or h ea tin g air in the container a,b constants in data-fitting e qu at i o n s b tim e co n s t a nt of thermocouple C heat capacity C P heat capac it y at co n sta nt pre ss ur e C v heat capacity a t co n sta nt vo lum e Gr Gra shof Number h he at-tra n sfe r coefficient h s pecific e nth a lp y k C/C v m ma ss m m ass flow rate N u Nusselt numb er P pressure Pr Prandtl numb e r Q heat quantity R uni versa l gas co n s t a nt T t em p era tur e t tim e u specific int e rn a l ene r gy a h NCv REFERENCES 1. de Nevers, N ., Fluid Mechanics for C h emica l Engin ee ring 2nd e d McGr aw -Hill, Ne w York, NY 109 ( 1991 ) 2. Wisniak J. "Discharge of Vessels: Thermodynamic Analy sis," J. Chem. Ed. 74 301 ( 1997 ) 3. R ya n J T ., R.K. Wood and P J Crickmore, An Inexpensive and Quick Fluid Mechanics Experim ent," Chem. Eng Ed ., 27 1 40 ( 199 3) 4. Cutler, S ., J Feichko, an d R. Waldron, "Bot tle Filling and Emptying E xperiments," Senior Process Engineering Labo ratory R eports, Dep art ment of C hemical and Fuels Engi n eer ing University of Utah ( 1997 ) 5. Kreith F ., Pr inciples o f H eat Transfer, 3rd ed., IEP New York, 395 ( 1973 ) 6. Omega Engine eri ng Co ., Catalog 29 Pag e Z-43 Stamford CT ( 1995 ) 7. Forr ester S E ., and G M. Evans The Importance of Sys tem Selection on Compressible Flow Analysis: Filling Ves se l s," C h em En g. Ed. 32, 3 08 ( 1998 ) 0 3 1

PAGE 34

Random Thoughts .. FAQS RICHARD M. FELDER, REBECCA BRENT North Carolina State University Raleigh, NC 27695 A t the teaching workshops we give, we propose a variety of instructional methods that deviate from traditional teaching practice. We recommend, for example, that instructors break up their lectures at frequent intervals with brief individual or small group exercises. We suggest using formal cooperative learning, in which students work on assignments in instructor-formed teams under con ditions structured to assure individual accountability for all of the assigned material. We caution against giving tests that only the best students in the class have time to finish, and we argue strongly against curving grades. Predictably, critical questions are raised about these rec ommendations and others we offer. In a series of columns beginning with this one, we want to review some of the most frequently asked questions (FAQs) and our responses. We have two reasons for doing this First, the suggestions we offer at the workshops are far from unique with us: they are being made with increasing frequency by educational re searchers, national study commissions, employers of engi neering graduates, and accrediting bodies like ABET. If you have not already been exposed to them, you almost certainly R i c h a rd M Fe lde r is Hoechst Celanese Profes sor of Chemical Engineering at North Carolina State University He received his BChE from City College of CUNY and his PhD from Princeton He has presented courses on chemical engineering principles, reactor design process optimization and effective teaching to various American and foreign industries and institutions. He is coauthor of the text Elementary Principles of Chemical Pro cesses (Wiley 1986) Rebec ca Brent is an education consultant spe cializing in faculty development for effective uni versity teaching classroom and computer-based simulations in teacher education, and K 12 staff development in language arts and classroom management She co-directs the SUCCEED Coa lition faculty development program and has pub lished articles on a variety of topics including writing in undergraduate courses cooperative learning public school reform and effective uni versity teaching will be before long, and some of our responses may be helpful as you consider the ideas being advanced. Our sec ond objective is to offer those of you who are already using the new methods some answers to give your colleagues, administrators and students, who are certain to raise the same questions with you. Here, then is our top ten list of questions frequently asked at teaching workshops. 1. Is there an y real evidence that these methods work? 2. I have a lot of material to get through in a semester. Can I use these methods and still have time to cover m y s y llabu s? 3. I teach a class of 175 s tudents in a fixed-seat auditorium Will these methods work in large classes ? 4. I'm teaching a course by distance education How can I get students active when I'm not in the same room with them ? 5. I tried putting students to work in groups but some of them hated it and one complained to my depart ment head. Why are some students so hostile to cooperative learning and what am I s upposed to do about the hostility ? 6 Many of my students are (a) unmotivated, (b) self centered (c) apathetic (d) lazy (e) materialistic, (f) unprepared, (g) unable to do high school math, (h) unable to write, (i) unable to read (j) spoiled rotten (Pick any subset.) How can yo u teach people who don t hav e th e right background or the willingness to work or even the Copyright C hE Divi s ion of ASEE 19 99 32 Chemical Engineering Edu c ation

PAGE 35

desire to learn? 7. Engineers constantly have to face deadlines. What's wrong with giving tests that only the best students have time to finish? 8. What difference does it make if my test averages are in the 50's, since I'm going to curve in the end? 9. My department head says that we can't count teaching too much in promotion and tenure deci sions because we don't know how to evaluate teaching. I s there a meaningful way to evaluate teaching? 10. The people who go to teaching workshops are mostly excellent teacher s-the ones who most need to change wouldn't go to a teaching workshop at gunpoint. How can I persuade my traditional colleagues to do some of the nontraditional things you' re recommending ? The workshop participant s who ask these questions are doing what they have been trained to do as scie ntist s and engineers and educated people, which is to ask for hard evidence before changing the way they 've always done things. We applaud them for asking. In this column we'll offer an answer to the first question, and s ubsequent columns will deal with the others. Q: Is there any REAL evidence that these nontradi tional methods work? A: Tons of it. Cognitive and educational scientists have learned a great deal about learning in recent years. The near-unanimous consensus is that we learn mainly by doing things and re flecting on the outcomes, taking in relatively little of what we just see and hear (e.g., in lectures) and retaining even les s. Countless studies have compared the academic perfor mance and attitudes of students taught using active and cooperative methods with the performance and attitudes of st udents taught more traditionally. The evidence for the ef fectiveness of the nontraditional methods is overwhelm ing (Specific references will be cited shortly.) Unfortunately, most professors have never seen a mono graph, paper or seminar on research into teaching and learn ing and would be hard pressed to name a journal or confer ence where such research might s how up When the "Prove it! card is played at our workshops (and even if it isn't) we therefore urge our questioners not to take our word for anything we say but to approach the matter scientifically and check the literature. We point them to a ser ies of three papers in Chemical Engineering Education written by Jim Haile ,C 11 which collectively provide the best summary we've ever seen of what cognitive scie nce has discovered about the learning process and the implications of this knowledge for teaching. We introduce them to the classic Teaching Tips, 121 in which Wilbert McKeachie offers an abundance of practi cal suggestions about every aspect of college teaching along with citations of the research that backs up the suggestions. We tell them about What Matters in College, l 3 4 l Alexander Astin's monumental s tudy of nearly 25,000 students at over 300 institution s that powerfully demonstrates the deficien cies of the traditional instructional model. We cite refer ences on cooperative learning (e.g., Johnson Johnson, and Smith l 51 ) that in turn cite hundreds of research studies attest ing to the effectiveness of thi s approach, and we discu ss the results of a longitudinal st udy one of us carried out of the effectiveness of cooperative learning in chemical engineer ing education. l 6 71 Brow se these references," we urge Th en decide whether the research and the methods we're advocat ing are worthy of serious consideration." More to come. REFERENCES 1. Haile J.M., Toward Technical Understanding'' : (a) "Part 1. Brain Structure and Function," Chem Engr. Education 31 (3), 152-157 ( 1997 ); ( b) Part 2. Elementary Levels, Chem Engr. Education, 31(4 ), 214-219 ( 1997 ); (c) Part 3. Advanced Levels ," Chem. Engr. Education, 32 (1), 30-39 ( 1998) 2. McKeachie, W J.,Teaching Tips : Strategies, R esearch, and Theory for College and University T eachers, 10th Edition Boston Houghton Mifflin Co. ( 1999 ) 3 Astin, A.W., What Matters in College San Francisco Jos sey Bass (1993) 4. Felder R.M., ''What Matters in College," Chemical Engi neering Education, 27 ( 4 ), 194-195 (1993) (View at< http:/ /www2 .n csu edu/unity/locke rs /use r s/fl (elder I public I Columns I Astin.html > .) 5. Johnson D.W. R.T Johnson, and KA. Smith, Active L earn ing: Cooperation in the College Classroom, 2nd Edition. Edina, MN, Interaction Book Co. ( 1998 ) 6. Felder R.M ., A Longitudinal Study of Engineering Stu dent Performance and Retention IV. Instructional Methods and Student Respons es to Them," J. Engr. Education, 84 ( 4 ), 361-367 ( 1995 ) (View at < http: I I www2. ncsu.edu I unity I lockers I users If I (elder I publi c I Pap ers I long4 .html > .) 7 Felder, R.M ., G.N. Felder, and E.J Dietz "A Longitudinal Study of Engineering Student Performance and Retention : V. Comparisons with Traditionally-Taught Students," J. Engr. Education, 87 (4), 469-480 (1998) (View at < http: I I www2. ncsu.edu I unity I lockers I users I f I (elder I public I Paper s I long5.html >.) 0 All of the Random Thoughts columns are now available on the World Wide Web at http:/ /www2 ncsu.edu/effecti ve_teaching/ and at http://che.ufl.edu/ ~cee/ Wint e r 1999 33

PAGE 36

.t3_..ijllllil._1a_b_o r,_a t_o r .:. y _______ __,) EVALUATION OF COMPUTER-SIMULATION EXPERIMENTS I N A SENIOR-LEVEL CAPSTONE ChE COURSE SCOTT R. WHITE, GEORGE M. BODNER Purdue University West Lafayette IN 47907-1393 I n 1986 the School of Chemical Engineering at Purdu e University began a revi s ion of its se nior-level capstone laboratory courses, including the development of a se ries of computer-simulation experiments de sc ribed else where. [1l For each computer si mulation the students are given a budget (i.e., $35,000) that is the amount they can spend on experimental run s, wages, and consultation fees. The computer also keep s track of the "vi rtual time the students use for each run and charges extra for work that has to be done on weekends. This paper describes the results of an evaluation of the effect of using these simulations. It is ba se d on three as sumptions: Wh en we c hang e what we teach, or how we teach, we change what the students l earn. A systematic eva luation should be done whenever major cha n ges are made in an established cur ri cu lum Systematic evaluations should look behind the facade of answers to the questi ons, "Do the students like it?" toward deeper questions such as "What w ill students l earn that they were not learning before?" and If we cou ld provide students w ith a voice to express their opinions and concerns, what changes would they recommend ? The basic research question behind this study was: How do the students' experiences with computer simulations com pare with their experiences with traditional laboratory ex periments ?" Corollary research questions included: "What did the students perceive as a valuable experience in both laboratory formats?"; How did the students' decision-mak ing processes and other group-related interactions differ be tween the two formats?"; "What do the students believe makes the computer-simulation experiment a legitimate exercise to include in the chemical engineering curriculum?" DEVE L OPING EVALUA TI ON METHOD S The study was based on a collaboration between members of a chemical education research groupl 8 1 and faculty and s taff from the School of Chemical Engineering who had developed and implemented the computer s imulation s. We began by scrutinizing a list of questions generated by Profes sor R. G. Squires and Dr. S. Jayakumar for use in a quantita tive study of student attitudes toward the simulations. Some of the que s tion s were retained and others were modified to make them either less complex or les s leading. The result of this review was a 15-item five-point Likertsca le ques tionnaire that included space for students to write additional comments and/or suggestions. The questionnaire was given to the students after they h a d completed both a traditional S co tt R. W h ite is a PhD graduate student in Science Education at Purdue University He re ceived his BS in Chemistry and Secondary Edu cation Certification in 1992 from Harding Univer sity. He received his MS in Chemistry from Purdue University in 1996 with G.M. Bodner. His research interests are in teaching and learning in science and curriculum reform. G eo r g e M. Bod ner is Professor of Chemistry and Education at Purdue. He received his BS in Chemistry from the State University at New York, Buffalo (1969) and his PhD in inorganic and or ganic chemistry from Indiana University (1972). His research interests are learning theory over coming barriers to curriculum reform and under standing the conditions for appropriate use of technology in teaching and learning chemistry. Copyright ChE Division of ASEE 1999 34 Chemical Engineering Education

PAGE 37

experiment and a computer-simulation exTABLEl perim e nt. Re s ult s of thi s s urvey for s tuSurvey Percentage Responses* dent s from two se me s ter s a re summarized in Table 1. Item Statement At:!_ee Neutral Disagree Avg. I. I like using computer simulations. 91 3 6 4.2 The authors developed a qualitative component of the eva luation C 9 1 based on s truc2. When using the computer simulation, I worried that my 22 16 62 2.4 tured interviews with individual students or data would be lost or that the program would fail. with groups of s tudent s; observations and 3. Time and budget constraints made the computer 86 3 11 4.2 experiments more reali stic. field note s collected in the laboratorie s, writ4 The conventional lab experiments worked bener that the 5 ll 84 1.9 ten comments from the s urvey s de sc ribed computer experiments. above, a nd interaction s w ith the s tudent s 5 The video tour of the plant added linle to the value of the 30 30 40 3.0 in th e l abs. A s those fa miliar with qualicomputer experiment. tati ve t ec hnique s mi g ht expect, the quali6. It was easy to learn and operate the computer sim ulation. 95 0 5 4.3 tative component provided the richest 7. Computer-simulation experiments intimidate me 6 8 86 1.8 source of data for thi s s tudy 8 The speed of data acquisition in the computer experiments 6 5 89 1.7 Collection of qualitativ e data began with makes me uneasy. 9 Computer experiments allowed me to focus on the prin82 10 8 4.2 the researcher si ttin g in a comer of the traciples to be learned rather than on the details of ditional l a b takin g field notes as he oboperating a particular piece of equipment. serve d w h at was h a ppening The s tudent s IO Computer experiments are more interesting that conven40 38 22 3.3 would frequently s tart co nver s ation s with tional experiments. th e researcher, aski n g w hat he was doin g 11. One di sa dvantage of computer experiments i s that I do 60 24 16 3.6 th ere and relating w h at they thought about not gain experinece with the real plant equipment. th e expe riment th ey were doing or what 12. I would like to see more computer-simulated experiments 73 14 13 3.8 in the chemical engineering curric ulum the y thou g ht or had heard about the com13. I would rather work on a computer simulation because it 27 27 46 2 .8 puters imulation experiments. Frequently the is l ess h azardous than a conventional experiment. st udent s would ph ys i ca ll y point out thing s 14 Conventional experiments give me a bener sense of the 54 22 24 3.6 that were working or not working with their kinds of problems likely to be encountered in industry. tradition a l experiments, which helped the 15. My group cooperated better during the conventional lab 13 43 44 2.6 researcher ga in an under s tanding of the exexperiments periment s the st ud e nt s were performin g. 16. The design problem imposed by the computer simulation 30 22 49 2.8 is not as challenging as those encountered during convenAs th ese interactions co ntinued the retional experiments. searc h er found it u sefu l to sw itch from the 17. A higher percentage of our time was spent planning the 86 3 ll 4.1 role of a n objective observer sitting in a design of computer experiments comer of the room taking notes to that of a 18. The computer s imulation allowed me to st ud y problems 78 19 3 4.0 participant-ob se rver li s tening to and talk that are more complex and realistic than the conventional ing with st udent s while they worked The experiments. s tud e nt s a l so see med more comfortable with 19. Computer simulations allow me to make more effective 95 5 0 4.4 use of time by reducing the amount of time needed to thi s a pp roac h. Th e result was an environrun experiments. ment in w hich a goo d rapport was devel20. The conventional lab experiments were easier to learn 6 8 86 1.9 oped b etwee n the researcher and the s tuand operate than the computer experiment. dent s pri or to the s tructured interview s Thi s 21. Computer simulations are a good way to learn new 79 22 0 4.1 approach also pro vi ded the r es earcher with processes and concepts. a set of ex perience s that allowed him to 22 Computer simulations work bener than the conventional 60 35 5 3 8 prod the s tudent s' memories during the experiments subsequent interview s when they were 23 Computer simulations are more likely to work" than 81 14 5 4.1 conventional experiments. asked to co mpare the two different labo24 Overall, I think the present combination of computer 57 14 29 3.4 ratory formats simulations and conventional experiments is appropriate. Ob se r va tion s collected while s tudent s This table summ arizes the results of two semesters. We combined the "Strongly were working in th e co mputer lab did not Agree and Agree" responses into one category-"Agree "Strongly Disagree and prov e u sefu l becau se mo s t of the deci s ion"Disagree" have been combined into the Disagree category. The "Undecided" makin g proc ess h a d already been accomresponses are indicated as "Neutral (N ) pli s hed durin g group meetings before the s tudent s came to the lab and the s tudent s Wint e r 1 999 35

PAGE 38

were less likely to involve the researcher in their activities while they worked with the computer. Insight into these group meetings and the interactions between members of the group was provided by the structured interviews, however. The structured interviews were the core of the qualitative evaluation methods. The researcher developed a list of ques tions that he wished to cover during the interviews covering many of the same topics as the Likert-scale surveys so that the researcher could triangulate his conclusions from differ ent data sources.[1 1 Using the structured topic list pro duced interviews that followed a s imilar pattern but the students had ample opportunity to bring up any subject they felt appropriate. The interviews were recorded, transcribed, and then ana lyzed using the method of inductive analysis.[ 111 The analysis consisted of reading the transcripts multiple times and con densing the students comments to common and uncommon categories by literally cutting and pasting together similar comments obtained in different interviews. QUANTITATIVE RESULTS AN D DISCUSSION The results of the Likert-scale survey indicated that the students liked using the simulations (91 % ; Ql) ; found the simulations easy to learn and operate (95 %; Q6); reported that the computer simulations did not intimidate them (86 % ; Q7); would like to see more of them (73 % ; Ql2); believed that the computer simulations allowed them to study more complex and realistic problems (78 % ; Q18) ; valued the bud getary constraints included with the simulations which made the simulations more realistic (86 % ; Q3); and believed that they spent a higher percentage of their time planning the design of the computer simulation (86 %; QI 7), which sug gests that the simulations provide the students with an expe rience that is different from the traditional lab. The students liked the simulations for a variety of reasons, including the fact that they were more likely to work than the traditional experiments (81 % ; Q23) thus giving the students reason able and workable data. The computer simulations were very different from tradi tional labs because of the speed with which data could be acquired. This did not bother the students or make them feel uneasy about the computer experiment (89 %; Q8) In fact they felt that this made more efficient use of their time (95 % ; Q19) The students felt the simulations allowed them to focus on the principles involved in an experiment (82 % ; Q9) and therefore were a good way to learn new processes and concept s (79 % ; Q21). But a majority (54 %; Q14) of the students felt that the traditional experiments gave them a better sense of the problems likely to be encountered in industry. Thus it is not surprising that a majority (57%; Q24) felt that the present combination of computer and simulation experiments was appropriate. The quantitative results produced a sense of conflict, or 36 dualism in the students opinions They simultaneously be lieved the computer simulations are a good instructional technique that helped them better focus on the principles they were expected to apply and at the same time that the traditional experiments gave them a better sense of the problems they might encounter in industry The s ource of this dualism cannot be extracted from the results of a Likert-scale survey but they can be obtained by triangu lating this data source with the results of qualitative research techniques. As we will see, the students s imultaneously regarded the computer simulation s as both good and "bad ." They are good because they allowed students to tackle more complex problems in which they were compelled to proceed with realistic budgetary and time constraints and because these experiments were more likely to "work," providing the stu dents with data that allowed them to complete a realistic scale-up The simulations are "bad" because they are not real ; they cannot fail in the same way a traditional experi ment would fail. Even though the students tended to value the ability to focu s on important conceptual engineering issues in the simulation experiments, they recognized that this "ability" has little to do with the world in which they actually live. QUALITATIV E RESUL TS AND DISCUSSI ON Twelve students were interviewed after they had com pleted one experiment of each type The theoretical frame work for this portion of the study falls within the domain of hermeneutics[ 121 in the sense that we are trying to give stu dents the opportunity to be heard to have a voice," through interpretations of the meanings of their statements and ac tions. The interviews were used to probe more deeply into the students' experiences, opinions, and beliefs about tradi tional versus computer-simulation experiments; to probe how students constructed the knowledge they gained from doing the lab experiments; to examine how they perceived com puter-simulation experiments (e.g ., as just one long equation to be worked out with data generated by the computer or as a chance to do meaningful engineering work similar to that done in industry); to explore their opinions on whether the computer simulations were more (or less) realistic than tra ditional experiments; to discern whether the simulations re quire a particular teaching style from the instructor; and to determine the aspects of the computer simulation that make it more (or less) difficult than the traditional experiments In some ways the students felt the computer simulation was more realistic, and perhaps more difficult, than their other experiments. (In the following vignettes "I" stands for the interviewer and the names are nicknames given to pro tect the students' identities.) I: You were talking about the computer simulation b e ing more in-depth. What did y ou mean b y that ? Ch e mi c al Engin ee rin g E du c ation

PAGE 39

Andy: Inst e ad o f d e alin g with th e unit y o u d e alt with m o r e o f what y ou d deal with in th e r e al plant .. the c omput e r int e fa ce d y ou to multipl e ty p e s of e quipm e nt and m o r e "r e al e quipm e nt than yo u wo uld u se in indu s t ry rath e r than ju s t th e s mall g lass tub e that we u se d f or th e c ati o n exc han ge And I thought that w as b e tt e r b ec aus e y ou ge t more of a full v i ew of th e operation rath e r than ju s t o n e s mall aspe c t of it Jody: In addition t o that to o we had a bud ge t that we had to follow Whi c h i s go nna b e tru e in r e al lif e on ce we g raduat e and do w hat we n e ed t o do t o ge t data and s tuff lik e that. The time and budgetary constraints imposed on the com puter-simulation experiments had the tendency to change the st udent s' decision-making process by forcing them to reflect on their decisions before taking actions as illus trated by comments made by Adam and Don who were in separate gro up s A dam: It made it m o r e r e alwo rld I g u ess. B efo r e, o n th e oth e r ex perim e nts if y ou w ant e d t o a s k th e profe sso r a qu e tion we d just go up and ask; e ven if it was just a stupid question Now if we wanted to talk to th e professor it w ould c o s t us $ 500 for a c onsultin g f ee It mad e y ou stop and think about it instead of just runnin g up and a s king th e prof e s s or when y ou could have figured it o ut y our se lf if y ou d just ha ve thou g ht about it. Don: It w as g ood t o ha ve a bud ge t. If th e r e wa s n o a c tual planning involv e d, with no bud g et w e wo uld just hav e run it for hours and hours and had sta c k s of paper for r e sult s. W e wo uldn t have th o u g ht about w hat we we r e doin g. These comments are echoed by the results of the survey which showed that the majority of the students felt that use of budget and time constraints made the simulation more realistic than the traditional laboratory experiments. Darrin and Laura found the realism introduced by the budget/time constraints intimidating I: L e t me ask yo u about th e co mput e r s imulati o n Wh a t di d y ou think of it w h e n y ou fir s t s a w i t ? Darrin: Hehl Intimidating. I: How ? Darrin: W e ll, ev en thou g h w e s at thr o u g h a wh o l e l ec tur e, I f e lt that I r e all y didn t kno w w h e r e t o b eg in and r e all y I w a s r e ad y to ge t anoth e r apparatu s ex p e rim e ntal probl e m W i th thi s [ c omp s imulati o n} I had n o id e a h ow t o s tart I w a s afraid that I was go in g to mak e a mistak e .. And plus th e r e's thi s thing that if y ou ask a qu es tion it wo uld c o s t yo u lik e $ 500 o r s o m e thin g. [ c on s ultati o n f ee ] S o yo u r e kinda t e nt a ti ve. Laura provided insight into why her group felt intimidated by the computer experiment when s he re s ponded to a que s tion that asked for her impre s sion of the computer simula tion Laura: I was s c ar e d b ec aus e it w a s n t lik e an y of o ur o th e r lab s w e r e, e v e n if yo u lik e t o t a ll y ge t b a d d a ta ... yo u d o n t ha ve an y thin g t o l ose. Y o u c an s till w ri te up y our r e p o rt and sa y that y our re s ults ar e no g ood But on this lab [ c omput e r simulation] y ou hav e to find y our c on s tant s. Wint e r 19 99 The problem was simple-there was no place for the stu dent s to hide. They could not gloss over or fudge poor data collected during the computer experiment the way they said they could when di s cu s sing traditional experiments Darrin and Laura s comments are not representative of the perceptions of the group of students who comp leted the computer experiment during the evaluation but their com ments rai s e an important issue in evaluation. Historically, evaluations of curriculum-reform projects have been based on what we have called a sports-mentality approac h _r 13 J Statistical techniques such as at-test on the mean scores of some measure of performance of studen t s in experi mental versus control sections of the co ur se are u se d to answer the questions "Is the new curriculum b ette r or worse than the old curriculum?" Darrin and Laura's comments remind u s that any subs tan tive change in curriculum will have both positive and nega tive effects. Some students will benefit, but others will be hurt. Evaluative studies s uch as this one, a llow one to search for both effects and then probe what additio nal c h anges could be made to maximize the positive effect and minimize the negative effect. Darrin and Laura's interviews identified another source of difference s in student s' perceptions of the computer sim ula tions-the amount of success the students felt they had en joyed In general, the students who were interviewed felt that they had enjoyed success with the computer simulation Darrin and Laura s group did not share this perspective, however, as illustrated by the following comme nt s: / : What w as the c omput e r s imulation supposed to do and w hat did it r e all y do a s y ou look back on it now ? What was it s upposed to r e pr e s e nt ? D a rrin : I think that it w a s s uppo se d to r e pr e s e nt a better w ay o f so l v in g a lar ge pr o bl e m that we c ould n e v e r have so l ve d o n a lab o rat ory sca l e With th e amount of trials w e ran it wa s supp os ed t o d e monstrate how much work y ou c ould g et d o n e, ... ho w man y trials y ou could ge t done on the co mput e r. But w hat it turn e d out to be wa s just trial and e rr o r. Laura : I think that w hat th e .. s imulation w as doing was to s h ow us h o w we c an u se a co mputer to s imulate something a nd th e n t o optimi ze co nditi o n s And then appl y them to an ac tual plant o r what eve r And .. I g ue s s it did it, ... I don t kn ow! I d o n t r e all y kn ow b e caus e I still don t r e ally under s tand h ow o ur v alu es co rr e lat e d t o th e a c tual runnin g of th e s imul a tion and th e runnin g o f pilot plant .. I don t r e all y think that I l e arn e d an y thin g fr o m it. I just l e arn e d to manipu lat e w hat we we r e t ry in g t o d o I m still a littl e uncl e ar on so m e thin gs. The interviews provided useful information abo ut the st dent s' perception of the role of the comp ut er in their re s pon s e to the question Which experiment gave you the b est experience ? Dallas: [Th e comput e r ex p e rim e nt] was the best. Granted, 3 7

PAGE 40

that th ey wanted t o try to g iv e us some sort of r e al-lif e simula tion what it s like in r e al indust ry ... W e w e r e a c tuall y abl e to ge t our numbers and do our scale-ups and do our actual e ngine e rin g work w ithout tr y ing to mess around trying to g e t something to work Or tr y in g to g e t s om e data or makin g up data .. w e w e r e a c tuall y able to do e ngin ee ring. Don: [ In th e c omputer e xperiment] y ou c ould kind of e stimate sort of wh e r e .. what would happ e n under c ertain c ir c um stan ce s Whereas in th e exp e rim e ntal part of th e r e gular lab, y ou would g et such confusing r e sults It was so diffi c ult to try to extrapolat e that onto an y lar g e s c ale. We basi c all y said We'll just hav e to throw this out and we ll se e what some body e lse did or make up something These comments reflect the common perception among students that the simulations were more likely to work than the traditional experiments; that they could acquire feasible data from the computer simulation But other students questioned whether the computer simulation gave them the best experience Adam: The b e st experien c e was probabl y with the water cation e xc han g e just because w e did a lot more r e s e ar c h with that to l e arn how to get the ri g ht data and stuff like that The computer simulation was int e resting but pretty much all the data was ri g ht there in front of y ou. Th e computer simulation was pretty neat. But it was a lot of wasted time for three people to sit there and do it, because onl y on e person c ould get on th e c omput e r and run it Ruth and Tina provided further insight into the computer experiment. R uth: With th e c omputer e xp e rim e nt all we did wa s c alcula tions ... with the spreadsheet and stuff That s all w e did. The whole lab was on e big lon g c al c ulation. Tina: Th e w hol e lab was just findin g numbers y ou had to put in the simulation You had to wo rk through a bun c h of e qua tions. EFFECT OF THE PROFESSOR S TEACHING STYLE Unlike surveys which can only provide answers to ques tions that are explicitly stated interviews often provide data on topics or questions one might not have anticipated Con sider the role of the professor s teaching style for example. This topic was not covered in the survey, but the interviews showed that it had a significant impact on the students' experiences. It was clear from the interview data that the professor's hands-on" teaching style during the planning ses sions had a direct impact on the students perception of the computer simulations and the success of these simulations. CONCLUSIONS The results of this study suggest that it would be a mistake to ask which laboratory format is better" for students. They indicate that computer simulations and traditional experi ments have different roles in the curriculum because they emphasize different aspects of engineering and require both 38 different levels and types of expertise. Students who were frustrated with traditional lab equip ment seemed to enjoy "actually doing the engineering re quired to tackle the complex problems provided by the com puter simulation. They did not have to worry about making up data ... or seeing what someone else did when the traditional lab failed. For these student s, the computer s imu lations were more "realistic" than the traditional lab that gave results students described as .. I'd turn it in for a grade but I certainly wouldn't buy it!" For other students, the simulations were less "realistic because they cannot fail the way a traditional lab fails This study provided insight into the role of the environ ment in which computer simulations are implemented Our results clearly indicated that budgetary and time constraints played an important role in making the computer simula tions seem "realistic"-so realistic that a few students felt intimidated by this aspect of the simulations. This study also suggests that computer simulations by themselves are not magic bullets that provide instructional and pedagogical benefits for the students in the absence of a human interaction between the students and the instructor. They are best thought of in terms of being a tool for instruc tion rather than a replacement for the instructor. The authors hope that this study leads others to recognize the importance of asking the correct questions when evaluat ing curriculum reform projects as required by ABET and NSF, and the importance of collecting qualitative inter view data to both reinforce quantitative data collected in anonymous surveys and to provide a deeper understand ing of the effect of curriculum change s on students atti tudes and opinions. ACKNOWLEDGMENT S The authors wish to thank Professor Robert Squires and R. S. Jayakumar of the School of Chemical Engineering for their support throughout this evaluation study. REFERENCES 1. Squires R.G., G.V. Reklaitis N.C. Yeh J F Mosby I.A. Karimi and P.K. Anderson, Purdue-Industry Computer Simulation Modules : The Amoco Resid Hydrotreater Pro cess Chem. Eng. Ed ., 25 ( 2 ), 98 ( 1991 ) 2. Squires R.G. P.K. Anderson G.V. Reklaitis S. Jayakumar and D.S. Carmichael, Multimedia-Based Educational Ap plications of Computer Simulations of Chemical Engineer ing Processes ," Comp. Appls in En g Ed., 1 ( 1 ), 25 ( 1992 ) 3 Jayakumar, S R G. Squires G V. Reklaitis P.K. Anderson K.R. Graziani and B.C. Choi, The Use of Computer Simu lations in Engineering Capstone Courses: A Chemical Engi neering Example-The Mobil Catalytic Reforming Process Simulation ," Int e rnat. J. of En g Ed. 19 ( 3 ), 243 ( 1993 ) 4 Jayakumar S ., R.G Squires G V Reklaitis P K. Anderson and L.R. Partin, Purdue-Industry Computer Simulation Modules 2 : The Eastman Chemical Reactive Dis t illation Process Ch e m. Eng Ed 27 ( 2 ) 136 ( 1993 ) Ch e mi c al En g ineerin g Edu c ati o n

