Chemical engineering education ( Journal Site )

Material Information

Chemical engineering education
Alternate Title:
Abbreviated Title:
Chem. eng. educ.
Physical Description:
v. : ill. ; 22-28 cm.
American Society for Engineering Education -- Chemical Engineering Division
Chemical Engineering Division, American Society for Engineering Education
Place of Publication:
Storrs, Conn
Publication Date:
annual[ former 1960-1961]


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


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:
oclc - 01151209
lccn - 70013732
issn - 0009-2479
lcc - TP165 .C18
ddc - 660/.2/071
System ID:

Full Text

Maurice A.^ Larson


UNITED SATE Statement of Ownership, Management, and Circulation
(Requied by 39 USC 3685)
I Pubamtin Tila 2 PubEaton Number 3 FUln Dat
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.
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
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)

Chemical Engineering Education
Department of Chemical Engineering
University of Florida Gainesville, FL 32611
PHONE and FAX: 352-392-0861
Web Page:

T. J. Anderson

Phillip C. Wankat

Carole Yocum

James O. Wilkes
University of Michigan
William J. Koros
University of Texas, Austin

E. Dendy Sloan, Jr.
Colorado School of Mines

Gary Poehlein
Georgia Institute of Technology
Klaus Timmerhaus
University of Colorado

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

2 Maurice A. Larson, of Iowa State University

6 Washington State University

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

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

26 Non-Adiabatic Container Filling and Emptying, Noel de Nevers

32 FAQS, Richard M. Felder, Rebecca Brent

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

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

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

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




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


a balance of


research, and

service ....



alumni, faculty


across campus,




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

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

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




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 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

Website at


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

Lee, James
PhD. 1978: Kentucky
Bioprocessig, miring

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

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

Miller, Reid
PhD. 1968, Califora. Berkeley

Petersen, James
PhD, 1979: lowa State

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

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

are taught in
one senior class
by using the
Ethics Challenge
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.

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.

;: -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
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-

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."

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.

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.

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

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.

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

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

gasdynamics chemical equilibrium chemical kinetics

* Education
* Combustion
* Pollution
* Explosions

fax: 403-529-2516

Combustion Dynamics Ltd

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 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 ...





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


changes in



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


Winter 1999

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
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
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,
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


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.

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 Commodities


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


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'
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

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)

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

"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

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.

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.
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
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



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

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

d(mv)=F (1)

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

rate of
momentum out
of the control

rate of
momentum into
the control

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.

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.

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
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

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:
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

_.._ _


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
Assuming that the integrand is continuous and noting that
the limits of integration are arbitrary leads to the continuity
+ +V.(pv)=0 (16)
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

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)
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
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
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.

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)
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

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

p-p0o=02 (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
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
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-

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
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,

or P Po

when the following constraints are satisfied:

gL/c2 <<


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.


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)
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


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)

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
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

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

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

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)
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)

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)
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)

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


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.

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

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-


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/
(CAy) superficial volume averaged concentration of species A,
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:, Chemical Engineering Depart-
ment, University of Michigan, Ann Arbor, MI 48109-2136.



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?

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) -(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 = 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)
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

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


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

k-l (0.4)
Tfinal Pfinal k ) 86 kpa ) =0563
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

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 -sst -mi out kT (T Turroundings) (17)

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

Winter 1999

initiall houtt)- = -rmoutT(k- 1)- o(T Tsurroundings) (18) _
and the equivalent of Eq. (14) is
[-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
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

To atmosphere 15
Pressure and 10
temperature TPI
indicators a
connected to Quarter turn E
data logger ball valves 0
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
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.


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
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


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


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)
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)
with maximum value

Tthermocouple Twall
Tendofstep- wall Jmaximum

( b a
Y lb-a)

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.

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.

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

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


. 5

- 5

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

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.

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.

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

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


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...


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
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

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
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
How can I persuade my traditional colleagues to
do some of the nontraditional things you're

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.

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
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:/ / /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:/ /
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 <
fielder public Papers /long5.html >.) D

All of the Random Thoughts columns are now available on the World Wide Web at and at

Winter 1999

,f lIaboratory





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-
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?"

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

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
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
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
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.

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.

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-
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
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
Dallas: [The computer experiment] was the best. Granted,

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
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-


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.

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.

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.

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


An Electronic Version

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.

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.

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 (
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

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


Welcome to Chemical Engineering 210

Boad Chet


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



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
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.


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-
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.
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.


0.50(x,,6F6+x ,7F7+0.50F 2)=x6,.F,
x,,6F6+x,7,F7+0.50F 2=X,8F,
X3.8+X5.8+X6.8+X7.8+X8 8=
x, 8F,=0.50Fo,
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



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.

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

Bubble point: f(T.p)=I- K(i,T,p)z,..
Dewpoint: g(T,p)=l- Zi.I
I K(i,T,p)

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}

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

le(0,1) Solution is {0=.37704}

Figure 4

Winter 1999


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


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
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,

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 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 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.

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-
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

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.

1. Bungay, H., and W. Kuchinski, "The World Wide Web for
Teaching Chemical Engineering," Chem. Eng. Ed., 29(3),
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
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




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.

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




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.


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.

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.

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.



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.

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

Influence of Power Level
on Critical Moisture
Content for Drying Sand

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

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

2 0.0032
o nStepsize= 1 min

a 0.0024 Stepsize= 2 min

." 0.0016


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.

o 0.015 m-"=.
SEqn. 2


= 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.

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




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


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 in the USA and 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.


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.


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.


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).




E 25



10 -


0 5 10 15 20 25
Day Number

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

Chemical Engineering Education

S,. .



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
treated as a local
workplace can be
valuable in
students to
health, and
impressions of

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.


3 To use the undergraduate thesis
experiment on distillation to introduce
students to the concept of medical surveil-
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.

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



Ethanol Fermentation

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
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.

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

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


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)
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

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.

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).

Experimental Results Obtained in ASSAYS 1 and 2

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

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




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 ( and CO, evolved (nco,) during

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-
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.

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.

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





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.

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
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)
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.

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)
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

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


valve 9




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-


35 298 K



2 '- -u B -- -- -- --p- a
: 473 K


0O ..


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.

16- 298 K

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.

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

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.


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.

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

Continued from page 25.

(cA/Y intrinsic volume averaged concentration of species A,
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

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
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





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
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.

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-


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

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 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."

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
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


| Water
inlet flow

Level measurement




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.

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.

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.

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
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

[^] 1 laboratory




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.

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
that can be consulted for additional details on the course

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.

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.

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
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.

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.


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


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

Chemical Engineering Education

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


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.

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-


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

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

Winter 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 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.

a gas-liquid interfacial area per unit total volume [m2/m3]
Gx liquid-phase flow rate per unit cross-sectional surface area
Gy gas-phase flow rate per unit cross-sectional surface area
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
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)

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.

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 ']
specific growth rate [h' ]
Pmax maximum specific growth rate [h-']
0 refers to the time t=0
i refers to the time t=i
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.,
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.,
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
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-
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.


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




a. 40

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.

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

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)

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
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


Alternative Measures of Quality

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
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.


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

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.

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


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

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

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
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-
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


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

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


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 -

R =0.97




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)

(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



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:
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





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

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

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

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

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

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

344.2+ (312.1)= 515.8

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

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

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

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

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
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

30 A

S* 6

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

= 0

*) 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

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

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





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


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



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



tZ t=

0 1111


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


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-
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
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.

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
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.



Motivating Students to Learn in Industry

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-
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

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
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
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.

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


a firn

irm I

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.

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.

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
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.

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.

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


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

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

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

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