PAGE 41

5. Jayakumar, S., R.G Squires, G V. Reklaitis, P.K. Anderson and B.K. Di etrich "T h e Purdue-Dow Styrene-Butadiene Polymerization Simulation," J. of Eng. Ed., 84 ( 3 ) 271 ( 1995 ) 6. Jayakumar S. R.G. Squires, G V. Reklaitis, and K.S. Grassi, "Simu lating the Air Product s Cryogenic Hydrogen Reactive Cooling Proces s," Chem. Eng. Ed ., 29 (1 ) 26 ( 1995 ) 7. Squires, R.G., K. Kuriyan S Jayakumar, G V. Reklaitis, M. Evans, B. Morra to, and R. Gutwein, "The Procter and Gamble Decaffeination Project: A Multimedia Instruction Module," Comp Appls. in Eng. Ed., 4 ( 4 ) 269 ( 1996 ) 8 Bodner G.M., and J.D. Herron, Comp letin g the Program with a Division of Chemical Education ," J. of College Sci. Teaching 14 (3 ) 179 (1984 ) 9 ij book review ) .. _._..._ _______ Alternative Fuels by S. Lee Taylor & Francis, Bristol PA ; 485 pages, $83.95 ( 1996 ) Reviewed by Thoma s R. Marrero University of Missouri-Columbia Knowledge of chemical proce sses i s important in the de velopment of more environmentally friendly fuels because of the implementation of stricter constraints on energy utili zation by almost all nations The main objective of Alterna tive Fuels i s to comprehensively describe the scie nce and technology of various proce ss treatments for the clean use of coal and coal products, synthesis gas, alcohols, s hale oil crude, biomas s, and solid wastes. This ambitious objective is presented in eleven topical chapters that include current reference s to the state-of-the-art for each type of fuel pro cessing. Dr. Lee ha s successfully compiled a comprehensive collection of pertinent data and information that were scat tered throughout the literature. Alternative Fuels i s necessar ily lengthy but neat clear and consistent. It can well serve as a chemical engineering text and as a reference book for practicing engineers and researchers. Alternative Fuels, a book in the Applied Energy Technol ogy Series has 485 pages 172 one-line process diagram s graphs, and sketches, and 96 tables of data The index li sts 470 s ubject terms excluding numerou s s ub-term s. All these features succinctly provide a wealth of informative data that i s easily accessed by the reader. In addition each c hapter ha s a se t of problems ( u sefu l for st udent s), and a so lution manual is available. It has 586 reference s, with 250 of them pub li s hed si n ce 1990 Reference s to relatively inactive clean coal technologies such as oil shale, s hale oil, and tar sands, are primarily taken from studies published prior to 1980. The first chapter present s a global overview of energy production, consumption, and reserves for coal, gas, and oil. Additional data are presented for electric power generation Wimer 1 999 9 Patton, M.Q., Qualitative Evaluation and R esearch M et h ods 2nd ed., Sage, Newbury Park ( 1990 ) 10. Ibid pp 187-189 11. Goetz J.P ., and M.D. LeCompte Ethnography and Qualita tive Design in Educational Research, Academic Press San Diego, CA ( 1984 ) 12 Gadamer, H.G., Philosophical H ermeneutics, University of California Pre ss Berkeley, CA ( 1976 ) 13. Bodner, G.M., Overcoming the Sports Mentality Approach to Program Evaluation: Action Research as a Metaphor for Curriculum Evaluation ," paper presented at the National Association for Research in Science Teaching Meeting At lanta, GA, April ( 199 3 ) 0 from renewable energy sources: biomass, geot herm al, h droelectric solar, and wind. Thi s chapter summar i zes the global energy situation with 18 graphs and 13 tables. Chapter 2, in 60 page s, focuses on three major topics that could produce environmentally clean so lid and liquid fuels from proce sse d coal. First, the ba sic properties of coal are presented along with safety issues related to coal mining and enviro nmental issues related to coal combustion. In the sec ond part, many development s in coal technology are de scri bed for u se as a means to clean fuel. The third part of Chapter 2 present s environmental issues and regulations, particularly related to coal mining. Chapter 3 deal s with coal gasification, which includes a series of processes that convert coal containing C, H and 0 as well as impurities suc h as S and N into fuel and/or sy nthe sis gas. A total of 10 gasification processes are s ummarized in about 30 page s. Then the equations are presented for sto ichiometry, thermodynamics and reaction kinetics rela tive to coal gasification. Chapter 4 pre se nt s more than two dozen processes to develop alternative liquid fuels from coal by pyrolysis, di rect and indirect liquefaction and severa l other known es tablished chemical-process techniques. Thi s material does not include process economics. The ne xt topic, Chapter 5, is the development of gas fuel s from coal. This material s ummarize s pertinent advances in the DOE (m ultibillion dollar ) Clean Coal Technology Pro grams and an extensive di sc u ss ion of Integrated Gasification Combined Cycle (IGCC) systems. The IGCC technology eco nomic s are discus se d. Advantages and disadvantages of com bined-cycle sys tem s are delineated as potential sources of fuel Chapters 6, 7, and 8 are presentations of more establis h ed technologies (coa l slurry, oil s hale and tar sands) as poten tial so ur ces of fuel. The coal s lurry focuses on transportation a nd handleability, but no economics. De scr iptions of oil sha le and tar san d are focused around process diagrams and pertinent chemical reactions. --------------continu ed on page 83. 39

PAGE 42

.ta .. 6_._c_u_ _rr_i_c_u_l_u_m _____ __,) PROCESS ANALYSIS An Electronic Version GEORGE B DELANCEY Stevens Institute of Technology Hoboken, NJ 07030 H ere at Stevens we adhere to the widespread proposi tion that using the Internet and the Intranet can be substantially beneficial to undergraduate chemical engineering education, both administratively and pedagogi callyY 2J There is also evidence that using software such as spreadsheets and equation solvers l 3 5 J not only develop s the skills and flexibility necessary for ready adoption of differ ent software packages l 61 for professional activities in indus try but also substantially supports learning. We believe that effective use of these tools requires a "c ultural change, or enhancement, on the part of both faculty and s tudents Full course integration is desirable l 61 and requires broad faculty participation which comes slowly in many cases_l7l The skill level and motivation of the students, however, can signifi cantly stimulate the faculty culture. At Stevens our initial focus has been on the students. We have attempted to forge a paradigm for a departmental electronic culture in our first chemical engineering course, Process Analysis, that comes in the second term of the sophomore year. Our efforts to do this are summarized in this article. COURSE SUMMARY The objectives of the course are to introduce students to chemical engineering to chemical processing equipment and chemical processes, and to apply material balances and basic phase equilibria to processing systems and the design of equipment from the equilibrium-stage point of view. At the same time, the course is being developed in an electronic environment in order to prepare students for this emerging characteristic of the workplace to enhance their learning and to establish a basis for distance learning and asynchro nous delivery By electronic environment" we mean the ub i quitous use of software for problem solving and tran s mission, communication between constituencies, presenta tions introduction to process simulators graphics, and com puter-aided instruction and learning. The course which uses the Luyben and Wenzel text / 8 1 is summarized in Table 1 Throughout the course the physical and chemical bases of the process or equipment being di cussed are emphasized and class discu ssio n s are often based on the problem assignment s and examples in the text. The students are required to complete a project involving con struction or enhancement of a website devoted to a chemical process or to a class of chemical processing equipment and each group pre se nts its project later in the semester. Through out the semester, students are encouraged to use the com puter-aided instruction modules developed at the University of Michigan and di s tributed by CACHE ( Computer Aids in Chemical Engineering). It should be noted that the integrity of individual and group work is s ubje ct to the Steven s Honor System. Thi s system is managed by the st udents through an Honor Board Individual cases are inve stiga ted and tried on the ba sis of well-defined procedures approved by both faculty and st dents Penalties range from warnings through grade loss to expulsion from the Institute. All homework and examina tions in the course are sig ned by the students and attest to the fact that they have adhered to the Honor Code. SOFTWARE Student s are required to purchase computers when enter ing Steven s. The software s uite that comes with the comput ers includes MS Office Pro 97, Mathcad Scientific Note book X-win32 WinQVT Netscape Java and Matlab. Software used in the course that is not included in thi s package is available for purchase on campus An extenGe o rge B Dela n cey is Professor of Chemical Engineering at Stevens Institute of Technolog y He received his BS MS and PhD degrees in chemical engineering from the University of Pitts burgh. He has presented undergraduate courses in process analysis transport phenomena reac tor design process control separation processes and thermodynamics. His current research inter ests are in biotechnology Copyrigh t ChE Divisio n of ASEE 1999 40 Chemical Engineering Edu c ation

PAGE 43

... the course is being developed in an electronic environment in order to prepare students for this emerging characteristic of the workplace, to enhance their learning, and to establish a basis for distance learning and asynchronous delivery. s ive application of Ethernet s upport s co mmunication s from t h e academic buildin gs, computational facilitie s, and residence hall s, a nd a gateway from the Et h ernet i s provided for access to the Internet. WebCT de scribe d at ( http : //homebr ew.cs. ub c .ca/webct ) i s u se d to organize the s ite s s umm ar i zed in Figure 1 to make grades, examples and so lut ions avai l a bl e to the s tudent s, to provide E-mail and Bulletin Bo ard co mmunic atio n too l s as well as a Ca l endar. The underlyin g html files Subject TABLE 1 Course Summary Overview of Chemical Indu s tr y a nd Pro cesses } Proce ss Equipment: Construction and Op era ti o n Proce ss Ana l yses: F l ow Sheet s a nd Pro cess Cond iti o n s Ma ss Ba l a n ces, De g re es of Freedom and Chemica l R eac tion Eq uilibrium Separation s: McCabe-Thie l e A nal yses Mult i -Component Pha se Eq uilib rium : I so therm a l Flash C l as s Pr esentations Welcome to Chemical E ngineering 210 0 C ourse Infonn&tion m I C hemical P ro c en u Chann Pauword ti}> Schedula F,nmplu andArngnmmts i Process Eauiprnent Click o n one ofthe icons above IIIMN U Weeks 3 4 4 2 R efe r ences I Group M e mb ers h ips I Grnding Po licy I Schedule and Assignmc n tsl Prep a ration a nd Submi ss ion o r Solutions I Pr ocess Equipment! Chemical Pr ocesses I Project I M i cro mentor I Scientific No t ebook I User Notes for Software I Scanning I Mailing Fi l es I Zipping Fi l es I Downl oa din g/Viewing Solutions and Examplt!s I McLean Comp ut e r R oo m I Virtual Library l Chemical Engineering New s Group I Ame ri ca n ln st i1ute of Chcmic:.11 En gi ne ers I Course l-l omcpagc Figure 1. Cours e hom epage and information h yper link s Wint e r / 999 were create d wit h Microsoft Word except po ssib l y for the s tudent proj ect pages Th e Ca l endar i s u se d on a cla ss ba s i s to post important events s uch as examination and assignment due-date s, clas s activi ti es, vaca tion s, etc. The Calendar can a l so be u se d on a p e r so nal ba sis by individual st udent s or the instructor with out others bein g able to view their entries Examinations are po s ted on the Bull etin Board an d the s tudent s ca n download/view the examina ti o n file. All que s tion s co n cerning the examination conte nt s are po s ted o n th e Bulletin Board a lon g with the responses of the instructor. Po s ting s from both partie s may take place at any convenie nt time and from a n y convenient lo ca tion and ma y b e viewed b y th e entire cl ass Thi s function is very useful espec iall y w h e n the exa m s are not due for severa l days or over a weekend. Que s tion s involving problem assig nm e nt s are handled in a si milar fashion except that viewi n g m ay b e limited to the gro up in que s tion and the grader for the clas s may respond as well. The Bulletin Board is a l so u se d on a gro up ba sis to m a n age gro up discussions on project and problem assignments a nd to transfer files associated with the se activities within the group E-mail s erves the common communication needs as well as being the vehicle for tran sfe rring h omework, exami n tions and project files between two partie s. The s ubj ect entry on the e-mail message is important for quick over v iews of previous messages a nd for the sea rch function s upplied by WebCT Scientific Notebook ( SNB ) found at htt p : // sc inotebook.tci sof t.com is u se d primaril y for so l ving sys tem s of nonlinear equation s arising from material balance s and in pha se e quilibrium ca c ulation s SNB ha s a u ser -friendly front end w ith a Maple kernel and s upp orts a word proce ss ing format. No s p ecial code is required and SNB can i nt erface wit h th e web a nd serve as a brow ser for Tex files. U se r s can quickly perform sy mbolic computations integration, differentiation matrix a nd vec tor operations and many other more co mpl ex com putations in vo l ved in calc ulu s, linear algebra differential eq uation s, and s tati s tic s For the se reasons, we c ho se to adopt SNB rather th a n Mathcact l 5 .7 l or Mathematica Y 1 The so lut io n to s imultaneous nonlinear equations a ss oci ated with material balance s i s illu s trated in Example 1 while Example 2 i llu s trate s how SNB i s u se d to s olve equilibrium fla s h calc ul ations Micro sof t Excel i s u se d for graphical (McCa be-Thiele ) so lutions of equilibrium-stage problem s : str ipper wit h a 4 1

PAGE 44

reboiler, gas absorption and stripping, distillation and liq uid-liquid extraction. We have elected to preserve the grap hical construction modality available in Excel for this introductory course rather than the spreadsheet solu tions to the material balance equations used at some other institutions. t 4 J The stage steps are constructed with the drawing tool in Excel. Distillation is illustrated in Example 3 and extraction in Example 4. Calculations can be done on the spreadsheet or SNB may be used saved with a screen capturing software (e.g Snagit) and pasted into Excel. Equilibrium data may be genera ted from a function (Example 3) or plotted directly as an input data series (Example 4). Stream points are entered as a new data series for convenience when zoomi n g in for stage stepping (Examp l e 3) or zooming out for the operating point in extractio n (Example 4). The course also makes u se of the MicroMENTOR system for deli very of the ed u cational software modules developed at the University of Michigan and distributed by CACHE. The modules suggested for use during the class are: UM Units, UMB eer, MATBAL UM -H awaii UM-POP and UM-McCabe The modules have been required in some cases and left to the discretion of the students in others. Distill is a program available from CACHE that does multicomponent flash and distillation calculations. It consid ers both the liquid and gas to be nonideal and it includes a database of 98 compo und s. The vapor-phase fugacity is calculated with the Redlich-Kwong equation of state The liquid-ph ase activity coefficients are based on the Hildebrand solubility parameter. This program is used in several assign ments to illustrate the effect of assuming ideal conditions in flash and bubble/dew point calculations for hydrocarbons Chem Windows is made avai l able to the student s for draw ing sc h ematics flow sheets (see the figure in Example 1, for example) and chemica l formulae. They have not yet been required to use this software and may use others if they wish. EXAMPLE 1--------------------------~ F J c,H, C hlorination Reactor inerts F31---- o, Ox y hydrochlorination Reactor Figure 1 ,....F_1_2 c ,H,c1 ,--.-F -1_0_ F 11 Inert" Fs HCI Sepa rator F9 C 2 H 3 CI 11 ,0 The flow diagram in Figure 1 illustrates a simplified version of the main steps in the production of vinyl chloride (C 2 H 3 Cl) from e th ylene (C 2 H 4 ). The reactions which occ u r separately in the dif ferent reactors are: C hl orination: C 2 H 4 + Cl 2 C 2 H 4 Cl 2 Oxyhydrochlorination: C 2 H 4 + 2 HCI + 1/2 0 2 C 2 H 4 Cl 2 + Hp Pyrolysis: C 2 H 4 Cl 2 C 2 H 3 Cl + HCI The ethy l ene feed, F 1 is 90 mole % ethylene and the remainder is inerts. The ch l orine and oxygen feeds F 2 and F 3 respectively are pure All of the ethylene, oxygen and chlorine react and all of t h e hydrochloric acid (HCl) fed to the oxyhydrochlorination unit re acts Only 50% of the total dichloroethane (C 2 H 4 Cl 2 ) fed to the pyro l sis reactor is converted with the remainder being separated and recycled with inerts in stream F 1 2 The inert concentration in the recycle stream is 50 mole %. Pure hydrochloric acid (HCI) i s r ecycled in stream F 1 3 The final product stream F 9 consist s only of viny l chloride and water. 42 Determine a ll of the unknown flow rates F i, and mole fractions x;j (mo l e fraction i in stream j). Set F 1 =1 mole/hr as a basis The species are labeled as shown in the following Table 1. TABLE I Labeling of Co mponen ts Species llld ex Species Index C 2 H 4 I C 2 H 4 Cl 2 5 Cl 2 2 C 2 H 3 Cl 6 HCI 3 H 2 O 7 o 4 Inerts 8 SNB solves the material balance equations given in Figure 2 in SNB format (single column) in le ss than one minute The s olution is shown in Figure 3 after some reordering for convenience. F 2 =O 9OF 4 F 3 = ( O 9O/2)F 5 F 4 +F 5 =F 1 x 5 6 F 6 =O 9OF 4 x 8 6 F 6 =O.10F 4 x ,.6 +x s.6 =1 X 1.1 F 1 =0 90F 5 x 5 7 F 7 =O.9OF 5 F 13 =2 ( O 9OF 5 ) x 8 7 F 7 =O IOF 5 O 5O ( x 5 6 F 6 +x 5 7 F 7 +O.5OF 1 2 ) = x 6 8 F 8 O.5O ( x 5 6 F 6 +x 5 7 F 7 +O 5OF 12 ) =x 3 8 F 8 X 7_7 F 7 =X 7.8 F 8 x 8 6 F 6 +x 8 7 F 7 +O 5OF 1 2 =x ,. 8 F 8 x 3 8 +x 5 8 +x 6 8 +x 7.8 +x 8 8 = I x 3 8 F 8 =F 13 x 5 8 F 8 =O 5OF 10 x 6.s F =x 6. F X 1., F =X 1,9 F 9 x 8 8 F 8 =O 5OF 1 0 X 5.7 +X 7 7 +X 8.7 =l O 5O(x 5 6 F 6 +x 5 7 F 7 +O 5OF 2 ) =x 5 8 F 8 x 6_9 +X 1. = l F io =F11+F ,2 F i g ure 2. M a t e rial Balan ces F 2 = .5 F 3 = 2, F 4 = 55556 F = 44444 F 6 = .55556 F 7 = 84444 F 8 = 3 6 F 9 = 1 2 F io = l .6 F 11 = .2,F 1 2 = l.4 F 13 = 8 x 3 ., = 22222 x ,. 8 = 22222 x 6 8 = 22222 x 7 8 = l l l l l x 8 8 = .2222 2, x 5 6 = .9 x 8 6 =.I x ,. 7 = .47368 x 7 7 = .47368 x ,. 7 = 5 26 3 2 X I 0 2 x 6 9 = 66667 x 7.9 = .33333 F = 1.0 Figure 3. Solution s Ch e mi ca l En g in e erin g Edu c ati o n

PAGE 45

e-EXAMPLE2 A petrochemical s tream consisting of 30 mole % propane 10 mole % nbutane 15 mole % n-pentane a nd 45 mole % n-hexane is to be flashed to 200 kPa. (a) Determine s tri ct bound s on the operat in g temperature. { b) Find the vapor flow rate per unit of feed and the product compositions for a temperature mid-way between the se limit s The equilibrium data are given in Table I TABLE I Phase Eq uilibrium Data Equi librium dat a: y = Kx T = 0 R, and p = psia: In K ( T p ) = (a/T 2 ) + a 3 + b 1 In p + ( b / p ) Species a, a, b b, C 3 H 8 -970,688.5625 7.15059 -0.76984 6.90224 n -C 4 H 1 0 -1 ,280 557 7.94986 -0 .9 6455 0 n -C 5 H 1 2 -1,524 ,8 91 7.33 129 -0.89 1 43 0 n-C 6 H 14 -1 ,778 90 1 6.96783 -0.84634 0 The solution procedure in SNB i s as follows: The function s in Figure I are defined in SNB format. Bounds for the bubble and dew points of the feed are at unity K va lu es for propane (i= I ) and n-hexane (i=4 ). These va lu es are obtained from the roots of the K function at the prevailing values of p (2 00 kPa ) and j The solutions are shown in SNB format in Figure 2. The bubble and dew points of the feed are roots of the associated function s at the prevailing p The re s ult s in SNB format are s h own in Figure 3. The vapor flow at th e midpoint temperatur e, 56l.4 R i s obtained similarly a nd i s s hown in Figure 4 E Tb (' ) ( ai I bi 3 J qui I num: K 1 T p =exp ~+ai ,J +bi 1 n p+--;4 Bubble point: f(T p )=lL K ( i T,p)zi ,J i=I 4 Dew point: g(T,p)=lL K(~ i; p) 1 = 1 I Z;1 [K ( i T ,p}-l ] I sot h ermal flash: h(t} T p)= [ (' ) ] J+t} K t,T p -1 i = I Figure 1 aJ l b j,3 ? +aJ 3 +bj 1 f np + =0 Tp Solution is {T=657 .27},{ T=449.88} T E( O, oo) Figure 2 f(T,p )=0 g( T p)=O TE(449,658) Solution i s {T=509.27} TE ( 449 658 Solution i s {T=613.54} Figure 3 I I h(t} T p) = ol I I t}E(O l) I Solution i s {t}=.37704} Figure 4 Wint e r 1999 EXAMPLE 3------------. Find the number of s t ages the best feed l ocation, and th e minimum refl u x ratio for a distilJ a tion column that separates ethano l and propanol at 101. 3 kPa. The ratio of th e vapor pressure of e thanol to th e vapor pre ss ure of propanol i s approximately co n s tant at 2. 10 The feed i s 48 mole % ethanol and 40 mol e% liquid The di s tillate and bottoms compo s ition s a re to be 96 mole % and 4 mole % ethanol respectively There i s to be a total condenser overhead with no sub-coo lin g. The reflux ratio is 3.0. The graph i ca l construction i s shown in the figure in Excel format. '-000 0 000 f 0 700 S 0 600 i j 0 500 X O AOO i 0 300 0 200 0 100 0 000 Ethan9I Pr9p1nol Syatm II 101.3 kP1 I I I I I ----------( --:::1 ::;::: :)/ --"6: ;v _v ;&, ... I _v ~ f:: ..V .,/ / v I .I V A I I ~ ,? 0 000 0 100 0.200 0 300 0 4 00 0 500 0 600 0 700 0 800 0 900 1 000 Mole traction ethanol in U ou l d x EXAMPLE 4-------------. Acetic acid (s pecie s I ) i s to be extracted from water (s peci es 2) u s in g isopropyl ether (s pecie s 3) as the solvent at 20 C a nd 1.0 atm The feed rate i s l000 kg/hr and contains 35 wt % acid in w a ter. The s olvent flow rate i s 1475 kg/hr and i s essentially pure ether. The raffinate is to co ntain no more than 10 wt % acid Find (a) the minimum so l ve nt flow rate (b) the number of equilibrium s t ages required for the separat ion a nd (c) the o utlet concentrations a nd flow rates. Equilibrium data at the operating conditions are given in the Table and the graphica l construction in Excel is s hown in the figure Water Layer x, x, 1.2 0.69 1.5 1.41 1.6 2.89 1.9 6.42 2.3 1 3.3 3.4 2 5 5 4.4 36.7 10.6 44.3 16.5 46.4 Ether Layer X, x, 99.3 0 18 98 9 0.37 98.4 0.79 97 .1 1.93 93.3 4.82 84.7 11.4 7 1. 5 21.6 58.1 31.l 48.7 36.2 TABLE Phase Eq uilibrium Data T FIGURE Excel Constuction for Example4 1gt1c tskHl9Rf9PYI .,,,,....., m 1 ,rn 1 19 c I llm .la I I I I I : I 1 f':~ .. I ,... r,-, I
PAGE 46

FILE HANDLING An extraordinary amount of time is required to open files to learn the source and/or the contents. This is especially true if more than one course is underway. Consequently, labeling of files is very important. Points are deducted from assign ments and examinations (which may not even be accepted in some cases) if the established procedures are not followed. No files are accepted unless adherence to the Honor Code is pledged by the students at the end of the file. Homework files are prepared weekly by students working in groups. The assignments are numbered and are generally specific problems in specific chapters of the text. Solutions to individual problems are labeled as gxpy-zc210 (tex, rap, doc, xis, ... ) where x is the group number, y is the chapter, and z is the problem number. The individual solution files are then assembled in a single archive, with WinZip for example. The archive is labeled gxayc210.zip where xis the group number and y is the assignment number. The archive is attached to an e-mail message with an informative subject entry and mailed to the instructor. The archives are placed in a directory named with the group numbers. Solution files for each problem are posted for downloading after the due date for the assig nment. Examination files include a reference to the individual student: exmc210qnabz.(tex, rap, doc, xis, .. ), where mis the examination number, n is the question number, ab are the student initials, and z is the group number. We have not encountered students with the same initials in the same group yet, but many modifications are possible to handle such an event. The examination files are then grouped in an archive labeled exmc210abz.zip and attached to a relevantly labeled e-mail message to the instructor. The archives are then placed in a directory named with the examination number. Copies of files received by the instructor are forwarded by e-mail to the grader. The grader is not permitted to accept files directly from the students. The grader has a set of directories similar to those used by the instructor, described above. The grader uses several additional s ubdirectorie s as sociated with the grading process. One directory is for graded files, one for ungraded files, and one for the archive in process of being graded. After the files are graded and the comments and points included, the files are returned directly to the st udents the grader retains a copy, and a copy of the file is sent to the instructor. The same file name is retained so that the instructor can replace the previously ungraded file with the graded one in his directory. It has always been possible to track down "missing" files with this procedure. The grader is responsible for posting grades in WebCT, but a separate grade file is maintained by the instructor and updated with each new assignment or examination. The grade file maintained by the instructor is the official one so that grades, once recorded, cannot be changed without the 44 instructor updating this file. Grade access is limited to the student, the grader, and the instructor but statistical infor mationis generally available PROJECT The objectives of the project are for the student to gain a special familiarity with a specific chemical process or a category of chemical process equipment and a significant experience with web-based presentation of technical materi als. The students are also required to present their projects to the class, electronically if desired The bases on which the sites are initiated or enhanced are summarized below. A) Process equipment (nine sites at present covering major equipment categories): l) Purpose and operating principle(s) 2) Historical background 3) Construction-schematic/illustrations 4) Pictures of actual equipment 5) Range of duties-sizes 6) Maintenance required 7) Utilities required 8) (Some) design equations 9) References B) Chemical Processes ( nine sites at present spanning range of top 50 chemicals): 1) Chemical formula and form/state of product 2) Uses and market price 3) Historical industry development 4) Common methods of production with raw material sources and side products 5) Major companies production methods, and production lev els 6) Details for a common production method: process chemistry with implications for most favorable process conditions; flow sheet(s); operating conditions and problems; environ mental considerations ; production costs; utility require ments; hazards; handling of waste and side products 7) References C) Grading (E ngineers from EXXON Research and Engineer ing and experienced faculty outside the department contribute to site assessment): 1) Content (see above) 2) Layout/format 3) User friendliness 4) Use of colors and graphics 5) Use of relevant hyperlinks and other resources Html files are prepared or edited with software that is generally available to the students, such as Microsoft Word, or that is the individual preference of the students Scanning hardware is used for some picture files, which can be edited with Photopaint. The completed project files, including im ages (gif, jpeg, .. ) are collected in an archive and mailed to the instructor as described above. Chemical Engin e ering Edu c ation

PAGE 47

DISCUSSION The students very quickly become accu s tomed to the elec tronic communication features of the course and the file handling procedure s It is, however helpful to deduct points for improper file labeling as well as to not answer exam questions online unless posted for viewing by the entire class Off campus s tudents find the electronic version help ful and students generally like the option of transferring files and posing question s at their convenience. Such item s ap pear at all hours of the day. For the solutions to assignments the availability of examples for downloading by the students i s crucial for each new application. The use of SNB for material balance s has greatly im proved the sophistication and ability of the s tudents to write an independent set of material-balance equations and to be acutely aware of the degrees of freedom. They are also able to spot-check their solutions with quick calculations u s ing time components, inerts etc Most students begin with doing the solution on paper and tran sferri ng it to SNB. They soo n progress some to doing the so lution completely online. A series of user sessions would be helpful at the beginning for learning the software Some s tudent s have prior expe rience in mathematics classe s, which is beneficial. Stu dents have voiced a preference for SNB to similar sof ware packages but these alternatives have not been ex plored in a formal way. The use of Excel for graphical multi stage constructions has been very successful and well accepted by s tudent s Error s can be quickly corrected The students cannot imag ine doing liquid-liquid extraction constructions by hand. When not required, the computer-aided instruction mod ules in MicroMENTOR are not extensively used. Use of the modules is greatly increased if the s tudent s are held account able, for example in class di s cussions on the Bulletin Board of WebCT. The modules once u se d are considered helpful. Distill is viewed as inconvenient by the s tudents although the results are very elucidating for comparison purpo ses. For this reason and for broader objectives it may be worthwhile to introduce Aspen in a limited way at thi s stage ChemWindows is viewed primarily as a cosmetic add-on for problem assignments when not required. It i s used for all examination drawings and for so me project work. As it significantly enhances the quality of the pre se ntation we are introducing some limited requirements for it s u se The class projects have been most successful and well received by the students. The project and equipment sites they have developed are being made an integral part of the course. The students are extremely creative and enthusiastic about finding relevant material s on the Web and using other resources to develop or enhance their site. The technical content of the class presentations attests to the learning com ponent that i s present in the projects. The students very Wint e r 1999 easily adopt a full range of software to s uit their personal needs for developing the required html files and exploit the extensive range of picture files available on the Web. The electronic approach used in Proces s Analysis has been repeated with much les s effort by the author in Transport Phenomena a course offered in the following semester. The approach ha s not yet been extended further across the chemi cal engineering curriculum. Software i s, however used in all of the courses to some extent and coding is required in several classes. The students have been observed using the specific skills (es pecially the graphical solution to multistage equilibrium se paration s) and exhibiting the electronic dexterity they gained in Process Analysis in subsequent courses The time commitment and the background that must be acquired on the part of faculty in order to develop an elec tronic version of a course as defined here are important related issues and tied very much to the available support and the electronic environment. It must be generally recog nized, however that significant effort is required which must be explicitly recognized and supported within the local edu cational context and the electronic capabilities of its con s tituenci es. The benefit s include presentation of a course in an environment that anticipates the workplace, that is more adaptable to the different life styles of the s tudents, and that can be conveniently modified and maintained with no more effort than a conventional course. The electronic version i s more supportive of the learning proces s and allows more spontaneity and interaction especially when the students are networked with their own computers. A greater focus on problem s tructure and formulation can be realized as well as a pronounced increase in the specific and general electronic skills of the s tudent s. REFERENCES 1. Bungay H ., and W Kuchinski The World Wide Web for Teaching Chemical Engineering ," Ch e m Eng. Ed ., 29 ( 3 ), 162 ( 1995 ) 2 Marr D W M., and J.D. Way "Using the Intranet in ChE Instruction Chem. En g. Ed. 31(2 ), 110 ( 1997 ) 3 Al-Dahhan M H., Computing in the Undergraduate ChE Curriculum, Chem Eng. Ed. 29 ( 3 ), 198 ( 1995 ) 4. Burns MA. and J .C. Sung, Design of Separation Units Using Spreadsheets ," Ch e m. Eng Ed ., 30 ( 1 ), 62 ( 1996 ) 5. Sandler S.I ., Spreadsheets for Thermodynamics In st ruc tion: Another Poin t of View ," Ch e m Eng Ed ., 31 ( 1 ), 18 ( 1997 ) 6 Davis J.F. G.E. Blau and G.V. Reklaitis Computers in Undergraduate Chemical Engineering Education," Chem. En g. Ed. 29 ( 1 ), 50 ( 1995 ) 7. Harb J N ., A. Jone s, R.L Rowley and W W Wilding ''Use of Computational Tools in Engineering Education ," Ch e m. En g. Ed ., 31(3 ), 180 ( 1997 ) 8. Luyben, W.L. and L A. Wenzel Chemical Process Anal ysis: Mass and Energy Balan ces, Prentice Hall Englewood Cliffs NJ ( 1988 ) 0 45

PAGE 48

5 Q laboratory ) llil-111111-lii------=----DEMONSTRATING SIMULTANEOUS HEAT AND MASS TRANSFER WITH MICROWAVE DRYING CHERI C. STEIDLE, KEVIN J. MYERS University of Dayton Dayton, OH 45469-0246 I n a recent article, Nirdosh and Baird [ 1 l emphasized the critical role that the laboratory plays in the undergradu ate engineering experience. It is crucial that laboratory experiments provide practical reinforcement of the theoreti cal chemical engineering concepts developed in lecture courses, but due to budget constraints it is often also neces sary to develop inexpensive experiments or to make use of existing equipment. This paper describes an effective and inexpensive microwave drying experiment that can be used on a variety of levels. On the introductory level, the data analysis associated with this experiment illustrates the nu merical approximation of derivatives from discrete data, while on the advanced level the experiment develops an understanding of simultaneous heat and mass transfer. B ACKGROUN D Of all the chemical engineering unit operations, drying is one of the most widely used with applications in various processing industries such as food processing pulp and pa per, pharmaceuticals etc. Because of its widespread use and the fact that it may account for up to ten percent of industrial energy consumption ,l2 1 it is therefore essential that the fun damentals of the drying process be encountered and under stood by undergraduate chemical engineering students. There are a variety of drying techniques such as vacuum dry i ng and spray drying 1 31 each with its own operational characteristics Conventional dryers include both direct and indirect methods of heat transfer [ 4 l Direct dryers (also called convection dryers) use contact between the wet solid and a hot gas to accomplish heat transfer with the vaporized liquid being carried away by the drying gas. In indirect dryers heat for drying is transferred through a wall that separates the wet solid and the heating medium with the vaporized liquid being removed independently of the heating medium. Indirect dryers are also called conduction, or contact, dryers. The driving force for heat transfer in both direct and indirect dryers is the temperature difference between the drying medium and the wet solid. In the case of microwave drying a magnetron produces a pulsing electromagnetic field. Polar molecules such as water align with this field and as the field direction changes the molecules are forced to re align. These molecular oscillations create friction that gener ates heat, raising the temperature and causing liquid evapo ration.l51 Thus, although supplied from an external source the energy for microwave drying is often thought of as being generated within the wet solid Although the mechanism of delivering energy to the wet solid in microwave drying is different than in conventional drying since microwave energy does not penetrate very far below the surface of the exposed material ,l 6 1 the transport processes occurring during microwave drying are very simi lar to those that occur during conventional drying. As liquid evaporates from the surface of the solid-liquid mixture liqCheri C. St e id le received her Bachelor of Chemi cal Engineering degree in 1995 and her MS in Chemical Engineering in 1998, both from the University of Dayton She is currently working as a process engineer in the glass industry and is responsible for implementation of new pro cess control equipment for the manufacture of glass tumblers and tableware Kevin Myers is a professor in the Department of Chemical and Materials Engineering at the Uni versity of Dayton. He received his BChE degree from the University of Dayton and his DSChE from Washington University in St. Louis His re search interests are in multiphase agitation and ._ __ __ __, chemical reactors Co p yr i g ht ChE D iv i s i o n of AS E E 1999 46 Ch e mi c al En g in ee rin g Edu c ation

PAGE 49

This experiment demonstrates the drying process effectively. It is an extremely flexible, safe, and inexpensi v e experiment that can be incorporated into the undergraduate laborator y curriculum. The experiment is easy to set up and run. Typicall y, meaningful experimental data for higher power settings can be collected in about thirty minutes. uid from within the so lid migrates to the s urface becau se of co ncentration gradients. Thus dr yi ng either b y conventional or microwave device s, inherently invo l ves sim ult a n eous heat and ma ss transfer. The primary data obtained during a drying experime nt are the moi s ture conte n t as a function of time ( on a dry ba sis the moisture co nt e nt eq u a l s the l iquid mas s divided by the so lid mas s). Curve 1 in Figure 1 pre se nt s a typical moisture con tent curve. The moi st ure content data can be differentiated to yie ld the drying rate curve, Curv e 2 in Figure 1. Dr yi n g R a te= ( ~: J( ~7 j (I) Typically, drying rate curves for materials with a thor oughly wetted surface exhibit three periods 171 The first stage is the warming -u p period This s ta ge is characterized b y increasi n g the material temperature to that of the evapora tion temperature of the wetting liquid. Also, the drying rate increa ses as the liquid begin s to va porize This i s fo ll owed by a period of constant-rate drying where the s urface mois ture evaporates and moi s ture is s teadily brought to the s ur face to maintain a contin u o u s liquid film over the s urface. The third, and l ast, period in the dr yi ng-rate curve, called the falling-rate period is c h aracterized by a nonlinear decre ase of the drying rate due to the increasingly uneven moisture distribution over the s mface The constant-rate and falling rate stages of the drying-rate curve are separated by the point of critical moisture content. This point mark s the instant point defining the critical moisture Ji content Time (I) ro 0::: O> C ~ 0 Figure 1. T yp i ca l moisture conte nt and drying-rate c urves for a solid with a thoroughly wetted s urfa ce. Winter 1 999 when the liquid no longer forms a contin u o u s film over the entire s u rface because the rate of moi s ture tra n s p ort to the s urface is le ss than the rate of evaporation from the s urfac e. The critical moi s ture conten t is not a material property. It varies with drying rate, thickness of the materia l particle s ize and other factors that affect moi s ture movement. Criti cal moisture content is be s t determined by experimen t. APPARATUS The equipme nt required for thi s experiment is inexpensive and easy to operate. Specifically an off-the-she l f Frigidare MC-1 lO0M microwave with ten power se ttin gs ranging from 72 W to 720 W was u se d. The power se ttings of the oven are in terms of percentage of the maximum power. Thi s micro wave oven is rat h er o ld but a comparable c urr ent model cou ld be purcha se d for le ss than $200. Other required items include a digital balance acc u rate to the nearest gra m a stopwa tch (or the microwa ve timer) microwaveable trays or Pyrex g l ass beakers and a thermocouple (specifica ll y, a type T was u sed). Sand wetted with water was s tudi ed, wit h th e nominal diameter of the sa nd particle s being 600 microns. METHOD Wet san d in the ratio of approximately 1 kg of sand to 0 .2 kg of water was u se d. Thi s ratio pro v ided the conditio n that all of the sand was completely wetted, but no sta ndin g pools of wa t er were present. The sa nd and water were thoroughly mixed together. The mixture was weighed sprea d into a n even layer approximate l y 0.025 m deep in a 0 15x0.15x0.05 m tray that was placed in the ce nt er of the microwave oven The microwave was started at a desired power l evel. The weight of the mixture was recorded every minute until the sa mple was dry. In addition, the s urface temperature of the sa nd could also be monitored with a thermocouple whe n the sa mple was removed from the oven for weighing. The effect of removing the sample from the oven for weighing was examined by u si ng various drying intervals s u ch as 30 seco nd s, 1 minute 5 minutes and 10 minutes Only mi nor changes in the sa mple weight were observed for the se different drying intervals. Appropriate safe ty pre ca utions incl ud e wearing safe t y glasses and oven mitts when handling the hot tray RESULTS Figure 2 pre se nts experimental moi s ture co nt e nt and dry ing rate data for the highe s t power setti n g of the mic rowave oven (720 W ), while Figure 3 compares the drying rate 47

PAGE 50

curves obtained at three power settings (720 W, 540 W and 360 W). All of the experimental data are very similar to the idealized data of Figure 1, clearly illustrating the three stages of drying. Table 1 demonstrates the influence of power setting on the critical moisture content. Recall that the criti cal moisture content is not a material property; rather it is dependent on operating conditions. The trend of increasing critical moisture content with increasing power setting is typical,C 4 l As the drying rate is increased (by increasing the power setting), it becomes progressively more difficult for the rate of moisture transport to the surface to remain as high as the rate of surface evaporation Thus, the falling-rate period, which begins when the critical moisture content is reached, occurs at a higher moisture content. During the constant-drying-rate period a pseudo-steady state is achieved, with the power input going to the latent heat of evaporation of the liquid and energy losses. Table 2 presents the efficiency of the constant drying-rate period assuming a latent heat of evaporation of 2300 k:J/kg (corre sponding to approximately 80 C). The efficiency is calcu lated as the energy required for water evaporation (equal to the evaporation rate, kg/s times the latent heat of evapora tion, kJ/kg) divided by the rate of energy input (kW = k:J/s) The observed efficiencies are rather high, ranging from 74 % to 87 % Given the advanced age of our microwave oven, its power output is probably less than the nominal value (we accepted the nominal power outputs and did not actually measure the oven's power output) If this is true the actual efficiencies are higher than those calculated here. The high efficiencies of microwave drying are par tially due to the energy-transfer mechanism. Since en ergy is supplied directly to the wet solid, there is no large energy requirement for heating the drying medium as in conventional drying devices. The efficiency decreases with increasing power level. Ini tially this was thought to be due to increased energy losses to the lower portions of the wet solid by conduction At low power inputs, the temperature of the wet solid might be 48 TABLE 1 Influence of Power Level on Critical Moisture Content for Drying Sand Critical Moisture Power Level Co11te11t now 7.4 % 540W 7.1 % 360W 6.2 % 144W 5 7 % Literature value 14 1 5 9 % (unknown power level ) Power Level (kW=kJ/s) 0.360 0.540 0.720 relatively uniform at the start of the constant-drying-rate period as conduction has had sufficient time to transport energy from the surface to the interior of the wet solid At high power inputs, conduction may be unable to transport energy from the surface to the interior regions of the wet solid rapidly enough to achieve temperature uniformity at the start of the constant-drying-rate period. If this were the 0.2 0 ~ . -----------~ 32 0.020 0 II) Cl 0.15 5 g :;::0 .10 C 2 C 0 0 0 0 5 .. .. ... ........... ..... ... Moisture Content 0.015 ... Drying Rate ... 0.010 .. .. 0.005 ::::, ... ... C) U) C o .. -~ ,. ,. ,. Cl 0 00 +--+--+--+--+----+--r---+-~ ,,_.. -- 0.000 0 5 10 15 Time (min) 20 25 Figure 2. Moisture content and drying-rate curves for a power level of 720 W. 0. 0 20 ~-----------~ c E 0 .015 II) =o g o 0 10 Cl C, o:: 0 005 Cl C ............... ..... ... ... X ... X X X ... = ... 720W 540W X 360W -~ Cl ... X ......... ~xXx 0 000 'F-, i--+-+-~+-+-+-++--11----+ =.-+--1--+ -"'+'""F'i: 0 5 10 15 20 25 30 35 40 45 Time (min) Figure 3. Comparison of drying-rate curves for power levels of 720 W 540 W and 360 W. TA B LE2 Influence of Power Level on Constant Drying-Rate Period Efficiency (from Figure 3 drying-rate curves) Evaporation Rate (kgls) 1.36 X l0 4 1.90 X l0 4 2 33 X [Q 4 Evaporation Energy (kJ/s) 0.313 0.437 0.536 Efficiency 87 % 81 % 74 % Ch e mi c al En g in ee ring Edu c ati o n

PAGE 51

case, conduction would s till be removing so me of th e energy input from the surface to th e interior durin g the co n s tant drying rate p er iod. Thi s energy lo ss wo uld lead to de creased efficiencies at high power level s. Thi s explanation, however lo ses it s appeal when th e energy inputs durin g the warming-up period are compared. Th ese e n e rg y inputs are a bout 240 kl independent of pow er l eve l. This constancy of e nergy input during the warming-up period indicates th a t energy lo sses d u e to conduction are likel y to be the sa m e for a ll power le ve l s The rea so n for decre as ing efficiency with increasing power le v el requires another ex planation that re quire s further in ves tigation The drying-rate curve s of Figures 2 a nd 3 were generated from t h e moi s ture-content curves through numerical differ entiation u s ing a central-difference method. The central difference method is best u se d in cases involving lar ge time intervals c s i as is the case with thi s experiment. Assuming evenly spaced data, at so me time ~ the ce ntral difference 0 20 40 6 0 80 100 120 Tim e ( m in) Fig ur e 4. Effect of step size on numeri c al differentia tion at a power l eve l of 144W. 0 020 ,--------~ '2 E "O g 0 015 Cl .I<: u s g o 0 1 0 Cl :::, 2 0 005 Cl C ~ ... Eqn 2 ... Eqn 4 Cl 0 000 +----+---+--+---
PAGE 52

I 3.5_.._i_a_b_o_r._a_t_o_r..:.y ___ _____ ) MEDICAL SURVEILLANCE AND THE UNDERGRADUATE THESIS IAN A. FURZER University of Sydney Sydney, New South Wales 2006, Australia T he Department of Chemical Engineering at the Uni versity of Sydney has a thesi s as a core unit of study in the final year of the chemical engineering curricu lum. Students are required to complete fifty credit units in their final year and the undergraduate thesis is worth eight of them. It normally takes place in Semester 1 and may overflow into the vacation break before the second semester begins The thesis discussed in this article is an experimental thesis concerned with the operation of a distillation column to collect composition data on the trays for a ternary mixture of ethyl acetate-ethanol-water. The same thesis however can incorporate a significant new component that is de signed to make students more aware of occupational expo sure to hazardous substances This insight can also be valuable in other chemical engi neering courses, including risk engineering, hazards and hazops environmental pollution and chemical engineering design The awareness of the dangers of human exposure to hazardous substances is becoming evermore important due to the long term health effects on workers in the oil, chemi cal, a n d bio t echnology industries as well as on the general public Chemical engineering students need to be aware of alternative process flowsheets that avoid hazardous substances and a general chemical reduction-use program Several popu lar books have brought the effect of chemicals on human health into the public consciousnessY 1 MEDICAL SURVEILLANCE The World Health Organization has a medical surveil lance program for the early detection of occupational dis eases It is a prevention program that should be brought to the attention of chemical engineering students. The informa tion from such program can help students reduce the ri s k of exposure to hazardous substances While medical tr e atment procedures are still not well developed for exposure that leads to cancer several decades ahead numerou s medical treatment procedures are outlined on Internet web s ite s, in cluding osha gov in the USA and work s afe gov. a u in Au s tralia There is also a CD-ROMr 2 1 that contains outline s of medical treatment and surveillance programs for a wide range of hazardous substances Another CD-ROM search for ethyl acetate provides a wide range of information including an eight-hour time weighted average ( TWA) exposure of 200 ppm in many countries and an outline of the proper medical treatment for it. This information confirm s the low risk to s tudent s of exposure to ethyl acetate during a s eries of distillation ex periments conducted for their undergraduate thesi s But chemical engineering students should be made aware of important and relevant sections of the subject's toxicology and epidemiology Benzene is a substance that ha s under gone a reduced TWA over the years and i s now recognized as being carcinogenic to humans Details on background levels of benzene a series of epidemiological studie s, and cancer mortalities can be found in reference 3. The environI an F urze, has been a faculty member in the Department of Chemical Enginering at the Uni versity of Sydney for over twenty-five years He has extensive teaching and research interests that include computing process simulation and chemical engineering plant design. He is the author of over eighty research publications and the textbook Distillation for University Students Co p yr i g lit ChE D i v i sion of AS E E 1999 50 Ch e mi ca l En g in ee rin g Edu ca ti o n

PAGE 53

mental effects of benzene [ 4 l in wastewater and aircraft en gi ne exhausts and their connective pathways to human s need to be carefully considered by c hemical engineering s tudent s in the design of plant s that produce ben ze ne or fuel s which produce benzene upon combustion. Can we expect a reduc tion in the TWA value for ethyl ace tate in the future if there is medical evidence of an adverse health effec t ? In New South Wales 1966 le g islation on Occupational Health and Safety ( Hazardou s Substance s) Regulation ha s a separate divi s ion on an employer 's dutie s relating to health surveillance The employer mu s t provide workplace health surveillance if any exposure to a hazardou s s ub s tance re s ults in a reasonable likelihood of a di sease or other effect on health. Thi s health s urveillance mu s t be under the s upervi sion of an approved medical practitioner. The type of s ur veillance is listed in so me detail for eleven hazardou s s ub stances, including acrylonitrile asbestos i socya nate s, orga nophosphate pesticide s, polycyclic aromatic s, and vinyl chlo ride The medical tests include the s tandard re s piratory func tion te s t s s uch as FEVl and FVC. OCCUPATIONAL HEALT H ANDEXPOSUREINTHELABORATORY The distillation column for the undergraduate the sis ex periment involves medium quantities of so l ve nt. The choice of solvents to be distilled by s tudent s s hould be ba se d on occupational health safety, ease of analy s is s hape of the x-y diagram for binary sys tem s, and the ability to extract knowl edge on se paration sys tems. Alcohol-and-water is a com monly used sys tem The occupational health ri s ks to s tu dents of exposure to ethanol b y inhalation need to be con s idered. How often are laborator y demonstrators advising s tudent s of the need to minimize the inhalation of ethanol? For an undergraduate thesi s or s pecial project, it i s u se ful to consider the ternary mixture of ethyl acetate, ethanol, and water. This mixture provides a good working environment to extract knowledge on ternary and binary azeotropes di s tillation paths, and the appearance of a two-liquid-pha se region in the distillation column. The following method has been u se d to actively reduce the ex po s ure to ethyl acetate and ethanol over th e s tudent s' extended period of work for the undergraduate the s i s The most important aspect of occupational health i s to introduce the subject to the students who will be involved to notify them of the low hazard of ethyl acetate and ethanol, and to di sc us s with them the method s of reducing the ma ss inhaled. Thi s introduces to them the p syc hological compo nent of occupational health, leading each s tudent to develop hi s or her own concern about the toxic nature of the inh a led s ub s tance s Thi s component will play an important part in Wint e r 1 999 The awareness of the dangers of human exposure to hazardous substances is becoming evermore important due to the long-term health effects on workers in the oil, chemical, and biotechnology industries, as well as on the general public. assessing their own exposure estimates. An approach to primar y pre ve ntion of exposure is to increa se ventilation in the laborator y, thu s dilutin g the inhaled air and reducing the composition of the so l ve nt in th e air The mas s of inhaled so lvent per breath can be reduced and the risk factor in occupational health could also be expected to be reduced Thi s invol ves s tarting exhaust fans and opening door s and windows to increase air c irculation. The next aspect of occupational health is associated with safe t y and i s co ncerned wi th the release of so lvent va por s due to a laborator y fire Students s hould be advised on the safe ty procedure s to be u se d in the event of a fire. They s hould be told to evacuate the laboratory immediately and not to attempt putting out the fire but to notify the safe ty officer in s tead Thi s i s extremely important for minimizing the risk of exposure. Good laborator y practice is important for occupational h ea lth. Th e r e i s a direct association between hazard identifi cation hazard analysis a nd good occupational health. When release of so l ve nt va por s i s the h azar d to be minimized a haz a rd identification pro gra m mu s t be carefully conducted on the di sti llation equipment, identif y ing all possibilitie s of vapor release and li s ting all actions needed to prevent s uch a relea se For example, vapor could be rele ase d by a failure of the coo ling water, a m ec hanical failure of a glass column piece, a gasket failure, or a low level in the reboiler. Special atten tion s hould also be given to the cooling water supply. Safety labels on the cooling water supply swi tche s are essential to prevent an in s ufficient s upply of cooling water and to mini mize the release of so lvent vapor. Al so associated with goo d laborator y practice i s the need for the s tud e nt to be pre se nt in the laboratory while the di s tillation equipment i s operating Student s s hould be well trained in emergency procedures such as turning off the s team s uppl y to the reboil er in the eve nt of a so lvent release. STUDENT EXPOSUR E T O ETHY L AC E TATE Students can be exposed to ethyl ace tate in a number of the 5 1

PAGE 54

laboratory experiment phase s. The first i s concerned with calibration of the gas-chromatograph ( GC) equipment for analysis of the ethanol-water-ethyl ace tate mixture Thi s can invol ve ex po s ure while tran s ferring ethyl acetate from a 20-L drum to s maller glass ve ss els the preparation of standards, and dur ing the running of the GC Student exposure to ethyl ace tate during thi s phase in generally low. An inter nal standard is used in the GC calibration. In thi s case it was 1-propanol, and s tudent s s hould be aware of it s occupational health characteristics. F i lling the di s tillation column with the ternary mixture involves transferring and mea s uring about 20 L of the mixture and often involve s mild expo s ur e to ethyl acetate. The di s tillation column, which had been tested for leaks with water will now be found to have a s mell of ethyl acetate, but no visible liquid leak s The laboratory i s equipped with a gas alarm sys tem with a se n so r adjacent to the di s tilla tion column. Thi s alarm i s co ntinuou s ly on but is not activated by this mild s mell of ethyl ace tate. Running the di s tillation column with good ve nti lation until s tead y state i s reached ma y require one hour and re s ult in mild exposure to ethyl acetate Additional exposure could occur when liquid sa mple s are withdrawn from the nine column trays. Further mild exposure would take place for eac h di s tillation run BLOOD TESTS AND STUDENT PRIVAC Y Student s may be exposed to ethyl ace tate before entering the unit operations l a boratory. Ethyl ac e tate i s a well-known solvent and i s often used in the cosmetic industry as a nail poli s h r e mover so fe male s tudent s who u se thi s s ub s tance might be ex pected to have a higher ethyl ace tate co ntent in their blo o d. The body may also generate ethyl aceta t e from the complex biochemical pathwa ys in the body. T h e initial ethyl acetate content of s tudent s' blood i s import a nt before the y enter the so l ve nt environ ment of the unit operations laboratory but thi s can po se so me problems with privac y. Student s have a right to priv acy concerning analysis of th e ir blood Voluntary agreement mu s t be obtained from the s tudent for a blood test for ethyl ace tate. Student s a re advised to contact the s tudent medical center for thi s blood te s t. Th e background le ve l of e th y l acetate in the blood s hould be around 0 5 mg/L. One female s tudent however, had an initial level of I .4 mg/L of ethyl acetate Thi s highlights the importance 52 OCCUPATIONAL HEALTH: EXPOSURE 35 I 30 ... ,...... 25 ... -0 g 20 q i:i "' ::, 0 15 e LEl 0 0. 10 ... 5 ... 0 111 I I I 0 5 10 15 20 25 30 Day Number Figure 1. Estimated impulse expos ur e to et h y l acetate ( mal e s tud e nt ]. OCCUPATIONAL HEALTH : EXPOSURE 40 I 35 30 ::0B "' -~ 25 "' e!, _q 20 ::, 0 e 5l 15 L0 10 s 0 I I I I I I 111 0 5 10 15 20 25 D ay Number Figure 2. Estimated impulse exposure to ethy l acetate (female student]. Chemi c al Engi n eer in g Edu c at i on

PAGE 55

of a blood test before conducting the distilla tion experiment. Students were then asked to estimate their exposure to ethyl acetate in the qualitative terms of low, medium and high and to estimate the number of hours at these exposure levels on each exposure day. This raw data represents a series of impulses of exposure to ethyl acetate, the height of the impulse being a measure of the perceived intensity of exposure. The unit operations laborato ry en v ironment treated as a local w orkplace can b e v aluable in drained and the solvents returned to the solvent store, marks the end of the occupational health exposure period Students are then advised to have a second blood test for ethyl acetate. Tab l e l shows the ethyl acetate content of t h e blood before and after the distillation experiments. The blood results are most encouraging, with both students having background levels of ethyl acetate at 0.5 mg/L. One might conclude that student awareness of the occupational health study reduced their exposure or that the ven tilation of the laboratory was adequate, or that the interval between exposures was suf ficient for the body to remove any excess ethyl acetate. The body reacts to these impulses by exhal ing ethyl acetate by generating enzymes to con vert ethyl acetate to other substances for rejec tion, or by storing ethyl acetate in fatty tissue. The comp l ex response is similar to the simple impulse response taught in chemical engineer ing mathematics and process control. The process could be made more quantitative if the inhaled dose of ethyl acetate per impulse could be measured This would require per sona] monitoring devices attached to the stu dent and subsequent analysis of the sample tubes These devices are useful in providing an inte gration of the impulse of ethyl acetate. Other methods, which are not practical in a unit op erations laboratory would involve a composi tion measurement of ethyl acetate in the air and a profile of the rate of inhalation of air i ntroducing students to exposure occupa ti onal health, and perceived impressions of exposure. TA B LE 1 Blood Analyses Ethyl Acetate (mg/L) Male Femal e 0.5 1.4 0.5 0 5 OCCUPATIONAL HEALTH GOALS () To use the undergraduate thesis experiment on distillation to introduce students to the concept of medical surveil lance. () To introduce students to the psychological response of exposure limits to inhaled chemicals when introduced to the Material Safety Data Sheets (MSDS). Medical surveillance is introduced to the students through blood tests both before and after the full laboratory period covering the undergraduate thesis. Students conducted a cali bration distillation runs and analyses over a period of four months, with the number of exposure days limited to about thirty days. The period of zero exposure to ethyl acetate between exposure days may have provided time for the body dynamics to remove excess ethyl acetate from the blood. () To obtain essential information on the previous history of exposure to a chemical through a voluntary blood test before the experiment begins. () To monitor the daily exposure to a chemical in a qualitative manner, thus introducing perceived exposures to a chemical and an improved awareness of occupational health. () To introduce ventilation as a key measure in reducing exposure to a chemical. () To have a repeat blood test after the laboratory experiment to ensure that preventive methods for reducing occupation health exposure have been successful. () To expect students to perform better in the occupational health area of chemical engineer ing design and that they have an improved concept of occupational health, both in the workplace and in general. REFERENCES 1. Crumpler, D Ch e mical Crisis, Scribe Publication (1994 ) Figures l and 2 show the estimated impulse exposure to ethyl acetate on the exposure days for a male and a female student, respectively. The data for these figures were ob tained from the student's own impression of the level of exposure as low, medium, or high, and the hours of expo sure. The differences in the two figures is due mainly to the different perceived exposure by each student. The two fig ures provide important qualitative information on the prob lem of estimating exposure without a personal monitor. The chemical engineering student will meet this type of occupational health problem when employed in the work place. The unit operations laboratory environment treated as a local workplace can be valuable in introducing stu dents to exposure, occupational health and perceived impressions of exposure. 2 CD-ROM Micromedia Environmental Health and Safety" The end of the laboratory experiments, with the column Wint e r /999 3 Furzer, RI., and I.A. Furzer, CHEMECA 96, Sydney Sept 30 Oct 2, 4, 69 ( 1996 ) 4 Furzer, I.A. Chem Eng Aust., ChE 21, 22-25 ( 1996) 0 53

PAGE 56

.f3 ... fi_.._1a_b_o ,-,_a_t_o_r ::_y ___ ____ ___,) LABORATORY EXPERIMENT IN BIOCHEMICAL ENGINEERING Ethanol Fermentation ALBERTO COLLI BADINO, JR. CARLOS OsAMu HoKKA Uniu ersi dade F e deral de Sao Carlos Sao Carlos, Brazil T he need for didactic experiments that will prepare our s tudent s for their professional future s, together with the importance of the ethanol indu s try in Brazil led u s to design and construct an experimental b e nchsca le kit for determining the kinetic parameter s related to the ethanol fermentation proces s The experiment s de sig n was ba se d on the principle s that guided construction of the Didactic Laboratory of Chemical and Biochemical Reactions at our university which included a laboratory composed of didactic kits for short experiments prepared by the students that would complement classroom knowled ge. A set of three bench -sca le fermentors was d esig ned and constructed that would support groups of five s tudent s per fermentor. To standardize the experiments, one of them was u sed in two type s of cultivation: the first at low substrate concentration ( = 16 7 gL 1 of glucose) and the A lb ert o Colli B a din o J r., is Associate Profes sor of Chemical Engineering at the Universidade Federal de S Carlos He received his Master 's from the Federal University of S. Carlos in 1991 and his Doctorate in Biochemical Engineer i ng from the State University of S Paulo in 199 7 His research interests are in mass transfer in con ventional fermentation processes rheology of fer mentation broths, and power requirements in con ventional fermentors. Carl o s Osamu Hokka is Associate Professor of Chemical Engineering at the Universidade Fed eral de S. Carlos. He received his Master s from Osaka University in 1976 and his Doctorate in Biochemical Engineering from the State Univer sity of Campinas in 1983 His research interests are in ethanol production in non-conventional con tinuous processes and beta-lactam antibiotic pro duction processes in non-conventional reactors. second at a concentration of 60 gL1 to a llow fitting of traditional kinetic model s without and with inhibition by the product respectively. THEORY We know that in a favorable environment, sim ple sugars ( monosaccharide s) are transformed in ethanol and carbon dioxide (CO 2 ) by the action of yeasts as follows : C6H 1 206 2 CO 2 + (hexose ) ye ast (3 0 C) (carbon dioxid e) 2 C 2 H 5 0H (e thanol) It can be seen that for each mole of hexo se, equal molar amounts of ethanol and carbon dioxide are produced In fermentation processes that sy nthesize primary metabo lites such as ethanol, cell growth and product generation take place s imultaneou s ly. Here the cellular growth and the prod uct sy nthesi s are directly related. Therefore, the ethanol pro duction can be predicted from the cellular growth kinetics. From the hypothesi s that the concentration of cells is a good measure of the enzymatic sys tem responsible for the trans formation of the substrate into product, it is convenient to define the s pecific growth rate ( as I dX =xctt (I) where dX/dt is the variation of the cellular concentration (X) with time (t). Several model s have been proposed to relate the specific growth rate ( ) to the limiting nutrient and inhibitor concen trations. A cla ssic kinetic model relating the specific growth rate to the limiting s ubstrat e was proposed by Monod. < 1 2 ) Working in a continuous process, he obtained the following relationship also applied to batch processes: Cop y ri g ht ChE Divi s ion o f ASEE 1999 54 Chemical Engineering Education

PAGE 57

The e qu ipment al lo wed us to perform sh o rt-duration didactic experime n ts (4.5 and 3.0 hours) using low-cost raw m a teria l s. s ax K 5+S (2) where ax is the maximum specific growth rate achievable when S >> K s, S is the concentration of growth-limiting nutrient, and K s is the saturation constant or the value of the limiting nutrient concentration at which the s pecific growth rate is half of ma x It is known that thi s model is valid for cultivations using low initial limiting substrate concentra tions (S 0 ). In ethanol fermentation, high initial substrate concentra tions (S 0 ) generate high product concentrations (P) that in hibit cellular growth and, consequently, the production of ethanol. Various relationships relating the effect of ethanol concentration (P) to the specific growth rate ( of the organism have been reported P51 One type of relationship that presents simi larity to the non-competitive inhibition in enzyme kinetics ha s been pro posed1 31 for modeling ethanol inhibition of Sa cc haramyces cerevisiae A term accounting for ethanol inhibition is added to the simple Monod kinetic model giving S KP ax K s+S Kp+P (3) where P is the ethanol concentration and K P is the product inhibition constant. From Eq. (3), we see that the higher the ethanol concentration, the higher its negative effect will be on the s pecific growth rate ( ). Still, in a fermentation proces s, the concentration of cells (X) and product (P) can be related to the limiting substrate concentration (S) by the yield coefficients dX X-X 0 Yx s =-=-(4) dS S 0 -S y _dP _P-Po (S) P I S dS S 0 -S MATERIAL AND METHODS : () M i c r o o q:a n is m Sac c harom yces cerevisiae (commercial Fleischmann's baker's yeast) was grown in two different media for ethanol production () Me di a Two different media have been used in the experiments: media l and 2 in ASSAYS 1 and 2, respectively cold water temperature controller s oleno i d v alve 5.0 ; MgSO 4 .7 H 2 O 0.4; yeast extract, 3.0; (NH 4 ) 2 SO 4 1.8; commercial antifoam ( dilution 1: 10 ), 5 drops; pH=4 6; sol ve nt distilled water. () Med iu m 2 (high substrate concentration) in gL 1 : com mercial corn glucose 66.0 ; KH 2 PO 4 5.0; MgSO 4 7 H 2 O, 0.4; yeast extract 3.0; NH 4 Cl 2.5; commercial antifoan (d ilution 1: 10), 5 drops ; pH=4 .6; so lvent distilled water. () Ex p eri m e n ta l Eq uipm ent The fermentors were made of glass adapting 1000-mL kettle -type (Pyrex) recipients with flat bottoms. Fermentor lids were made of technyl (nylon) and adapted to the opening of the recipients. Agitation was accomplished using magnetic s tirrers adapted to the base of the fermentors. The temperature was controlled through so lenoid valves activated by bulb and capillary controllers. Wells for controlling and monitoring temperature, and tubes for heat transfer ("U" tubes ), samp ling inoculation, and gas exit were made of stainles s s teel and connected to the lid s of the fermentors Devices for determining the volume of CO 2 liberated by the fermentation were constructed of PVC (poly vinylchloride) pipes with internal diameters of 10 cm and height s of 100 cm and connected to the gas exits of the fermentors based on equipment proposed by Nilsson et al. 1 6 1 A scheme of the experimental apparatus is shown in Figure l () Ana l ytical Met h o d s Cell concentration was evaluated as dry weight. Broth samples were centrifuged, washed twice, and resuspended with di st illed water. Aliquots were diluted and the absorbance of the suspension was measured at 650 nm with a spectrophotometer (Micronal B-395). The concenso leno id valve hot CO 2 Meter water O ___ r Yo.Iy ... 20 gas exit ---+ drain 40 _---=80 magnetic stirrer v al v e~1 ~1.!~====J I L~ drain () Me dium 1 (low substrate concentra tion) in gL 1 : commercial corn glucose (90% w/w in glucose), 20.0; KH 2 PO 4 Winter 1999 Fig u re 1. Schematic of the experimental apparatus. 55

PAGE 58

tration of the yeast suspension (X) in gL 1 was related to the absorbance according to the following calibration curve: X=0.817 Abs valid for X < 0,45 gL 1 Glucose concentration (S) was determined as total reduction sugars by the colori metric Somogyi methodl7 1 using the same spectrophotom eter. Ethanol concentration (P) was determined by the method of oxidation by potassium dichromate after distillation and indirectly relating to the volume of CO 2 formed. 0 Measurement of C0 1 Volume Produced by Fermenta tion According to the stoichiometry of the microbial reac tion for each mole of consumed glucose, two moles of ethanol and CO 2 are produced. Therefore, if it is possible to measure the amount of CO 2 produced by the reaction, this can be related to the concentration of ethanol present in the fermentation broth. Such a procedure facilitates the didactic experimental routine as ethanol analysis by the method of oxidation by potassium dichromate or other conventional method is time consuming, requiring the supernatant of the centrifuged sample to be distilled and diluted for titration. Measurements of the CO 2 volume produced by the fer mentation process were accomplished in the following man ner. Initially, the pipe was filled with water from the bottom (valve 1) up to the level Yo measured from the top of the pipe, maintaining valve 2 open. Before cultivation was started, valves 1 and 2 were closed. The CO 2 produced left the fermentor through the gas exit tube and entered from the top of the pipe, pushing the water level down. The water level was monitored by a transparent tube connected vertically to the PVC pipe According to hydrostatic principles, in order to maintain constant atmospheric pressure in the head of the fermentor, the water level in a thin tube connected to the ba se of the PVC pipe was controlled manually by draining the water. The number of moles of CO 2 produced during fermenta tion can be calculated at any time by the ideal gas law where n; number of moles of CO 2 evolved at time t=i V; gas volume at time t=i R gas constant [R=82.04 atm cm 3 /(g-mole K)] T gas temperature [K] D internal diameter of the PVC pipe [D=lO cm] (6) Y ; distance from the top of the PVC tube to the water level at time t=i y O Distance from the top of the PVC tube to the water level at t=0 0 Experimental Procedure Two fermentation assays were carried out at 30 C using the culture media previously pre sented. First, the 1000-mL fermentors containing 700 mL of culture media were sterilized in an autoclave at 121 C for thirty minutes. After sterilization, the temperature was main56 tained at 30 C. Then, 100 mL of inoculum activated in a shaker with different cell concentrations (X 0 ) was added, completing 800 mL of initial working volume The inocu lum is activated prior to its addition to the culture medium in order to eliminate the adaptation stage of the microorganism to the culture medium (the "lag" phase of the process). Samples of 10 mL were withdrawn approximately every half hour right after inoculation up to complete depletion of the glucose (end of fermentation) At the moment of sample withdrawal, the distance between the top of the PVC pipe and the water level (y) was measured. Samples were divided into two aliquots of 5 mL for analysis of the concentrations of substrate (S), ethanol (P), and cellular mass (X). The assays carried out at low and high initial substrate concentra tions (S 0 ) were designated as ASSAY 1 and ASSA Y2, re spective ly. Ethanol concentrations were determined analyti cally only in ASSAY 2 to be related to the CO 2 volume produced. In ASSAY 1, the ethanol concentrations were determined indirectly by the CO 2 volume produced. RESULTS AND DISCUSSION The results obtained in ASSAYS 1 and 2 are shown in Table 1. Usually, fermentation processes are relatively too time-consuming to be used in didactic experiments, but stan dardization of the assays provided reasonably short experi ments of 4.5 and 3.0 hours, respectively. From the experimental data of ASSAY 2, it was possible to relate the numbers of moles of ethanol and CO 2 generated by the ethanol fermentation. The number of moles of CO 2 evolved ( n co 2 ) was calculated by Eq. (6) and the number of ethanol moles (ne,hano,) formed was estimated as nethanol = PVMethanol (7) where Pis the ethanol concentration in the broth (in gL 1 ), V is the broth 's volume (in L) and M et h a n o l is the ethanol mo lecular weight (Me,hano1 =46 ). TABLE 1 Experimental Results Obtained in ASSAYS 1 and 2 ASSAY I ASSAY2 time ( h ) S(gL ') x(g L ') P (gL') time(h) S (g L ') x(g L ') P (gL 1 ) 0 0 16.66 1.08 0 00 0 00 60 00 18 .38 0.00 0 5 15 .78 1.09 0.46 0.58 48.25 19.80 2 79 1.0 14 .9 4 1.15 1.03 1.17 35.73 22.5 1 7 18 1.5 12 .97 1.37 1.8 3 1.75 26 60 23 .78 l0.37 2.0 11.19 1.56 2 40 2.33 15 .38 24 .4 3 15.16 2.5 9.68 1.76 3 66 2.92 5.86 25.83 18.35 3.0 6 69 2 06 4 .69 3 .5 3.52 2.52 5.72 4.0 0 .88 2.72 7.10 4 5 O IO 2 84 7 44 Chemical Engineering Education

PAGE 59

Figure 2 illustrates the good linear relationship between values of n et h a n o l and n c o 2 The linear regres s ion of the ex perimental data re s ulted in a slope of 1.006 and a regres s ion coefficient (R 2 ) of 0.994, showing that the theoretical-ex perimental methods, proposed to determine the ethanol con,...., 30 "' 11) ] ..__, N 20 0 0 ON Jl 10 10 20 30 40 nethanol 1 a2 (moles) Fig u re 2. Relationship between the number of moles of ethanol formed (n e ,h ano il and CO 2 evolved ( n c o J during ASSAY 2. .:::-en 18 9 s 0 X t:,. p 15 "t:,. >< 12 6. t:,. 'i:, fjq' 9 r-: ._:; 6 3 3 0 4 5 time (h) Fig ur e 3. Fit of Monad's model to the experimental data of ASSAY 1. 80 ~-~--~--~-~--~-~ 30 S o X t:. P 60 20 time (h) 2 >< 20 '"._, 10 Figure 4 Fit of the Aiba et al ., model 131 to experimental data of ASSAY 2 Wint e r 1999 centration indirectly in the broth by the volume of CO 2 produced generated very good results. Yield coefficient s Y XJs and Y r i s were determined by ex perimental data (Table l) as being: Y XJs =0.11-Y r is =0.46 (ASSAY 1) and Y XJs =0.14-Yr s=0.35 (ASSAY2), respec tively. Experimental results were analyzed based on two classical kinetic models Initially Monad s model (Eq 2) was fitted to the experimental data of ASSAY 1. The specific growth rate ( was estimated by Eq. (1) where dX/dt was calcu lated from the polynomial equation fitted to the curve X(t) The kinetic parameters, max =0.32 h 1 and K 5 =0.63 gL 1 were obtained by nonlinear regression of and S values according to Eq (2) using Marquardt's algorithm. In the same way, the kinetic model proposed by Aiba, et a1. r 3 J was fitted to the experimental values of ASSAY 2, considering the kinetic parameters ax and K 5 obtained previously (ASSAY 1) The product inhibition constant (KP) was estimated as being 6.29. Figures 3 and 4 illustrate the good fits of the kinetic models ofMonod and Aiba, et al., to the experimental values of the assays carried out at low and high initial substrate concentration ASSAY 1 and ASSAY 2, respectively. A fourth-order Runge-Kutta technique was applied to simulate the predicted curves of S, X and P as a function of time. For comparison, Table 2 shows the values of the kinetic parameters evaluated from the data of other workers. l 8 1 It can be seen that the values of the kinetic parameters obtained by the present work are within the range of values encountered in the literature showing that the proposed methodology can be of great value for educational and research purposes CONCLUSIONS The equipment allowed us to perform short-duration di dactic experiments (4.5 and 3.0 hours) using low-cost raw materials. The relationship of ethanol/CO 2 produced by the fermentation was very close to unity (1.006), showing the precision of the equipment. Use of the CO 2 meter in didactic Continued on page 70. TABLE2 Kinetic Parameters E g amb e rdi ev and Ieru s alim s kii' 81 28 Aiba and Shod a 181 3 0 Pironti 1 81 C ys ewsk.i 1 81 Bazua and Wilke 181 Hoppe and Han s ford 181 Thi s work 3 0 35 35 30 30 0 .3 1 0.4 3 0.26 0 58 0 64 0 64 0 32 15 5 4 9 0 24 3 3 0 63 Y ,~ ( ) 20 6 0.39 55 0.35 13 7 0.47 5.0 0.44 40 0.52 5 2 0.43 6 29 0.35-0.46 57

PAGE 60

.tA .. fi-3._c l a s s _r_ o o m __ _____ ___.) PERMEATION OF GASES IN ASYMMETRIC CERAMIC MEMBRANES CARLOS FINOL, JOAQUIN CORONAS University of Zarago za 50009 Zarago za, Spa in I ntere s t in and applications of inorganic membranes h ave been growing exponentially for the la st fifteen years. 111 Inorganic membrane s can be classified into two catego ries: den se and porous. Den se membrane s are dominated by pall a dium and it s alloys ( met a l m e mbrane s) and so lid elec trolyte membrane s (co mmonly sim ple or complex oxides or oxide-solid so lution s) Porou s membrane s are made of ox ides carbon, glass, metal s, zeolites, etc The most unu s ual commercial porous ceramic membranes h ave an asymmetric structure consisting of a s upport layer (ge nerally, alpha alumina) with large pore s and a se paration la ye r made of a d iffe rent material (g amma-alumina zirconia sil ica etc.) that controls the permeation flux. The industrial development of inorganic membranes started in the 1940s with the purification of nuclear fuels Thi s proce ss is ba se d on the se paration ofU 2 3 5 FJU 238 F 6 by Knud se n d iff u s ion Knud se n diffu sio n occurs when the mean-fre e path of the gas molecule s is much larger than the average pore dimensions of a porou s material through which the molecules diffu se, and permeate will be enriched in the molecule of the lower molecular weight. In thi s short article, we will describe an easy experiment that help s to understand the transport through ceramic mem brane s when Knudsen diffu sio n occurs THEORY In general, transport through porous ceramic membrane s can be related to the pore diameter dp accor ding to IUPAC definition s : 1 2 1 macropore s with dp > 50 nm where basically viscous flow (i n this case, no se paration of mixtures is pos sible) and Knudsen diffu s ion occur; mesopores with dp be tween 2 and 50 nm where Knudsen diffu sio n and multilayer diffusion/capillary condensation take place; and micropore s with dp < 2 nm, where molecular sievi ng effec ts can be expected Surface-multilayer diffu s ion and capillary con densation are achieved when the permeating molecule pref erentially absorbs on the membrane pore s or condenses within the pore s due to capillary forces re s pectively Both trans port mechani s m s which allow the se paration of mixtures with very hi g h se lecti vities, a re especially important at relatively low temperature s and with sma ll pore s, mesopore s and even micropore s. When permeance s through mesoporou s membrane s are studied for permanent gas or vapors at low relative pre s ures the transport mechani s m is controlled b y the Knud se n diffusion Sometimes laminar flow (visco us flow at laminar regime) can appear if the membrane has defects ofmacropore size, decrea s ing the separation power of the membrane Con s idering both the laminar and Knud se n flow contributions the transport eq uation can be written as Carlos F in o/ is a PhD student in the Depart ment of Chemical and Environmental Engi neering at the University of Zaragoza His research interests are in using special reac tors for partial oxidation of hydrocarbons. Joaq uin C o r onas is Assistant Professor in En vironmental Engineering at the University of Zaragoza He received his PhD in chemistry in 1995 from Zaragoza Univers i ty His research interests are in developing membranes for sepa ration and ceramic membrane reactors Co p yr i g ht C h E Di v i s i o n of A S EE 1999 58 Ch e mical En g in ee rin g Edu c ation

PAGE 61

F=F+FP=i er+ Erp H 2 T K ov 3 8 L-cRT where FT tota l permeance [mol/(m 2 s Pa )] F K is the Knud se n contribution (!) F 0 v factor that multiplied b y P gives the l ami nar contribution to the total permeance P average pressure across the membrane [Pa] E porosity of the membrane r pore radius [m] L membrane thickness [ml -c tortuo s ity of the medium R gas constant [8 .3 14 m 3 Pa/ ( mol K ) ] T absolute temperature [Kl viscosity of the gas [Pa s] A plot of FT vs P allows one to estimate the relative contri butions of both transport mechani s m s: the higher the slope of such a plot the higher the laminar contribution and vice versa. On the other hand, for a certain membrane material if the laminar contribution is not important, or P=0 it is po s sible to calculate the Kn u dsen separation factor for two gases, A and B, by using the first term of Eq. (1) aA / B = F~ = FK MA (2) Since Knudsen diffusion occurs when the mean-free-path of the molecules O.) is much larger than the mean pore radius (r) of the membrane the Knud se n number defined by the equation Kn=~= 16 r SnP 2M (3) is a convenient dimensionless group to quickly estimate if Knudsen diffusion will likely be the dominant transport mechanism When Kn>l, the molecule s collide with the pore wall much more often than with each other. EXPERIMENTAL The membrane used was a commercial ceramic tube 10 cm long and 0.7 cm i.d. purchased at SCT (Societe des Ceramique s Techniques, a s ubsidiary of US Filter). The ceramic tube had an asymmetric st ructure consisting of a support layer of a -alumina and an inner layer of y alumina with 5-nm diameter pore s. The ends of the mem brane were non-permeable, restricting the permeation area to a length of 5 cm. Figure 1 shows the permeation system. The membrane was sealed in a stainless-steel module by silicone o-rings, and the pressure at the feed site (o r retentate ) was measured by a pressure transducer. At the permeate side, the perme ation flux was measured using a bubble flowmeter and, finally an electric furnace allowed modification of the memWint e r 1999 A very simple system can be used with a commercial ceramic membrane of y -alumina to measure permeances of single gases (e.g., N 2 and He) at several total pressures and temperatures. brane temperature The permeate side was open to the atmo sp here and permeate pre ss ure was always the atmospheric one. The permeance Fin mol/(m 2 s Pa ) was calculated as where Q molar flux [mol/s] F=_g_ ALiP (4) A permeation area referred to the internal side of the ceramic tube (l.l xl0 3 m 2 ) LiP pressure drop [bar] between retentate and permeate sides RESU LT S AND DISCUSSIO N The variation of He and N 2 single-gas permeances vs the total pressure (calculated as [P 1 + P 2 ]/2 where P 1 and P 2 are the pressur es at the permeate and retentate sides, respec tively) through an alumina membrane was s tudied at 298 an d 473K. Figure 2 s how s that the He permeances at 298 and 473K did not change with total pressure and the same can be sta ted for the N 2 permeances (F igure 3), at least in the pressure range tested, which means that the viscous con tribution is not important and in the equation (Eq. 1) that governs tran s port through the membrane, the second term Needle valve C: QJ 00 g ;z E .:! .; :t Pressure mea s urement Oven Temperature control : .. 11~ 111 Membrane module Bubble flowmeter Figure 1. Experimental system. 59

PAGE 62


PAGE 63

Discontinuitie s Continued from pag e 25. intrinsic volume averaged concentration of s pecie s A, molelm 3 c Ay spatial deviation concentration molelm 3 c P constant pre s sure heat c apacity J/kgK c speed of s ound mi s D dispersion ten s or m 2 l s F force acting on a particle N g gravity vector m/ s 2 k thermal conductivity JlmsK L vi s cou s length m L P inertial length m L length associated with the pressure change, ~p m M u / c Mach number m mass of a particle, kg n number of moles n unit normal vector n yo unit normal vector direct e d from the y pha s e toward the o phase p pre s sure Nlm 2 P m pressure for an incompre s sible flow Nlm 2 r po s ition vector m R gas constant Nm/mole K Re pu 0 L Reynolds number R A molar rate of production of s pecie s A owin g to homoge neou s reaction molelm 3 s R Ay molar rate of produ c tion of s pecie s A in th e y pha s e owing to homogeneou s reaction molelm 3 s T temperature, K T rn temperature for an incompres s ible flow K time s t (n) s tre ss vector Nln i2 T stres s tensor Nlm 2 u 0 characteri s tic velo c ity mi s v ma ss average v elocity v ector mi s v 111 velocity for an incompres s ible flow mis v A velocity of s p e cie s A mi s v Ay v elocit y of s pecie s A in th e y ph as e, mi s v y mass average velocity in the y pha s e mi s ( v y ) superficial mas s average velocity in the y phase mi s ( v y) Y intrinsic mass average v elocity in the y phase mis v y spatial deviation velocity mi s V volume m 3 o/ averaging volume m 3 Y y (x t) volume of the y pha s e contained in the a veraging volume m 3 w arbitrary velocity mis x position vector l ocating the centroid of an averaging volume, m Winter /999 Gr ee k L e tt e r s E y Y y (x t)l o/, v olume fraction of the y phase '}._ unit v ector fluid v i sc o s ity N s lm 2 p ma s s den s ity kg/m 3 REFERENCES 1. Whitaker S. The Dev e lopment of Fluid Mechanics in Chemical Engineering ," p 47 in On e Hundred Y e ar s of Ch e mi c al En gi n ee rin g, edited by N. Peppas Kluw e r Aca demic Publishers Dordrecht ( 1989 ) 2 Gre s ho P M ., and R.L Sani In c ompr e ssibl e Flow and the Finit e El e m e nt M e thod John Wiley & Sons Inc., New York NY ( 1998 ) 3 Birkhoff G ., H y dr o d yn ami cs: A Stud y in Logi c, Fa c t and Similitud e, Princeton University Press Princeton NJ ( 1960 ) 4. Whitaker S. Fundam e ntal Principles of Heat Transfer R.E. Krieger Publishing Company Malabar FL ( 1983 ) 5 Bird R.B. W.E St e wart, and E.N Lightfoot, Transport Phenom e na John Wil ey & Sons Inc. New York, NY ( 1960 ) 6. Whitaker S. A Simple Geometrical Derivation of the Spa tial Averaging Theorem Ch e m Eng Ed. 19 18 (1985 ) 7 Truesdell C ., E ss a y s in th e Hi s to ry of M ec hanics Springer Verlag New York NY ( 1968 ) 8 Aris R. V ec tor s, T e n s or s, and th e Basi c Equations of Fluid M e chani cs Prentice-Hall, Englewood C l iffs NJ ( 1962 ) 9. Whitaker S Introdu c tion to Fluid Mechani c s R.E. Krieger Publishing Company Malabar, FL ( 1981 ) 10 Pucciani O F ., and J. Hamel Langu e e t Langage, 4th ed., Holt Rinehart and Winston New York NY ( 1983 ) 11. Stein S.K., and A. Barcellos, Cal c ulus and Analyti c G e om e t ry, McGraw-Hill Inc. New York NY (1992 ) 12 Majda A ., C o mpr es sibl e Fluid Flow and S y st e ms ofCons e u ation Law s in S e v e ral Spa ce Variable s, Springer-Verlag N e w York NY ( 1984 ) 13 Whitaker S ., L e vels of Simplification : Th e U s e of Assump tions Restrictions and Constraints in Engineering Analy sis ," Ch e m. En g Ed ., 2 2, 104 ( 1988 ) 14. Whitaker S. Laws of Continuum Physics for Singl e -Phase Single-Component Systems ," in Handbo o k of Mult i phase S ys t e m s, edited by G H e tsroni Hemisphere Publishing Cor poration New York, NY ( 1982 ) 15 Anderson T B. and R. Jackson A Fluid Mechanical De s cription of Fluidiz e d Beds ," Ind Eng. Ch e m Fund ., 6 527 ( 1967 ) 16 Marle C M. Ecoulement s monophasique en Milieu Poreux R e u. Inst. Fran r; ai s du P e trol e, 2 2 ( 10 ), 1471 ( 1967 ) 17. Slattery J C ., Flow ofViscoelastic Fluids Through Porous Media ," AIChE J ., 13, 1066 ( 1967 ) 1 8 Whitaker S ., Diffusion and Dispersion in Porous M e dia ," AIChE J. 13 420 ( 1967 ) 19 Whitaker S ., Th e M e thod of Volum e A ve raging, Kluwer Academic Publishers Dordrecht ( 1999 ) 20. Gray W.G ., A Derivation of the Equations for Multiphase Transport ," Ch e m. En g Sci. 3 0 229 ( 1975 ) 21. Carbonell R.G. and S Whitaker Dispersion in Pulsed Systems II: Theoretical Developments for Passive Disper s ion in Porous Medi a," Ch e m En g. S c i ., 38 1795 ( 1983 ) 22 Paine M A. R.G Carbonell and S. Whitaker, Di s persion in Pulsed Systems I : Heterogeneous Reaction and Revers ible Adsorption in Capillary Tub e s ," Ch e m. Eng S c i. 38 1781 ( 1983 ) 23 Quintard, M ., and S. Whitaker Convection, Dispersion and Interfacial Transport of Contaminants: Homogeneous Porous Media ," Adu. Water Resour. 17 221 ( 1994 ) 0 61

PAGE 64

f.A ... fi 111111 311-c_u_rr_i_c_u_l_u_m ________ __,) A JOINT CHEMICAL/ELECTRICAL ENGINEERING COURSE IN ADVANCED DIGITAL PROCESS CONTROL JOSEPH J. FEELEY, LOUIS L. EDWARDS University of Idaho Moscow, ID 83843 T his article describes a course in the digital control of industrial processes jointly offered by the Depart ments of Chemical Engineering and Electrical Engi n eering at the University of Idaho. The course grew out of a perceived need for engineers with the a bility to design multi input multi-output (MIMO) digital controllers for indu stria l processes, especially in the pulp a nd paper indu s try Evi dence for thi s need lies in the numerou s opportunities for improved process performance offered by advanced multi variable digital control technique s. The primary goals of the course are to Teach students how to design MIMO controllers for industrial processes Give students laborator y experience with represen tative industrial processes and control systems Introdu ce students to the use of state -of-th e-art comp ut er-aided design software for control-system design Pr ovide an opportunity for chemical and electrical engineering students to work together in interdisci plinary teams The following sectio ns de scri b e how the course attempts to achieve these goals, highlights so me of the distinctive fea ture s of the course, and di sc u sses some plans for the future. COURSE DESCRIPTION Digital Proce ss Control is a three se me ster -credit hour course jointly offered by the c hemic a l and electrical engi n eeri n g departments It is a required course for chemical engineering (C hemE ) seniors and an e lecti ve course that satisfies a breadth requirement in the co ntrolsystems area for e l ectrical engineering (EE) se nior s. ChemE stude nt s h ave previously had a co u rse in chemical process control u sing classical control-system design methods. The prerequisite for EE st udent s is a junior-level course in sig n a l s a nd sys tems. Some EE students, however have had a prior course in classical co ntrol-s ystem design as applied, primarily, to e l ectromechanical systems All of the EE st ud e nt s and many of the ChemE stude nt s have had a prior course in linear algebra or have had so me exposure to vectors and matrices in previous courses. Computer-aided de s ign is an essential part of the course, and all st udent s have had prior experience with Matlab, 111 the main computer program u se d in the course. The course is taken by about twenty-five ChemE st udent s and fifteen EE st udent s each year and is jointly taught by a faculty member from each department. Joseph J Feeley is Associate Professor of Electr ical Engineering at the Uni versity of Idaho. He received his BS from the New Jersey Insti tute of Technology and his MS and PhD de grees from the University of Idaho all in electri cal engineering He is cu rrentl y writing a text book on digital process control Louis L. Edwards has been Professor of Chemical Engineering at the University of Idaho since 1971 He teaches process con trol and his research focuses on simulation of pulp and paper processes. He has published over fifty papers and has consulted with more than thirty pulpand paper-related compa nies worldwide Co p yr i g ht C h E D ivis i o n of ASEE 1 999 62 Ch e mi c al Engin ee rin g Edu c ation

PAGE 65

The course is organized around three major laboratory experiments of progressive difficulty Each of the experi ments is a laboratory-scale simulation representative of a typical industrial process. The lecture material and the laboratories are coordinated so that when a lecture se quence describing a new design technique is completed it is followed immediately by a laboratory experiment using that technique. The choice of an appropriate design method or methods for this class is not easy Current practice in the process control community is focused on some form of internal model control (IMC), r 2 31 of ten model predictive control (MPC) or dynamic matrix control (DMC). On the other hand, methods based on the linear quadratic regulator Inlet flow Stabilization tank laboratory experiment. The model is linearized at an operat ing point and discretized at a selected sampling rate to form the basis for the contro ller design. The control problem is presented as an optimization problem, trading off control effort for state performance. The dynamic programming ap proach is used to develop the Riccati equation and its steady state solution is used to find the optimal DLQR gain. By beginning with this scalar example students get a good understanding of the basic DLQR approach before moving to the multi variable problem with the additional complexity of vector-matrix notation, studied later. The effect of state variable feedback in changing the system time constant is also clearly seen in this simple example Stu dents solve this control prob lem by hand and then use (LQR) 1 41 and its extensions to incorporate state estimation via linear quadratic estimation (LQE) recover stability mar gins via loop-transfer recov ery (LTR) and design for ro bustness directly through H ~ and -synthesis rsi have been more popular in the electro mechanical control commu nity. The distinctions are not of course that clear Bio reactor Matlab to verify their calcu lations. Simulink is then used to simulate the continu ous-time nonlinear process model under DLQR control. Vent Outlet flow ~---~----Valve ,__ __ Measured level Digital Controller Actuator Experiment 1 Control signal Figure 1. Experiment 2 : a simulated bioreactor system consisting of two cascaded tanks. The first experiment is a simple waterl evel control problem that illustrates the application of DLQR state variable feedback control and there are many successful LQR applications in the process industryf 61 and many PID applications in the electro-mechanical area. Pl A compromise of sorts has been reached by using the discrete-time version of the linear quadratic regulator (DLQR) as the main design tool used and pointing out the relation ships among DLQR DLQG, PID, and MPC as the occasions arise. DLQR is used because of its generality, the availabil ity of Matlab design tools and the familiarity of most of the students with Matlab, and the simplicity of the resulting control algorithm. Once a suitable process model is devel oped the design can be efficiently completed in an interac tive computer session using Matlab Simulink, 181 and the Control Systems ToolboxJ 9 1 Students quickly learn the rela tionship between performance index weightings on the one hand and state performance and control effort on the other hand, and are able to find control gains that yield satis factory system performance. After an introduction to sampling and development of a discrete-time model, the DLQR method is introduced through a scalar single-input single-output ( SISO) example. The ex ample concerns a simple nonlinear process model associated with the liquid-level control problem addressed in the first Wint e r 1999 methods to a first-order system It allows the chemical engi neering students to become familiar with the notion of sam pling and digital-control-system hardware and familiarizes the electrical engineering students with the relatively long time scale and nonlinearity of process systems. The relation ship of scalar DLQR control to traditional proportional con trol is pointed out, and the need for integral augmentation to eliminate steady-state error is clearly illustrated. The inte grator-augmented DLQR controller is then shown to be equivalent to a traditional PI controller. Experiment 2 The second experiment involves a second-order process consisting of two cascaded tanks simulating a bioreactor system. The objective of this experiment is to control the level in the second tank, in spite of disturbances caused by a fluctuating flow from the first tank into the bottom of the second tank Both tank levels are measured so that full state feedback can be used. The control input is the position of an outlet valve on the second tank. A schematic of the experi mental setup is shown in Figure 1 The system is modeled as a two-state discrete-time SISO process and the vector-ma trix form of the state and output equations is introduced. The controller for this process is designed using the integrator6 3

PAGE 66

augmented DLQR method 1101 extended to accommodate level set-point changes and zero steady-state error. The necessity of using vector-matrix notation and computer-aided design tools begins to become evident to students when solving this relatively simple second-order problem. The relationship of this simple DLQR controller to the classical PID controller is also examined. Experiment 3 The third experiment requires the design of a two-input/ two-output controller to simultaneously control the water level and total head in a laboratory-scale model of a paper machine head box. A schematic diagram of the process is shown in Figure 2 and details of its construction are avail able from the second author. Zero steady-state error and set point control are required for both controlled variables. This is the most challenging of the contro l prob lems and involves four Air inlet flow A ir STUDENT REACTION Student reaction to the course has been quite favorable. The ability to test controller designs in the laboratory is men tioned by most students as the most positive feature of the course Many students comment on the power of the com puter-aided design tools to facilitate control system design A smaller number view the computer approach as "magic." DISTINCTIVE FEATURES OF THE COURSE This course includes a number of distinctive features The most obvious is the combination of chemical and electrical engineering students in a joint class. This helps bridge the disciplinary gap and teaches the students how to work with specialists in another field on challenging interdisciplinary Water level measurement states, two inputs and two outputs. Nevertheless, most students are now familiar with the DLQR design-it eration procedure and are able to complete the con troller design with little dif ficulty. In the laboratory, students are encouraged to attempt this MIMO control manually to fully appreci ate the difficulty of the task and the power of the DLQR Figure 2. Experiment 3: a laboratory-scale model of a papermachine headbox problems The student to-student learning pro cess generally involves the ChemE students teaching the EE students about process dynamics and modeling, while the EE students take the lead on computer and digital controller topics The fact that the class is jointly taught by a faculty mem ber from each department also gives the students complementary views of a given control or mod eling issue. Both instrucdesign method. The control design is carried out on a full state feedback basis COURSE STRUCTURE Daily homework assignments are focused on the upcom ing laboratory experiment. Each assignment addresses a dif ferent phase of the design effort with the last assignment before a given lab producing the controller algorithm re quired for that experiment. Student performance is assessed through the homework, two one-hour tests, a two-hour final exam, and reports on each of the laboratory experiments The lab reports are a crucial part of the course requirements and account for about one-third of the course grade Lab groups generally consist of three students two from chemical engineering and one from electrical engineering. Students pick their own lab partners but "bonus points" are awarded to interdisciplinary lab groups to encourage interaction between the electrical and chemical engineering students Individual lab reports in the form of a technical memo with supporting appendices are required of each student. 64 tors attend each class, with the EE instructor primarily addressing the control theory and design topics and the ChemE instructor focusing on process modeling and the relationship of the labs to actual industrial processes. A second key feature of the course is that it approaches the digital control-design problem from the "direct digital de sign" perspective and uses the DLQR design method as the basic design approach This approach is taken for a number of reasons. First, the DLQR method approaches controller design as an optimization problem, trading off control effort for state performance. This is a crucial concept that is often overlooked, or addressed after the fact, in other design meth ods. Second currently available CAD tools such as Matlab, Simulink, and associated toolboxes simplify computations so that students can focus on design objectives and not become bogged down in mathematical complexities Stu dents are taught the advantages of testing controller designs by nonlinear simulations on a desktop computer before try ing them out on the actual process. Third, the DLQR ap proach seems more direct in that a good design can be accomplished almost entirely in the discrete-time domain, Ch e mi c al Engin ee rin g Education

PAGE 67

avoiding, for the most part the complexities introduced by transform mathematics. Finally we realize that s tudents com pleting this course will not be experts in DLQR design methods. They will, however have had a practical introduc tion to the topic that will alert them to the power and poten tial of advanced control techniques They will also be aware of some of the relationships among three control methods representative of the electro-mechanical, process and tradi tional control communities, i.e., DLQR, MPC and PID The final important aspect of the course is the laboratories and their integration with the lecture material In a sense, the course is driven by the laboratory experiments and the stu dents are motivated to master the lecture material so that they can perform the laboratory experiments satisfactorily. Homework assignments are carefully chosen so that satisfac tory comp l etion of the last assignment before a lab should fully prepare the student to carry out the experiment properly PLANS FOR FUTURE DEVELOPMENT A number of plans are underway for further development of the course. New laboratory experiments are always under consideration. We plan to include an experiment involving system identification to develop an empirical model of the headbox process next year. Longer-range plans include add ing a t h ird i n put and a third output to the headbox experi ment and developing a new experiment using a distillation column as the process to be controlled. The lecture material is also being continually refined and updated to find more logical and consistent ways to introduce this relatively ad vanced material to students at this stage in their development. More emphasis will be given to IMC methods such as MPC and DMC in future versions of the course The course notes currently being used as the course text are being readied for pub l ication and should be submitted to a publisher this year. ACKNOWLEDGMENT The authors are very grateful to Barry King for construct ing and maintaining the laboratory experiments and for his assistance with the students in the laboratory REFERENCES 1. Matlab: High Performance Numeric Computation and Visu alization Sofrware, The Mathworks Inc ., Natick MA ( 1993 ) 2 Garcia C E ., D .M. Prett and M. Morari Model Predictive Control: Theory and Practic e, A Survey ," Automatica, 25 335 ( 1989 ) 3. Morari M. and E. Zafiriou Robust Pro cess Control, PTR Prentice Hall Englewood Cliffs NJ ( 1989 ) 4. Kuo, B C. Digital Control S yste ms Holt, Rinehart, and Winston, Inc. New York, NY ( 1980 ) 5. Zhou, K. J.C Doyle, and K. Glover Robu st and Optimal Control, Prentice Hall, Upper Saddle River NJ ( 1996 ) 6 Ramirez W F ., Proc ess Id entification and Control, Academic Press, Boston MA ( 1994 ) 7 Franklin G F J D. Powell and A. Emami-Naeini, F eed back Control of D ynamic Sy s t e ms Addison-Wesley ReadWinter 1999 ing MA ( 1994 ) 8. Simulink: D y nami c S ys t e m Simulation Softwar e, The Mathworks, Inc. Natick MA ( 199 3) 9 Control S ys t e ms Toolbo x, The Mathworks Inc ., Natick MA ( 199 3) 10. Kwakernaak H. and R. Sivan Lin ear Optimal Control Sys tem s, Wiley-Interscience, New York NY ( 1972). 0 BOO K R EV I EW: Nume r ic Co mputatio n Continued from page 11 The next three chapters are devoted to function interpola tion and differentiation (C hapter 6), numerical integration ( Chapter 7) and approximation of functions, lines and sur faces (Chapter 8). In the final chapters, the author discusses the solution of ordinary differential equations and initial value problems (Chapter 9), boundary-value, eigenvalue, and free-boundary problem s (Chapter 10), and finite-differ ence methods for partial differential equations (Chapter 11 ). Finally the author provide s four appendices. In Appendix A, he provide s a brief calculus refresher with a table of various series approximations to a variety of functions. In Appendix B he discus ses orthogonal polynomials Again, he includes a number of tables of various orthogonal polyno mials, including Legendre Chebyshev, Jacobi, Radau, Lobatto Laguerre Hermite and Gram and their properties. Developing the ability to proficiently use the computer and interface with the operating system i s one of the main hurdles often encountered by students in a numerical meth ods course. Appendix C provides thi s essential background information, including an introduction to the c-shell to the unix file syste m to the vi editor and to the compilation and linking of typical program s. Finally, Appendix D provides a Fortran primer, while an index to the publicly available program s discussed above is provided in Appendix E Although they are covered briefly in Chapter 10, I would have liked additional coverage of weighted residual tech nique s. These methods provide a powerful tool for the solu tion of ordinary differential equations and provide the basis of finite element methods for the solution of partial differen tial equations. In devoting only seventeen pages to this topic the author has given the students only a brief glimpse of the power of s uch methods and may leave students thinking that they are not as useful as others that are covered in much more detail. Further, as i s often the case, the coverage of technique s for solving partial differential equations is quite limited. I assume that thi s is because the author expects this materials to be covered in a later course. In general Pozrikidi s ha s met his goals and has produced a usable text in which he covers the fundamentals of numeri cal methods while at the same time enab l ing t h e reader to understand how to use the various techniques to solve physi cal problems in science and engineering. 0 65

PAGE 68

.ta_..51111iiii3.._1a ____ b_:o_r,..:.a.:..:t..=o..:.r~y-------) INTRODUCING PROCESS-DESIGN ELEMENTS IN THE UNIT OPERATIONS LAB CHRISTINE L. McCALLUM, L. ANTONIO ESTEVEZ* Cornell University Ithaca, NY 14853-5201 0 ne of the main problems of the classical chemical engineering or unit operations laboratory is the dan ger of its becoming repetitive and boring This pa per will address that problem and show how a simple labora tory experiment can be turned into an interesting experience for both the students and the instructor by giving it a slight flavor of process design This was accomplished by adding to the requirements of the final lab report the solution of a design problem related to the experiment done in the lab The problem combined scale-up elements and changes to the material 's chemical nature, or to the physical characteristics of the laboratory setup. This paper describes our approach to a classical stripping experiment where air was u se d to remove toluene from water in a packed column. Although the approach is de scribed in t h e context of this specific experiment, it can be applied to any classical experiment. THE UNIT OPERATIONS LAB AT CORNELL To better understand the approach presented in this paper, we will first give a brief description of how o u r lab is organized The senior class is divided into five groups (Mon day Tuesday, Wednesday Thursday and Friday), with four teams in each group. Six experiments are carried out in four rotations During the first week, all groups work on a short, introductory experiment. Then the first rotation starts on the second week and l asts four weeks, as do all rotations: a rotation involves preparation experiments (2 weeks), and a report. (The report on one rotation and the preparation in next rotatio n may overlap.) Two shorter experi m ents are done in one rotation. The stripping experiment (described in thi s paper) is done in one full rotation The Unit Op erations Laboratory at Cornell has a page on the world Pr esent Address : University of Pu e rto Rico Ma yaguez, Pu e rt o R ico 00681-9046 wide web, currently at http : // s peedrcr.cheme.comell.edu/Uo/ that can be consulted for additional details on the course organization ABOUT THE EXPERIMENT The essence of this experiment is to remove a trace amount of pollutant (a volatile organic compound) from water by stripping it with air The experimental apparatus and basic operating procedure are as follows. A ppara t u s In this experiment, air and a toluene-con tami n ated water stream are brought into counter-current con tact in a stripping column The experimental apparatus i s shown in Figure 1. The column is 15 cm (0.5') in diameter packed with 5/8 plastic Pall rings The approximate pack ing h eig h t is 1.15 m (3.75'). A 0 5-m 3 (135-gal) tank with an electric mixer i s used to homogenize the feed supply. Rota meter s are used to read liquid and air-flow rates, and the thermocouple s are u se d to read tank and column tempera tures. The liquid feed flows by gravity at low flow rates and is driven by a centrifugal pump at higher flow rates. Sample s are collected at the water-stream inlet and outlet and ana lyzed using a Gow-Mac gas chromatograph equip-ped with a flame ionization detector and Spectra-Physic s integrator. Chr i stine L. McCal/um received her MS in chemical engineering at Cornell University where she conducted research in the area of molecular simula tion and thermodynamics She received her BS from Bucknell Universit y, also in chemical engineering. She currently works for Intel Corporation in Ph oenix Arizona L. Anton i o Estevez received his PhD from the University of California Davis H e holds a BS degree from the University of Santiago Chile and a M S from the Central University of Ve n ezuela He has been on the faculty of the U niversity of Puerto R ico since 1987 having previously taught at the University of Santiago Chile and the Simon Bolivar University in Caracas H e was on sabbatical leave at Cornell University during the academic year 1996-97 H is research in t erests are supercritical fluids separation pro cesses and multiphase reactors. Copyright ChE Division of ASEE 1999 66 Chemical Engineering Edu cation

PAGE 69

Procedure The tolu ene/water feed is prepared at least one hour before the lab begins by filling the feed tank with water and adding an amount of tolu ene to produce the desired concentration The mixer runs until the lab period begins After verifying proper valve settings air is allowed to flow through the system by opening valve V 4 and its flow rate is regulated with valve V3 The water is then intro duced by opening valve V2. A water seal is main tained by keeping the wa ter depth at the bottom of the column within 3 to 5 cm Steady-state condi tions are reached after about five minutes. Water s amples are collected at ports SI and S2 (drain pipe) in glass vials. Corks wrapped with aluminum foil are used to seal the samples, which are then la beled and stored on ice. The samples are analyzed using the gas chromato graph. Pressure drop gauge pressure at the bot tom of the column, and column temperature are also measured. THE NEW APPROACH As explained previously the students have four weeks to complete the ex periment and submit a fi nal report. In the first week the TA meets with the group for a brief ex planation of the experi ment's objectives and op eration of the apparatus A memo with the require ments is distributed at this Win te r 1999 This paper presents a simple way to enhance the learning experience in the unit operations laboratory by adding a related design problem to the T lab requirements. Vl V2 Figure 1. Apparatus for the stripping e xperim e nt. TABLE 1 Design Problem Clearwaters and A s sociate s is a company that produce s bottled drinking water on a large scale. In their process they chlorinate their source u s ing standard processes and therefore the level of chloroform (and of other trihalometh a ne s, THMs ) goe s beyond specification s. They are considering reducing the THM le v el by stripping with air in a packed column. The water after chlorina tion ma y c ontain up to 400 ppb ( by mole ) of THM, 30 % of which i s chloroform 70 % bromodichloromethane and negli gible amounts of bromoform and dibromochloromethane. Their typical daily production i s about 60 000 5-gal bottles A s the engineering team of Clearwaters and A ss ociates your as s ign ment is to design the THM-removal system using stripping. They plan to use a column packed with 5/8 plastic Pall rings The final THM level must be le s s than 50 ppb (by mole). T y pically the operation wiU take place at 77 F. meeting The students are also referred to the experiment s reference manual which is available on the world wide web at http : //cheme. cornell.edu/ ~lae/432/stripping.htm. Using the traditional ap proach the students would figure out what variables to measure to determine the overall mass-transfer coef ficient and would probably present some plots showing the effect of say, flow rates on mass-transfer coefficient and on the height of a trans fer unit. With this new ap proach, the students must also solve a design problem and include the solution in their final lab report. In this particular design problem, the s tudents use the infor mation collected in the lab to compute the height of the column. In addition to scale up they must also consider a difference in the chemical composition/nature of the inlet stream or the size/type of the column packing. The students need to determine whether or not that differ ence is important, and if so, how to account for it. A SAMPLE DESIGN PROBLEM Table I shows a sample design problem given to the students in one of the fall 1996 rotations. Each team was responsible for provid ing a final design for the company. This task required determination of the pack ing height and the column diameter. The students were given guidelines to help achieve this goal, and in this particular case they were re f erred to Velazquez and 67

PAGE 70

Estevez 111 to find the needed THM's properties. Addition ally, they were asked to s ubmit three piece s of documenta tion: a pre-experimentation memo a preliminary-calcula tions memo and a final report. Each of these parts had a specific objective and helped break up the student workload during the four-week lab period. The three parts are dis cussed in detail in the following sec tion REPORTING REQUIREMENTS Pre-Experimentation Memo To help meet the final objectives of the design problem the students were asked to first prepare an experimental plan including a factorial design of experiments, to measure the mass-transfer coefficients and the height of a transfer unit in the laboratory. The plan was to be submitted to the TA in a short memo at least two days before the day the experiment was to begin. The next day, a copy of the memo would be returned to the team with comments. Some of the questions to be answered in thi s memo were: What i s the problem being addressed? What variables will be measured? Which flow rates will be run? How much toluene is needed to prepare the feed solution? What proce dure will be followed? What calculations will be performed ? How will the results be presented ? The pre-experimentation memo was an important part of the lab as it required the team to become familiar with the apparatus and to carefully plan the experiment. The group also needed to figure out how to do the necessary calcula t io ns ahead of time so they could record all the relevant data. Preliminary Calculations On the second lab day each team had to s ubmit calcula tions of the height of a tran sfe r unit and the mass-tran sfer coefficient for each experimental point taken on the first lab day. Results had to be presented in a short memo (2 or 3 pages ) along with a table showing Land V in lab units and G ,, G Y' m S, x/xb, N 0 H 0 and K a in SI units. A brief description of the procedure that would be u se d to solve the design problem also had to be included in the memo In the preliminary-calculations memo, the s tudents had to use the actual data they collected on the first lab day. At this point, they will realize if they are doing so mething wrong, or if they have failed to mea s ure a variable they need (e.g. atmospheric pressure column temperature). This also takes part of the load of writing the lab report from the la st week and spreads out report writing over two weeks rather than ju st one. Final Report In addition to the traditional final lab report components, the s tudents were asked to u se a correlation for their data of the form 68 (I) where o/ o. and~ are constants. To find these constants, they used the Linear-regre ss ion option of a ny standard sprea s heet, with three contiguous columns containing log K a, log G ,, and log G yEach team also had to prepare a plot of the height of an overall transfer unit and the overall ma ss-trans fer coefficient versus the liquid ma ss velocity (liquid mass flow rate per unit cross-sectional area) at various gas ma ss velocities. Experimental results, and the values predicted by Eq. (1) must also be included. Finally, the students had to prepare a parity plot with their results This is a plot of the experimental K a (or H 0 ,) versus the values computed by Eq. (1). Included as a reference were the diagonal line and constant-error lines (e.g., +5 % and -5 % lines). The solution to the design problem was to be discussed in a separate section of the laboratory report. Using the correla tions proposed in the earlier section of the report duly cor rected for chloroform and bromodichloromethane the re quired packing height and column diameter (for a gas-flow rate equivalent to 1.6 times the minimum gas-flow rate) had to be computed. In the calculation of the diameter a gas flow rate equivalent to 50 % of the flooding value s hould be used After performing all the calculations, the team had to report on the recommended design and the criterion or crite ria they used A company-type format of about ten pages was s uggested for the final report. Thi s required the s tudents to evaluate the proces s by which they u se d actual lab data to so lve an industrial de s ign problem and to justify their final de s ign proposal. RESULTS OF THE EXPERIMENTS We found it interesting to present the results of this experi ment. Although this is not directly related to the main objec tive of the paper, we realized that correlations for the se type s of systems are le ss abundant than originally thought. To present the results in a compact manner all the data col lected during the course of the se mester (fo ur rotations, 219 point s) were consolidated in a s ingle sp readsheet. Then, a correlation of the form s ugge sted above in Eq. (1) was obtained. The result was K ,a = 8.7 G~ 778 G -;1 05 (2) The correlation coefficient ( R 2 ) for this fit was 0 908 With Eq. (2), a parity plot was prepared to s how the goodness of fit for the correlation obtained (shown in Figure 2). In a parity plot, the points s hould align along the diagonal which represents the perfect fit. Although the exponent for G y is s mall, it s standard error determined that it is stat i stica ll y significant. Chem i cal Engin ee ring Education

PAGE 71

ON THE SOLUTION 10,000 OF THE DESIGN PROBLEM I The design problem presented here requires I T the comp ut ation of mass"' .., transfer coefficients for E :::a trihalomethanes in a sys0 !. tern si milar to the one c:i' u sed in the l a b The ca l1,000 culation of the packing "O .. height a nd column diam.... CJ :a eter is relatively straight.. .. forward If the data obQ., tained for toluene are --,... I V I V ,. i I / I u sed however the st udents have to determine h ow to correct for the fact 100 V ,, I I / I I Yt 11 ---i1 I i / I l / 1 I I I / : 1 1 1 I I I "' +---I I I -! I I I I I I I I I, I I I I I I I I I _._l__ stripping system to remove toluene from a wastewa ter stream The packing available was 1.5 P a ll rings. Gas flow rate to b e used wa s 1 5 times the minimum value. The stu dents had to figure out how to correct for the different packing size Problem 2. Addition ally the s tudents had to find the value gas-flow rate ( relative to the mini mum va lu e) that mini mized the a mount of pack ing in the co lumn. that the so lute s in the lab 100 and in the problem are different. To stim ul ate 1 000 Experimenta l K x a [mol/m 3 s] 10,000 Problem 3. Similar to the problem presented in ext e n s o but with just one trihalomethane (chloro form ) creativity, no guidelines were given o n how to ap Figure 2. Parity plot for th e ma s s-transf e r c oeffi c i e nt proach this aspect of the problem P er h aps the most e l egant way to do this is to obtain a dimensionless corre l atio n to fit the experime nt al data, i.e ., Sho x = cp Re ~ R e~ Sc I ( 3) The exponent for Sc cannot be obtained from the experimen tal data s in ce only one sol ut e is used. A val u e taken from the lit erat ure can be used instead. The students can determine the exponent s a and from the experimental data and then use the correlation to obtain the mass-tran s fer coefficient for the design problem. The five groups in thi s rotation used a similar but somew h at lengthier approac h. They took a di mensionless corre l ation from the literature 12 1 and u sed it to figure o ut a corrected" va lu e of 'V in Eq (1) (one for c hl orofor m and one for dichlorobromomethane) Then they used the "corrected Eq (1 ) for the corre s ponding trihalomethane. The important thing here is that all the grou p s co uld find their way in solvi n g the problem cor rectly This contrib ut es more to the learning process than guiding them through a sol uti on that could be more to the instructor s liking. OTHER DESIGN PROBLEMS Durin g the fall 1996 semester, we prepared four design problems, one for each rotation. The goal was to encourage independent thinking abo ut h ow to use a routine lab experi ment to so l ve a real-life problem. To provide some ideas of how this approach can be implemented we highlight here the main features of the other three problems Problem 1. Here the st ud e nt s were asked to de sig n a Wint e r 1999 STUDEN T FEEDBAC K No formal specific s urvey to explore the students' impres sion of the approach wa s conducted because the idea of writing thi s paper came after the term wa s over. But a three page cour s e-evaluation que s tionnaire wa s distributed to give the st ud ents an opportunity to comment genera lly on the course. One of the que s tion s was "Please comment on each experiment in term s of it s v alue as a chemical engineering learning experience ." We al s o informally asked s ome stude nt s what they thought of the n ew approac h The feedback was highly positive. The most praised fact was the opportunity to connect an isolated lab experiment with a real-world problem Some of the actual re s ponses were: Thi s was th e most v aluable learnin g e xp e rience I got a lot out of it in t e rms of learning w hat it takes to d es i g n a r e al c olumn I r e all y lik e this lab. It was compli c ated but not impossible. Ov e rall, I l e arned a lot. I lik e d this experiment. The lab itself is simple, but und e rstanding and utili z ing the data is very comple x Of all th e e xp e rim e nt s, I fe e l that this one was th e one I l e arned the most in. Good learning e x perien c e in terms of scale-up design." 69

PAGE 72

CONCLUSIONS This paper presents a simple way to enhance the learning experience in the unit operations laboratory by adding a related design problem to the lab requirements. This idea can be implemented in any classical unit operations experiment suc h as distillation, absorption liquid-liquid extraction, or even in a heat-exchanger experiment. By implementing this approach, the students connect lab work to rea l -life problems, their creativity is stimulated, and the concepts are learned enduringly. The feedback from the students confirms this. It is apparent that they felt more comfortable the following term when they had to apply these concepts to design absorption, stripping, and even distilla tion columns as part of the task in the senior design course. NOMENCLATURE a gas-liquid interfacial area per unit total volume [m 2 /m 3 ] G x liquid-phase flow rate per unit cross-sectional surface area [ml G Y gas-phase flow rate per unit cross-sectional surface area [m] H 0 L height of an overall transfer unit based on liquid-phase concentrations [ml Ethanol Fermentation Cont inu ed from page 57. experiments of ethanol fermentation dispensed with the dis tillation of the sample and the chemical analysis of ethanol. The experimental data obtained in the assays allowed fits of a simple kinetic model (Monod) and of one that takes into account the inhibition by the product Y 1 Comparison of the kinetic parameters obtained in the present work with others in the literature showed that the experimental device has attained the expected aim. NOMENCLATURE D internal diameter of the PVC pipe [D=lO cm] K P product inhibition constant [gL 1 ] K s saturation constant (gL 1 ] M e than o l ethanol molecular weight [M ethanol =46] n c o number of moles of ethanol formed n ; number of moles of CO 2 evolved at time t=i product concentration [gL 1 ] p p local atmospheric pressure or barometric pressure [atm] gas constant [R=82 04 atm cm 3 / (g -mole K)] 70 a t m R R 2 regres s ion coefficient [-] S limiting substrate concentration [gL 1 ) T gas temperature [Kl t time [h] V broth s volume [L] V; gas vo lum e at time t=i [mL or cm 3 ) X cellu l a r concentration (gL 1 ] y 0 distance from top of PVC tube to water l evel at t=O [cm] K x overall mass-transfer coefficient [mis] L liquid-phase flow rate [kg/s] m Henry 's constant expressed as a ratio of mole fraction s N 0 x number of overall transfer units based on liquid-phase concentrations Re x liquid-phase Reynolds number Re y gas-p h ase Reynolds number S stripping (desorption) factor Se x liquid-phase Schmidt number Sh 0 x liquid-phase Sherwood number based on KL V gas-phase flow rate [kg/s] x mole fraction of solute in the liquid pha se at the inlet x b mole fraction of solute in the liquid phase at the outlet Greek Symbols a exponent in Eqs. (I) and (3) exponent in Eqs. ( 1) and (3) y exponent in Eq (3)


PAGE 73

Microwave D rying Continu e d from pag e 49 temperatures show trends similar to the drying-rate curves (see Figure 6). The temperature profiles demonstrate the warming-up period, a consta nt temperature period, as well as a rapid falling-off period. Due to the difficult nature of temperature measurement in microwave drying these data are app li cable only for representing trends in the tempera ture profile. It should be noted that the surface-temperature va lu es of Figure 6 did not reach 100 C, the boiling point of water at atmospher i c pressure This can be accou nt ed for by the effect of s urf ace evaporation 151 and by the measurement technique that allowed for some cooling of t h e sample upon removal from th e microwave. In ge n eral during microwave drying, the s urfa ce temperatur e falls between the wet -bulb temperature in th e oven and the boiling temperature. 141 Our microwave oven was not equipped with a carouse l and given the uneven nature of the energy field in a microwave oven the temperature probably var i ed with position. We did not st ud y thi s phenomenon however and only measured the surface temperature near the middle of the sample ll Students ca n be asked to design an experiment that illustrates that mass transport effects are respons ibl e for the falling-rate period. Evaporation of water with no solid present can be used to do thi s In this instance the fal lin rate period i s eliminated, with the constant-dry in g-rate period ending abruptly as the la s t of the water i s evapo rated. ll Particle-size effects can be examin e d by using s and that ha s been sifted into different particle sizes. Critical moisture co nt e nt increases with decreasing particle size since smaller particles pack closer together slowing the movement of moisture to the surface. 41 ll Datta 1 5 1 sta te s that geo metry plays a role in microwave processing. This effect ca n be exp l ored by c h anging the s h apes of the containers and the thicknes s of the bed height while maintaining a constant ma ss ll The a nal ysis presented here doe s not yield detailed information abo ut the fundamental mechanisms of heat and mass transfer during the drying process It is possible however to es timat e interphase heat and mass transfer coefficients and effective liquid diffusivities using methods presented in th e lit erature_l 4 l Also, comp l ex mathematical models of the drying process have been developed. 1 9 1 ll Drying characteristics of different materials can be compared Various materials have been used in our lab orato r y, including calc ium carbonate sponges, bread and fruit. CONCLUSIONS This experiment demonstrates the drying process effec tively. It is an extremely flexible, safe, and inexpensive experiment that can be incorporated into the undergraduate laboratory curriculum. The experiment i s easy to set up a nd run. Typically meaningful experimental dat a for higher power Winter 1999 G' v Q) .... ::J ro .... Q) C. E Q) I1 00 ~------------~ 8 0 60 40 2 0 ~ 0 -l----4-----+----~--"-+----l----+----+---+---I 0 5 1 0 15 Tim e (m i n) 20 Figure 6. Surfa ce temperature profile for a power level of 720 W. 25 settings can be collected in about thirty minutes. This experi ment also involves data analysis that introduces students to the various methods of treating data and the errors associ ated with each method ACKNOWLEDGMENT S We would like to thank the reviewers of this paper for their positive comments and pertinent suggestio n s that im proved it. NOMENC L ATUR E index for numerical differentiation (unitless) m mass of liquid ( kilogram s) m mass of solid (kilogram s ) time (minutes) t 0 time of first data point (minutes) t 0 time of last data point ( minutes) REFERENCES 1. Nirdosh I. and M.H.I. Baird Low-Cost Experiments in Mass Transfer," Chem Eng. Ed. 30 ( 1 ) 50 (1996 ) 2 Mujamdar A.S. "Drying Neglected ," Letter to Chem. Eng Prag ., 93 ( 10 ), 9 ( 1997 ) 3. Oakley D.E. "Produce Uniform Particles by Spray Drying," Chem. Eng. Prag. 93 ( 10 ) 48 ( 1997 ) 4. Porter H.F. P.Y. McCormick R.L. Lucas, and D.F. Wells, Gas-Solid Systems Chap. 20 of Perry s Ch e mical Engi n ee rs Handbook Fifth Ed. McGraw-Hill New York, NY ( 1973 ) 5 Datta A.K. Heat and Mass Transfer in the Microwave Processing of Food Ch e m. Eng Prag ., 86 ( 6 ), 47 ( 1990 ) 6 Keey R B ., Drying: Principl e s and Practic e Pergamon Press, Oxford p. 3 ( 1972 ) 7 Williams-Gardner, A., Industrial Drying, Leonard Hill, Lon don, England p. 45 ( 1971 ) 8 Ray M Engineering Experimentation, McGraw-Hill, New York NY p 169(1989 ) 9 King, C.J and J.P. Clark, eds., Water Removal Processes : Drying and Concentration of Foods and Other Materials," AIChE Symposium Series 73 ( 163 ) ( 1977 ) 7 1

PAGE 74

.ta ... 5 ... 311-.....a_s_s_e_s_s_m_e_n_t _______ ) RANKING GRADUATE PROGRAMS Alternative Measures of Quality JOHN C. ANGUS, ROBERT V. EDWARDS, BRIAND. SCHULTZ Case Western Reserve University Cleveland, OH 44106-7217 A ssessing the relative quality of graduate programs is of great interest to policy makers, academic admin istrators, prospective students, employers of gradu ates, alumni, and the general public Rankings by federal advisory panelsr 11 and the popular pressr 21 are widely quoted, but despite the importance assigned to the rankings, there ha s been little critical, detailed analysis of their relevance and accuracy. In this paper we will present an analysis of the most prominent of these reports,r1 1 especially as it relates to chemical engi neering programs. Although our discussion is confined to chemical engineering, we believe that the gen eral conclusions and methods are also relevant to the other engineering disciplines. The National Research Council (NRC) in 1995 released a massive studyl 1 1 ofresearch doctorate programs in the U S. It was the product of a committee of eighteen academics from Editorial Comment ... A strength of th e engineering education s y s tem in the United States is it s diver sity. It is evident in such characteristics as student demographics college mission s and sources of funding enrollment levels research strengths collaborations and curricula A positive outcome of our system is the diverse pool of graduates pro duced to meet the varied and dynamic workforce needs of the world Choices however are being made that can significantly impact programs industry selects schools for recruiting visits; s tudents commit to pursue graduate or undergraduate degrees at specific schoo l s, and private and public foundations and agencies award grants contracts and gifts to selected institutions. The perceived quality of an institution i s often an important factor in these decision s, and rankings by institution college, or degree program contribute to defining perceptions In recent years chemical engineering departments have been asked to asse ss the quality of their programs to direct improvement strategies. This movement is being driven internally as well as externally by regional and national accreditation entities and by funding agencies (e g. state government s) Ideally rankings would directly assess the quality of graduates and the improvement in students while they were enrolled. Since no method to do this has been devised, ranking schemes typically use a combination of numerical program data and peer ranksing to determine a score instead of doing a direct assessment. Although the efficacy of this approach is still being debated it ha s increased the importance of peer comparison and the availabil ity of program data. This issue of CEE presents a paper by Angus, Edwards and Schultz that proposes a ltern ative measures of graduate program qua lit y. Not sur prisingly, an extensive review process revealed that the subject of rankings is a various disciplines and is a follow-on study of a similar report issued in 1982. The report contains reputational rankings, based on a survey of graduate faculty, as well as an impressive amount of factual data from several independent sources. Unfortunately, most atte ntion has been focused on the survey results-in particular the reputational ranking based on perceived faculty quality. This is apparently due, at least in part, to the method by which the data were displayed in the report and the normal tendency to simplify complex data sets into a single easily understood quality index. Several aspects of the NRC report have caused concern. (See, for example, the summary article by Mervis. [ 3 l) One striking feature of the results, noted by the authors, was the "remarka ble stability among programs rated in the top and bottom quarter" between the 1982 and 1995 reports. An other striking feature was the heavy reliance on the survey contentiou s issue The reviewers as well a s the author s identified many of the shortfall s in assessing the quality of something a s complex multidimen s ional and subjective as graduate programs. ldentified is s ues included the establishment of false goals, for example publishing papers in smaller segment s or of lesser quality or hiring faculty in publication-intensive research areas simp l y to increase the publica tion count all of which would increase ranking but could likely decrease program quality. Another issue wa s that inaccurate or inappropriate data would undermine the conclusions Examples include the counting of non-competitive research funding (e g state funding) and the use of a limited set of journals for citation searching or a limited set of societies in award counting that would not recognize program s with an emphasis on emergin g research areas. This paper presents a sound analysi s of the recent NRC rankin g, and many of the conclusions as well as the a nalytical approach can be extended to other ranking schemes. Although the proposed alternative measures of quality may be open to criticism the authors clearly show the sensitivity of rank order to the selected s et of measures and to weighting algorithms ( e. g. intensive vs. extensive ) We hope that this paper will increase awareness of the shortfalls of any approach to ranking and help one gauge what can and cannot be ascerta in ed from a ranking We do not believe th at a composite index can be developed to accurately measure the relative quality of chemical engineering degree programs in our complex graduate education system, nor is it desirable to drive programs to conformity. We recognize, however that ranking schemes have increased their presence in our profession and to ignore their impact wou ld be a mistake. We hope that this article will stimulate serious discussion in the community. D Copyright ChE Division of ASEE 1999 72 Chemical Engineering Education

PAGE 75

results, which were used to generate reputational rankings rather than quantitative measures of quality Another concern, quite apparent to engineers and scie nti sts, was the minimal distinction drawn between intensive (size independent) and extensive (size dependent) measures of quality Although the report included a number of sta tisti cal tests of the data no detailed analysis of so urces of error in the data se ts was provided. Finally, there was no assessment of program quality based on st udent outcomes in their subsequent professional life. The committee was TABLE 1 Survey Questions Used for the NRC Reportl 1 1 Bl Familiarity with work of Program Faculty 1. Considerable familiarity 2. Some familiarity 3. Little or no fa miliarit y B2 Schola rl y Quality of Program Faculty 1. Di st inguished 2._Strong 3. Good 4. Adequ a te 5. Mar g inal 6 N o t s ufficient for doctoral education 9. Don t know well enough to evaluate B3. Familarity with Graduates of this Program 1 Considerable familiarity 2 Some fami liarit y 3. Little or no familiarity B4 Effectiveness of Program in Educating Research Scho lar s/Scientists 1 Extremely effective 2. R easonable effective 3. Minimally e ffective 4. Not effec tive 9. D o n t kn ow well e nou g h t o evaluate BS Change in Program Quality in Last Five Years I._ Better than five years ago 2. Little or n o c han ge in th e la s t five ye ar s 3. Poorer th an fiv e years ago 9. Don t know well e n o ugh to evaluate John C Angus is Professor of Chemical Engineering at Case West ern Reserve University He received his BS MS and PhD degrees from the Univers ity of Michigan He worked on thermoelectric materials at the 3M Company for three years before joining the faculty at Case. He has worked on the growth of diamond by chemical vapor deposition and various electrochemical problems for almost forty years Robert V. Edwards received his PhD from Johns Hopkins University in 1968 and took a post-doctoral position at Case Western Reserve University to work on the then-new field of laser light scattering for transport measurements He joined the Case faculty in 1970 and has subsequently made numerous contributions to the theory and practice of laser light scattering with collaborators both here and abroad. Brian D Schultz obtained his BS in chemical engineering from Case Western Reserve University in 1977 He won a National Science Foundation Fellowship which he used to pursue his Master s degree at Case and in the fall of 1998 he began working toward his PhD at the University of Minnesota Research interests include ternary phase diagrams for /ow-pressure crystal growth as well as the thermochemi cal behavior of group Ill nitrides. Wint e r 1999 aware of many of these concerns and, in fact, was unable to address so me of them for lack of time and resources The commit tee was also aware that the report might be used in s uperficial ways that were not intended. The NRC report is bein g used by deans, legislators, and founda tions in the allocation of resources and in other critical decisions. It is therefore u seful to understand the report and to critically exam ine its conclusions In this paper we will give an analysis of the data for the chemical engineering programs covered in the report. We will also give alternative rankings using data from the NRC report and other so urces We emphasize that the rankings pre sented here are meant only to illustrate the methods employed and to reach general conclusions. B ec ause of limitations in the data available to us th e position of a particular individual program in the rankings should be treated with caution. PART1 THE NATIONAL RESEARCH COUNCIL REPORT Methods The most discussed part of the NRC report is the faculty survey conducted in 1993 Questionnaires were sent to randomly se lected faculty and each participant was asked to rank approximately fifty programs. Other than a list of faculty, the participant s were provided no other information about the pro grams. The survey que stio n s are s hown in Table 1. Forty-one graduate fields of s tudy were covered in the NRC report, one of which was chemical engineering For chemical engineering program s, 206 usable responses were obtained from 361 questionnaires, a 57 % response rate. Within chemical engi neering 93 of the 121 engineering departments awarding PhD degrees during the 1986-92 time period were included These 93 departments produced 96 % of the chemical engineering PhD s awarded during that period The results of the survey were tabulated in the Appendices to the NRC report. The programs were listed in order in the tables ac cording to the results of the first survey question (the average ranking of faculty quality ). This procedure was used in response to complaints that data in the 1982 report were difficult to interpret becau se programs were listed alphabetically. The result, however ha s been to focus on this one sing le measure of quality, despite the fact that rankings in the other categories (e.g. program effective nes s and visibility) are also provided in the report. One of the purposes of the committee that compiled the 1995 NRC Report was to expand the objective mea s ures developed by prior committees. Some program s tatistics were provided: num ber of faculty, number of PhDs granted, number of PhDs awarded to female and minority students and non-citizen s, and the average length of time to receive a PhD Quantitative measures of quality were also provided: 1) percentage of faculty with research support (% SUPP ), 2) percentage of faculty publishing during 1986-92 (% PUB ), 3) total publications during 1986-92 ( PUB ), and 4) total citations to publi s hed work during 1986-92 (TC). The latter two were also reported on an intensive ( normalized per faculty) basis, i.e., PUB/TF and TC/TF (see Table 2 for a description of 73

PAGE 76

the terms in the NRC report ) Survey Results In Table 3 we give the acronyms u s ed in s ubsequent tabl e s for id e ntification of universities and in Table 4 w e li s t the g radu a te program s in chemical engin e ering a s the y wer e rank-ord e red by perceived faculty qualit y ( 93Q ) in the NRC report. Thi s is the order in which the program s are listed in Appendix P of the NRC report. A s triking but not widely appreciated feature of the NRC report i s s hown in Figure 1 ,* which i s a plot of the survey res ul ts for faculty quality (93Q) versus program effectiveness (93E) A very s trong correlation is evident. For example R 2 =0 97 when the data a re fit with the equality (93Q)=(93E) This s trong correlation can arise simply because high-quality faculty will produce effec tive graduate programs We believe it i s far more likely that the respondent s did not discriminate between faculty quality and program effectiveness and treated both ques tions the same This strong correlation was noted in the NRC report in Appendix 0-8 where a Pear s on product moment correlation coefficient of 0.98 between 93Q and 93E was given for chemical engineering. Similar strong correlations between 93Q and 93E were observed for the In Fi g ur e 1 and subsequ e nt figur e s w e indi c at e th e squar e of th e d eg r ee of c orr e lation by R 2 th e c o e ffi c i e nt o f d e t e rmina tion Th e magnitud e of R 2 i s s impl y d e scrib e d a s th e fra c tion of th e raw varianc e i n the data s e t account e d for b y using th e fitt e d e quation. Th e plots and valu e s of R 2 w e r e obtain e d u s in g an Ex ce l s pr e adsh ee t TABLE2 Definition of Terms Used in this Paper NO T E: /11 t h e N R C re p o r t, the sy m bo l s for the variab l es r efe rred to bo th the rank order, a n d, w h ere a ppli cable, to t h e average score of t he rati n gs on th e scale of I to 5. We give bo th defi niti ons in th e li s t below T he defi n itio n s a r e t aken from Appe n dix P page 469, a n d T a bl e 2-4, p age 25 of th e N R C r e p o rt No t all of th e ca t ego r ies u sed i n th e N R C r e p o rt we r e u sed in t h is p aper. Th e te rm s b e l ow th e d as h ed li n e we r e n ot u sed in th e N R C repo rt. 93 Q R a n k o rd e r of"sc h o l ar l y qu a lit y of p rog ram fac ult y." (Average sco r e o n a sca l e o fO t o 5 w ith 5 r e pr ese ntin g Di s tin g ui s hed ) 93E R a nk o rd e r of p rogra m e ff ec ti ve n ess i n e du ca tin g r esearc h sc h o l a r s a nd sc i e nti s t s." (Ave ra ge sco re o n a sca l e of O t o 5, w ith 5 re pr ese ntin g E x tr eme l y Eff ec ti ve ) VIS R a nk ord e r o f v i s ibilit y o f th e d oc t o ral program ( P e r ce nta ge o f th e qu es ti o n naire s that r e port e d s om e knowl e d ge of th e pro gra m b y a n a n swe r o th e r th a n Don t kno w w e ll e nou g h to e valu a t e" o r Littl e o r n o fa miliarit y" t o o n e o r m o r e o f th e fiv e qu es ti o n s.) T C Rank order of th e t o t a l numb e r o f c it a ti o n s a ttribut e d t o p rog ram fac ult y in th e p e riod 198 8-92. ( T o t a l numb e r of c it a ti o ns attribut e d t o p rogra m fac ult y.) S o ur c e : In s titut e of S c i e ntifi c In fo rm a ti o n. C/F R a nk ord e r of th e c it a ti o n d e n s it y fo r th e pro g ram fac ult y ( T o t a l numb er of cit a tion s ( T C) di v id e d b y th e numb e r of pro gra m fac ul ty ( T F)). S o ur ce: In s titut e of S c i e ntifi c In fo rm a ti o n PUB Total numb e r o f publi ca ti o n s attribut e d t o pro gra m fac ult y fo r th e p e ri o d I 98892 TF N umber of pro gra m fac ult y in fa ll 1 992 N RC R e p o rt fo r calc ul a tin g P U B ff F a nd TCffF ; AICh E 191 for c al c ul a tin g HONffF a nd S U PPffF. HON Number of h o n o r s re ce i v ed b y fa c ult y. 16 81 See t ex t fo r d e tail s SUPP T o t a l resear c h s upp o rt fr o m a ll so ur ces Sourc e : Na ti o n a l S c i e n ce Fo und a ti o n 141 TABLE3 Ac ron y m s Used to Identif y Universit i es A S U Ar i zo n a St a t e U ni ve r s it y NEU N o rth eas t e rn U ni vers it y UA Z Uni ve r s it y o f Arizo n a UM OC U ni ve r s it y of M isso uriCo l u mbi a A U B Auburn U ni ve r s it y NITT N ew J e r sey In s t. of T ec hn o l ogy U CB Uni ve r s it y o f Ca li fo rni a -B e rk e l ey UMOR U n ive r s it y o f Mi sso u r i R o ll a BY U Bri g h a m Y o un g U ni ve r s it y NWU N o rth wes t e rn U ni ve r s it y UCD Uni v er s i ty o f Californi a -D av i s UMS U ni ve r s i ty o f Mi ss i ss ippi C IT C a lif o rni a In s t. o f Te c hn o l ogy OH U Oh io U ni ve r s i ty U CI N U ni ve r s it y of C in c i n n a ti U OK Univers i ty o f Oklah o m a C L A R Cl arkso n U n ive r s i ty OKS U Okl aho m a S t a t e U ni ve r s it y UC L A U ni v of Ca li forn i a Los A n g el es U P A U n ive r s i ty o f P e nn sy l van i a C L MN C l e m so n U ni ve r s i ty ORS U Or ego n St a t e U ni ve r s it y U CO U ni vers it y of Co l orado U RI U n ive r s i ty o f Rh o d e I s l a nd C M U C a rn eg i e M e ll o n U ni ve r s it y os u Ohi o St a t e U ni ve r s it y U CSB U n iv of Ca l ifo rni aS a nt a B ar b ara U S C U n iv. of So uth e rn Ca li fo rni a COL Co l umbia Uni v er s it y P I TT U ni ve r s it y o f Pitt s bur g h U CT Univer s it y of Co nn ec ti c ut U T A U n ivers it y o f T exas-A u s tin C ORN Corn e ll U ni ve r s ity PL YU P o l ytec hni c U ni ve r s it y UDE U ni v er s it y o f D e l aw ar e UTN U n iv. of T e nn esseeKn oxv ill e C SM C ol ora d o S c h oo l of Min es P R U Prin ce t o n Unive r s it y U FL U ni v er s it y o f Fl o rid a U T U L U ni ve r s it y o f Tul sa CUN Y C UN Y G ra d S c h & U ni v Ce nt e r PS U P e nn sy l v ani a St a t e U ni ve r s it y U H U ni v er s it y of H ous t o n UU T U ni ve r s it y ofU t a h CWR U Ca se W es t e rn R eserv e U ni v. PUR Purdu e U ni ve r s it y U I A U ni v er s it y o f I owa U V A U ni ve r s it y o f Vir g i n i a D U KE Du ke U ni ve r s it y RICE Ri ce U ni ve r s it y U I C U ni v er s it y o f Illin o i s C hi cago U WA U ni ve r s it y o f W as hin g t o n GIT G eo r g i a In s titut e o f T ec hn o l ogy RO C H U ni ve r s it y o f R oc h es t e r U ID U ni v er s it y o f I da h o U WI U ni v. of Wi sco n s in Madi so n llT Illin o i s In s titut e o fT ec hn o l ogy RPI R e n sse l ae r P o l y t ec hn ic In s titut e U IL U ni v o f lllin o i s,U rb a n a-C h a mp a i g n U WY U n ive r si t y of Wyom in g IS U I owa St ate U ni ve r s it y RS U Rut ge r s S t. U ni v-New Brun sw i ck U KS U ni ve r si t y of K a n sas VAND Vander bi lt U n iversity JH U John s H o p ki n s U ni ve r s it y SIT Steve n s In s titut e o fT ec hn o l ogy U KY U ni ve r si t y of K e n t u cky VP V ir g ini a P o l y t ec h In st & S t ate Un i v KS U Kan sas St a t e U ni ve r s it y ST AN S ta n fo rd U ni ve r s it y U LV U ni ve r s it y o f L o ui sv ill e WAS U Was hin g t o n St a t e U ni ve r s it y L E H L e hi g h U ni ve r s it y S UN Y St a t e U ni v o fN ew Y o r kBu ffa l o U M A U ni v. o f M assac hu se tt sAmh e r s t WPI Wo r ces t e r P o l y t ec hni c In s titut e L SU L o ui s i a n a S t. U & A& M Co ll ege S YR S yrac u se Univ er s it y U MD U ni v. o f M ary l an dCo ll ege P ark ws u Way n e Sta t e U ni ve r si t y M IT M assac hu se tt s I ns t. o f T ec h TAM T exas A& M U ni ve r s it y U M E U ni ve r s i ty of M a in e WU Was h ing t o n Un i v er sity M SU Mi c hi ga n S t ate U ni ve r s it y TU L Tul a n e U ni ve r s i ty U M! U ni ve r s i ty of Mi c hi ga n wv u W est Vi r g ini a U ni ve r s it y NC S U No rth Caro lin a St a t e U ni ve r s it y UA KR Unive r s it y of Akr o n U M N U ni ve r s it y of Minn eso t a YAL E Y a l e U ni ve r s it y N D U ni ve r s it y o f N o tr e D a m e 74 C h e mi ca l E n g in ee rin g E du c ati o n

PAGE 77

other engineering program s Thi s level of correlation s trongl y s uggest s that the se two of the five s u rvey question s gave indi s tin guishab l e re s ult s. In Fig u re 2a, we show a plot of perceived T A BL E 4 R a n k Ord e r of C h E Fac ul ty Qu a lit y S u rvey R es ul ts ( 9 3 Q ) Give n in t h e R C R e p o r t Rank Orde r I 2 3 4 5 6 7 8 9 10 II 12 13 1 4 U n ivers i ty U ni vers it y of Minn eso t a Ma ssac husetts ln s titut e of Technology University of California-Berkel ey University of Wi sco n s in-M a di son Univ of Illinoi s Urbana-Champaign California In s titut e of T ec hn o l ogy Stanford University Unive r si t y of D e l aware Pr inceton University U ni ve r s ity of Texas at A u stin Un i versi t y of P ennsylvania Carnegie Mellon U ni vers it y Corne ll Unive r s it y Un i v of California-Santa B arbara 15 Northwestern University 16 Purd u e University 1 7 U ni ve r s it y of Hou s t on 1 8 U ni ve r sity of Michigan 1 9 CUNY-Grad S c h & Univ Center 20 Un i versity of Washington 2 1.5 U ni v of Massachusett s Am her st 21.5 Rice Unive r si t y 23 P ennsy l vania State University 24 U ni ve r sity of Notre D ame 25 North Carolina State U ni ve r sity 26 University of Colorado 27 L e hi g h U niver s it y 28 University of California-Davis 29 State U ni ve r s i ty of ew Yark -Buffalo 30.5 Unive r s it y of Virginia 30.5 Georgia In sti tut e of T ec hn o l ogy 32.5 Y a l e U ni versi t y 32.5 I owa State Unive r sity 34 Unive r s it y ofFlorida 35.5 R ensse l aer P o l ytec hni c In s titut e 35 5 J o hn s H o pkin s Universi t y 37 T exas A& M University 38 W as hin g t o n U ni vers it y 39 Unive r si t y of California-Los Ange l es 40 U ni ve r s it y of R oc he ste r 41 Ohio State University 42 Virginia P olytec h ln st & State Un i v 43 Rut gers State Un i v.-New Brun swick 44 Universi t y of Pitt sb ur gh 45 Michigan State Un i versity 46 Case Western R eserve University 47.5 Syracuse U ni versity 47.5 Illin o i s In s titute of T ec hnol ogy 49 Clarkson University 50 Br ig h am Young Unive r s ity Wimer 1 999 faculty quality (93 Q ) versus faculty s ize ( TF ). The value of R 2 i s 0.40 s ugge s t ing that the s ur vey re s ult s for faculty quality are influ e n ce d to s ome extent by program size, b u t th a t other factors are also important. The program visibility ( VIS ) was defined as the percenta ge of respondents who reported some knowledge of the program. In Figure 2b we p l ot the fac u lty qua li ty (93Q) versus the visibility ( VIS ). A s trong corre l ation, R 2 = 0 .8 4, i s observed. One cannot prove cau se -and-effect relation s hip s through correlation alone; however the se resu l t s s uggest that the perceived faculty quality (93Q) sco re s arise, at lea s t in part because respondents rate highly those faculty with 5 00 4.00 0 g_ 3.00 g ;; (5, l;> ] 2.00 "' 1.00 0.00 +---0.00 1.00 2 00 3 00 4 00 5.00 Progra m E ff ective n ess (93E) F i g ur e 1. Surv ey results of program effectiveness (93E) vers u s facult y quality score (93Q} R2=0.97 w h en fit wit h (93E)=(93Q}. F igur e 2 (a ) Surv ey results of faculty quality (93Q} ve rsus total faculty (TF). R 2 =0.40 for lin ear fit y =mx+b (b) Surv ey results of faculty quality (93Q} versus visibility (VIS). R 2 =0.84 for lin ea r fit, y =mx+b a 5 00 4 00 J 6 g 3 00 -< (5, a 2 00 ] t!: 1.00 0 0 0 00 ---------------~ 6 5.00 1 4 00 I g 300 I : t -I l o I a 2 00 l o I I "' 1 00 J IO 15 20 25 30 35 40 Tota l Fac ul ty (TF) 0.00 --------------30 40 50 60 70 80 90 100 Vi.,ibilily (VIS) 75

PAGE 78

iliarity arises so l e l y whom they are familiar. If thi s fam becau se of the true quality of the fa culty, this result is benign ; otherwise it is not. ved between rank orA r e latively stro ng correlation i s obser der of visi bilit y and rank order of total ci a weaker correlation between rank order number of faculty ( R 2 =0.39 ) The se re s becau se they s how that s maller depar imp act by virtue of their re sea rch outp l a tion s are also found between perce ( 93Q) and the number of publications ( pub l ications per faculty (R 2 =0.64 ), to t a l citations ( R 2 =0 7 l ), and citations tation s ( R 2 =0 .6 7 ), and of v i s ibility and total ult s are encouraging tments can have an ut. Significant co rreived faculty quality R 2 =0.73), number of TABLES Programs not Considere d i n Fina l Ra n kings Beca u se of Lack of Data on Research S u pport Brigham Young U ni ve r s it y Un i versity of Louisville Ci t y University of New York Un i versi t y of Maine Duke U ni ve r s it y U ni vers it y of Mississippi Illin o i s In s titut e of Technology U ni vers it y of Notre Dame Northeastern Un i versity Un i vers it y of Rhode I s land Ri ce U niv ers it y Univers it y of Tulsa U ni versity of Akron University of Wyoming University of Idah o Washington U ni ve r sity University of Kan sas Worcester Polytechnic U ni ve r s i ty T A BLE6 p er faculty (R 2 =0.56). In s ummary it appears that the respon dents made no distinction between the survey questions on faculty quality ( 93Q) and program effectiveness ( 93E ) To some extent, sheer size influenced the quality rankings and respondent s gave high ranks to programs with which they were familiar. Strong po s itive correla tions exist between the survey results of faculty quality and the publication and Ra n k Ord e r a nd Sca l e d Scores of ChE Gra du ate Programs citation rates. PART2 ALTERNATIVE METHODS OF RANKING A lt e r n a tive Me a s ur es o f Qua lity Four extensive measures of program quality are used : 1) Number of pub l ications, 2) n u mber of citations to publications, 3) research funding and 4) faculty honors. In addition, each of these extensive mea su re s is normalized by the number of faculty to provide four intensive mea s ures of quality We use these data to develop alternative ranking s of programs based on both the extensive and inten sive criteria. We also provide a final com po s ite ranking based on the extensive and intensive rankings. We are quite aware that the se so-cal led "o bjective mea s ure s are imperfect, and we will at tempt to point out potential problems with each of the me as ures we u se. Quantitative mea s ures of quality are not new. Some were used in the NRC Report as mentioned above. Also the often-maligned U.S. News and World R eporr21 used a lumped score in which 76 U nivers i ty UMN MIT UTA UC B UDE UWl TAM PSU UCLA UCSB PRU NCSU NWU PUR STAN UMI C IT UIL CORN LEH UPA CMU GIT SUNY C WR U UMA LSU UCO UTN UWA UCO JHU UUT UOK UH osu RPI PITT !SU UVA Us in g Exte n sive Cr i teria Publicatio11s Citatio11s Support Ho11ors Exte11sive Comv.osit e Score Rank Score Rank Score Rank Score Rank Score Rank 100.0 I 100 0 I 76.6 3 77.9 2 354.5 I 79.2 2 65 0 3 100 .0 1 100 .0 I 344.2 2 76 .6 3 76.6 2 55 .6 5 72.6 3 28 1.4 3 50.3 4 45.2 4 1 3.4 41 56.1 6 1 65 0 4 34.2 6 23.3 8 30.7 12 68.3 4 156.6 5 34. 1 7 16 .2 19 41.6 7 58 .6 5 150.6 6 22.8 24 6.4 4 5 77.8 2 21.8 27 128 7 7 25.7 1 7 1 3.0 24 23.0 19 54.0 7 115.7 8 43.4 5 39.6 5 12.2 44 1 8.7 35 114.0 9 27.6 13 24.9 7 28.2 13 32.5 16 113.2 1 0 2 1.0 27 17 .5 16 25.7 16 46.9 10 111.1 11 22 6 25 16 .8 17 35.9 9 35.0 1 2 I 10.4 1 2 24 7 20 1 3.7 23 13.1 42 52.8 8 104.3 1 3 34 .0 8 21.4 9 26.0 15 20.3 32 101.6 1 4 27.6 1 3 27.7 6 25.4 17 20.3 3 1 101.0 1 5 28.4 II 1 8.8 13 2 1. 2 24 32.0 1 7 100.4 1 6 24.0 2 1 17. 9 15 2 1. 7 23 29.8 20 93.4 1 7 25.7 1 7 1 8.3 14 15 .4 34 34.0 15 93.4 18 30.2 9 1 9 0 12 2 1. 9 22 2 1. 8 27 92.9 1 9 30.2 9 20.3 IO 23.9 18 1 7 2 37 91.5 20 26. 1 16 1 9.3 II 8.4 56 35.0 12 88.7 2 1 1 8.9 30 9 6 38 10. 7 48 47 .6 9 86.8 22 28.2 1 2 10 .3 35 26.7 1 4 20.8 29 86.0 23 23.8 22 14 .8 22 10.0 50 34. 1 14 82 .6 24 1 2.0 50 15. 8 21 32.2 II 22.4 25 82.3 25 1 9 .1 29 1 2.1 26 11.9 45 35.3 II 78 3 26 24 9 19 II.I 3 1 22 .1 2 1 1 6.5 39 74 8 27 1 6 5 33 9.4 39 16.9 3 1 3 1. 9 1 8 74 7 28 3.7 87 0.5 9 1 6 1.6 4 8.5 53 74.3 29 27.6 1 3 1 6.5 1 8 14 .4 37 13.8 43 72.3 30 23.5 23 1 5.8 20 8.8 53 24.0 24 72.1 3 1 1 5.1 36 I I.I 32 34.4 JO 11.1 48 71.6 32 22.5 26 II.I 32 21.2 25 1 6.5 39 7 1 .3 33 11. 8 51 7.9 44 5 1. 6 6 0 0 83 71.3 34 12.2 48 5.3 52 23.0 20 27.2 2 1 67.7 35 18.5 3 1 9.2 40 11.2 47 26 7 22 65 6 36 1 7.3 32 9.9 37 15 .3 35 20 2 34 62 7 37 1 6.4 34 5.8 48 18 .9 29 18.1 36 59.2 38 1 2.6 44 1 2.2 25 8.7 55 22.4 25 55.9 39 1 4.3 38 10.0 36 1 0.4 49 20.3 32 55.0 40 Average of Extensive Composite Scores for All Universities 71.85 Chemical Engineering Educa ti on

PAGE 79

40 % was based on a reputational survey of deans and mem bers of the National Academy of Engineering. The remain ing 60 % was derived from quantitative measures of research support faculty honors, and student selectivity The U.S. News and World report also included both extensive and intensive measures of quality. We have included only the top forty programs in the extensive, intensive, and composite rankings. Our purpose is to focus on alternative methods of ranking rather than the rank order itself. We have no wish to identify any program as being of low quality. Publications and Citations We use both the total number of faculty publications (PUB) and citations to published papers (TC) from the NRC report as extensive measures of research quality The same variables normalized by the total number of faculty (TF) are used as intensive measures of research quality. We recognize these are imperfect mea sures. For example, research with the longest range and most profound impact may go unnoticed for decades. Also it is difficult to agree on what constitutes a publication, and there is a proliferation and duplication of research papers of mar ginal merit. The number of times a research paper has been cited is a summary judgment, albeit imperfect, of its relevance and importance. But papers with classic errors (for example cold fusion) may attract numerous citations More signifi cantly a single review paper or a paper describing a widely used test or procedure can generate an inordinate number of citations not closely related to research quality. Finally certain sub-fields within chemical engineering may more easily pro duce publishable results than others. Research Support The NRC report contained data on the percentage of faculty that received research support ( % SUPP) and the percentage of faculty that published ( % PUB). We found these variables provided little discrimination, especially between high-ranked programs and we did not use them in our analysis. Instead, we elected to use total research support from all sources (SUPP) collected by the National Science Founda tionr41 as an extensive quality measure. These figures were used without modification .* We emphasize that the compilation reported by the NSF is meant to be complete; it includes state support and support from other federal agencies indus try, and foundations We also note that total research support is one of the primary measures used in recent scholarly studies of the relative quality of research universities. 151 The data in the NSF report are reported by the individual institutions and may not be reported on a similar basis; research support from ancillary research institutes or unrelated programs may be included in some cases. Also, the amount of We made one exception to this g e n e rali z ation. For our own uni versity, w e removed th e expenditur e s of th e Ma c romol e cular S c ence Department from the NSF figures. This lowered the CWRU e xtensive ranking and left the intensive ranking unchanged. W e w e re unabl e to mak e a similar corr ec tion for oth e r programs. Winter 1 9 99 state support for research may not be uniformly reported. Eighteen chemical engineering programs in the NRC re port were not listed in the National Science Foundation report. There is no indication whether this is because no data were submitted by these programs or whether they had too little research income to appear on the table (only the top 100 engineering programs were included in the table). Rather than estimate the research support from other sources, we excluded these programs from our rankings The programs that were excluded are listed in Table 5. Faculty Honors Inexplicably faculty honors were not used as a quality index for engineering and science programs in the NRC report ; they were used however in the NRC rankings of programs in the arts and humanities There are certain categories of honors and awards for chemical engi neering faculty that can easily be tabulated. For junior fac ulty we used the number of recipients of NSF Career Devel opment Awards, NSF Young Investigator awards, and Presi dential Investigator awards over the period 1988 to 1996/ 61 for mid-career faculty we used winners of the principal AIChE awards from 1987 to 1996/ 71 and for senior faculty we use the sum of the current number of National Academy of Engineering members l 81 plus one-half of the number of Fellows of the AIChE P 1 This arbitrary choice is based on the observation that there are approximately twice as many AIChE fellows as NAE members in the departments sur veyed. Retired and emeritus faculty were excluded The three categories (junior, mid-career senior) were scaled to give each equal weight. We believe that including only AIChE honors and awards over-emphasizes the traditional areas of chemical engineering. In future rankings we suggest including honors and awards from other professional organi zations (e. g the American Chemical Society the Electro chemical Society and the Materials Research Society) Alternative Rankings The numerical data in each exten sive category (Publications, Citations Support, and Honors) were scaled so the maximum value in each category was 100. The total extensive score for each program was ob tained by summing the four scores for each extensive cat egory The overall extensive rank order was determined from these summed scores (see Table 6). The programs are listed in Table 6 in the order of their total extensive score The intensive score s in each category for each program were obtained by dividing the extensive scores by the appro priate number of program faculty. For calculation of PUB/ TF and TC/TF, all data were taken from the NRC report. For calculation of HON/TF and SUPP/TF, the data were taken from references 4,6,7,8, and 9. The intensive scores were also scaled so that the maximum value in each category was 100. A total intensive score was obtained for each program by summing the scaled intensive scores of the four catego ries The programs are listed in Table 7 in the order of their total intensive score. 77

PAGE 80

It is tempting to use the intensive rankings in Table 7 as a value of Ll In ( TC) is the fractional change in number of total measure of the average, individual quality of the program citations required to change one place in the rank order. In faculty. But one should be cautious in doing so, especially the middle range the average fractional change required to for the s maller programs. In so me cases the average intenmove one place in the rank order of citations is approxisive score is heavily influenced by the activities of one or mately 0.03; however greater fractional changes (over 0.30) two particularly strong individuals. This effect was meaare required to move one place in the rank ordering at either s ured in the NRC report by the Gini coefficient, which is a extreme. Similar behavior is observed for the other extenmea s ure of the non-uniformity of the di str ibution of scores s ive variables, also s hown in Figure 3. These results s how among the individual s. Since we did not have access to the that while it is relatively easy to move in the middle range of raw data we could not make thi s estimate. rank orders, it will be mor e difficult for programs to move A composite extensive plu s inten sive ranking was also calculated. A simple s ummation TABLE7 of the total extensive plu s intensive scores Rank Order and Sca l ed Scores of ChE Graduate Programs gave undue weight to the inten sive scores. Using Intensive Criteria We rescaled the intensive total scores to give Publications / Facnl!J. Citatio11slFacu/ty_ Su(l/!.ortlFacu/!J_ Honors/Faculty llltensive Composite the same average score as the extensive University Score Rank Score Rank Score Rank Score Rank Score Rank scores. A composite extensive/intensive score UMN 100 .0 100.0 I 56 .7 II 69.5 13 326 2 I was calculated for each program using the UTA 81.7 4 81.7 3 58.3 9 9 1. 8 4 313.5 2 total extensive score and the rescaled inten MIT 81.7 4 67 .1 5 74.1 6 89.2 6 312 .1 3 sive total score For example, for MIT the STAN 88.3 2 88 7 2 63.9 8 61.5 19 302.4 4 composite extensive/intensive score was obUCB 84.8 3 76.2 4 19 .8 47 100.0 280.8 5 tained from CIT 77.0 6 57.3 6 54.7 12 90.2 5 279.1 6 UWI 57.4 10 27.4 27 58.2 10 98.8 2 241.8 7 344.2+ 7 l.S 5 (312. 1 )=5 1 5.8 JHU 60.4 8 44.3 10 96.2 4 37.4 32 238.3 8 130.66 CWRU 38.3 33 50.4 8 73.7 7 61.6 18 224 .0 9 UIL 54.8 II 39.0 12 32.3 24 86 .0 lO 212 .0 10 This procedure gave the same overall weightPRU 44.8 25 37.3 14 40 .5 15 88.9 8 2 11.5 II ing to the extensive and inten s ive sco res UDE 54.8 II 37.3 14 30.9 26 82.9 12 205.9 12 The composite extensive/intensive scores and SUNY 54.3 13 33.8 19 22.9 40 93.9 3 205 .0 13 rankings are given in Table 8. UCLA 60.4 8 55 .1 7 30.8 27 56.7 22 203.1 14 UCSB 49.1 20 44 2 II 41.7 14 58.0 21 193.0 15 It is most appropriate to compare programs NWU 46.5 24 25 9 31 18.3 53 89 0 7 179 .6 16 using the se parate extensive and intensive TAM 36.5 37 10.2 62 97.9 3 33.0 38 177.7 17 mea s ure s in Tables 6 and 7. Nevertheless, SYR 61 .3 7 36.6 16 39.2 16 40.6 31 177.6 18 the composite extensive/intensive ranking in UPA 52 2 14 38.6 13 13.2 61 66.4 16 170.3 19 Table 8 has value. For example, when makCORN 50 9 1 7 32. 1 21 36.8 19 44.0 29 1 63.8 20 ing a choice of a graduate program a proPSU 30.4 48 15.4 45 30.5 29 86.2 9 1 62.5 21 spective student will make an integrated asUMA 47.0 22 29 9 24 18 7 51 66.8 15 1 62.3 22 UOK 34.3 43 22 9 35 100 0 I 0 0 83 157.2 23 sessment of both extensive and inten sive meaNCSU 42 6 27 31.7 22 37.7 18 44 .2 28 156.3 24 s ures Small programs that are rated very UMl 50.4 19 33.5 20 25.4 34 46.1 27 155 .6 25 highly on a per-faculty ba sis may have a YALE 5 I .3 16 48.0 9 19 6 49 36.1 34 155 1 26 l imite d range of course work and research UCO 35.2 41 20.0 40 30 5 28 69 .2 14 154.8 27 options; large program s with high extensive CMU 35.7 40 18.0 41 15 .9 56 84.8 II 154.4 28 scores may not have the desired level of indiLEH 50.9 17 34. 1 18 33.4 21 28 9 42 14 7.3 29 victual faculty quality A composite score also osu 42 2 28 21.1 39 21.7 45 62 .3 17 147.3 30 permits comparisons with other lumped OKSU 20.9 70 7.7 71 98.0 2 19 .5 53 146 .0 3 I scores-for example, the U.S. News and UVA 38.3 33 26 .7 30 23 7 38 55.9 23 144.6 32 RPI 39.6 30 22.5 36 27.5 30 43.8 30 133.4 33 World report rankings UWA 49.1 20 29.4 25 22.6 43 26.1 44 127.2 34 Sensitivity and Error Analysis The se nUH 21.7 68 9.4 67 38.6 17 55.0 24 124 .7 35 sitivity of the rank ordering to changes in the PUR 43.5 26 27.4 27 26.2 32 24.6 47 121.6 36 extensive data sets ( Publication s, Citations, COL 31.7 45 9.8 65 43 1 1 3 36.3 33 120 9 37 UCO 47.0 22 31.7 23 9.2 69 30.3 39 118.2 38 Support and Honor s) can be calculated by UTN 7.8 92 I.I 92 86.2 5 14.4 60 109.5 39 calculating Ll In X( = Ll XIX) for each of the LSU 29 6 52 13. 2 52 32.8 23 29.5 41 105 1 40 rank ordered data sets We show plots of Average of fllte11sive Composite Scores for all Universities 130 .66 these results in Figure 3. For example, the 78 Chen1ical Engineering Education

PAGE 81

TA B LE S Examp l e of Rank O r der of C h E Graduate Program s u s i n g a Single Compo s ite Extensive/Intensive Criterion C.o mp_ os i te Sco r e Co ma os it e Sco r e U 11i ve r s ity R esc al etf'! No rmali ,e d D ev iati o 11 W Ra11k U 11 ive r s ily R esc al et! -'1 N ormali ze d D ev iati011 W Ra11k U ni versity of Mi n n eso t a 533.9 1 00 0 4.9 I Pu rd u e U n ivers i ty 1 68 5 3 1. 6 1 3.3 25 Massac hu se tt s I ns t. of T ec hn o l ogy 515 8 96 6 2.1 2 U ni v. of Massac hu se t ts, Amhe r s t 1 67.6 3 1. 4 1 3 8 26 U n ive r s i ty of T exas a t A u s tin 453 8 85.0 2.2 3 U nj ve r s i ty of Co l o r a d o 1 59.9 29.9 I O. I 27 U ni ve r s i ty of Ca li fo rni aB erke l ey 3 1 9 4 59.8 1 8.8 4 (,) U n ive r s it y of Oklahoma 1 57.7 29.5 30.4 28 ( d ) U n jve r s i ty of Wi sco n s in M a di so n 283 5 53. 1 8.2 5 Syrac u se U n iversity 1 5 1. 4 28 4 1 4 2 29 U ni ve r s it y of D e l awa r e 269.8 50.5 6.7 6 Oh jo St a t e U ni vers it y 1 46.6 27 5 1 0.9 30 S t a n fo rd U ni ve r s it y 267 3 SO. I 9.9 7 Unjve r si t y of Was h i n g t on 1 42.2 26.6 1 2 8 3 1 Ca l ifo rnia In sti tut e of Tec hn o l ogy 246 9 46.2 7.3 8 University of Ca l ifo rni aD avis 137. 1 25 7 1 8 .l 32 Prin ce t o n U ni ve r si t y 227 4 42.6 6.6 9 U n jve r s i ty of H o u s t o n 136.2 25.5 2 1. 6 33 T exas A&M U ni ve r s it y 226.4 42.4 20.4 IO ( d ) R e n sse l aer P o l ytec hn ic In s titu te 1 36.0 25.5 2 9 34 U n iv. o f C ali fo rni aL os A n ge l es 225 6 42.3 1 5.2 11 U n ive r s it y of Te nn essee -Kn oxvi ll e 1 34 6 25.2 37.7 35 (d) U ni v o f Cali fo rni a-Sa nt a Barb ara 2 1 9.3 4 1.1 4.6 1 2 U n ive r si t y of Vir g inj a 1 34.5 2 5 .2 7.5 36 U ni v o f Illin o i s, U rb a n a-C h am p a i g n 2 1 0.0 39.3 8 .1 1 3 Georg i a In s titu te of Tec hn o l ogy 1 34 1 25. 1 1 6.4 37 Case W es t e rn R ese r ve U n iversity 205.5 38.5 1 4.5 1 4 L o ui sia n a S t a t e U n jve r s i ty 1 32.5 24.8 1 3.3 38 P e nn sy l v ani a St a t e U n ivers i ty 205. 1 38.4 1 5.2 1 5 O k lahoma S t a t e U nj versity 1 3 1 .0 24.5 3 1. 0 39 ( d ) N o rth wes trn U ni ve r s i ty 203. 1 38 0 1 5 8 1 6 Ya l e U ni ve r s i ty 1 25.8 23.6 1 8. 1 40 John s Hopkin s U ni ve r s it y 202.7 38.0 1 6.4 1 7 N o rth Carolin a Stat e U ni ve r s it y 1 96.4 36.8 7.0 1 8 M xtens i ve sco r e p l us r esca l ed i ntensive score as desc ri bed i n text Sta t e U ni v. of New Y orkBu ffa l o 195.3 36.6 1 5.2 1 9 lb/ Sta n da r d dev i ation of the rank order nu m bers of t he eig ht qua l ity U n ive r s i ty of Mi c hi ga n 185 9 34.8 7.5 20 m e asures as des c ribed in text. Corne ll U ni ve r s it y 183 0 34.3 6.8 2 1 1 '! R anki n g ma y be low because of diff e r e /1/ basis or erro r in R esea r ch Unive r s it y of P e nn sy l va n ia 182.4 34 2 20.9 22 (,) Support category Le hi g h U ni ve r s it y 1 72.5 32.3 11. 9 23 /di R anki n g m a y be h i gh because of differe/1/ basis or error i n R esea r ch Ca rn eg i e M e ll o n U ni ve r s it y 1 71.7 32.2 1 6.7 24 Sup p ort ca t ego y. Figure 3. ( a ) 0 Fra c tional 0.6 0 c h a n ge in I 0. 60 ci tations I 0. 50 0. 50 t,. ( TC) l (TC) 0 40 0.40 vers u s rank = G o rd e r ; top;. t. 0 30 !::. 0 30 .E rank e d .E
PAGE 82

into the first decile or out of the tenth decile. Without knowing details of the data collection it is not possible to make a rigorous assessment of the uncertainty in the rank ordering. However a heuristic assessment can be made. We assume, based on experience and for purposes of argument, that there are independent errors of 10 % in each of the four extensive numerical data sets: Publications, Cita tions, Support and Honors. In the middle range of each of the extensive data sets, the fractional change, 6. XIX, re quired to move one place in the rank order is approximately 0 03 (see Figure 3). A fractional error of 0.10 therefore corresponds to approximately 0.10/0.03 "" 3 places in the rank ordering Also, if the errors are independent one would expect an error in the composite rank ordering of extensive criteria to be approximately = 6 places. If similar argu ments are used for the four intensive data sets we also find an approximate error of six places This error is not indepen dent of the error in the extensive rank ordering. We conclude that, in the middle range, programs within 5 to 10 places on the composite extensive/intensive rank ordering are essen t i ally indistinguishable from each other. This estimate is consistent with our common-sense interpretation of the rank ordering e g., programs in the second decile are probably superior to programs in the third decile, and so on. The eight separate (though not completely independent) measures of quality give the rankings a certain degree of robustness that a single criterion would not have. The inter nal consistency of the eight measures of quality is estimated by computing the standard deviation of the rank order number of the eight separate quality categories for each program (see Table 8) For example, for MIT the average of the eight rank orders is 3.5 and the standard deviation of rank orders is just { (2-3 5)2 + (33 5)2 + = 2.07 8-1 Large values of the deviations indicate programs where indi8 0 TA B LE9 Summary of R 2 Values for Linear Fits Between Ra n k Order of Faculty Quality (93Q) and Rank Order in the In d ivi du al Rankings Ranking Category Fit R' PUB y=mx+b 0.7469 PUB/FAC y=mx+b 0 6005 CIT y=mx+b 0 7482 CIT/FAC y=mx+b 0.6028 SUPP y=mx+b 0 2484 SUPP/FAC y = mx + b 0.1351 HON y = mx + b 0 6993 HON/FAC y=mx+b 0 6370 0 I I 70 4 a 1 ; 60 -~ so = 0 40 ;; .. '"' ... 30 "' .. .. 1! 20 0 ... 10 0 0 70 a .., e 60 ~ = so 0 t 40 'ci 30 .. ? 20 0 ... : 10 "' 0 0 0 I R =o n l o ,, 10 20 30 40 so Ran k Ord er o f Ex t ensive Meas ur e j R 2 = 0 65 1 0 0 e G 0 0 ,, <-, Cl _.,,. 10 20 30 40 so Ra n k O rder o f I ntensive Meas u re 0 0 0 e 60 60 IO 20 30 40 50 60 Ra n k O r d er of C om p o s it e M e as u re 0 70 70 70 Figure 4. ( a ) Composite extensive rank order versus rank order of faculty quality (93Q}; top-ranked programs are near origin. R 2 =0.72 for linear fit, y=mx+b (b) Composite intensive rank order versus rank order of faculty quality (93Q); top-ranked programs are near origin. R 2 =0.65 for linear fit, y=mx+b. (c) Composite intensive/extensive rank order versus rank order of faculty quality (93Q}; top-ranked programs are near origin. R 2 =0.72 for linear fit, y=mx+b Ch e mi c al Engine e ring Edu c ation

PAGE 83

a 20 e IS "" '8 ] 10 El :z: b e 0 25 20 IS "" ... .. i 10 I :z: s 0 20 e IS Cl, '8 ] 10 El :z: 0 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 N orm a llud E xt en sive Sc ore Ran g e 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 Nor m allu d In te nsive Sc o re Range 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 N orm aliu d C om posite St o re Range Fig u re 5 ( a ) Histogram showing number of programs ver sus scaled total extensive score; top-ranked programs are at the right. (b) Histogram showing number of programs ver sus scaled total intensive score; top-rank ed programs are at th e right. (c) Histogram showing number of programs ver sus scaled total c omposite extensive/intensive score; top ranked programs are at the right. Winter 1999 vi ctual quality rankings are the lea st internally consistent. In many cases, large de viat ion s are associated with small pro grams that rank higher in intensive than in extensive catego rie s. However in so me cases, large deviation s may indicate problems in the data For example, four programs have much higher rankings in re searc h s upport than in the other quality categories: Texas A&M, Oklahoma State University, the University of Oklahoma and the University of Tennessee. On the other hand the University of California at Berkeley and the University of Penn sylvania h ave much lower rankings in research s upport than they have in the other categories. We believe that the se di s parities likely arise from different reporting ba ses and may not reflect true differences in re search support. Three programs have a much higher ranking in Honors than in the other quality categories: Pennsylvania State University, Carnegie Mellon University, and North western University. The rank order of perceived faculty quality (93Q) was correlated with the rank order of each of the se parate quality categories (see Table 9) The weakest correlations were found with the two s upport categories, SUPP and SUPPffF con sistent with our belief that so me of these data are not re ported on a consistent basi s. Nevertheless, we are reluctant to exclude research s upport from the quality measures Re search support is probably a better current and leading indi cator of qualit y than the other categories. Also, total research support is a primary criterion used for assessing quality of re searc h universitie s .l 5 l Rather than re-ranking the programs excluding SUPP and SUPPffF, we believe it is more reason able to identify program s where different bases for reporting s upport may have s trongly influenced the rankings. Plot s of the rank order from the faculty quality survey (93 Q ) from the NRC report versus the overall extensive, overall intensive, and composite rankings are given in Fig ure 4. In the figure, the high-quality programs are near the origin. The figure clearly s how s how the quality survey and our method s identify the sa me several programs as the highest quality. PART3 D I SCUSS I ON A N D RECO MME NDA TI ONS Interpretation of Rankings There is no calibration stan dard for quality against which any methodology can be tested Neverthele ss, we find it very suggestive that our composite extensive/intensive ranking and the NRC reputational s urvey identify the same few top programs. For example, comparison of Table s 4 and 8 shows that the same top two programs, MIT and Minne so ta and nine out of the ten top-ranked program s are the sa me in both the NRC reputational ranking and our numerical ranking. But only three out of ten program s in the seco nd decile and two of ten in the third decile are the same The NRC reputational 8/

PAGE 84

rankings, which rely heavily on anecdotal word-of-mouth information, will be most accurate for the few high-profile, extremely good programs, and will be less accurate for smaller lower-profile, and second-tier institutions. But the numerical measures of quality should remain useful in as sessing the relative quality of all institutions W e conclude, subject to the c av e ats given about th e a cc urac y of th e data itself, that our simpl e numerical measures do c orrelate w ith program quality as it is normally understood. Further comparison of Tables 4 and 8 leads to additional insights. One may divide programs into three broad catego ries. First are the programs that are highly rated on both the NRC reputational survey and the numerical ranking. Prime examples are the University of Minnesota Massachusetts Institute of Technology the University of Wisconsin, the University of California at Berkeley, and Stanford Univer sity Second are programs that rank significantly higher in the numerical ranking than in the survey. These programs are often (but not always) associated with smaller, research intensive programs. Examples are the University of Califor nia at Los Angeles, Case Western Reserve University and Johns Hopkins University Finally, there are well-known programs, which do well in the reputational ranking that do not do as well in the numerical measures These programs may be relying on past, rather than current, performance. Further insight can be obtained from histograms of the final scores, shown in Figure 5. For ease of interpretation, in the figure the scores from Tables 6, 7, and 8 were scaled to give maximum values of 100. Figure 5a and Table 6 clearly show that three programs (MIT Minnesota and Texas) have extensive scores well above all other departments. This dis parity is lessened somewhat when the intensive scores are compared (Figure 5b and Table 7) This same uneven distri bution of scores is found in Figure 5c, which shows the distribution of composite extensive/intensive scores The top half of the composite score range contains only seven pro grams; the remaining programs fall in the lower half. The summary shown in Table 8 and Figure 5 indicates that the highest quality chemical engineering programs are relatively few in number and significantly higher in quality than the rest. Below the top five or six programs there is a wide range of programs with relatively similar quality. Finally while we believe that for most programs the rankings given here are an accurate reflection of quality, we emphasi z e on c e again that one should be cautious in draw ing conclusions from the absolute position in the rankings of a single program Limitations of Ranking Systems Respondents to the NRC questionnaire were asked to rate fifty separate pro grams. An individual respondent will only have personal detailed knowledge about a small fraction of these. The resulting reputational rankings will inevitably be influenced by the network of informal contacts and acquaintances of the 8 2 respondents. This will lead to a bias against smaller pro grams and will also make the reputational rankings a lagging indicator of program quality Another major problem with the NRC report recognized by the committee was the lack of data on the performance of graduates from the programs. We were unable to find any direct quantitative measure for asses s ing the performance of graduates of chemical engineering graduate program s Since one of the principal goal s of a graduate program is the education of the next generation of researchers, this is a serious omission indeed. Personnel departments of major corporate employers of PhD chemical engineers often main tain internal ratings of programs based on the performance of their employees Perhaps this information can be pro vided in some suitable blind format to future NRC com mittees. This is a project that could be addressed by the AIChE and the other engineering societies. Another pos sible measure of performance is the number of graduates that obtain tenure-track appointments at research univer sities other than their own. We suggest that future ranking systems also include some measures of the effectiveness of technology transfer. To partially accomplish this, the Publication category could be expanded to include patents issued to faculty and graduate students. Similarly, Citations could include papers or patents cited within patents. More difficult to count but very useful, would be the number of new businesses formed as a result of activities within the program. The quality measures used in the NRC report and in this paper are appropriate for doctoral-level research-based gradu ate programs. However, master's-level programs, especially practice-oriented programs, are of growing importance Fu ture ranking s ystems should attempt to separately measure the quality of these programs. The difficulty in accounting for the rapidly changing in terdisciplinary nature of modern engineering is another prob lem encountered when developing ranking systems. Tradi tional academic boundaries do not always reflect the reali ties of engineering practice. The NRC report addressed this problem by ranking "programs rather than "departments For chemical engineering these two categories are usually commensurate but this may not be the case for chemical engineering programs with strong efforts in biotechnology or advanced materials Ranking programs with major com mitments in these fields can be difficult when the academk administrative units do not correspond to the categories used in the ranking scheme Very strong, interdisciplinary efforts may not appear in the data set or conversely, remote extra neous efforts can be included Obtaining a reliable data set, based on uniform criteria is a formidable task. The NRC committee had great difficulty in defining program bound aries in modern biology and molecular biology, where the pace of change is particularly great. Ch e mi c al En g ine e rin g Edu ca tion

PAGE 85

Neither the NRC report nor this paper uses any mea sures of the quality of graduate teaching The lack of quantitative measures of teaching performance is a con tinuing long-term problem The very long time between the NRC reports (1982 to 1995 ) is yet another problem. Waiting more than a decade for an assessment is s low even by the s tandards of academia. Some form of continuing assessment, for example on a triennial basis would be more useful. This would give more timely results and would also permit running averages of several years to average out fluctuations in the data Concluding Remarks With all of these difficulties, one can legitimately ask why bother with rankings at all? We believe that universities will be under ever-increasing pres sure to justify tuition rates and the cost of performing re search. Whether we like it or not, ranking of academic pro grams will continue and will likely increase. It is in the profession's interest to see that the rankings are based on rational measurable criteria. But there is little reason to continue relying on surveys. Reputational rankings only con firm the obvious about the top few programs, permit declin ing programs to remain co mpla cent, and fail to recognize increasing quality where it occurs. ( Conclusions and Recommendations ) 0 Alternative mea s urable quality indice s exist that correlate well with graduate program quality as it is normally under stood. 0 The profe ss ional soc ieties the National Academy of Engi neering and the National Science Foundation should take the lead in developing the se quantitati ve measures of pro gram quality and appropriate data bases to s upport the se measures. 0 Special attention s hould be paid to developing methods for assessing the performance of students after they receive their graduate degrees ; this shou ld include using information from employers of graduates. 0 Methods of assessing the effectiveness of technology trans fer and impact on indu s try s hould be developed 0 Assessments should be made on a mor e frequent schedule perhaps triennially REFERENCES 1. Goldberger M.L. B.A. Maher, and P E Flattau Eds. Re search-Doctorate Programs in the United States: Continuity and Change National Academy Press Washington, DC (1995 ) 2. America 's Best Graduate Schools," U.S. News and World Report, 2400 N St NW Washington, DC 20037-1196 ( 1996 ) 3. Mervis J ., Sci e nce 168, 1693 ( 1995 ) 4. National Science Foundation, Academic Science and Engi neering R&D Expenditures: Fiscal Year 1993 NSF 95-332 ( Arlington VA 1995 ) Table B-48; and National Science Foundation, Academic Science and Engineering R&D ExWinter 1999 penditures: Fiscal Year 1994, NSF 95-332 ( Arlington, VA, 1996 ) Table B-48 5. Graham, H D ., and N. Diamond, The R ise of American Re se arch Univ e rsiti es, Johns Hopkins University Press Balti more, MD ( 1997 ) 6 National Science Foundation web site, April 1997 7. A c ti vi t ies D irectory 96 American Institute of Chemical En gineers AIChE New York NY ( 1996 ) 8. Dir e ctory of M e mb e rs and Foreign Associates National Acad emy of Engineering Washington, DC, September ( 1996 ) 9. Ch e mical Engin ee ring Faculty Dir ectory : 1996-1997 45, AIChE New York NY ( 1996 ) 0 BOOK REVIEW: Alternative Fuels Continued from page 39 The u se of geothermal energy is presented in Chapter 9 This topical discussion notes that at depths of about six miles from the earth s surface the temperature is greater than 100 C. Thi s equates to more energy storage than the total thermal energy in all the nuclear and fossil fuel resources only solar energy is comparable Along with scientific and technological update s, the advantages and disadvantages of geothermal energy utilization are outlined; this alternative so urce of energy will potentially become a larger part of the world's energy consumption in the near future because geo thermal energy is both available and economical. In the United States approximately 3 GW of electric power is produced in 20 power plants from geothermal reservoirs. Geothermal energy also has great potential as a practical provider of heat to local areas. The overall conversion routes of biomass are described in Chapter 10 They include thermal (combustion, gasification, liquefaction and pyrolysi s), anaerobic digestion, and fer mentation to liquid ethanol fuel. The descriptions in clude 15 process diagrams and several tables of data. A selected amount of cost data is provided for ethanol pro duction from lignocellulose. Chapter 11 presents a comprehensive overview of rela tively recent development s in the generation of energy from municipal solid wastes, including spent tires and polymeric materials Proce sses include incineration anaerobic diges tion and landfill gas reco ve ry pyrolysis thermal cracking, and partial oxidation via s upercritical fluids. In s ummary Alternativ e Fuels s uperbly achieves its pur pose by bringing together a wealth of practical information required for a thorough understanding of those chemical process technologies urgently needed for the development of fuel s for future u se. Dr. Lee is to be commended for his extraordinary efforts in sy nthesizing all these facts and sys tem s in a clear and consistent manner. Possibly his next book could focus more on the use of biomass geothermal, and solid waste resources-three areas that are undergoing rapid development. 0 83

PAGE 86

.ta_..5 1111111 3._l_e_a_r._n_ i _n.:g:...._in_, n_d_u_s_t....:ry:...._ ___ ) This co l umn provides examples of cases in which students have gained knowledge insight and experience in the practice of chemical engineering while in an industrial setting. Summer internships and co-op assignments typify such experiences; however reports of more unusua l cases are a l so welcome. Description of the analytical tools used and the skills developed during the project should be emphasized. These examples should stimulate innovative approaches to bring real world tools and experiences back to campus for integration into the curriculum. Please submit manuscripts to Professor W. J. Koros, Chemical Engineering Department, University of Texas, Austin, Texas 78712. FROM THE CLASSROOM TO THE WORKPLACE Motivating Students to Learn in Industry A. CHRISTIAN FRICKE Rensselaer Polytechnic Institute Troy, NY 12180 W hat makes a successful engineer? No one would deny that technical expertise is critical to master ing real-world engineering problems. Yet techni cal mastery is only half the battle; there are also many com plex social skills that must be learned in order to make success ful use of technical knowledge in a workplace setting. The purpose of this article is twofold First, it will note a few examples of specific nontechnical skills that can be useful in managing the day-to-day workplace realities of a BS-level engineer. These skills and strategies are taken from the author's personal experience in working as a process engineer for three years at a mid-sized manufacturing con sulting firm, as well as from conversations with and observa tions of dozens of col l eagues working in varied chemical and mechanical product design and manufacturing settings. The majority of these engineers were within six years of graduation and were in the process of learning the social skills necessary for moving up the corporate ladder from technically oriented process positions to more business-ori ented managerial functions. It is hoped that these observa tions will prove useful to engineering professors who have not worked for an appreciable amount of time at the BS level and who therefore have experienced the industrial setting in a much different context. Second, this article will outline several ways in which the need for acquiring these informal skills can be communi cated to the vast majority of students who will end their education at the BS level. The conveyance of technical con cepts, skills, and information is undoubtedly what the under graduate experience is all about, but by suggesting some of t h e social contexts within which these skills will be mobi8 4 lized, engineering educators can increase students effec tiveness in putting this technical material to use in the work place. There exists a real need to alert undergraduates to the fact that excelling in the classroom, although critical, is only half the equation in preparing to be an effective professional. Otherwise naive students need to be explicitly made aware of the seemingly commonsense notion that one must indeed "learn in industry" in order to be a successful corporate engineer. STUDENT PERCEPTIONS T h is need is illustrated by a recent survey of seventy-six undergrad u ate engineers at Rensselaer Po l ytechnic Institute (RPI) in which 95 % indicated they had a "very firm" or at least a "somewhat firm" idea of what the daily work experi ence of an average engineer is like (see Figure 1) A close look at the numbers indicates that students likely do not have the firm grasp of engineering workplace realities that they profess For instance there was no correlation between students A. Chr i stian Fr i cke graduated from North Caro lina State University with 88 degrees in chemi cal engineering and biochemistry He has worked in a Merck Pharmaceuticals production facility as a research assistant in a molecular genetics laboratory and as a process engineer for CTC a manufacturing consulting firm with headquarters in Johnstown Pennsylvania He is currently a doctoral candidate in the Science and Technology Studies program at RPI. Copyrig h t Ch E Di v i s i on of AS E E 1 999 Ch e mi c al En g in ee rin g Edu c ation

PAGE 87

There exists a real need to alert undergraduates to the fact that excelling in the classroom, although critical, is only half the equation in preparing to be an effective professional. .. Success on the job ... entails learning many complex social behaviors in addition to those necessary for classroom success. It also entails developing an entirely new perspective on what constitutes "engineering." reported knowledge of "w hat it is that engi neer s do on a daily basis and their personal relationships. Students who had no close rela tives or acquaintances with engineering back grou nd s (more than 60 % of those surveyed) were just as likely to indicate a firm knowl edge of daily working realities as those with engineers in the family." In the absence of actual engineers to talk with and observe, student conceptions of "wo rkplace realities" are vague and simplistic at best. TABLE 1 Overall Survey Results (76 respondents ) multiple potentially "correc t solutions In hi s book Designin g Engineers, [ 21 Louis Bucciarelli characterizes these fundamentally soc ial aspects of engineering de s ign and prac tice in the following manner: Response to the (ll/estion: "Do you hav e a.firm idea of what it is that engineers do 011 a daily basis? [ P ]arti cipan t s in design work wit hin a ri c h, multidimensional environment that reaches we ll beyond the narrow co nfin es of their own ob j ect worlds. A custome r s needs are not given or discovered, but must be cre ated; an operator's capabilities must be de fined; building codes need interpretation; Very Finn Idea: 30% Somewhat Firm Idea : 65 % No Firm Idea: 5 % No Idea Whatsoever : 0 % This simplistic view can undermjne an e ngin eer's effectiveness in accomplishing personal, profes sio n a l and socie tal goa l s in th e workplace. In addition, the apparently prevalent student attitude of believing they al ready know what professional working realities are all about can serio u s l y limit the benefits to be gained from intern and co-op ex peri ences. These experiences provide the ideal se ting for observing the practical day -to -day social s kill s and strategies necessary for BS-level s uccess. In order to realize this benefit, however students mu st be actively lookin g for these potential lessons in the first place. UNDERGRADUATE SOCIALIZATION The majority of undergraduate s tudents form their first concrete conception of "e ngineering through survey courses and introductory seminars that are s tructured to help fresh man an d first-semester sophomore s tudents choose a par ticular discipline. At RPI, for example, second-semester fresh man stude nt s take a course, titled "Engineering Seminar ," that is designed to provide the s tudent with information relative to the various engineering field s and curricular areas ."[ 11 These t ypes of survey courses generally focus on the e nd products of engi n eering work. In other words they emp ha size what it is that the vario u s disciplines accomplish. They leave und ergraduates with a feeling that they understand what it is that engineers do but without an appreciation for the soc ial realitie s of how these tasks are accomplished. This distinction i s significant. Undergraduate coursework fosters the perception that the engineering working experi e nce is one of so lving highly conceptual, well-defined, sc ence-based problems in a largely individualized setting, with emp ha sis on arriving at a single, objectively "c orrect solu tion But the reality that working engineers enco unt er is one of solving highl y practical, undefin ed, procedural-based prob lem s in an extremely tight-knit social setting, resulting in Wint er / 999 costs must be tried out; budget limits must be agreed upon. The task must be organized into subtasks; suppli ers must be coaxed to co mmit t o a price and d e liv ery date; the dropout problem at Phot oqu ik must be con structed. All of this is designing. In all of this, cho i ces are being made, decisions foreshadowed, and possibilities discounted. In other words, working engineers mu st create and manage formal and informal social s tructure s in order to generate built product s. Success on the job therefore entails learning many com plex social behavior s in addition to those n ecessary for class room success It also entails developing an entirely new perspective on what constitutes "e ngineering ." Without re structuring the entire undergraduate experience to incorpo rate these workplace lessons engineering educators can nev ertheless prepare s tudent s for this impending paradigm shift b y at least bringing it to their attention. In addition, there are many s pecific exercises that can be easily incorporated into the existing undergraduate curriculum to reinforce some of the nontechnical social skills necessary for s uccess in the corporate workplace. WORKPLACE REALITIES According to one early '80s study, technical profession als typically s pend over a third of their work week writing, editing, or preparing report s "[ 3 J If you also include compos ing letters propo s als, drafting schedules and procedures, taking field notes, and generating other more informal modes of written communication, then "w riting easi l y occupies more than half of the typical engineer 's work experience. Oral communication al so occupies a major portion of the engineer's time. Thi s can include time spent in meetings or on the phone with vendors or customers, time s pent on the shop floor interacting with technicians and workers, etc. 85

PAGE 88

Taken together, these two activities comprise by far the bulk of an engineer 's work week. In the workplace setting that the vast majority of graduates will enter, typical engi neers will likely use only 10 % of their technical background 10% or so of the time. Of course, the specific 10 % will vary widely for each individual, making technical breadth within the curriculum essential. But the fact remains that most of a working BS-level engineer's time will be spent not actively solving technical problems, but instead communicating po tential technical solutions to others. Engineers' effective ness, reputation, and career success will be based on techni cal expertise, yet determined by how well they manage to translate this expertise into action through mastering such nontechnical workplace skills as effective communication, organization, and persuasion This is of course, not a or,e-way flow of information For every memo that is written or presentation that is given, someone (presumably) reads and listens. The successful en gineer also has to take in and interpret an enormous amount of written and verbal information Organizing and making effective use of this information requires good critical read ing and listening skills. Given the enormous amount of in formation generated in the typical corporate workplace, quickly and effectively separating the wheat from the chaff i s an important skill in itself. The vast majority of a BS engineer's time is taken up with both taking in and communicating information. Success re quires possessing the nontechnical skills necessary to first recognize and then convince and organize others to act on information that is important. CONVEYING WORKPLACE REALITIES So, how can you alert engineering undergraduates to this reality? One strategy is to suggest that effective communica tion is an essential engineering skill-one that can be just as important as any technical ability. In the words of historian Henry Petroski 14 1 "some of the most accomplished engi neers of all time have paid as much attention to their words as to their numbers, to their se ntences as to their equations, and to their reports as to their design s Pointing out to students the vital importance of mastering effective writing, reading speaking, and listening reinforces the notion of engineering practice as a social activity. Yet there are many other "nontechnical" skills that are also important to success at the BS level. If presented to students at all, these skills are most often communicated in the most general of terms, with successful engineers de scribed as possessing "curiosity," perseverance ," "self-con fidence," "common sense," and so forth. What undergradu ates need is a resource that highlights the importance of specific skills, motivated by a concrete social context and picture of the day-to-day realities of corporate engineering practice. The key is to motivate students to appreciate the 86 complex social realitie s of engineering practice by giving them a tangible feel for the workplace setting that most will find themselves in. Using Popular Culture One resource for accomplishing this is the comic strip "Dilbert." In many respects, "Dilbert i s an entirely accurate ethnographic account of the typical BS-level engineering experience. According to one le a ding management consultant, 1 51 It 's not a comic strip, it's a docu mentary-it provides the best window into the reality of corporate life that I've ever seen." It therefore provide s an excellent resource for undergraduates (or anyone else, for that matter) interested in the daily interactions of practicing corporate engineers. Dilbert can be read as providing very specific, contextualized examples of the many workplace issues and challenges that BS-level engineers must confront and overcome in the process of applying their technical knowledge to real-world problem s. Of course, illustration doe s not imply prescription Dilbert should certainly not be taken as illustrating a social ideal or model for how engineering professionals ought to navigate these issues. It can, however offer a view of what so me of these issues are and motivate students to contemplate how they would manage similar circumstances in a more con str uctive manner. Dilbert provides an alternative insider 's perspective that if presented as se rious social satire and critique, can be a valuable learning tool for preparing for the reality of the engineering workplace. Taking Dilbert as serious social commentary can pre pare students for making the leap from viewing corporate engineering as a purely technical activity to seeing it as a technically mediated, yet essentially so cial endeavor. It can also prevent the di s illusionment commonly generated by the experience of realizing that daily workplace realities are quite different from naive undergraduate preconception s Discussing Specific Strategies A truly comprehensive list of specific social skills useful for managing Dilbert like situations in a constructive manner would be almost infinite in length This section merely presents five strate gies that can be particularly critical to workplace success. "Newly minted BS-level corporate engineers usually learn these stra tegies only after a so metime s painful and poten tially damaging period of trial and error. Discus s ing these important nontechnical ski ll s within the context of appropri ate undergraduate coursework can benefit graduates by ac celerating their on-the-job learning curve enormously. 1. Save everything that crosses your desk. Undergradu ate education reinforces the notion that when some thing is done ," it 's over with. With the end of each semester, textbooks are returned to the bookstore and class note s are relegated to recycled paper bins. But in the real world, project s never really come to an end. You never know when, say, a cost analysis done for a long-forgotten proposal might come in handy. Chemi c al Engineering Education

PAGE 89

Saving old work (even draft work) can prevent future duplication of effort. 2. Document ever y thing in writin g There is no such thing as an unambiguous verbal order. 3. Learn to use a daily planner Some corporations provide a standard dayplanner system free of charge to technical employees, but even if their use is not officially encouraged, dayplanners are an essential tool for maintaining order in personal and project schedules (planning meetings, scheduling travel, keeping notes, maintaining contacts, etc ) Under graduates are used to having order imposed for them-tests, project deadlines class times course materials, etc. are all organized in advance. This leaves novice engineers completely unprepared for the job of creating their own order from the chaos of daily events. A good dayplanner system is an indispensable tool for managing this process. 4 Use the "plus a fifth" rule. One of the most difficult things to learn in managing complex technical efforts is how to account for the unexpected. Even the most detailed, well-researched proposal or project plan can be subject to unanticipated delays, setbacks, cost overruns, and instances of Murphy's Law in action. Planning for the unforeseeable is a management skill that can only be learned through experience. In the meantime, beginning engineers can instead simply assume that all but the most straightforward tasks will take 20 % longer and cost at least 20 % more than expected. Even if nothing goes wrong, coming in significantly under budget and ahead of schedule can be much preferable to the alternative for all involved. 5 Learn where and when to c ompromi se A critical skill for managing working relationships is knowing when an issue is important enough to battle over. One engineer who is employed by a large tool and home appliance manufacturer characterizes this as realizing that "the sun doesn t rise and set on a toaster oven Maintaining an uncompromising stance on say, the color of a new product can make collaboration impossible. But no compromise should ever be made on any aspect of engineering design production, or management that infringes on the health or social welfare of others Students should recognize that they will have to make such distinctions for themselves and that the answers will rarely be clear-cut and obvious. Practicing Workplace Strategies in the Classroom In addition to simply discussing the aforementioned strategies there are also many relatively simple, straightforward teach ing techniques that can be employed to help students de velop positive social working skills like good communica tion, organization, planning, patience etc. Once again pre senting a truly comprehensive list would be impossible Wint e r 1999 this is merely a collection of seven specific activities to illustrate the breadth of possibilities. l. Pe,forming peer e v aluations for student oral presen tations. Having undergraduates evaluate one another's presentations in a structured manner would focus listening skills and give students practice in recogniz ing key points, initiating critical discussion etc. 2. Practi c ing giving oral and written equipment operation and sampling procedure directions. Unit op labs provide a plethora of opportunities for sharpening interpersonal communication skills. For example students could actively direct and observe each other rather than passively following TA instructions. Also, students could be required to generate written operation and sampling procedures for subsequent lab groups to follow. These experiences would highlight the importance of precision and clarity in giving both written and verbal direction. 3. Swapping notes for lab reports. Another pote n tial unit-ops exercise would be to require groups to exchange notes and generate reports based on each other's data. This would highlight the importance of preserving data and other information for unexpected future uses while also stressing the necessity of precision and clarity in all forms of engineering communication. 4 Writing and presenting standard business communi cations. Practice in writing and speaking can be combined with aiding students in their job search. First, students could be asked to research and produce a short report and presentation on a particu lar industry or market. Then students could generate a resume and letter of application to an appropriate company based on their research. This would give students practice in evaluating what is important to individuals working within other organizations, while at the same time reinforcing proper business communi cations etiquette sharpening business research skills, etc. Students could also gain valuable experience from attending departmental seminars and producing short memos or similar communications detailing key information presented, summarizing discussions, etc. 5. Producing detailed project plans. Senior design courses also provide a wealth of opportunities for practicing nontechnical "real-world," social and organizational skills. For example, students could be required to generate detailed project proposa l s outlining specifically what is to be done and how it will be accomplished, complete with a breakdown of activities, timelines for completion etc. This would give students experience in organizing work and delegating responsibility in a formal and considered fashion. Coupling this activity with a proposal 87

PAGE 90

pre s entation would al s o g ive s tudent s practi ce in clearly articulating and a dvo ca ting a propo se d cour s e of a cti o n. Al s o requiring periodic proj ec t updat es would prompt s tud e nt s to mana ge d e l ays a nd s etba c k s in an organiz e d fa s hion 6. D e alin g w ith ve ndor s. Anoth e r good s enior de s ign experience would be to as k s tudent s not ju s t to model a proce ss, but to al s o locate s pec out and price the s pe c ific equipment nece ss ary to make th e proce ss run Thi s could be done by s imply g iving student s access to a Thoma s Regi s ter (now avail a ble on the Internet ) and a telephone. This would expose students to the realities of uncertainty and would likely require management of time and ( imaginary ) cost overrun s 7 Rotatin g g roup memb e rs and r e sponsibiliti e s. Per forming well on corporate-engineering project team s mean s responding constructively to change This can be s imulated in the cla s sroom by requiring students to periodically reshuffle project lab and homework groups Similarly s pecific roles such a s coordinator note taker etc ., can be rotated within g roups. Having to con s tantly fill new roles and interact with different individual s would s harpen both leadership and col laborative cooperative interpersonal s kill s. In general implementing these activitie s would require some extra TA work and additional time s pent in providing more qualitative feedback on as s ignment s But all of the s e sugge s tion s can be incorporated into exi s ting coursework with a minimum of curricular di s ruption Outlining the Positive "Qualities of Success" Finally, s imply presenting a few s pecific context s of engineering practice can also motivate student s to begin thinking about the contra s t s between their undergraduate educational expe rience and their impending BS-level working reality. For example successful engineers (whether employed in sales processing, design management or any other capacity) are often called upon to Recognize problems that aren't apparent ( F o r ex ampl e b e ing abl e to wa lk throu g h a pr o du c tion fl o or and see o pportuni t i es f o r cos t-sa v in gs o r r ec o g ni ze s ubtl e e thi c al qu es ti o n s that o th e r s o ve rl oo k .) Define problems that are nebulous ( Pr o bl e m s in th e r ea l wo rld ra r e l y c om e n umb e r e d fo r easy refe r e n ce ) Choose solutions that are realistic (A s k i ll that d oes n t n ecessa ril y me an limit i n g th e ra n ge o f pos s ibl e solutions -o ft e n th e most su cces sful e n g in e er s a r e o n es w ho r ecog ni ze th e pra c ti c al p oss ibiliti es of see min g l y impra c t ica l a pproa c h es ) Plan how to make solutions work 88 (A pr ocess that includ es m a r s h a llin g r eso ur ces, m o t iv at in g o th e r s, k ee pin g p eo pl e o n t a sk r ecog ni z in g p o t e ntial pi tfa lls an d a co mpl ex co mbinati o n of man y o th e r di scre t e so c i a l s kill s.) Convince others to follow ( Th is i nvolves recogn i z i ng t h a t t h e i mp o r ta n ce a n d ur ge n cy o f a probl e m a nd th e feasi bili ty of a so l t io n a r e dir ec tl y pr o por t i o n a l t o t h e s kill a n d cl ari ty w ith whi c h th ey a r e p r ese nt e d a nd d efi n e d. ) Cooperate in dealing with contingencies ( Thi s in vo l ves r ea li z in g that exe r c i s in g p a ti e n ce a nd und e rstandin g w ith o n e's g roup m e mb e rs ge n e rat es th e coo p e rati ve so lid a ri ty n ecessary fo r ove r co min g c r ises.) Undergraduate s should view them s elve s a s continually s triving to meet thi s po s iti v e ideal through the con s tant ac quisition of constructive practical s ocial s kill s CONCLUSION In summary, undergraduate engineering students likely confuse familiarity with what engineers produce with what professional engineer s actually do on a daily basi s Thi s confusion is reinforced through introductory engineerin g sur vey cour s e s and an o v erall curriculum that empha s i z e s the built product s and technical aspect s of engineering over the s ocial proces s e s through which these products are generated. Although student s are given an opportunity for direct expo s ure to engineering workplace realities through intern and co-op experience s, the aforementioned preconception s are counterproducti v e to u s ing these experiences in the context of developing genuine con s cious in s ight into the e ss ential s ocial aspect s of engineering practice. Engineering educator s would benefit students by simply alerting them to the fact that the creative, challenging pro cess of learning to do engineering will not end but only begins at graduation. Likewise undergraduates would ben efit from being presented with contextualized examples of the daily workplace realities of corporate engineering life. Taking a few moment s to illustrate the social side of engi neering practice along with devoting s ome curricular effort to reinforcing the s e aspects of engineering work would help motivate s tudent s to think about their profe ss ional future s in concrete terms and provide undergraduates with a construc tive context for developing positive professional social s kill s In the end, thi s would result in more reflexive more effec tive engineering professionals REFERENCES 1. Engineering Seminar R e n sse la e r C a tal og 1996 97 RPI Publication s, Tro y, NY 317 ( 1996 ) 2 Bucciarelli Loui s L ., Designin g En g in ee rs MIT University Press Cambridge MA 149 ( 1994 ) 3 Lampe D.R. Engin e ers Invi s ibl e Activity : Writing ," T ec h nolog y R ev 73, April ( 1983 ) 4 Petroski H Engineers as Writers, Am e r. S c i e nti s t 423, September-October ( 1993 ) 5 Levy S ., Dilbert s World," N e w s w ee k 53, August 12 ( 1996 ) 0 C h e mi c al En g in ee rin g Edu c ati o n

PAGE 91

AUTHOR GUIDELINES This guide is offered to aid authors in preparing manuscripts for Chemica l Engineering Education (CEE), a quarterly journal published by the Chemical Engi n eering Division of the Amer i can Society for Engineering Education (ASEE). CEE publishes papers in the broad field of c h em i cal e n g in eering ed u cation. Papers ge n era lly describe a co ur se, a l aboratory, a ChE department, a ChE educator, a ChE c urri cu lum research program, machine co mput ation, special instructional programs, or give views a nd opinions on various topics of interest t o the profession. Specific suggestions on preparing papers TITLE Use specific and informative titles They s h ou ld be as brief as po ss ible cons i ste nt with the need for defining the subject area covered by the paper. AUTHO R SHIP Be consistent in authorship designation. Use first name second initial and surname. Give comp l ete mailin g address of place w h ere work was cond u cted. If current address is different, includ e it in a footnote on title page. ABSTRACT: KEY WORDS In c lud e an abstract of les s than seve nty-five words a nd a li st (5 or less) of keywords TEXT We request that manuscripts not exceed twelve doubles paced typewritten pages in l ength. Longer manu scr ipt s ma y be returned to the a uthor (s) fo r r ev i s ion/ s hort ening before being reviewed. Assume yo ur reader is not a novice in the field Include only as much history as is needed to provide background for the particular material covered in your paper. Sectionalize the article and insert brief appropriate headings. TABLES Avoid tables a nd graphs which involve duplication or s uperfluous data. lf yo u can use a grap h do not includ e a table If the r eader n ee d s th e table, omi t the grap h Substitute a few typical results for lengthy tab l es whe n practical. A void computer printout s NOMENCLATURE Follow nomenclature sty le of Chemical Abstracts; avoid trivial names If trade names are used define at point of first u se Trade names shou ld carry a n initial capita l only, w i t h no accompany in g footnote. Use cons i stent units of measurement and give dimen s i ons for all terms Write all equat i ons and formulas clearly, a nd numb er imp or tant eq uati ons consecutive l y. ACKNOWLEDGMENT Include in acknow l edgment only such cred it s as are essential. LITERATURE CITED References should be numbered and listed on a separate sheet in the order occu rring in the text. COPY REQUIREMENTS Send two l egible copies of t h e typed (do ubl e-spaced) manuscript on standard l etter-size paper. Submit original drawing s (o r clear prints ) of graphs and diagram s on se parate sheets of paper and include c l ea r glossy prints of any photographs that will be u se d. C h oose graph papers with blue cross sect i o nal line s; other co l ors interfere with good reproduction. Label ordinates and abscissas of grap h s a l ong the axes a nd outside the grap h proper. Figure captions a nd le ge nd s will be set in type and n eed not be lett ered on th e drawings. Number a ll illustrations consec utiv e l y. Supply a ll cap ti o n s and l ege nd s typed on a separate page. State in cover letter if drawings or photographs are to be returned. Authors should a l so include brief biographi ca l sketc he s and recent ph otographs with the manuscript. Send yo ur manuscript to Chemical Engineering Education, c/o Chemical Engineering Department Univers it y of Florida, Gainesville, FL 32611-6005


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EGWZ4BYDZ_XWTHW0 INGEST_TIME 2012-02-17T15:48:02Z PACKAGE AA00000383_00141
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES