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 )
periodical   ( marcgt )
serial   ( sobekcm )


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



Aw Steve LeBlanc

Featuring ...


and Chemical Engineering at ...


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

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

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

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

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

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

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 mea-
surement and give dimensions for all terms. Write all equations and formulas clearly, and number important equations consecu-

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 reproduc-
tion. 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 biographical
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

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

Tim Anderson

Phillip C. Wankat
Carole Yocum
Christina Smart
James O. Wilkes, U. Michigan

William J. Koros, Georgia Institute of Technology


E. Dendy Sloan, Jr.
Colorado School of Mines

Pablo Debenedetti
Princeton University
Dianne Dorland
Rowan University
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
William J. Koros
Georgia Institute of Technology
David F. Ollis
North Carolina State University
Ronald W. Rousseau
Georgia Institute of Technology
Stanley I. Sandler
University of Delaware
Richard C. Seagrave
Iowa State University
C. Stewart Slater
Rowan University
James E. Stice
University of Texas at Austin
Donald R. Woods
McMaster University

Chemical Engineering Education

Volume 36

Number 2

Spring 2002

82 Steve LeBlanc, of the University of Toledo, G. Glenn Lipscomb

88 Tulane University, Kyriakos Papadopoulos

94 Economic Risk Analysis Using Analytical and Monte Carlo Techniques,
Brendan R. O'Donnell, Michael A. Hickner Bruce A. Barna
108 Teaching Product Design Through the Investigation of Commercial Beer,
Stephanie Farrell, James A. Newell, Mariano J. Savelski
130 Using Written-Answer Questions to Complement Numerical Problems.
Case Study: A Separation Processes Course, Simon M. Iveson
160 Spreadsheet Solutions to Two-Dimensional Heat Transfer Problems,
Ronald S. Besser
170 An Undergraduate Course in Applied Probability and Statistics,
Thomas Z. Fahidy

100 Internet Resources for Chemical Engineers,
Rahmat Sotudeh-Gharebaagh

102 Web-Based VR-Form Virtual Laboratory, Dong Yabo, Zhu Miaoliang
122 Developing Troubleshooting Skills in the Unit Operations Laboratory,
Aziz M. Abu-khalaf
138 Fluidized Bed Polymer Coating Experiment, Robert P Hesketh,
C. Stewart Slater Stephanie Farrell, Michael Carney
144 Metal Recovery from Wastewater with an Electrochemical Method,
Der-Tau Chin
150 A Holistic Unit Operations Laboratory,
Laureano Jimenez, Josep Font, Josep Bonet, Xavier Farriol
156 Mass Transfer Experiment: Determination of Liquid Diffusion
Coefficients, Francisco Ruiz Bevid, Maria del Mar Olaya
166 A Virtual Unit Operations Laboratory,
Patrick J. Fleming, Michael E. Paulaitis

114 The Effective, Efficient Professor, Richard M. Felder

116 Can We Teach Our Students to be Innovative? Stuart W Churchill

134 Use of an Integration Technique to Trace Phase Equilibria Curves,
James P. Russum, Donald P Visco, Jr

128 ASEE, ChE Division Program

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 0 2002 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 Chemical Engineering Education, Chemical Engineering Department., University
of Florida, Gainesville, FL 32611-6005. Periodicals Postage Paid at Gainesville, Florida and additional post offices.

Spring 2002

]w educator



of the


of Toledo

University of Toledo
Toledo, Ohio
lose your eyes and envision Johnny Carson. As
Carnac, he is wearing a turban and holding an enve-
lope to his forehead. Johnny's face contorts as he
struggles to divine the answers to a question written on a slip
of paper hiding inside. "What are...a diploma, the signature
of Linda Furney, a photo of Bruce Lee, and a comb?"
Johnny holds the envelope between the thumb and forefin-
ger of his left hand, rips off the end, blows gently into it, and
extracts the paper. "The question is: Name three things in
Steve LeBlanc's office and one that is not." Steve's office
obviously contains much more, but these items tell a lot about
his contributions to the chemical engineering community and
his life outside it.
A Diploma
Steve was born and raised in Toledo, Ohio. His mother
Nellaine, a Toledo native, worked as a sales representative
while his father Martin, a San Francisco native, was a shoe
repairman. The two first met at a shoe repair store where they
were employed. Both were high school graduates, but nei-
ther had gone to college. Steve has one sibling, his sister,
Sandra Meinzer, also of Toledo.

After graduating second in his class from high school,
Steve's parents encouraged him to further his education. Based
on his academic performance, an essay, and an interview,
Steve was awarded the Stephen K. Mahon Scholarship from
Toledo Edison to support study at the University of Toledo.
This prestigious award covered tuition, books, supplies, and
even parking expenses. What caught the interviewer's atten-
tion about Steve? His grades? His essay? Nope. The inter-
viewer was most interested in why he read Reader's Digest!
Steve was an avid reader in high school, but the interviewer
thought it was unusual that a teenager would enjoy Reader's
Digest (obviously a favorite read of the interviewer).
Steve enrolled in the chemical engineering program at To-
ledo. He was attracted to chemical engineering because of
his high school chemistry teacher, Raymond Squire. Raymond
was a University of Toledo chemical engineering graduate,
the first recipient of the Toledo AIChE Section Outstanding
Graduate Award in 1957, and it was his inspiration and en-
couragement that sent Steve down his career path.
Steve took two classes from Raymond: general chemistry
Copyright ChE Division of ASEE 2002
Chemical Engineering Education

... how do you measure how long it takes to get a PhD?
Steve's unit of measurement is miles. He put 65,000 miles on
his Toyota Tercel commuting between Toledo and
Ann Arbor to complete his degree.

and advanced placement chemistry. In general chemistry, Steve was introduced to
technical problem solving. He was particularly impressed by Raymond's problem-
solving skills, organization, and attention to detail. Steve also appreciated the time
Raymond took to talk with students about chemistry and life. These are the same
attributes that students commend Steve for today.
The Stephen K. Mahon Scholarship afforded Steve an opportunity to work part-
time during the academic year and full-time during the summer at Toledo Edison.
When initially approached about working, Steve turned down the offer-he already
had a job working at McDonald's. What Steve didn't realize was that McDonald's
was paying minimum wage, $1.65/hour, while Toledo Edison paid $3.75/hr.
Steve soon learned that working for Toledo Edison was more profitable, but he
doesn't regret having passed on the opportunity that first year. He met his wife-to- --
be, Molly McKelvey, while working at McDonald's. Because of his interest in Steve and Molly tie the ir
Molly, Steve continued to work part-time for McDonald's during the summer months knot, June 25, 1976.
after his day shift with Toledo Edison had ended.
Steve's job at Toledo Edison was first in Mechanical Engineering, then in Power
Engineering, the department involved with the design and construction of the Davis-
Besse Nuclear Power Plant. He helped document design revisions and maintain
construction records. The income helped instill an appetite for the latest electronic
gadgets. He spent most of his first paycheck on his first calculator-a whopping
$150! Steve had selected his first college physics lab partner based on the fact that
the guy was smart and had a calculator. Now, all he had to look for was smarts.
Steve was a bright student. His academic excellence led to receipt of the Perry's
Outstanding Sophomore Award and the EIT Award (for the highest grade on the
Engineer-in-Training Examination). His class also excelled-three of the nine gradu-
ates went on to receive Doctoral degrees. Charlie arrives, and.
After receiving his Bachelor's degree in June of 1976, Steve married Molly and
headed to the University of Michigan to pursue a Doctorate. He completed the
requirements for a Master's, passed the Doctoral-qualifying exam, and had started
on his research when he decided to forgo completing the degree requirements-
Molly was pregnant with their first child, Charlie, and had quit her job as a fifth-
grade teacher. Also, his father was seriously ill with cancer. Steve decided to return
to Toledo Edison as a full-time engineer, where he continued to work on the design
and construction of Davis-Besse in the Power Engineering Department.
Steve worked at Toledo Edison for three years as one of the three chemical engi-
neers they employed. He also remained active in the local AIChE section. Through
AIChE, he kept close contact with the faculty at Toledo and was given an opportu-
nity to change his career path when Joseph Boston left the Toledo faculty. Shortly
thereafter, Les Lahti, the Department Chair, asked Steve to take Joe's position.
The decision to become an academic was not easy: Steve and Molly now had ... before you can say t
two children (Susie was born in 1979), and he would have to complete his Doc- dynamics, here's Joe, S
toral degree requirements while simultaneously teaching a full load. But Steve had Emili, and Charlie (with
tired of the routine work he was doing and was sorry he hadn't finished his degree, at Snowbird), all grow
so he took the faculty position. As his current colleagues can attest to, Steve's
Spring 2002

willingness to take on such a formidable workload is charac-
teristic of his personality and commitment.
Steve worked with Scott Fogler studying solid dissolution
by acidization. Scott had developed an interest in the
acidization of sandstone oil reservoirs. Acidization opens
pores in the formation and thereby enhances oil recovery. A
mixture of hydrofluoric and hydrochloric acids does the job,
but either acid by itself does not-and this synergistic effect,
the catalysis of HF dissolution by HC1, intrigued Scott.
Steve's project was to look at the dissolution of other ma-
terials, in particular manganese oxides. He was to determine
if dissolution is catalyzed, what controls the rate of dissolu-
tion, and if any other physical processes were involved.
He found that upon dissolution, manganese oxides form
two types of structures: 1) a Swiss-cheese structure in which
a number of holes would appear and grow in diameter and
depth until they coalesced, and 2) a peeling orange structure
in which the holes would preferentially grow laterally with
little change in depth. The latter is not desirable in electronic
materials where one desires to etch sharp rectangular chan-
nels on silicon wafers. A better understanding of the physics
of manganese oxide dissolution might lead to improved manu-
facturing processes in the microelectronics industry.
Steve also studied the dissolution of a polydisperse collec-
tion of particles. He used population-balance methods to
model the evolution of the particle size distribution with time.
Steve demonstrated that use of the mean particle size severely
underestimates dissolution time relative to predictions that
account for the complete distribution. The deviations were
greater if dissolution was mass-transfer limited rather than
reaction-rate limited. This was due in part to a broadening
of the size distribution that occurred under mass transfer-
limited conditions.
When asked what were the most significant outcomes of
his Doctoral research, Steve gives personal and technical
answers. Personally, the non-technical education Scott pro-
vided was most significant: "Scott was a good teacher-he
taught you how to think and ask questions." Technically, Steve
says the fundamental framework developed to analyze pow-
der dissolution, dissolved structure morphology, and the ef-
fect of defects was most significant.
From 1980 until 1985, Steve worked three jobs: 1) gradu-
ate student, 2) instructor, and 3) husband and father. Most of
us had our hands full with graduate school, but Steve simul-
taneously taught two classes each quarter. During this time,
Molly and Steve also welcomed two additional family mem-
bers: Emily in 1982 and Joe in 1985. Clearly, Steve had de-
veloped excellent time-management skills.
By the way, how do you measure how long it takes to get a
PhD? Steve's unit of measurement is miles. He put 65,000
miles on his Toyota Tercel commuting between Toledo and
Ann Arbor to complete his degree.

Upon receipt of his diploma, Steve became an Assistant
Professor and started his progression up the academic ladder.
He didn't dawdle along the way. He became an Associate
Professor in five years, and four years later he reached the
pinnacle of academic success-Professor. In the fall of 1993,
he started his progression through the administrative ranks
when he became Interim Department Chair. "Interim" was
dropped from the title in 1995 and he has served as Chair
continuously since then.
The Department has flourished under Steve's leadership,
despite numerous challenges. The Engineering College moved
to a new building in the fall of 1995 and Steve was respon-
sible for overseeing all aspects of the Department's move,
from faculty office assignments to disassembly and reassem-
bly of the unit operations laboratory, including a two-story
glass batch distillation column.
After the move the Department lost three faculty members.
Steve helped fill the void left behind by serving as Depart-
ment Chair, Undergraduate Studies Director, and Graduate
Studies Director for almost a year. As he had done earlier in
his career, Steve also taught two classes each term.
Steve staged a four-year-long recruitment campaign to re-
place the lost faculty. His efforts were rewarded by the hiring
of seven faculty members (Martin Abraham, Maria Coleman,
John Dismukes, Isabel Escobar, Dong-Shik Kim, Arun
Nadarajah, and Connie Schall) who have become the research
and teaching core of the Department. In addition to the move
and faculty hiring, Steve oversaw conversion of the academic
calendar from quarters to semesters, incorporation of man-
datory co-op in the curriculum, and ABET 2000 accredita-
tion. Any one of these items represents a tremendous amount
of effort. The fact that Steve shepherded all of them during
the past seven years is truly Herculean.
In May of 2001, the LeBlanc family welcomed another
degree chemical engineer to its ranks. Charlie graduated with
a BS in Chemical Engineering and is currently a process en-
gineer with Central Soya in Bellevue, Ohio. In addition to
teaching Charlie, as Department Chair Steve had the plea-
sure of presenting Charlie his diploma at graduation.

The Signature of Linda Furney
Steve has always been an outstanding teacher. Like most
faculty members, he never received training in education, but
he possessed an innate ability to teach. His recipe for teach-
ing success involves four ingredients. They are
1. Try to remember what you had trouble learning.
Reactor design had been especially difficultfor Steve as an
undergraduate, and he looked back to those days for
inspiration when preparing classroom materials. Moreover
he had kept all of his notes from his undergraduate studies,
and reviewing them helped him identify the most difficult
material when preparing to teach a new class.
2. Learn the material with the students.
Chemical Engineering Education

Steve was out of academia for three years while he worked for
Toledo Edison. Upon returning to academia, he had to re-
learn much of the material at the same time the students
learned it. This helped him identify difficult concepts and
develop novel approaches to explain them.
3. Look at the material from the student's perspective.
A concept that is obvious to a Professor may not be so for a
student. Imagine yourself with the background of a typical
student and envision where you might have difficulty.
4. Continuously educate yourself.

Steve was able to
use what he learned
in graduate school
in Ann Arbor
directly in the
classroom in
Toledo. The
processes of
learning and
teaching can be
synergistic (even if
you're not a
graduate student in
chemical engineer-

Steve's teaching
ability was recognized
in his second year

place team entry in the 1999 AIChE Student Design Compe-
tition. Their team prepared the winning solution to the
"Dicyclopentadiene Recovery from Byproduct Naphtha
Steam-Cracking" problem given to students that year. Steve
places receipt of this award near the top of his list of most
rewarding moments as a faculty member.
Steve has never lost his desire to excel in the classroom.
He still receives the highest teaching evaluations. Addition-
ally, he has pursued his educational interests through a vari-
ety of activities outside the classroom.

h- _- -

Sensei Tom Nehring (of Kempo Martial Arts), Joe, and Steve.

(1982) with the high-
est teaching award given by the University: The University
of Toledo Outstanding Teaching Award. This award is given
to three members of the University faculty each year in rec-
ognition of their teaching abilities. Steve's receipt of the award
at such an early stage of his career is remarkable. It reflects
the respect, admiration, and inspiration he generates in the
classroom. Student comments on class evaluations clearly
support such a conclusion: "Dr. LeBlanc is one of the great-
est teachers I have ever had. He is courteous, interesting, and
able to communicate the subject to students very well. I have
great respect for him and feel he is a great asset to this de-
partment." "He has sparked my interest in chemical engi-
neering since day one."

Often, good teachers are accused of being easy graders.
This is not the case with Steve. Design and Unit Operation
Lab reports often drip red ink. Students know they will fail
his class if they don't perform. While this often makes for
more work, especially when seniors fail design and want
to make it up, Steve doesn't relax the high standards he
sets for his students.

The greatest rewards for your classroom efforts come in
the accomplishments of your students. For Steve, the most
recent came when Jeff Burhenne, Dan Gastaldo, and Robert
Kasprzak won the William Cunningham Award for the first-
Spring 2002

Steve has been
an active member
ofASEE for many
years. In addition
to numerous tech-
nical presenta-
tions, he has
served as Program
Chair, National
Chair, Executive
Committee mem-
ber, and Nominat-
ing Committee
member for the
Chemical Engi-
neering Division.
Steve's most

memorable ASEE
session was an evening session devoted to computer use in
the undergraduate laboratory. According to Steve, "It was a
roundtable discussion where everyone talked about what they
were doing with computers in the undergraduate laboratory.
It was a great way to pick up tips in the earlier days of inter-
facing." In his 1991 presentation "Computerized Data Ac-
quisition in the Unit Operations Laboratory: An Experiment
in Thermodynamics," Steve discussed how to interface com-
puters and equipment, especially the construction of inter-
face boxes and connecting cables. Although routine now, these
issues required much greater technology savvy then. Steve
continued to explore "bleeding" edge applications of tech-
nology, including the use of NetMeeting in the classroom
for remote interaction and providing access to real experi-
ments through the Internet. A second presentation at the 1991
meeting, "The Use of MathCAD and Theorist in the ChE
Classroom," garnered the J.J. Martin Award from the
Chemical Engineering Division, given to the best presen-
tation in the Division.
Steve has developed an interest in interacting with high
school students to recruit them into chemical engineering.
He obtained NSF funding from the Young Scholars Program
to run a Chemical Engineering Summer Workshop for five
years (1991-1995). The program was a continuation of a lo-
cal version originally started by former colleague Jim

Lacksonen. High school students from around the country
attended the workshops. The objective of the program was to
attract them to science and engineering, and preferably to the
University of Toledo chemical engineering program.
The workshops were a success in many ways. They attracted
nearly 200 academic superstars: the average GPA hovered
near 4.0/4.0. One student even received a perfect 32 on the
ACT. The groups were also well-rounded: about 60% seniors,
40% juniors, 40% female, and 10% minority.
The workshops also offered many new challenges. How
does one deal with hormonally super-charged teenagers? How
does one prevent students from publishing inappropriate
material on the Web sites that you taught them to create?
The rewards outweighed the problems; about 75% of the
seniors indicated that they would major in engineering,
and Toledo's chemical engineering program received
twenty of them.
While Steve was in graduate school, Scott developed an
interest in problem-solving strategies. Steve became keenly
aware of this when Scott brought problems to group meet-
ings for discussion and solution. Codifying problem-solving
techniques and teaching them in the classroom intrigued
Steve, and in 1991 he decided to spend his first sabbatical with
Scott exploring the idea. Scott was thrilled to have "one of the
most creative persons that I have ever worked with" join the
project and knew they could be on to something big.
Initially, they asked what is industry doing-what types of
problems were encountered and what types of skills were
required to solve them. Steve and Scott took two or three
students from a group of about ten recruited from the reactor
design class to visit a number of local companies. During the
visits, they looked for problems that could be used to illus-
trate problem solving throughout the curriculum.
As material was gathered, Scott and Steve realized that they
had more than several good problems to use in class-they
had enough material for a problem-solving class and an in-
struction book. This was the genesis of Strategies for Cre-
ative Problem Solving. Writing commenced in early 1992,
Prentice Hall agreed to publish the book, and it was finally
published in 1995. In its fourteenth printing, it has sold over
38,000 copies. The rapid adoption of the book by educators
across the country led to the authors receiving the ASEE
Meriam-Wiley Distinguished Author Award in 1996.
Steve's collaboration with Scott as a textbook author con-
tinued with the publication of Open Ended Problems in
Chemical Reaction Engineering by Prentice Hall in 1995.
The problems are intended to complement "normal" example
and homework problems by providing an opportunity to syn-
thesize creative solutions. The solutions require application
of the fundamental principles taught in class, but also pro-
mote learning at the highest levels of Bloom's Taxonomy:
synthesis and evaluation.

Steve is currently working on yet another book. In 2000 he
received a call asking for suggestions on how to update Donald
Coughanowr's Process Systems Analysis and Control.
McGraw-Hill was interested in publishing a third edition of
this popular undergraduate controls text, and with his typical
thoroughness, Steve prepared an exhaustive list of modifica-
tions. This led to a request from the publisher for a prospec-
tus on how he would change the book and ultimately an offer
to write the third edition collaboratively with Donald. Steve
quickly accepted the offer to work on what is one of his fa-
vorite texts. The third edition is scheduled for publication in
the summer of 2002, and Steve promises the book will not
lose its current feel.
Steve has a knack for preparing 1-2 page technical sum-
maries that students find extremely useful when solving prob-
lems (check out Scott Fogler's Chemical Reaction Engineer-
ing web site for a medical diagnosis of this condition at /> and follow "The Knack"
link). These summaries of equations and concepts let stu-
dents quickly find the information they need and then direct
them to where it comes from in the text. Two examples are
reproduced here, and Steve would gladly provide more if
asked. Don't be surprised if you find such summaries at the
end of each chapter in the third edition of Process Systems Analy-
sis and Control and develop a desire to use them yourself.
Steve's prodigious efforts in education were recently rec-
ognized with the ASEE North Central Section Outstanding
Teacher Award for 2000. The Senate of the 124th General
Assembly of the State of Ohio recognized this achievement
with a proclamation signed by the President of the Ohio Sen-
ate, Richard Finan, and the 11th Senatorial District Repre-
sentative, Linda Furney. We may not have to provide jus-
tification for Steve's salary to the State this year...thank
you Linda Furney!

Photo of Bruce Lee
In May of 1998, Steve developed a serious illness. He had
a fever that would come and go and was high enough to send
his heart racing. During one episode, when the rate was so
high he couldn't count it, he went to the emergency room and
spent his first night in a hospital bed. Doctors were unable to
diagnose his mysterious illness and decided to place him on
a therapeutic course of antibiotics. It worked. In two weeks,
the symptoms disappeared and he ended treatment.
Unfortunately, the infection had not been eliminated. The
weaker elements may have been wiped out, but the stronger
ones survived and grew. The original symptoms returned-
fever and heart arrhythmia-as well as new ones, pneumo-
nia. Steve was very ill once again. The doctors, still unable to
diagnose his condition, decided to prescribe a stronger anti-
biotic and a longer course. This time it worked. After going
off the medication, the symptoms did not return and he re-
mains healthy to this day.
Chemical Engineering Education

/f C -rAE- ^ C e- T&.JSS-Et. bu ?Sk C4YU GcL

ertrunoii. c*aMMol~, oc tt r*.itif Mobto-t MW a-rfc^ o^l t> torp-rf
Aoa*t S..At VOL ktcA S.ZGCIAgDt 0. ACDQr -
CCOCA___ .AAI -C-ifc ____ 010____________C J O.0
icu mei o\ -^i m n a~n oraomf C
J'TiA&-WO.%j V ,A 'JS'_______
VWoi.ACP 7'- 7
rLr aT. r \

s^'-a Try ^. --, 1 ^eV1 ferfultnt PYLAI L~Cek-l~ f^Lfe OF -TVI oJJu~t& )

FiiiA-- T r -

_o I_ ^
-~~~~ ~ ^-^^r^ F

-l -\__^^ UC
Ar1~91 A'ACj~L 4

t- 3 *tC> e )
'.0CC. C ^ S t- cc
J ("Wt ~ >=^Ws'MT.cWv)

/4T~'\ r

A TIo examples of Steve's knackfor
V presenting difficult concepts.

Re(s-_ oe._e ssre is

K (0) = X- )

PtSI Ao (I) .E iCYiAC
d8Y, Y,= 1,T

I H 7iO S c Cc 6 I1
Sp ringC .S.e.. V,- i (C(5' CCS) 2( 5 'C B -
Spin 20 F02 tC0 (c) ("-( i L -

Spring 2002


Steve thought a vigorous running schedule during the Spring
semester might have weakened his immune system and given
the illness purchase. Consequently, he gave up running as a
regular form of exercise. This might have helped his immune
system, but he missed the exercise and started looking for a
more acceptable alternative.

He didn't have to look further than his youngest son Joe
for an answer: martial arts. After taking Joe to Kempo karate
classes for three years and seeing the excellent workout it
provided, Steve decided to take it up himself. During the past
four years, Steve has progressed up the Kempo ranks. He
currently holds a 2nd Kyu Brown Belt and is only two
levels from the coveted Black Belt, the belt that Joe earned
in June of 2001.
Karate provides exercise, but it also provides ample op-
portunity for injury. Steve has sported an impressive array of
black, blue, and red marks since taking up the sport. With
little prodding, he is more than willing to demonstrate on you
how he got them.
Karate has increased Steve's interest in Asian culture, art,
and philosophy. He has acquired for his office a Zen rock
garden, a Bruce Lee calendar (illustrating many of the master's
moves), and a water wall. Steve shares his love of martial
arts with one of our latest hires, Dong-Shik Kim. Dong-Shik
holds a 3rd Dan Black Belt.

Steve even brought Molly into his pursuit of alternative
exercise activities. During the summer of 2001, Steve and
Molly celebrated their 25th wedding anniversary by taking a
trip to Hawaii and cruising through the islands. They enjoyed
a helicopter ride over an active volcano, ocean kayaking, and
a bike trip from the summit of Mt. Haleakala to the coast. At
the end of this 40-mile bike trip (mostly downhill, as you would
expect), their legs were not overly tired, but their hands would
need a few days to recover from gripping the brake levers.


Steve's current hairstyle doesn't require a comb. So why
would he have one in his office?

As a physically similar short bald man, I (GL) can appreci-
ate Steve's choice of hairstyle. Unfortunately, our close re-
semblance has led to some confusion among students. On
more than one occasion, a student has entered my office-
which has a sign outside that reads G. Lipscomb-and asked
"Dr. LeBlanc" for help with a problem from one of his classes.
I don't mind the chemical engineering problems-it's the
karate problems that I hate.

While Carnac may have provided a glimpse into Steve's
office today, it is an ever-evolving landscape of paper files,
diverse objects, and memorabilia. The next time you see him,
ask him about the latest addition-even better, come visit
and take a tour yourself. Steve would gladly show you around
the department that he has poured his energy and soul into
for many years. 0

i3~t ------
V/,-^ -(;^CA .JC& J F oCA CA CC.J AeCli.AT
AOL icCC C. oCAi., w --

CTriS~oI To C'0:T
ts Y s) Y(r) KTXC)

(~ 77cL0536~i~ ToM1
Syeo) -----


*I A -A. ccCscC 053 EJ
^yf/ / ^

/ !4.
SC'.I / \ SAS., oiltMK. r--

m FL*O ~


Tulane University

Tulane University
New Orleans LA 70118
T ulane University's "Uptown
Campus," the home of chemi-
cal engineering, is located in
a scenic section of New Orleans, one
of America's most charming cities. It
is a selective private research univer-
sity and has educated chemical engi-
neers for well over a century. In 1979,
at a joint congress of the American
Chemical Society and the Chemical
Society of Japan, a symposium was
held on the history of chemical engi-
neering, the proceedings of which
were published in Advances in Chem-
istry Series. J.W. Westwater provided
an article that traced the beginning
of chemical engineering education
in the U.S.,'m and in it, after recog-
nizing MIT as the first chemical en-
gineering program in the nation and
the University of Pennsylvania as
the second, he said
Tulane University was the first school
in the South and apparently the third
in the United States to have a four-
year curriculum labeled "Chemical
Engineering." This was in 1894 (ref).
The degree label was a BE in
Chemical Engineering and the first
recipient was B.P. Caldwell in 1895
(ref). There were no courses labeled
Chemical Engineering and no
"Professor of Chemical Engineering."
The most pertinent courses, three in
industrial chemistry, were taught by
John Ordway. In 1893 this man bore
Copyright ChE Division of ASEE 2002

Streetcars typify the flavor of New Orleans, home of Tulane University

Chemical Engineering Education

three titles: (1) Professor of Applied
Chemistry and Director of Manual
Training School; (2) Professor of
Industrial Chemistry; and (3) Professor of a gr
Industrial Chemistry and Acting Professor and
of Civil Engineering. The courses in
industrial chemistry and those in chemis- C
try were taught in the same building. Thus
at Tulane, chemical engineering seems to
have its roots in chemistry.
Others, however, disagree with some
of Westwater's statements. According to
Tulane Professor Emeritus Raymond V.
Bailey, " is a matter of record in the
history of AIChE that MIT had the oldest program by one
year and that Tulane was second....It is also part of that record
that Tulane has the oldest published curriculum."m'
In the early days, Tulane's chemical engineering was a pro-
gram in chemistry and did not have departmental or school
status. On October 28, 1981, then-Professor Emeritus Francis
M. Taylor, speaking to the Tulane Student Chapter of AIChE
on the history of the department, stated that Charles Samuel
Williamson, charter member of AIChE, was hired in 1913 to
be Head of the "School of Chemical Engineering." That was

Past Chemical Engineering Faculty at Tulane

1913-1917 A.L. Metz
1913-1947 Charles Samuel Williamson, Jr. (Head of Department 191
1938-1970 Francis M. Taylor (Head of Department 1947-1951)
1939-1943 Jack A. Geister
1944-1951 Charles G. Marshall
1950-1951 Chuk C. Ma
1951-1994 Raymond V. Bailey (Head of Department 1951-1974 and
(Currently Emeritus Professor of Chemical Engine
and Emeritus Associate Dean of Engineering)
1951-1961 Mack M. Gilkeson
1957-1960 James E. Kinard
1960-1981 Robert E. Weaver (Head of Department 1977-1981)
1961-1995 Samuel L. Sullivan, Jr. (Currently Emeritus)
1961-1977 Daniel B. Killeen
1962-1967 Charles H. Barron, Jr.
1963-1976 Dale U. von Rosenberg
1965-1975 Robert Chambers
1968-1974 Gordon H. Harris
1973-1977 Duane F Bruley (Head of Department 1973-1977)
1976-1984 Danny W. McCarthy
1976-1980 James M. Henry
1977-1980 Lynn J. Groome
1976-1978 Neil Larry Book
1979-1986 Richard W. Freedman
1975-1978 Thomas R. Hanley
1980-1982 Bert Wilkins
1981-1988 Aysel Atimtay
1981-1985 Henry Lutrell
1982-1986 Young Gul Kim
1986-1993 *Anil Menawat
1992-1997 John Y. Walz

Spring 2002

By providing the infrastructure for
eater number of collaborative efforts
proposals between the departments of
chemistry, biochemistry, and chemical
engineering, the Tulane Institute for
Chemical Sciences will increase
our research abilities
and opportunities.

also the year when chemical engineering was classified as a
separate school from chemistry. In the mid-1950s, the de-
partment officially took the name of "Department of Chemi-
cal Engineering." Until the 1950s, Tulane's chemical engi-
neering was primarily an undergraduate program, giving only
a few Master's degrees at the graduate level. Then, in 1960,
Ray Bailey started the doctoral program.'21
Past chemical engineering faculty over the course of the
20th century are listed in Table 1.i31

Tulane's Department of Chemical Engineering
has always been a small, select department. Our
current faculty is the largest in our history. Every
faculty member is active in research and is funded
3-1947) by external grants. Nine have federal funding, and
the National Science Foundation supports seven
of them. The recent average yearly output of the
department has been about four archival-journal
1981-1993) publications per faculty member. In pursuing our
ering educational and research mission, we are fortu-
nate to have the invaluable help of Ms. Belinda
Lacoste, Executive Secretary, and Mr. Paul
Lane, Lab Coordinator.
The current professor with the longest history in
the department is Vic Law, who was also the very
first PhD graduate in the department in 1963. Vic
became Associate Professor in 1966 and Profes-
sor in 1970. In 1975, he left chemical engineering
in order to start Tulane's Department of Computer
Science but returned in 1988. He teaches courses
related to design, computer methods, statistics, and
applied mathematics. Since 1990, Vic has had an
industrially funded research program targeted at
the demetallation of Cat Cracker catalysts. He has
also had DoE-funded projects on the use of Beach
Cones to reverse coastal erosion and a project on
modeling the emissions of methane from rice pad-
dies in order to predict global methane emissions.
An AIChE fellow and a fellow of the IChemE, Vic

Tulane's ChEfaculty. Seated, left to right, Vijay John, Richard Gonzalez, Victor Law, Daniel Lacks, and
Kim O'Connor; Standing, left to right, Yunfeng Lu, Kyriakos Papadopoulos,
Daniel De Kee, Brian Mitchell, and Peter Pintauro

took a two-year leave (1998-2000) to found a program in
chemical engineering at the University of Limerick in Ire-
land. He is a licensed chemical engineer in the State of Loui-
siana and a Charted Engineer in the United Kingdom and
Kyriakos Papadopoulos joined Tulane's faculty in 1981,
three months before defending his doctoral dissertation at
Columbia University, where he earned his MS with Gerald
Holder and his DEngSc with Huk Cheh. A member of the
Editorial Board of Colloids and Surfaces, his research is in
the stability of dispersions and their transport through porous
media. He has recently focused on development of a "capil-
lary microscopy" technique that has uniquely led to the visu-
alization of several new phenomena. He has taught a wide
variety of undergraduate courses in thermodynamics, engi-
neering physical chemistry, fluid mechanics, separation pro-
cesses, process control, and applied mathematics in chemi-
cal engineering, as well as several graduate courses in trans-
port phenomena, thermodynamics, surface and colloid sci-
ence, analytical mathematics, polymer physics, and advanced
separation methods. He has received departmental, engineer-
ing-school, and campus-wide teaching awards and served as
Chair of the department from 1998 until the end of 2001.
Vijay John came to Tulane in 1982, the year he obtained
his doctorate from Columbia University. In his research, Vijay
is interested in the use of self-assembly to template the syn-
thesis of nanostructured materials. Applications of his work

include field-responsive nanostructures, structural polymer-
ceramic nanocomposites, and functionalized nanoparticles.
He also does research in clathrate hydrate technology with
applications to desalination and materials synthesis. For two
years (1996-1998), he served as Program Director in the
Chemical and Transport Systems Division of the National
Science Foundation. His research is funded by NSF, the U.S.
Army, and private industry. He teaches courses in applied
thermodynamics, reactor design, and materials science and
has received departmental awards for teaching. Vijay is very
active in University efforts to build research collaborations
across the chemical sciences, and he is the current Chair of
the department.

In 1986, Peter Pintauro joined the department after re-
ceiving his PhD from UCLA and teaching at UCLA and
Manhattan College. His research interests are in electrochemi-
cal engineering and membrane separations, with funding by
the National Science Foundation, the Army Research Office,
and private industry. Currently, his research efforts are di-
rected in two areas: 1) development of new membrane mate-
rials for next-generation fuel cells, and 2) modeling multi-
component ion and solvent transport in ion-exchange mem-
branes. Since 1997, Peter has been the North American Edi-
tor of the Journal of Applied Electrochemistry. In 2001 he
was the first recipient of Tulane's School of Engineering
Outstanding Researcher award. He has taught undergradu-
ate courses in transport phenomena, numerical methods,
Chemical Engineering Education

and design, as well as graduate courses in electrochemi-
cal engineering, corrosion, membrane separations, and
advanced transport.

In 1990, we welcomed Kim O'Connor as
a junior faculty member. After a BS at Rice
and a PhD at Caltech (both in chemical engi-
neering) where she worked under the direc-
tion of James Bailey, Kim did two post-doc-
toral fellowships, one at Caltech in molecular
biology and one at Northwestern in cellular
biology. Her research is in cell and tissue en-
gineering. She has led and continues to lead
multidisciplinary research groups in the health
sciences. Kim's specialty is kinetic phenom-
ena in animal-cell culture, such as apoptotic
cell death, multicellular spheroid self-assem-
bly, and cell differentiation. Her teaching has
included courses in the Department of Chemi-
cal Engineering, the School of Engineering,
and the Molecular and Cellular Biology Gradu-
ate Program. Kim has courtesy appointments
at Tulane in the Department of Surgery, the
Molecular and Cellular Biology Program, and
the Cancer Center. She is the recipient of three
research grants from NASA, as well as one
engineering-school and two university-wide
teaching awards.

Also in 1990, we hired Richard Gonzales as the Herman
and George R. Brown Professor of Chemical Engineering.
Richard earned his PhD from The Johns Hopkins University
in 1965. A past president of The North American Catalysis
Society, he has been making research contributions to ca-
talysis, reactor design, and surface science for the last forty
years. His recent areas of interest include the design of ce-
ramic membrane reactors, catalysis by solid acids, and clay
chemistry. The National Science Foundation, the Depart-
ment of Energy, and private industry currently fund his
research. He has mentored twenty-five PhD students, ten
postdoctoral students, and several masters' students. He
has lectured extensively around the world and has taught
graduate and undergraduate courses in reactor design,
catalysis, and thermodynamics.
In 1994, Dan Lacks joined our ranks as a junior faculty
member. After earning his BS in chemical engineering from
Cornell and his PhD in physical chemistry from Harvard, Dan
did a post-doctoral fellowship at MIT. His research involves
the application of molecular simulations to problems of chemi-
cal engineering interest. Dan has taught process design, pro-
cess control, separations, thermodynamics, statistical mechan-
ics, and transport phenomena. He has received a CAREER
grant fron the National Science Foundations, as well as de-
partmental and campus-wide awards for his teaching.
In 1990, Brian Mitchell also came to our department as
Spring 2002

a selec
and h

for w


a junior faculty member. Brian earned his BS from the Uni-
versity of Illinois at Urbana-Champaign and his MS and PhD
from the University of Wisconsin. Prior to joining Tulane, he
was an instructor at Wisconsin and had done
an NSF/NATO post-doctoral fellowship at the
University of Karlsruhe, Germany. At Tulane
el is he has taught a variety of undergraduate
courses, including material and energy bal-
tive ances, heat transfer, probability and statistics,
fte unit operations laboratory, and materials sci-
ence. At the graduate level, he teaches a course
rch in materials design. Brian's research encom-
Si passes a variety of materials processing top-
ics, including the production of metal/ceramic
Sas nanocomposites and ceramic fibers, and the
characterization of ceramics using spectros-
ted copy and thermal analysis. He is active in the
cal department as advisor to the student chapters
of AIChE and Omega Chi Epsilon. He also
3ers serves the School of Engineering as a Uni-
eli versity Faculty Senator. He has been editor
of the AIChE Materials Engineering and Sci-
a ence Division newsletter since 1997, and he
is currently Vice-Chair of the AIChE New
*1y. Orleans Local Section.
Following a distinguished academic career
in Canada that led to several research awards,
Daniel De Kee (PhD University of Montreal, 1977) joined
the department in 1997. A past president of both the Cana-
dian Rheology Group and the International Committee on
Rheology, as well as a Fellow of the Chemical Institute of
Canada, Dan has coauthored four textbooks in the areas of
applied mathematics and rheology. He has also edited nu-
merous volumes in these fields. Dan chaired the XIIth Inter-
national Congress on Rheology and is a member of the Inter-
national Advisory Board of the Canadian Journal of Chemi-
cal Engineering. At Tulane, he became the driving force be-
hind the recent creation of a multidisciplinary research team
in the area of polymer engineering and science. A prolific
author, Dan has been invited to present numerous lectures
worldwide on topics in non-Newtonian fluid mechanics, dif-
fusion in polymers, and polymer rheology.
Prior to joining us in January 2001, our youngest and most
recent hire, Yunfeng Lu, worked at Applied Materials for
almost two years and at Sandia National Laboratories for one
year. His PhD thesis research at the University of New Mexico
under C. Jeffrey Brinker earned him the Victor LaMer Award
of the American Chemical Society in 2000. His current re-
search focuses on nanostructured materials, microelectronic
materials, and sol-gel materials.

Our undergraduate student body is currently composed of

about 40% women and 30% minority students. ...
We take pride in the small classes that we of-
fer within the major. The resulting low stu-
dent-to-faculty ratio allows for close student-
professor relationships, giving students supe-
rior advising and mentoring. Our proximity
to the petrochemical and chemical industry
offers our undergraduates opportunities for
summer internships as well as long-term em-
Curriculum Tulane's chemical engineer-
ing program has been accredited by ECPD or
ABET since 1954. In October of 2001, we had
our most recent accreditation visit, which led to a si \- ear accred -
tation. In recent years, our graduating classes had an a\ erage num-
ber of about twenty students, although in the pjs, there \ere in-
stances when this number was in the mid-forties ()ur erjduate,
have always been successfully employed by induir r\. and con trar
to the current national trend, the majority of them are still en mploy ed
by the petrochemical industry. About 15% go to graduate school.
and another 5% to medical school. A number ha\e also pursued
successful careers in law, finance, and other non-traditional a\ enues.
We offer a traditional chemical engineering curriculum. \\ h em-
phasis on design and in-plant experience. In vie\" of the changing
perspectives of the industries we serve, we are not stud ing the
possibility of implementing several non-traditional options. such
as biotechnology, environmental engineering, and business.
Practice School A unique feature of our undergraduate pro-
gram is the Tulane Practice School, which was started bh Ra', Baile\
in 1951 and offered its first session in the spring of 1952.1- The
idea was proposed by local industry (Pan-Am Southern, subse-
quently Amoco), and the faculty member in charge was Mack
Gilkeson. Since the time of its inception, the method of running the
Practice School has evolved according to the nature of the partici-
pating industries and the changing chemical engineering curricu-
lum in the country.
Presently, the Practice School is taught in the second semester of
the senior year, and groups of three or four students work on a
project at a local industrial facility, hospital, or governmental agency.
Projects must be of real and current concern to the organizations
and can range from study of an operating process to development
of a new process. Projects are open-ended and students are expected
to apply the principles of good design practice with realistic con-
straints, such as economics, safety, reliability, aesthetics, ethics, and
social impact. Students are normally assigned to a project that ful-
fills certain career goals and an effort is made to assign every stu-
dent to their top choice of available projects. One Tulane professor
is assigned to supervise each group, although the students interface
with practicing engineers and other industrial personnel. Students
are expected to develop the ability to learn from and earn the coop-
eration of such personnel. Practice School carries six credit hours,
and the students are expected to devote approximately twenty hours
a week to it. They use and develop leadership skills and the ability

to communicate effectively through reports to practic-
ing engineers as well as to the faculty and fellow stu-
dents. They have to deal with a real-world problem with-
out the guarantee of a successful solution, thus needing
to demonstrate and develop creativity in choosing those
tools that are important for the project and that have
been acquired in course work.
One professor has overall responsibility for the Prac-
tice School, selects appropriate projects from the local
companies, and assigns the students to a project in No-
vember. Typically, before the end of the Fall semester, the
students have had their first visit to their assigned plant,
agency, or hospital. Initiative in setting up appointments
and demonstrating leadership is encouraged from the be-
ginning. The students are required to devise a work sched-
ule based on constraints imposed by classes that they may
be taking in the Spring semester. Different companies may
require different schedules for working on location.

Chemical Engineering Education

Upon completion of the final report and presentation, the
hosting companies evaluate the group using a standard ques-
tionnaire. In addition to giving a final presentation at the work
site, a day is set at the end of the semester when all groups
give a presentation at Tulane-juniors are encouraged to at-
tend these presentations. After listening to all presentations,
the professors evaluate each group, and the seniors are asked
to do the same for their peers. Thus, two-to-three weeks be-
fore graduation, we are provided with multiple-source feed-
back on various aspects of the overall education Tulane chemi-
cal engineers have received.
Each Practice School group's faculty advisor makes deci-
sions about final grades, and different advisors put more/less
weight on different criteria. In addition to the technical merit
of a project, such criteria may include initiative, amount of
work, hours on the project, common sense displayed, colle-
giality, leadership, reliability, promptness, and meeting dead-
lines. The professor also considers very strongly the students'
impression of the participating company, agency, or hospital,
as well as the overall performance of the group during the
final presentations. Not every member in a given group nec-
essarily earns the same letter grade, and in fact, in some of
the groups, grades may be widely different.

Undergraduate Research Because of the low student-to-
faculty ratio and the fact that all members of the faculty are
active in research, we are able to provide every under-
graduate student with the opportunity to be involved in
research. About half of the students decide to spend vary-
ing portions of their undergraduate years doing research
in the form of part-time work, independent-study courses,
or senior honors theses. Those undergraduates who join
research labs as freshmen or sophomores and continue
through their upperclassmen years are able to accomplish
research that leads to conference presentations and even
journal publications and patents.


The graduate program's emphasis is clearly on the PhD. In
admitting graduate students, we seek those who have the abil-
ity and ambition to pursue a doctorate. With few exceptions,
in recent years all our graduate students have been full-time
and fully supported as Teaching Assistants, Research Assis-
tants, or Fellows. Our present graduate enrollment is twenty-
five, with an almost even distribution of U.S. and interna-
tional students. Women make up one-third of all the students.
A Louisiana Board of Regents fellowship restricted to supe-
rior American applicants carries a competitive stipend and
has helped recruitment of highly qualified U.S. students. In
the last decade, our doctoral graduation rate has averaged
over five PhDs per year. Our highest productivity was in the
1997-98 academic year when we graduated nine PhDs with
nine faculty members on board.

The average time to complete the PhD is about four-and-a-
half years. The students must take sixteen courses beyond
the BS, up to five of which may be independent studies, usu-
ally offered by the student's research advisor. A Master's thesis
counts for two courses. Because of the increased emphasis
on demonstration of research aptitude by PhD students, in
the mid-nineties we adopted a doctoral-qualifying-exam for-
mat that moved from testing the candidates on "textbook fun-
damentals" to discerning their ability to do research. All
graduate students are asked to take their qualifying exam
in the summer of their first year, after they have had two
semesters of coursework. For the exam, we give a research
problem to the students. They then have one month to write
a proposal and defend it in front of a committee of three
or four professors.


Over the years, we have established close ties with the de-
partments of chemistry and biochemistry. These ties recently
led to the creation of the Tulane Institute for Chemical Sci-
ences (TICS). By providing the infrastructure for a greater
number of collaborative efforts and proposals between the
departments of chemistry, biochemistry, and chemical engi-
neering, TICS will increase our research abilities and oppor-
tunities. Under the aegis of TICS and other multidiciplinary
programs currently in development, we will be applying for
research-center grants that will take us to a quantum-step
higher level of funding.

Our Department continues to forge ahead, in step with the
University's stated mission to enhance its research reputa-
tion. There are new challenges to be met. The University is
in transition to taking the "decentralized management cen-
ter" as its financial model, in which the individual units of
the University become responsible for their own finances.
The Department, in conjunction with the engineering school,
sees in this an opportunity to manage its finances and carry
out strategic decisions to enhance our enrollments, curricu-
lum, and research. A major aid in enhancing growth is
administration's recognition of chemical engineering as one
of its best science and engineering departments. We also ex-
pect our non-traditional options at the undergraduate level to
attract students. A welcome challenge in the years to come
will be increasing the Department's endowment.


1. Westwater, J.W., "The Beginnings of Chemical Engineering in the
USA," Adv. Chem. Ser., 190, 141 (1980)
2. Bailey, R.V., personal communications with the author (2002)
3. Records of the Tulane School of Engineering
4. Harris, H.G., and D.U. von Rosenberg, "A Chemical Engineering Prac-
tice Division," Chem. Eng. Prog., 69(6), 59 (1973)
5. Walz, J.Y., "The Chemical Enigneering Practice School Program at
Tulane," Chem. Eng. Ed., 29(4), 246 (1995) O

Spring 2002



Using Analytical and Monte Carlo Techniques


Michigan Technological University Houghton, MI
Investment decisions are typically based on some form of
cash-flow analysis, such as net present value (NPV) or
internal rate of return (IRR). The analysis is first per-
formed using predicted performance of the project over the
project life as if the predictions were deterministic. The sto-
chastic nature of these predictions can then be handled using
a variety of risk analysis techniques, such as: best case/worst
case scenarios; single-parameter sensitivity analysis (Strauss
plots); analytical error propagation; Monte Carlo simulation;
and decision trees. In this paper, we present the development
and application of a Microsoft Excel spreadsheet template
that facilitates analytical and Monte Carlo risk analysis
of investment decisions. We have found the template par-
ticularly useful in teaching risk analysis to senior students
in the design course.

Best/worst case scenarios calculate a return on investment
for the most profitable set of investment conditions and an-
other return for the worst possible set of conditions. This ap-
proach analyzes both ends of the spectrum in terms of return.
Generally, however, the worst case will not exceed the mini-
mum return and the best case will. Because the expected re-
sult is somewhere between the two extremes, most evalua-
tions will not be resolved by this method. The method is use-
ful for those few cases where the worst-case scenario is found
to be acceptable or the best case is found to be unacceptable.
Single-parameter sensitivity analysis tests the variability
of the result with respect to one economic variable. This type
of risk analysis is common because the calculations and in-
terpretation are simple. Only one variable is changed at any
given time, and the result (which is frequently linear) can be
shown graphically on a Strauss plot (NPV versus change in
the variable of interest). Because much information can be
derived from a small amount of work, some companies man-
date that all capital appropriation requests be accompanied
by sensitivity tests on key input values such as raw material

price, labor, utilities, etc. This technique can show the break-
even point for each of the critical economic variables.
Analytical methods use error propagation analysis to evalu-
ate the risk involved. This approach uses statistical identities
to relate the variability of each parameter to its distribution.
In order to define variability in the desired risk measure, the
relationship between the parameters and the desired measure
is combined in equation form. This equation combines all
facets of variability in the economic input values into a single
expression of variability in the desired risk measurement. The
inclusion of variability and multiparameter influence upon
the result makes the analytical method applicable for exam-
ining risk where the variability is well defined and the input
parameters are assumed independent.
Monte Carlo simulations have been used to simulate ran-
dom variation in sets of related variables. First, a statistical
distribution is specified for each input. Then the simulation
randomly selects one value for every input from the speci-
fied distribution for that item. The set of random input values
is used to calculate a result. This process is repeated a suffi-
cient number of times so that the distribution of outcomes
can be used to reliably predict the variability of the calcu-
lated result. The simulation can be run hundreds or even thou-
sands of times to explore every possible combination of vari-
ables. Monte Carlo methods have gained increasing atten-
tion due to the increased power and decreased cost of desk-

Brendan O'Donnell received his MS degree in chemical engineering from
Michigan Tech in the summer of 2001. His graduate research focused on
developing software tools to aid in process design, economic assessment,
and process improvement. He holds a BS degree in chemical engineering
from Michigan Tech and is currently working as a photolithography engi-
neer for IBM in Burlington, Vermont.
Bruce Barna is Professor of Chemical Engineering at Michigan Techno-
logical University He holds BS and MS degrees from Michigan Tech and a
PhD from New Mexico State University. He worked as a process engineer
for Reynold's Metals and Exxon and as a plant engineer and plant man-
ager for Kalsec, Inc., prior to joining the faculty at MTU.
Michael Hickner is currently a chemical engineering PhD candidate at
Virginia Tech. His research involves synthesizing and characterizing new
proton exchange membranes for fuel cells. He has conducted part of his
fuel cell research at Los Alamos National Laboratory. He received his BS in
chemical engineering from Michigan Tech in 1999.
Copyright ChE Division of ASEE 2002
Chemical Engineering Education

Currently with IBM, Inc., Burlington, Vermont
2 Currently at Virginia Tech, Blacksburg, Virginia

top computing. Once the distributions are chosen for each
economic input, repeating the calculations becomes trivial.
Decision trees employ a method of weighting an event's
economic impact by its probability. The use of decision trees
follows a left-to-right progression where each decision builds
upon the previous one until a final outcome is reached. Each
branch of the decision tree has a probability and an economic
value. The expected value of a decision can be calculated by
summing the product of the probability and economic out-
come to each decision node. Comparing the result of each
probability node will result in an ultimate, numerically based
decision. The advantage of this method is the ability to incor-
porate calculated probabilities and economic factors to give
a numerical result for a complex decision-making process."1
The logical place to teach these risk-analysis techniques in
most chemical engineering curricula is in the capstone de-
sign course. Best/worse case techniques and single-param-
eter sensitivity techniques are readily mastered by all stu-
dents. The analytical, Monte Carlo, and decision-tree tech-
niques can be more of a challenge, depending on the statisti-
cal background of the students and the time available for them
to write their own simulation routines. Recognition of this
was the impetus to develop a spreadsheet-based learning tool
that could be used to facilitate risk analysis using both ana-
lytical and Monte Carlo methods.

A cash-flow analysis was used to demonstrate analytical
and Monte Carlo techniques. A sample cash-flow table was
generated using net present value (NPV) as the result of in-

Years Constant Cash Flow
Net Present Value= (1)
i=l (1+MAR)'
Cash Flow
Constant Cash Flow (1 +Inflatio (2)
Cash Flow =
Income-Expenses-Investment-Working Capital-Tax (3)
Tax= Profit pre-tax *Tax Rate (4)

Profit pre-tax = Income- Expenses -Depreciation (5)

This cash-flow table was incorporated into a spreadsheet tem-
plate to facilitate analytical and Monte Carlo analysis of the
NPV. The template allows the user to tailor the analytical and
Monte Carlo analyses to a specific set of economic vari-
ables, including the distribution of each variable. The next
part of this article will first describe the theoretical basis
for each analysis and then its implementation in the ac-
tual Excel spreadsheet.
Spring 2002

In this paper, we present the
development and application of a
Microsoft Excel spreadsheet template
that facilitates analytical and
Monte Carlo risk analysis of
investment decisions.

The cash-flow analysis provides the relationship between
our result and the inputs: income, expenses, working capital,
investment, and inflation rate. In order to determine the vari-
ability in the result, we will use the method of error propaga-
tion.1231 When a variable, c, is a function of a number of vari-
ables, x x,,...x., it can be written
c= (x,x2,...Xn) (6)

It follows that if each xi is independent and o( represents the
variance of c, then

c =( aC a' )2 +...+ ac 2G (7)
+a ac 2 c (7)
Ixi dx2 ) -.+ xJ
By applying this equation to economic variation, with NPV
as c, fixed capital investment (Inv) as x, income (Inc) as x,,
expenses (Exp) as x3, working capital (WC) as x4, and infla-
tion (Inf) as x we arrive at the following expression for the
variance of the net present value:

2 NPV'2 22 (2NPV' 2
S- Inyv 2 \ dInc y
aNPV 2 2 ( NPV )2 F2 NPV 2 (2
Exp awC ain Inf

Using Eq. (8), if we can define all of the terms on the right
side of the equation, we should be able to calculate the vari-
ance of the net present value. The problem then becomes one
of calculating the components of the right side.
Approximating the partial differentials Partial differ-
entials represent the slope of the function with respect to the
variable of interest at a small increment. If we assume that
the function responds nearly linearly due to an incremental
change in the variable of interest, then we could approximate
the partial differential by changing the variable a small per-
centage above and below the base case value and calculating
the slope from the two resulting points. Figure 1 shows such
a change and resultant NPV with a line fit.
We can use this approach to calculate the partial deriva-
tives of NPV with respect to the rest of the variables follow-

ing a similar procedure. The task then becomes one of esti-
mating the variances for each variable.
Estimating individual variances Defining the variance
requires estimation based on past experience or future pre-
diction. For our purposes, we assumed that each of the five
input variables follows one of three distributions: normal,
uniform, or modified-beta. Although other distributions could
be added, these three distributions can represent most of the
types of distributions encountered. Each distribution requires
additional inputs to define the variance. The normal distribu-
tion requires the standard deviation as an input (the variance
is simply the square of the standard deviation). The mean
value is assumed to be the base case value, m. For the other
two distributions, we need estimates of the maximum, mini-
mum, and the most likely value for the beta distribution. Thus,
we define
a = minimum value
b = maximum value
m = most likely value (mode)
The variance for the uniform distribution can be calculated

2 (b-a)2 (9
1 (9)
The modified-beta distribution uses these maximum and mini-
mum inputs to calculate variance based on the following PERT
(Program Evaluation and Review Technique) simplified for-
mula: [5]


Modified-beta distributions can be skewed either positively
or negatively. The expected mean is different from the most
likely value and is calculated by
_t= (11)

Implementing the analytical approach on a spreadsheet
Once the variance for each variable and the partial differen-
tial of NPV with respect to each variable (or input) has been
approximated, the overall variance can be calculated. The
partial derivative for each variable is estimated by modifying
the individual parameter a set percentage (specified by the
user) and calculating the slope between the two perturbed
points. The variance of each parameter necessary to satisfy
Eq. (8) can be correlated from input values of uncertainty.
Equation 8 combines the parts to calculate the variance in net
present value. We have now quantified the uncertainty in our
economic decision variable using the analytical method.
The assumption of independence is one weak aspect of the
analytical method. Some of the variables are often interre-
lated. For example, expenses are often related to investment,

working capital is sometimes related to investment, etc. In
this respect, the Monte Carlo simulation may be more appro-
priate where such interrelations exist. The Monte Carlo tech-
nique does not explicitly account for interrelations either, but
more combinations of variables are explored, as is illustrated
in the following paragraphs.

Monte Carlo simulations can reduce error compared to the
analytical approach by performing random walks within
specified distributions and determining the results directly
from repeated trials. In this case, we are interested in finding
the variability in the net present value based on the variabil-
ity in the five economic parameters mentioned earlier.
Using assumptions for individual variability, we can pick
sets of random expenses, incomes, investments, etc., and cal-
culate a result for that set using the cash-flow equations (Eqs.
1-5) to find a net present value for the set. Variance in net
present value can then be extracted directly from the distri-

Single Parameter Strauss Plot Investment

> 0.80

10.2 10.4 10.6

9.4 96 9.8 10.0
Investment (MM $)

Figure 1. Determining the partial derivative ofNPV with
respect to investment, using a single-parameter
Strauss plot.






0 10 20


Figure 2. Normal probability plot for
mean = 10 and a2 = 4.
Chemical Engineering Education


Slope =-1.0

L 6

bution of net present value results.
Statistical derivation Normal, uniform, and modified-
beta distributions are used for the Monte Carlo simulations
as well. The normal distribution represents the standard nor-
mal or Gaussian curve. For such a distribution, approximately



) 0.10


0 2 4 6 8 10 12

Figure 3. Uniform probability plot for minimum 2 and
maximum 10.

0 2 4 6 8 10
Figure 4. Modified-beta probability plot for minimum O,
maximum 7, and mode 2.





0.40 -


0 0.2 0.4 0.6 0.8 1
Figure 5. Cumulative modified-beta probability distribu-
tion for minimum 0, maximum 7, and mode 2.
Spring 2002

67% of all values lie within one standard deviation of the
mean, 95% lie within two standard deviations, and 99% lie
within three. A normally distributed variable can be charac-
terized by its mean and variance (or standard deviation). Fig-
ure 2 illustrates a normal probability curve with mean of ten
and variance of four. Of course, the sum of the probabilities
for all occurrences is one.
The uniform distribution gives equal probability for any
occurrence between the minimum and maximum endpoints,
and is completely characterized by these values. Figure 3
shows a uniform distribution with a minimum and maximum
of ten and two, respectively.
The modified-beta distribution can be skewed in either di-
rection from the midpoint. It is characterized by its most likely
value (mode) and estimates of low and high values. Figure 4
shows the probability distribution for a sample modified-beta
distribution with a mode of two, a low of zero, and a high of
seven. In this case, there is a lower probability of values to
the left of the mode than to the right.
Using the distributions for a Monte Carlo analysis requires
programming our spreadsheet to generate random numbers
and then to extract a value from the normal, uniform, or modi-
fied-beta distributions established by the input of uncertainty
for the variables.

Generating random values within a distribution The
Microsoft Excel spreadsheet has some built-in statistical func-
tions. For instance, given a random number, a mean, a stan-
dard deviation, and a standard distribution, the function
will return a random value from that distribution. A similar
function exists for the uniform distribution. No such function
exists for the modified-beta distribution, however, so it had to
be programmed separately. The procedure below outlines the
calculation routine for the modified-beta distribution. The same
approach could be applied to any desired distribution.
Using the modified-beta probability distribution function
(Eq. 12, Figure 4), we integrate to get the cumulative prob-
ability distribution (see Figure 5). Then, Excel generates a
random number between one and zero corresponding to an
f(x) (Figure 5). The x-value is selected based on the random
f(x) and scaled to the minimum and maximum range speci-
fied. The result is a random value within the distribution

(a+b+l)! ba(lxb
f(x)= (a xa(1_ x)

a =
b =

minimum value
maximum value

Generating the simulations The randomized value of
an input variable (for example, expenses) is then combined
with the other randomized values of the variables in a set

using Eqs. (1-5) to calculate one NPV. In a Monte Carlo simu-
lation, the calculation of NPV is repeated multiple times, each
from a new set of random inputs. Initially, we anticipated
that 50 to 100 iterations would produce a stable and repro-
ducible NPV distribution. With this number of iterations we
found that the results were very sensitive to the bin size se-
lected for frequency analysis and often had not stabilized.
Considering that the computing power is readily avail-
able and that a calculation requires only one or two sec-
onds, we increased the iteration count to 500, which
proved to be sufficient.
The variability in net present value can be extracted di-
rectly from the dataset. The specific use of the spreadsheet is
discussed next, followed by a case study.

Modern spreadsheets have the usefulness of being program-
mable, extensible, easy to use, and good teaching tools. Un-
like various programming languages that hide the intermedi-
ate results, spreadsheets allow the user to see the inputs, the
dataset, and any calculations. The risk analysis template de-
veloped here provides all these aspects for both the analyti-
cal and Monte Carlo analyses.
To use the template, first the user must supply the inputs
on the first tab of the spreadsheet. Inputs include base-case
(mean or mode) values for investment, expenses, income,
inflation rate, and working capital. Cash-flow analyses also
rely on pre-set variables such as project life, minimum ac-
ceptable return, tax rate, and depreciation schedule. Mini-
mum acceptable return (MAR) and tax rate are the only pre-
set variables that can be modified by the user in this tool. The
depreciation is fixed at seven-year modified accelerated cost-
recovery system (MACRS) and project life is fixed at ten
years, all typical values for an industrial project.
Next, the user must select the distributions for each of the
variables. There is a pull-down menu for selecting the distri-
bution type: normal, uniform, or modified-beta. Below the
pull-downs are cells for specifying the variability of the dis-
tribution, selected by
Normal Requires standard deviation, o; uses base-
case values as mean
Uniform Requires a (minimum) and b (maximum)
Modified-beta Requires a (minimum) and b
(maximum); uses base-case values as most likely
These entries link to calculation tabs in the spreadsheet.
No further user input is required. The analytical result for
variance is shown on the INPUT tab, while the Monte Carlo
results are shown on the RESULTS sheet. On the HISTO-
GRAMS tab are histograms for net present value as well as
for each variable. Here the user can assess whether the simu-
lation adequately represents the input distribution. Because

the Monte Carlo calculation is dependent on random-num-
ber generation, recalculating the spreadsheet can result in
slightly different distributions.
Both analytical and Monte Carlo results include a confi-
dence interval for NPV. Sometimes it is more useful to calcu-
late the probability of a net present value greater than zero
(PNPv>0). Such a calculation would represent the probability
of the project meeting or exceeding the minimum acceptable
return given the expected variation in the variables. If the
distribution result is normally distributed, then a simple tech-
nique for calculating p-values in Excel is to use the
function. Setting VALUE=0 and subtracting the result from one
will calculate the probability of the project NPV exceeding
the minimum acceptable return. If the distribution is not nor-
mal, then probabilities must be determined directly from the
frequency distribution histogram.

A simple case study based on a hypothetical project can
help illustrate application of the template. Distributions for
the variables have been selected based on historical experi-
ence, probable error in cost estimation, etc. The minimum
acceptable return (MAR) is set at 20% and tax rate set at the
federal corporate level of 34%. Project life is 10 years. Table
1 summarizes the inputs necessary for computing an analyti-
cal and Monte Carlo risk analysis.

Base-Case Parameters for Risk-Analysis Case Study

NPV Histogram

I 60
L 40

NPV ($)

Figure 6. Histogram of net present value given con-
straints in Table 1.
Chemical Engineering Education

Max Min Std. Dev.

Working Capital

3% Normal

Base Case
$10 MM
$4 MM
$8 MM


The base-case values seen in Table 1 provide a starting point
for the study. The variability information permits study of
the variation in the net present value due to the predicted
variation of each parameter. With this information, the risk-
analysis tool has the necessary inputs to perform both the
Monte Carlo and the analytical determination of the variance
in net present value.
Using the Excel template, the expected or deterministic
NPV for the base case is found to be $1.18 MM. The tem-
plate also gives the analytical result for net present value vari-
ance as $4.455 MM2 with a standard deviation (o) of $2.11
MM. Assuming a normal distribution, the 95% confidence
interval can be generated by taking the mean NPV $1.09 MM
plus/minus 2 o, or $5.28 MM to $-3.09 MM. The p-value for
NPV greater than 0 is 0.7, signifying a 70% chance of the
project exceeding the minimum acceptable return.
Monte Carlo results are also generated by the template us-
ing 500 iterations. The data set is displayed using histograms.
For this case, the mean is found to be $1.46 MM with a stan-
dard deviation of $2.09 MM. A 95% confidence interval for
net present value is $5.50 MM to $-2.69 MM. The p-value
for NPV greater than zero is 0.77, signifying a 77% chance
of the project exceeding the minimum acceptable return. The
Monte Carlo results are close to but slightly different than
the analytical result.
The histogram in Figure 6 shows a sample of the simula-
tion. Again, because the results are based on random-number
generation, each recalculation could have slightly different
results. With 500 iterations, the mean and confidence inter-
vals are essentially constant between simulations, but the
shape of the histogram varies much more than we had antici-
pated. The central limit theorem suggests that the sampling
distribution of the mean can be approximated by the normal
distribution, regardless of the population. Therefore, we would
expect the outcome of a calculation involving large numbers
of input variables to be normally distributed, regardless of
the distribution of the inputs. This does not appear to be the
case for our cash-flow analysis and suggests that the Monte
Carlo results are probably a better measure of project risk
than the analytical results.

We have used this template for several years in the capstone
design course and have found it to be useful in teaching the
concept of risk analysis and analytical estimation of that risk.
The students perform a feasibility analysis of a new project
or plant in the fall semester. As part of this analysis, they are
asked to include an economic analysis and risk analysis of
the venture, both in their written report and their oral presen-
tation to the management of Fictitious Chemical Company,
their hypothetical employer.
For the economic analysis, students use a spreadsheet tem-
Spring 2002

plate that they have each been asked to generate in a prior
homework assignment. For the risk analysis, however, most
of the students were historically limited to the single-param-
eter sensitivity approach since not all of them had the statis-
tical background or time to conduct the more elaborate analy-
ses. Some of the more capable groups are challenged by ask-
ing them to perform an analytical or Monte Carlo analysis to
illustrate the techniques to the entire class.
With this template now available, we are able to ask all the
students to apply the full spectrum of risk analysis techniques
to their projects. They are provided with the spreadsheet file
and told that they are free to modify it or use it as they see fit.
We find that use of the template and this approach allows us
to concentrate more on the actual teaching of risk analysis
and less on the programming required to do the analysis. All
the students are able to successfully apply the software. Those
with a strong statistical background tend to do a better job of
interpreting the results.

Risk analysis is a critical part of any project decision. In-
creasing the minimum acceptable return or setting higher
breakpoints are simple methods of compensating for risk that
have been used as shortcuts in the past. The goal of this re-
search was to develop a spreadsheet template for quantifying
the risk in the discounted cash-flow measure, net present value.
Analytical and Monte Carlo methods were implemented in a
Microsoft Excel template for ease of use. Both methods result
in a mean and a standard deviation value. The template also
calculates confidence intervals based on the results.
The template is a work in progress. We hope, in future ver-
sions, to be able to streamline some of the Monte Carlo simu-
lations, to develop macros for group calculations, and to by-
pass some of the more computationally intensive tasks. We
also hope to add the ability to alter project life, iteration num-
ber, depreciation schedule, etc.
We have found the template to be quite useful in teaching
risk analysis concepts to our seniors in the plant design course.
Faculty who would like to try it in their courses may contact
Bruce Barna to get the latest version. Be advised that the file
is large (approximately 6.5MB). We would appreciate feed-
back on the experiences of other users.

1. Blank, L.T., and A.J. Tarquin, Engineering Economy, 4th ed., McGraw
Hill, New York, NY, pp. 574-580 (1998)
2. Jelen, F.C., and J.H. Black, Cost and Optimization Engineering,
McGraw Hill, New York, NY, p. 168 (1983)
3. Holland, F.A., FA. Watson, and J.K. Wilkinson, "Probability Tech-
niques for Estimates of Profitability," Chem. Eng., Jan. 7, pp. 105-110
4. Freund, .E., and I. Miller, Probability and Statistics for Engineers,
Prentice-Hall, Englewood Cliffs, NJ, p. 114 (1977)
5. Whitehouse, G.E., Systems Analysis and Design Using Network Tech-
niques, Prentice Hall, Englewood Cliffs, NJ, p. 42 (1973) 0

e, internet


For Chemical Engineers

University of Tehran Tehran, Iran

he amazing capabilities of the Internet have exponen-
tially improved the chemical engineering curriculum.
Through this medium, it is now possible to learn about
the latest process engineering techniques and novel technolo-
gies, to compare the capabilities of commercial simulators,
to obtain detailed information on research and education, and
to participate in discussions with colleagues from around the
world, among other things. The challenge that comes along
with this medium, however, is learning how to navigate its
various areas in order to optimize the informational gain.
The interconnection of millions of computers on the Internet
has caused some to worry that quality, stability, credibility,
and security cannot always be assured, so careful use of the
medium must always be a consideration. Fast circulation of
information between chemical engineers will most certainly
lead to new developments, i.e., building new body parts that
can function intimately with living tissue requires delivery
of a large amount of biomedical information that could be
acquired through the Internet.
In the past few years, the Internet has transformed the way
chemical engineers share information by removing the barri-
ers of time and distance.[1'21 The goal of this paper is to pro-
vide a brief listing and a short discussion of on-line resources,
discussion groups, and electronic mailing lists. In order to
assure the long-term usefulness of this information, a more
complete version can be found on the website of the Chemi-
cal Engineering Department at the University of Florida
where the links will
be updated as necessary over time.
Starting Points Good starting points for learning more about
the features of the Internet are shown in Table 1 and Table 2.

Rahmat Sotudeh-Gharebaagh is Assistant Professor of Chemical En-
gineering in the Engineering Faculty at the University of Tehran. He re-
ceived his BEng from Sharif University of Technology, Iran, and his MS
and PhD from Ecole Polytechnique de Montreal, Canada. His teaching
responsibilities include applied mathematics for chemical engineers, pro-
cess simulation and transport phenomena. His research interests involve
modeling and simulation of fluidized bed reactors, pharmaceutical engi-
neering processes and information technology.

Copyright ChE Division of ASEE 2002

Although many other pages are available, these have some
unique and useful features. Many reviews have been pub-
lished regarding search engine performances, ratings, popu-
larity and traffic, their trends over time, and search tech-
niques.13,41 Good starting points specifically for chemical en-
gineers are presented in Tables 3 and 4 (university web pages
are also good for obtaining information15,61).
Companies There are many sites for chemical companies,
and links to them are provided in Tables 3 and 4. Typical
information is product lists, literature, contact information,
submission forms, on-line catalogues, and reference data.
Societies and Institutions A compilation of Web pages for
chemical engineering societies is given in Table 5. In addi-
tion to contact information, membership applications, infor-
mation of meetings, grant and award information is also listed.
Several ChE society sites have member directories.
Educational Resources An increasing number of papers
have been devoted to evaluating the use of the Intemet for
educational purposes.17"8 Table 6 gives web sites that contain
lecture notes, tutorials, exercises, exams, etc., with additional
tutorials being listed in Table 7. Table 8 presents a series of
hypertext links related to ChE history.
Research A list of web pages of academic and government
research organizations and laboratories in given in Tables 9
and 10, and a patent search system is shown in Table 11.
Special and general reports for decision makers can also be
found on the Internet (i.e., the report on the strategy and struc-
ture of chemical engineering research in the USA that was
developed by a team of UK professorstg1).
Journals Table 12 lists web pages for technical and gen-
eral chemical engineering journals, and a compilation of sci-
entific ChE journals is shown in Table 13. Chemical-engi-
neering related journals are also available through Elsevier
Science web site by registration.[ 10
Software and Simulators In Table 14, a selected list of pro-
cess simulators and simulation software is listed. In addition,
a compilation of a general list of chemical engineering soft-
ware can be found on the Internet."1,121 Few web-based cal-

Chemical Engineering Education

[ TABLE 1. Selected Starting Points
Internet Information Center
Learn the Internet

I TABLE 2. Selected Internet Search Pages
Google (high relevancy, proximity, clustering, news)
Yahoo (subject directory, easy information location, etc.)

E TABLE 3. Selected Starting Points for Chemical and Process Engineers
Askache (general assistance, journal information, etc.)
Che-comp (relevant links, etc.)

[ TABLE 4. Additional Links of Interest to Chemical Engineers
Chemical Abstracts (searchable database)
ChE and Computer Resources

[ TABLE 5. Selected Sites for Chemical Engineering Societies and Institutions
American Chemical Society
American Institute of Chemical Engineers

[ TABLE 6. Selected Sites for Chemical Engineering Online Courses
Advanced Mathematics
Advanced Reactor Design

a TABLE 7. Sites for Additional Online Tutorials
Sugar Engineers' Library

E TABLE 8. Sites of Historical Information for Chemical Engineers
Chemical Engineering Timeline http://www.pafko,com/history//h_time.html
History of ChE

E TABLE 9. Selected Sites of ChE Academic Research Centers and Laboratories
Advanced Combustion Engineering Research Center
Biochemical Engineering Research Group

E TABLE 10. Selected Sites of Governmental Research Organizations and Laboratories
United States Department of Energy
Von Karman Institute (Fluid Dynamics)

[ TABLE 11. Selected Sites of Patent Information Resources
Delphion Intellectual Property Network
U.S. Patent and Trademark Office haap://

[ TABLE 12. Sites for Technical and General ChE Journals
Chemical Engineering Progress
Hydrocarbon Processing

[ TABLE 13. Selected Sites for Scientific Chemical Engineering Journals
American Institute of Chemical Engineers (AIChE)
Canadian Journal of Chemical Engineering

[ TABLE 14. Selected Sites for Chemical Engineering Process Simulators and Software

E TABLE 15. Usenet Groups of Interest to ChEs and Links to their Webs (energy)
Sci.engr.chem (chemical engineering)

[ TABLE 16. Selected Electronic Mailing Lists for Chemical Engineering
Cheme-1 (Chemical Engineering) .htm
Csche-l (Canadian Society of Chemical Engineering)

culators, which are useful for fast calculations (i.e., TCC),
can also be found on the Internet. CEP's Software Direc-
tory maintains comprehensive software information rel-
evant to chemical and process engineering."31
Mailing and Discussion Groups Discussion groups and
electronic mailing lists are given in Tables 15 and 16.
Electronic mailing lists are well suited to small groups
of users with specialized topics, and Usenet newsgroups
are suited for general topics and more users."'41

In the future, the Internet will see an increasing num-
ber of users, better use of existing resources, and imple-
mentation of new technologies. It is presently under-used
as a teaching and research tool and has enormous poten-
tial yet to be discovered in those areas. For chemical en-
gineers, the main challenge is to develop this technology
for remote collaborative research, collective education.
Use of its full potential will overcome many existing
barriers of time and distance. In addition, on-line educa-
tion to the full potential of the existing resources could
affect the chemical engineering curriculum in the com-
ing years. Chemical engineering studies will be affected
when experimental set-ups, pilot plants, and laboratories
can be shared by connection to the Internet with a
robotically controlled camera, remote Internet instrument
control, and a data-acquisition system.

The author would like to thank the Chemical Engineer-
ing Department at the University of Tehran for its help.
Financial support provided by the University of Tehran
and the Office of the Vice-Chancellor for Research is also
acknowledged. Special thanks also go to Dr. Dale Kirmse
for incorporating this paper on the web site of the Chemi-
cal Engineering Department at the University of Florida.

1. Rosenzweig, M.D., "The Net Makes a Mark," CEP, 96(11), 93
2. Mascone, C.F, "Engineering the Next Millennium," CEP, 95(10),
102 (1999)
7. Lancashire, R.J., "The Use of the Internet for Teaching Chemis-
try," Analy. Chim. Acta, 420, 239 (2000)
8. Johnson, D.R., M. Ruzek, and M. Kalb, "Earth System Science
and the Internet," Comp. & Geosciences, 26, 669 (2000)
14. Murray, K., "Internet Resources for Mass Spectrometry," J. Mass
Spect., 34, 1 (1999) 0

Spring 2002

H, laboratory



Zhejiang University Hangzhou 310027, China

Recent technological advances on the Internet have en-
able2d a multitude of applications to operate via the
World Wide Web. E-Learing is one such applica-
tion. To this end, various institutions and universities are now
offering online courses that students from all over the world
can subscribe to and attend. Various teaching methods and
tools have appeared, such as Web Courseware, On-Line An-
swer Machine, Web Classroom, etc.
Virtual experiments have also been developed. Some re-
cent articles describe the design of virtual experiment sys-
tems and their uses in academe.'" 4 Another important simu-
lation technology, virtual reality (VR), was introduced by
John Bell and Scott Fogler as a powerful new tool in en-
gineering education.51
We have developed an infrastructure for a Web-Based VR-
Form Virtual Laboratory (WBVL) to aid in the undergradu-
ate laboratory. We have successfully implemented several
virtual chemical engineering experiments, such as "Measure-
ment of Water's Degree of Hardness." In the following para-
graphs, we will describe the four basic features of WBVL,
give technical details, describe the set-up, introduce the vir-
tual experiment operational model, and list the benefits to be
gained from WBVL.

Educating students in engineering and related scientific
fields is made difficult by the complex ideas and phenomena
that are hard to demonstrate by conventional methods. For
example, it is hard to get students to understand the structure
of molecules through the use of 2-D graphics, and a virtual
experiment faces the same problem. In order to help the stu-
dents understand and master virtual experiments, they must
be presented in 3-D form, where they can observe any object
from any point of view and any angle. This important feature
of WBVL not only helps the student accomplish their objec-

tives, but it also aids in holding their interest.
When performing experiments, students run the apparatus,
observe phenomena, record data, and complete a report of
the experiment. This means that they must interact with the
experiment at every stage. For this reason, interactivity is
another important feature of WBVL. Each virtual experiment
is capable of presenting different reactions to differing input
by the students. This interactivity helps the student feel that
they were actually doing the experiment.
At times the students cannot access a laboratory and the
only way they can conduct an experiment is through the
Internet, so network basing is also important. With network-
based virtual experiments, experiments are not limited to the
laboratory environment.
Virtuality, another feature of WBVL, is a kind of virtual
experiment based on simulation and is called Simulation
Experiment (E). In some of the articles mentioned previously,
another kind of virtual experiment, called the Remote Con-
trol Experiment (RCE), is described in which the students
control actual experimental instruments via the Internet. This

Dong Yabo holds an MS in EE and is now a
PhD candidate in Computer Science & Engi-
neering. His research interests involve the
application of Internet-Based Virtual Reality to
distance education and embedded systems.

Zhu Miaoliang holds an MS degree in com-
puter science from Zhejiang University (1981).
He is presently a professor in the Computer Sci-
ence & Engineering Department at Zhejiang
University He has been a visiting professor at
several U.S. universities, including Maryland,
Missouri, Kansas, and RPI.

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

is not feasible in all cases, however, since some experiments
use apparatus that cannot be remotely controlled, or they take
a long time to complete, or they are too expensive. RCE also
limits the number of students who can participate. SE has
no such limitations and is a more realistic form for vir-
tual experiments.

A server/client model is used in WBVL, with the server
side being based on a web server (see Figure 1). All the vir-
tual experiments are kept on the server side, and when the
students want to load one they send a request to the server
and it delivers the corresponding experimental data. The ben-
efit of this approach is that updates and revisions are done
centrally on the server and the latest version is always avail-
able to the clients (students).
The architecture of the client side is shown in Figure 2.
Microsoft Internet Explorer is the user interface and all the
experimental material is interpreted and represented in 3-D
form. The 3-D scenes of the virtual laboratory and experi-
ments are described by VRML,16] the most common standard
for describing interactive 3-D objects on the web. A VRML
plug-in is needed on the client side to interpret the 3-D

Virtual User's Virtual Virtual
Campus Guide Laboratory Experiments

Web Server

Figure 1. Server-side system architecture.

Internet Explorer

Figure 2. Client-side system architecture.
Spring 2002

scenes. The following paragraphs describe all the com-
ponents of this system.
Web Server The Web server is used on the server side to
deliver virtual experiment curricula to the client side.
Virtual Campus The virtual campus is a 3-D university
environment where the students can go to different virtual
laboratories to perform different experiments. In this virtual
campus there are buildings and laboratories; the laboratories
have hyperlinks directing the students to the experiments.
User's Guide A user's guide teaches the students how to
use WBVL and includes information on the virtual experi-
ments (explained more thoroughly later in this paper).
Virtual Laboratory Framework The virtual laboratory
framework is the common framework of most virtual experi-
ments. Just as in the real world, many virtual experiments are
performed in the same laboratory and thus share a common
virtual laboratory framework, including a 3-D laboratory
scene and an interface to other parts of WBVL, etc. This ap-
proach has three benefits: it can reduce the design complex-
ity of the experiment (the designer need only concentrate on
the experimental contents, models, and private 3-D scenes of
each experiment); updates of the laboratory scene and func-
tion are much easier; and the shared laboratory framework
makes the graphics interface of all the experiments uniform.
This helps students master the use of WBVL quickly.
Proprietary Experiment Database Although WBVL has
a uniform laboratory framework, each experiment should have
its own proprietary characteristics. The experimental scenes
must be different because different experiments use different
apparatus and devices. Experiments' contents and models
are also different from each other. These private proper-
ties of each experiment are stored in the proprietary ex-
periment database.
Help Text Database The help text database contains the
text of online help for the experiments.
Internet Explorer and VRML Plug-In In order to use
WBVL, a general web browser on the client side is required.
We use Microsoft Internet Explorer (see Figure 3). ActiveX
is used to organize and gather several parts of WBVL to-
gether and Internet Explorer acts as an ActiveX container.
Internet Explorer is also the uniform graphics user interface
(GUI) for WBVL.
Realistic 3-D models of the laboratory and the experimen-
tal proprietary scenes are developed using VRML. The Vir-
tual Reality Modeling Language (VRML) is a file format for
describing interactive 3-D objects and worlds. VRML, used
on the intent, intranets, and local systems, is also intended
to be a universal interchange format for integrated 3-D graph-
ics and multimedia, so it fits WBVL's features well.
Although VRML is designed to describe the interactive 3-
D objects, to a certain extent its interactive features and 3-D


Help Text


modeling ability are still limited. For example, standard
VRML can only receive some simple mouse actions, such as
clicking, touching, and dragging. Also, it doesn't support
curved surface modeling. Such simple interactive and mod-
eling abilities are insufficient for WBVL, but fortunately, some
extensions have been developed that extend the VRML func-
tions, including keyboard-input node, drag-and-drop node,
geometric NURBS node, and geometric spline nodes that can
model more complex and higher quality shapes with fewer
surfaces and less file size. With
the VRML extensions, we can
manipulate WBVL easily and ..- **. a -.
more efficiently. --- --
A VRML plug-in is neces-
sary to interpret the 3-D Vir-
tual Experimental Environ-
ment (VEE). There are cur-
rently many VRML plug-ins
available, but not all of them
support VRML extensions or
provide a Software Develop-
ment Kit (SDK), so the selec-
tion of a VRML plug-in is very
In VEE, students can oper-
ate all kinds of virtual devices Figure 3. One WBVL e
Water's Degr
(detailed later in this paper),
such as beakers and test tubes,
which are shaped as 3-D objects by VRML. Through these
operations, the students can control the experiment's progress,
observe phenomena, and record data. Because a VEE con-
sists of many virtual devices and other 3-D objects, the 3-D
scene can be very complex, and a good VRML plug-in can
improve the system performance.
Considering all the above criteria, we have chosen Cortona
VRML Client,7,18' one of the best free VRML plug-ins, to
implement WBVL.
Script Nodes The Script node is an important VRML node
used to program behavior in a scene. Each Script node has
associated programming language code that is executed to
carry out the node's function. Using the Script node in WBVL
has four purposes:
1. Communication with objects outside VEE In order to
make WBVL work, cooperation between it and other
elements of WBVL is necessary. For example, VEE
should communicate with Online Help to change the help
text as the experiment progresses. Such communication
ability is achieved by the programming language code
in the Script node, together with EAI or SDK technol-
ogy, mentioned below.
2. Animation generation Many actions in WBVL, such
as object moving, coloring, shaping, etc., use animation.

ee of

Most animation is generated with the Script nodes,
which manipulate the attributes of objects to gener-
ate their animation.
3. Receiving students' commands In most cases,
students'commands are received by sensor nodes, which
is the standard way VRML interacts with the user. In a
3-D experimental scene many sensor nodes may be
used, such as TouchSensor, CylinderSensor, PlaneSensor,
etc.-all are able to receive the mouse actions and send
one or more events to the
L.. Script nodes. The Script nodes
_3 then interpret the students' or-
ders and carry out the cor-
responding action (typi-
cally, sending some events
to other nodes).
4. Simulation in WBVL The
Script nodes, which have
simple computational ability,
are used to carry out uncom-
plicated simulations. Although
the computational ability is
somewhat limited, the simu-
lation results can be used di-
rectly to change the repre-
ent: "Measurement of sentation of VEE, so simu-
Hardness. "
lations that have low compu-

data input/output can
Script nodes.

national complexity and large
be easily performed using the

EAI External Authoring Interface (EAI), designed to al-
low an external environment to access nodes in a VRML
scene, is part of VRML97 standard. Using EAI, we can ex-
tend the interactive and computational abilities of VRML.
VEE, together with another important and powerful element,
Java Applets, is embedded in the same web page. Most of the
numerical/symbolical computations, analyses, and simula-
tions of the virtual experiments are the responsibility of
Applets. EAI enables Applets and VEE to communicate
with eact other. Applets can exchange events with VEE
and be notified when the node fields in VEE are changed.
By means of EAI, WBVL can obtain extensible
interactivity. The students can interact with experimental
environments via the normal controls on a web page. For
example, students can adjust an experimental parameter
by inputting data in a textbox or by changing the view-
point using an Applet button.
SDK Although EAI can extend the interactive and com-
municative abilities of VRML, these features are still quite
limited. For example, EAI is incapable of creating some ad-
vanced interactive objects, such as popup menus, or of ac-
cessing all nodes in a 3-D scene. In addition, EAI is compli-
cated to use. SDK (Software Development Kit) is a better
Chemical Engineering Education

solution because it is easier to use and much more powerful.
It is more frequently used in the design of WBVL.
SDK provides an Application Programming Interface (API)
that enables the developers to integrate 3-D technology and
VRML into a web page. A 3-D scene is treated as an ActiveX
control, and outer VBScript, JavaScript, and Java Applet codes
use SDK to access and manipulate any object in it or use the
VRML event model to exchange events with it. At the same
time, some low-level functions provided by SDK are neces-
sary in building advanced interactive means, such as popup
menus, hints, toolbars, etc., which are very useful in building
a user-friendly interface. For example, when a student clicks
on a test tube, a popup menu will appear to allow selection of
what to do with it, i.e., dump it or move it.
Outer Programming Code An outer programming code
is requisite when using the Script node, EAI, and SDK. Co-
operation of each part of WBVL is also achieved by using it.
Java and Script languages are used for outer programming
codes and are transferred to and run on the client side.
The Script nodes of VRML can use Java classes as their
associated programming codes. Generally, Java Applets use
EAI to communicate with VEE, but SDK can also be used by
Java Applets and JavaScript/VBScript codes to collabo-
rate with VEE. For example, the on-line help system is
handled in this way.

Similar to the traditional experiment set-up, WBVL is also
composed of four parts: the user's guide, laboratory experi-
ments, on-line help, and experiment reports. The user's guide
tells the students the basic rules about the virtual experiments.
After mastering the necessary learning, students can perform
the experiment. The on-line help system provides informa-
tion to the student at any stage of the experiment. After fin-
ishing, students are required to fill out an experiment report,
which will be graded immediately by WBVL.
User's Guide Setting up the user's guide has two main
goals. First, students can get basic information about the ex-
periment they will perform. The user's guide is intended to
be an extension of the traditional curriculum held by teach-
ers. Second, through the user's guide students can become
familiar with WBVL. Although WBVL is designed to simu-
late the true world of experiment as much as it can, it is still
not easy for a novice to run the experiment. The user's guide
gives the students enough information to use WBVL.
The user's guide is divided into three parts:
1. Introduction to WBVL This part describes the basic
use of WBVL. Students can read this part and become
familiar with the whole system.
2. Virtual Devices Library Virtual devices are those de-
vices used in the experiments that have special functions
and that students can interact with, e.g., beakers, tubes,
Spring 2002

etc. The virtual devices library contains a basic intro-
duction to them and describes their uses. Because of the
differences between the virtual and the real worlds, with-
out learning about how these devices look and how to
use them, students would not be able to conduct the
3. Specifications of Virtual Experiments This part gives
the details of each experiment on WBVL, including the
experiment's goals, principles, contents, methodologies,
and emphases. It contributes greatly to the students' un-
derstanding of the experiments.

Laboratory Experiment In WBVL, one virtual labora-
tory environment and several proprietary environments have
been implemented so a variety of experiments can be per-
formed. Figure 3 illustrates an experiment ("The Measure-
ment of Water's Degree of Hardness") where students can
operate beakers, graduated cylinders, pipettes, burettes,
etc., to measure the degree of hardness of a water sample.
In the process of experiment, they observe the phenom-
ena and record the necessary data. The original data of
this experiment is given randomly to make the experi-
ment more authentic.

Online Help In performing experiments, students (espe-
cially novices) are sometimes confused by the numerous ex-
perimental steps and data and need a help mechanism that
can give them instruction whenever they need it. Online help
is an important feature of WBVL. Along with the experiment's
process, the online help loads help text from the database
and displays it in a help frame, as shown in the right side of
Figure 3. The help text tells the students what to do and how
to do the next step. Relative knowledge about the experi-
ments is also provided in the help text.

Experimental Report In conventional curricula, students
submit a report on the experiment as their last step. The re-
port includes the design of the experiment, phenomena record,
data record, analysis, and conclusions. It sums up what the
students learned, whether they reached their goals, and how
well they did. The report also allows the teacher to discern
the level of the students' knowledge.
Because of the importance of the report, it is also a part of
WBVL. Once the students finish their experiment, a report
that was prepared beforehand is presented for them to fill
out. WBVL then grades their experiment according to records
made during the experiment combined with their final re-
port. For computer grading, templates of the experiment re-
ports are compiled in a standardized form (i.e., the templates
are composed of groups of multiple-choice tests or blank
quizzes) that are carefully designed by the instructors. Be-
cause the standardized forms enable computers to check the
reports automatically, a teacher's workload is greatly reduced.
Using this method, however, gives the students little oppor-
tunity to write a complete experimental report, so written re-

ports are also required in order to give the student training in
report writing.

Virtual experiment operational models control the opera-
tional sequences of the experiments and can greatly affect
the difficulty or ease of performing the experiment. The choice
of this model should be carefully considered in the system-
atic analysis. Generally, the operational models can be di-
vided into three categories: concurrent models, serial mod-
els, and combined models.

Concurrent Model In the concurrent model, all virtual
devices in the experimental scene can be operated concur-
rently, i.e., the students can choose an arbitrary device to op-
erate at any time. Figure 4 shows the architectural structure
of this model. Skilled students usually prefer this model be-
cause of its unrestricted nature. Real-world experiments can
be best simulated with this model, but it also has the follow-
ing disadvantages:
1. The students get no information about what they should do
next since there are no restrictions for their operation. Some
inherent restrictions in the real-world experiments are not
noted, leaving so many choices for them that they sometimes
become bewildered.
2. Because actions are arbitrarily chosen, abnormal results-
even illegal operations-are inevitable. For example, an
experiment that requires reagent A to be mixed with B, then
with C, may have that sequence altered in a virtual experi-
ment, with uncertain results. Under the concurrent model, the
system has to cope with all the extra operations and give
appropriate responses. If numerous virtual devices are used,
there may be too many extra operations and combinations to
be dealt with. This is a key problem in using the concurrent
model. Because of this disadvantage, the concurrent model is
seldom used alone in WBVL. Only those virtual experiments
that don't restrict the sequence of each step use this model.

Serial Model An alternative is the serial model, in which
the student must obey a predetermined operational sequence.
In this model, the students can operate only one device at a
time, and which device they can operate is determined by the
sequence database. After they finish one stage of the opera-
tion, another device becomes operable while the other de-
vices remain "blind," i.e., cannot be operated. The students
are thus forced to perform the experiment in a predetermined
correct sequence.
Compared to the architectural structure of the concurrent
model, the serial model inserts a valve between the user in-
terface unit and the virtual device (see Figure 5). An opened
valve enables the students to control the corresponding de-
vice, while a closed valve makes the corresponding device
inoperable. The open and closed sequences are predetermined
and stored in the operational sequence database at the time of

the experiment's design.
Benefits of the serial model are
1. The experiment's process is clear to the students. They are
guided step-by-step through the experiment and can finish it
easily and more efficiently. At the same time, this model
focuses the student's concentration on the experimental
phenomena and data, and not just on how to finish the
experiment. This feature is especially usefulfor novices who
are not familiar with either WBVL or the experiment.
2. By limiting the student's operation to a single device, the
serial model avoids the possibility of illegal operations, so
the design complexity of WBVL is greatly simplified. This
can greatly shorten the development period of WBVL as well
as reducing its cost.
The drawbacks of the serial model are obvious. They are
1. A reduction of reality sensation. The students cannot choose
which virtual devices they control. This limits the indepen-
dent students.
2. It disables the experiment's variability. By making the

Virtual Virtual Virtual
Device 1 Device 2 Device n

User Interface Unit

Figure 4. Architectural structure of concurrent model.

Virtual Virtual Virtual
Device 2 Device 3 Device 4

Operation Virtual
Sequence Device 1

User Interface Unit

Figure 6. Architectural structure of combined model.
Chemical Engineering Education

Device n

students strictly follow a predetermined experimental step,
slight changes are impossible. Student initiative becomes
Since the virtual experiments using the serial model are
subject to the risk of being treated as mere television shows,
it is only used in exhibitive experiments.

The pure concurrent and serial models are rarely used in
the design of WBVL because of their disadvantages. An al-
ternative is the combined model, which combines features of
both the concurrent and serial models.
The concurrent model has no predetermined operational
sequence, while the serial model defines a straightforward
operational sequence without any branches. The operational
sequence defined in the combined model is much more com-
plicated than either of them.
The number of controllable virtual devices in the combined
model can be larger than one, as illustrated in Figure 6. It can
be seen that although the global structure is serial, the con-
current model is also sometimes presented. The global serial
operation model insures that the experiment can reach a cer-
tain end, while the partial concurrent model enables students
to select the best way to accomplish the experiment. The
student's selection and operation will affect the final score.
This model is suitable for most virtual experiments. For
instance, the experiment "Usage of Analytical Balance" uses
such a model. Opening the sliding door should occur before
placing the object on the pan. The serial model is used to
determine the operational sequence of these steps, but
since the selection of the poises is arbitrary, the concur-
rent model is also used.

Based on its four basic features, WBVL has a number of
benefits for engineering education. First, it is valuable for
simulating some special experiments, e.g., dangerous experi-
ments, very large or small experiments, expensive experi-
ments, or experiments that cannot be done in a traditional
laboratory. Second, WBVL can be used as pre-lab prepara-
tion. Before students perform a real experiment, they can
become familiar with it through WBVL, thus improving the
efficiency and avoiding potential damage of some vulner-
able apparatus. Third, WBVL may be a way to solve the ex-
perimental education problem of distance education. In dis-
tance education, students from geographically distinct loca-
tions can no longer go to a laboratory to perform an experi-
ment. Online virtual experiments are more realistic in this case.
In order to aid E-Learning of the basic chemistry course, a
set of virtual experiments has been realized in WBVL, in-
cluding "Measurement of Water's Degree of Hardness,"
"Usage of Analytical Balance," "Electrolytic Polishing," and
Spring 2002

"Ionic Equilibrium of Electrolyte in Water." Some students
use this system as a pre-lab preparation and feel that after study
on the WBVL, they can better master the basic concepts and
operations of the experiments.
A demo of WBVL is available at
It is a Chinese web site where a virtual campus and four vir-
tual experiments are provided. We recommend the following
minimum configuration for running WBVL: Windows PC,
PII 450 or higher, 128M RAM, TNT2 or better display card
(16 bits color), 1024x768 screen, MS Windows 98/2000,
Internet Explorer 5.0 or higher, Cortona VRML Client 3.1.m[

WBVL should be considered as a supplement rather than
as a replacement for conventional experiments. With its four
important features (3-D form, interactivity, networking bas-
ing, and virtuality), WBVL makes the students feel they are
performing a real-world experiment. WBVL is structured in
Client/Server architecture, with the experiment curriculum
material held on the server side. Because the only require-
ment on the client side is a general web browser and a VRML
plug-in, WBVL is easy to use. The set-up is similar to con-
ventional curricula, so students can quickly master it. The
operational model used in most of the virtual experiments is
the combined model, which can achieve a better compromise
between the virtuality and the system complexity.

The support of the project by the National High Technol-
ogy Research and Development Program of China (863
Program)(Project No. 863-317-01-04-99) is gratefully ac-

1. Schmid, C., and A. Ali, "A Web-Based System for Control Engineer-
ing Education," Proc. of Amer Cont. Conf., 5, 3463 (2000)
2. Apkarian, J., and A. Dawes, "Interactive Control Education with Vir-
tual Presence on the Web," Proc. ofAmer Cont. Conf, 6, 3985 (2000)
3. Shin, Dongil, En Sup Yoon, Sang Jin Park, and Euy Soo Lee, "Web-
Based Interactive Virtual Laboratory System for Unit Operations and
Process Systems Engineering Education," Comp. and Chem. Eng., 24,
1381 (2000)
4. Ball, J., and K. Patrick, "Learning About Heat Transfer: "Oh, I See"
Experiences," 29th Ann. Front. in Ed. Conf., 2, 12C5/1 (1999)
5. Bell, John T., H. Scott Fogler, "A Virtual Reality Based Educational
Module for Chemical Reaction Engineering, found at http://
6. The Virtual Reality Modeling Language, International Organization
for Standardization, ISO/IEC DIS 14772-1, found at http://
7. Cortona VRML Client, found at
8. Cortona Software Development Kit, found at http:// O

S=1 classroom


Through the Investigation of Commercial Beer

Rowan University Glassboro, NJ 08028-1701

Historically, design courses in the chemical engineer-
ing curriculum focus on teaching process design
rather than product design. A traditional program
may contain one or two design courses at the senior level-
the first generally addresses the design of unit operations such
as physical separators, distillation columns, heat exchangers,
turbomachinery, and other process components, while a sub-
sequent capstone design course provides an opportunity for
students to combine what they have learned in previous
courses such as thermodynamics, reaction engineering, and
transport phenomena. In the capstone design course, students'
efforts are usually geared toward designing a process to manu-
facture a commodity chemical, such as cumene or styrene.
This traditional design education originated and was driven
by the needs of the chemical commodity industry that domi-
nated the chemical industry during the twentieth century.
Recently, Cussler'" indicated the importance of including
product design in the capstone design course. His view is
consistent with a new industry reality where the traditional
oil and chemical companies are introducing major changes
to remain competitive. Process optimization, energy integra-
tion, and alternative raw materials are no longer sufficient to
provide chemical companies with a leading edge. This new
business reality suggests that producing shorter-life products
and being the first on the market is the "new" way to succeed
in business and stay profitable.
In the U.S., new start-up companies, mostly in the product
business, are constantly emerging. Cussler's statistics also
show that in the last twenty years, more chemical engineer-
ing graduates have gone to work in companies that manufac-
ture products rather than in traditional chemical plants.
Westerberg and Subrahmanian12' also address the importance
of introducing product design in the chemical engineering
curriculum and give a clear description of the differences
between process and product design. They list the main char-
acteristics that define chemical products as
SProducts that are chemicals, such as pharmaceutical drugs,
proteins, pesticides, and cleaning fluids


Products that require chemistry in the manufacturing
process, such as computer chips
Devices that involve chemistry in theirfunctionality, such as
asbestos-removal systems, fuel cells, and portable oxygen
Products that are produced in small volumes and that
possess a high added value
Such products have to meet certain customer needs and
can only be conceived and designed by a multidisciplinary
team that includes engineers. If chemical engineering stu-
dents are to be ready to participate in product design, the cur-
riculum must be adjusted to introduce product design.
At Rowan University, the first introduction to product de-
sign occurs in the Freshman Clinic, a two-semester sequence
that introduces all freshman engineering students to engineer-
ing. The first semester of the course focuses on
multidisciplinary engineering experiments using engineering
measurements as a common thread; the theme of the second
semester is the reverse engineering of a commercial product
or process. Previous reverse engineering projects have in-
volved products such as automatic coffee makers, 3'4' hair dry-
ers,m51 and electric toothbrushes.'16 We also incorporated the
design and reverse engineering of a process into our Fresh-
man Clinic through a brewing process.'7 The project de-
scribed in this paper focuses on the investigation of com-
mercial beer as a means of providing a first introduction
to chemical product design.

Stephanie Farrell is Associate Professor of Chemical Engineering at Rowan
University She received her BS in 1986 from the University of Pennsylva-
nia, her MS in 1992 from Stevens Institute of Technology, and her PhD in
1996 from New Jersey Institute of Technology. Her research expertise is in
the field of drug delivery and controlled release.
James Newell is Associate Professor of Chemical Engineering at Rowan
University Prior to joining the Rowan faculty in 1998, he spent three years
as an assistant professor at the University of North Dakota. His technical
research area is in high-performance polymers.
Mariano J. Savelski is Assistant Professor of Chemical Engineering at
Rowan University. He received his BS in 1991 from the University of Buenos
Aires, his ME in 1994 from the University of Tulsa, and his PhD in 1999
from the University of Oklahoma. His technical research is in the area of
process design and optimization.
Copyright ChE Division of ASEE 2002
Chemical Engineering Education

Many properties are important in determining the overall
character, flavor, and stability of beer. They include head sta-
bility, apparent carbonation, color, specific gravity, pH, alco-
hol content, sugar content, protein content, and viscosity. In
addition, packaging properties such as material, color, fill
level, and sound-upon-opening contributed to the overall sen-
sory experience and perception of the product.
Some of these properties can be evaluated by simple ob-
servation, while others can be evaluated only by using spe-
cialized instrumentation or chemical analyses. We evaluate
the packaging material, the fill level, the sound-upon-open-
ing, and head stability, the apparent carbonation, the color,
the pH, the alcohol content, the sugar content, and the cost of
three commercial beers. We also consider the broader picture
by addressing environmental issues and recycling, econom-
ics, marketing, and ethics.
Packaging is the final stage of the brewing process and rep-
resents the consumer's first impression of the product. Beer
packaging, therefore, represents a highly competitive mar-
keting focus that requires marketing creativity and techno-
logical advancement to fulfill consumer needs and build sales
and profits. In the United States, 11% of beer packaged is on
draft, 53% in cans, and 33% in bottles.'"' Since the first trials
of putting beer into cans almost seventy years ago (Kreuger
Brewing Company, Newark, New Jersey), beer companies
have continually striven to develop innovative packaging
materials and methods. Recently, two major U.S. brewers
(Miller Brewing and Anheuser-Busch) began market testing
beer in bottles made of polyethylene terephthalate (PET).[9'
Other recent packaging innovations include unique can and
bottle shapes, foam-inducing devices, and creative labeling.110'
Glass beer bottles are manufactured in a variety of shapes,
sizes, and colors. Clear or green bottles have become a popu-
lar marketing feature, but they provide absolutely no protec-
tion against light exposure-and beer can develop a skunky
flavor within minutes of exposure to light'8' as desirable iso-
a -acid bitter substances undergo light degradation to form
3-methyl-2-butene-l-thiol (MBT). Amber, brown, and black
are the only glass colors that provide protection against light.
Plastic beer bottles were recently introduced after engineers
overcame many challenges in developing suitable materials
for this application. Some of the desired characteristics of
plastic bottles are a shelf life of 120 days, low oxygen per-
meability (<1.0 ppm), minimal loss of carbon dioxide (<15%),
heat stability during pasteurization, protection against UV
light, recyclability, and cost-effectiveness.["' Research focuses
on developing polymers and treatments that improve these
features. For instance, innovations in reducing the oxygen
permeability of the package material include incorporating
oxygen scavengers into the plastic."21 Other challenges faced
by engineers are the economics of plastic packaging and the

recyclibility of the plastic materials. PET bottles are currently
20-50% more expensive to produce than glass bottles, and
the economics of plastic bottle production will become com-
petitive with glass only for a production in excess of 100
million bottles per year."31 The amber color that is added to
the PET to provide protection against UV light contaminates
the PET, reducing its recyclability."'1
Aside from the obvious marketing opportunities provided
by the beer label, government regulations require that all al-
coholic beverage labels must include the following informa-
tion: brand name, class and type designation, commodity state-
ment, name and address, health warning statement, net con-
tents, and country of origin.11' In 1995, the United State Su-
preme Court struck down a 60-year ban on listing the al-
cohol content on beer labels, claiming that the law vio-
lated free speech rights."16
The fill level of the liquid inside the bottle is important. If
the level is lower than 1.5 inches below the cap, oxidation
may produce off flavors and the carbonation may decrease,
causing "flat" beer. If the bottle is filled higher than 1 inch
below the cap, metallic off flavors may develop from inter-
action with the metal cap.
The ability for a beer to form foam, the stability of the foam,
and the uniformity of bubbles are all very important qualities
in beer. Upon pressurization in its container, typical beer is
supersaturated with between 2.2 and 2.8 volumes of carbon
dioxide per volume of beer."71 This carbon dioxide is released
in bubbles that form by nucleation on sites such as small ir-
regularities on the surface of the glass, particles in the beer,
or gas pockets that form upon opening.
The presence of foam in beer directly and positively af-
fects the release of flavor components from the beer. There
are substances in the beer that are vital to the flavor, and some
of these substances are surface active, preferentially distrib-
uting themselves on the surface of the foam."8' Thus, it is
desirable to achieve a nice foam in the beer and for this foam
to be stable over the time it takes to drink the beer.
Certain compounds are considered "foam negative" because
of their negative effect on foam formation and stability. Some
of these compounds occur naturally in the brewing process-
for instance, some amino acids and lipids involved in the
fermention are foam negative if they remain in the final prod-
uct. In addition, several external factors can interfere with
foam stability in beer. Improper cleaning of the beer glass
can leave a foam-negative residue, as can greasy food or lip-
stick on the rim of a glass.
Size uniformity of bubbles is a desirable characteristic of
the foam because it contributes to foam stability. Pressure
inside a small bubble is greater than that inside a large bubble,
causing a small bubble to "disappear" if it contacts a larger
one. This phenomenon, called disproportionation, can be re-
duced by adding a gas of low solubility, such as nitrogen, to

Spring 2002

the beer. Guinness is an example of a beer that uses nitrogen to
achieve small bubbles of uniform size, as shown in Figure 1. The
larger bubbles in most beers appear to flow upward through the
liquid to the surface; however, the small bubbles in Guinness ap-
pear to flow downward. Researchers performed a flow simula-
tion using FLUENT to explain this phenomenon: small bubbles
(<0.05 mm diameter) succumb to a downward drag force in the
boundary layer near the glass, whereas larger bubbles have suf-
ficient buoyancy to resist this force. In the middle of the glass,
all bubbles flow upward.r191
The widget is a device used to help create a long-lasting, stable
foam in certain styles of beer that do not easily form a foam. The
first commercial use of a widget was introduced by Guinness.[20]
Guinness patented their widget design and other beer producers
have since patented their own proprietary designs. Until 1999, the
widget found inside Guinness cans was a hollow plastic pod found
in the bottom of the can of beer. After 1999, a new design was
introduced. The new widget looks similar to a ping-pong ball and
should be cheaper to manufacture than the original design. Both
designs, however, are covered by the same patent and function in
the same way. There is a small hole in the widget through which
beer can flow in or out when exposed to a pressure difference. The
can is filled with beer and pressurized with carbon dioxide. A small
quantity of liquid nitrogen is added immediately before the can is
sealed; it quickly vaporizes and increases pressure inside the con-
tainer, forcing beer and gases into the widget through the small
hole. When the can is opened, the pressure is released; the beer and
the gases in the widget are forced out through the small hole at a
very high speed, and as this stream rips through the liquid in the
can, it causes foam to form inside the can. This produces a nice,
stable foam in a beer that otherwise would not have a very good
foam. In addition, the N2/CO2 foam that forms has smaller bubbles
that make it more stable than a traditional CO2 foam. The widget
designs are shown in Figure 2.
Beers are found in a wide range of colors, from very pale straw-
colored lagers to amber or copper-colored ales, to dark, almost black
stouts. The color is determined by the malt and other solid materi-
als that are used in the brewing process. Heat-induced Maillard
reactions between sugars and amino acids occur during the killing
of malt, to produce meladonins and color pigments. Higher kiln-
ing temperatures result in darker color malt and final product.
The meladonins produced during killing have an important
impact on beer flavor.1211
The pH is a very important factor influencing the flavor of beer.
Beers are acidic, with pH values typically in the range of 4.0 to 4.5.
As pH falls below 4.0, the flavor tends to be sharper and more
acidic, and the aftertaste is dry. Above 4.6, the taste is cloying, and
a chalky aftertaste occurs. The pH also affects the stability of the
foam and the clarity of the beer. Beers with pH above 4.5 have
poorer foam stability and also tend to form haze (protein particles
that cloud the beer).
The specific gravity of the liquid is monitored throughout all the
stages of the brewing process. A change in specific gravity during

fermentation occurs when sugars are converted to alco-
hol according to the reaction

C6H1206 2CO2 +2C2H50H
The initial specific gravity prior to fermentation is
high due to the starches and sugars dissolved in the liq-
uid; the specific gravity decreases as fermentable sugar
is converted to alcohol. The difference between the ini-
tial specific gravity (before fermentation) and the final
specific gravity (of the product) readings may be used
to calculate the total alcohol content of the beer.
In the U.S., alcohol content of beer is typically given
in weight percent (%w/w), that is, grams of alcohol per
100 grams of water. In other countries, it is much more
common to give alcohol content in volume percent (%v/
v). A beer that has an alcohol content of 5% v/v has an
alcohol content of only 3.95% w/w. These percentages
are related by the densities of alcohol and water, and
the conversion can be performed by using the equation

Figure 1. The small, uniform bubbles in Guinness
are due to nitrogen.

Figure 2. The Guiness widget (1999 on left,
2001 on right).
Chemical Engineering Education

%v / v =(1)
0.789 g alcohol 100 ml water
ml alcohol 100 g water

During a three-hour laboratory period, students can rea-
sonably analyze and compare three beers. In our laboratory,
we analyzed several commercial beers, but found the follow-
ing best suited for student experiments: 1) Budwiser, an
American lager that is quite light in color and is available in
cans or brown bottles, 2) Bass Ale, a slightly darker English
ale that is available in brown bottles, and 3) Guiness Draught,
an Irish stout ale that is packaged in an aluminum can with a
widget. Guinness was chosen to provide an opportunity to
explore the widget, stable foam, and small-bubble flow; Bass
Ale and Budweiser were chosen for the availability of pub-
lished information. This section describes the methods and
results of the commercial beer analysis.
The experimental procedure includes several standard tests
used in the brewing industry, modified as necessary to be
performed in a student laboratory by individuals without spe-
cialized training. The results presented in this section are typi-
cal experimental results obtained by students in an educa-
tional setting rather than in a research laboratory or a con-
sumer testing facility. They should not be interpreted as an
endorsement of any brand name or particular product.

Packaging Prior to opening the package, students should
The material of the packaging (glass or aluminum)
The color of the container if it is glass
The hardness of the container if it is a can. Does it dent easily
if squeezed?
Any labeling information, e.g., the type of beer (ale, lager,
stout), where it is produced, alcohol content, if it is pasteur-
ized, and patent information
The distance from the cap to the liquid level
To gain an appreciation for the governmental regulations
on labeling and advertising of alcoholic beverages, students
researched the laws on labeling information as a homework
Sound-Upon-Opening The next step is to open the beer
and listen to the sound as pressure is released. Students should
describe the sound as a high pitch (indicating a high level of
carbonation) or a low pitch (indicating low carbonation).

Head Retention After opening the beer, the next step is to
pour the beer and observe the foam stability or head reten-
tion. For the case of Guinness Draught, the beer must be
poured immediately after it is opened because the widget cre-
ates a generous amount of foam that must be captured in the
head; for other beers, it is less important to work so quickly.
Spring 2002

To compare the head retention of different beers, the fol-
lowing standard pouring procedure is followed for each prod-
uct. Approximately 200 ml of the beer is poured into a 500-
ml glass beaker. The beaker should be tilted and the beer
poured steadily onto the side so that approximately 1 inch of
head (foam) develops on top of the 200 ml of liquid beer. The
foam formation should be observed; where do the bubbles
appear to come from-the top or the bottom? The size and
uniformity of the bubbles should be noted.
A good rule-of-thumb for head retention is given by Fix.J221
A one-inch head should last for five minutes without the ap-
pearance of voids (spots where the surface of the beer liquid
is not covered by foam). Students use this guideline to evalu-
ate the foam retention of the beer.

Color After pouring the beer for the foam-retention test,
there will be enough beer left in the container to proceed to
the remaining analyses. There are several methods for mea-
suring color in beer, two of which are used in this experi-
ment. The first method uses a color comparison chart, called
a Davison Color Chart (available for about $6.00 at local
homebrew shops), to match the beer color to a standard color
on the chart. The chart assigns color values from 3 degrees
Lovibond ('L) to 19 (L). The second technique is a spectro-
photometric method standardized by The American Society
of Brewing Chemists. This method, called the Standard Ref-
erence Method (SRM), measures the absorbance of light with
a wavelength of 430 mm through a sample of one-half-inch
width. The color of the beer as quantified by the SRM proce-
dure is related to absorbance by Beer's Law (named after the
scientist, not the beverage)

A CSRM Degrees SRM (0.5 in) (2)

where A is the absorbance and CSRM is the color in degrees
SRM. In the case of Guinness Draught, which is very dark in
color, it was necessary to dilute the beer by a factor of 4 prior
to spectrophotometric analysis, subsequently including this
in the SRM calculation. This was not necessary with other
beers. In addition, the standard cuvettes available in our
laboratory had a path length of 1.0 cm, and the appropri-
ate conversion factor of 1 cm/0.3937 in was applied to
the path length in Eq. (2).
Degrees Lovibond are equivalent to 10 degrees SRM, and
the results of the two methods for color analysis can be com-
pared. Student results of 2.5 L for Budweiser, 12 "L for Bass,
and 24 L for Guinness compare well with the published val-
ues provided in Table 1 (obtained using the Davison Color
Chart and taken from Fixi221).

pH The pH of the beer can be analyzed using a pH meter
or a pH test strip. The results are then compared to published
values (taken from Fix'221 and shown in Table 1).

Specific Gravity Because commercial beers were used in

this experiment, it is impossible to measure the specific gravity
of the liquid medium prior to fermentation. The final gravity
was measured and typical values of initial specific gravity
for each style of beer were used to estimate the alcohol con-
tent. The specific gravity is measured using a hydrometer,
which can be purchased from a homebrew shop for approxi-
mately $6.00. Specific gravity is typically reported at 60'F,
and measurements taken at other temperatures can be adjusted
using the temperature correction factor

CF = 2 x 10-6 T2 0.0001 T + 0.0018 (3)
where T is in 'F. This equation was obtained using a polyno-
mial fit through manufacturer-supplied, tabulated specific
gravity values (True-Brew-USA) in the range of 320F to 860F.
The conversion of sugar to alcohol during fermentation is
accompanied by a change in specific gravity as expressed by

%w / w = 105(SGinitial SGfinal) (4)
where the factor 105 is dimensionless and accounts for the
change in density of a solution as sugar is converted to alco-
hol by the reaction described above. The initial specific gravity
is estimated using the typical values for various types of beer
shown in Table 2,r23] allowing the calculation of the alcohol
content. The estimated alcohol content is then compared to
the alcohol content obtained by direct measurement (see next
section) as well as published values.
AlcoholAnalysis The alcohol content of the beer was ana-
lyzed using a YSI 2700 Biochemistry Analyzer. These re-
sults are compared to the estimated alcohol content using the
specific gravity method, as well as published values in Table
3. An alternate technique for measuring alcohol content is to
use one of the commercial enzyme test kits, such as the Etha-
nol Test Kit from Boehringer-Mannheim. These require only
a spectrophotometer for analysis, but were a little difficult
for the freshmen to use.
The Widget The final step in the product analysis is to
investigate the widget in the can of Guinness Draught. The
Guinness can should be carefully cut apart. Inside the can is
the plastic widget, which should be examined. Students should
look for the tiny laser-drilled hole from which the liquid beer
and gases rush out upon opening to induce nucleation by
mechanical shear. For homework, students read the patent
and learn more about how the widget works.
Cost Through comparison of the different commercial beers,
students gain understanding of the desirable properties that
contribute to the overall quality of the product. An important
factor closely linked with these properties is the cost. Stu-
dents can obtain cost information on commercial beers in
local newspaper advertisements and by calling local stores
that sell beer. Bottles of Budweiser and Bass Ale are sold in
6-packs of twelve-ounce containers, while Guinness Draught
cans are sold in 4-packs of 14-ounce containers. Students
obtain pricing information and calculate the unit price per

ounce of the products. Typical results (for the Southern New
Jersey area, based on a single pack) are: Budweiser, $0.063/
oz; Bass Ale, $0.111/oz; Guinness Draught, $0.116/oz.

In addition to writing a laboratory report, the homework
assignment and additional out-of-class activities include in-
vestigation of the issues that contribute to the "broader pic-
ture" of product design: patents, environmental and recycling
issues, marketing, government regulations and taxation, eco-
nomics, and ethics.
Students research the Guinness widget patent to learn more
about the features, function, and production of this device.
The patent provides detailed information on materials of con-
struction, methods of manufacture, gas solubility, dimensions,
function, and pasteurization. This information is summarized
in Table 4. Students also search for patents on proprietary
devices related to the widget.
Students consider environmental issues as they are asked
to investigate sustainability of the brewing industry. After
learning about the brewing process, students investigate top-
ics such as reducing water use, waste minimization, and re-
cycling of containers.
Government regulations regarding production, marketing
and labeling, sale, and consumption of alcoholic beverages
is another aspect of product design that must be considered.
Students are asked to research the government regulations
regarding alcoholic beverage labels, as previously described
in the Background section of this paper. The importance of
marketing is emphasized by having students present a mar-
keting plan for a new product to potential investors. Ethics is

Analysis of Commercial Beer
(Properties were compiled from several sources.)

Alcohol % by Calories/100
Color volume mi11"
Beer (Lovibond) pH1;" (by weight)
Budweiser 2.0231 -4.40 4.66 (3.60) 40
Bass Pale Ale 10.0231 3.97 4.50 (3.60) 45
Guinness ~25.0[~ 4.27 (3.42) 43

Starting Specific Gravity of Various Beer Styles
Typical values of the starting specific gravity are given by
Papazian.I"23 These are given as ranges for general types
of beer rather than for each specific brand of beer

Style of Beer Example Starting SG
American Lager Budweiser 1.035 1.045
Classic Ale Bass 1.043 1.050
Stout Guinness 1.036- 1.055

Chemical Engineering Education

also emphasized in our Freshman Clinic, and there are many
possibilities for investigation related to beer and the brewing
industry. One controversial topic for investigation is the mar-
keting of alcoholic beverages to inner-city consumers and
economically disadvantaged minorities. [24'25


Commercial beer is used as a means of introducing fresh-
men to the concept of product design. Issues relevant to prod-
uct design are addressed, including packaging, properties of
interest to consumers, patent information, and the importance
of marketing the product. Student feedback indicates that this
approach is well-received by the students and presents a first
opportunity to consider the design of a chemical engineering
product. Overall course evaluations averaged 4.6 to 5.0 for
the three years the course has run. Student comments indi-
cate that the most important things learned in this course were
teamwork, presentation skills, the interdependence of engi-
neering and marketing, and the relevance of broader issues
such as intellectual property and ethics.


The American Society of Brewing Chemists provided in-
formation on methods of chemical analysis for beer. We are
very grateful to Mark Edelson, co-owner of Iron Hill Brew-

Sample Results of Alcohol Analysis
of Three Commercial Beers

Alcohol content Alcohol content Published
estimated using using YSI alcohol
Beer SG (%w/w) analyzer (%w/w) content"I
Budweiser Lager 3.8 3.71 3.73
Bass Ale 3.7 3.55 3.60
Guinness Stout 3.5 3.38 3.42

Examples of Information
Contained in the Guinness Patent (US 4832968)

Property or Feature Details
Method of foam production Shear-induced nucleation as liquid
and gases are released through a tiny
hole in the plastic pod at a high
Tiny hole 0.061 cm diameter; laser-bored
Material Polypropylene
Manufacturing technique Blow molding
Volume of liquid in plastic pod 15 ml
Gas Mixture for pressurization N, (2% vol/vol) and CO, (150%
vol/vol), supersaturated
Pasteurization After sealing, 600C for 15-20

Spring 2002

ery and Restaurant, for giving us a fascinating tour of West
Chester, Pennsylvania, brewery. Beercrafters in Turnersville,
New Jersey, provided invaluable suggestions and advice on
the brewing process. The Bureau of Alcohol, Tobacco, and
Firearms assisted us in making sure that we complied with
federal rules and regulations regarding possession of alco-
holic beverages.

1. Cussler, E.L., "Do Changes in the Chemical Industry Imply Changes in Curricu-
lum?" Chem. Eng. Ed., 33(1) (1999)
2. Westerberg, A.W., andE. Subrahmanian, "Product Design," Comp. & Chem. Eng.,
3. Hesketh, R.P, and C.S. Slater, "Demonstration of Chemical Engineering Prin-
ciples to a Multidisciplinary Engineering Audience," Proc. 1997Ann. Conf ASEE,
Seattle, WA (1997)
4. Hesketh, R.P, K. Jahan, A.J. Marchese, T.R. Chandrupatla, R.A. Dusseau, C.S.
Slater, and J.L. Schmalzel, "Multidisciplinary Experimental Experiences in the
Freshman Clinic at Rowan University," Proc. 1997Ann. Conf. ASEE, Seattle, WA
5. Marchese, A.J., R.P Hesketh, K. Jahan, T.R. Chandrupatla, R.A. Dusseau, C.S.
Slater, and J.L. Schmalzel, "Design in the Rowan University Freshman Clinic,"
Proc. 1997 Ann. Conf ofASEE, Seattle, WA (1997)
6. Ramachandran, R.P., J.L. Schmalzel, and S. Mandayam, "Engineering Principles
of an Electric Toothbrush," Proc. 1999 Ann. Conf. ASEE, Charlotte, NC (1999)
7. Farrell, S., R.P. Hesketh, J.A. Newell, and C.S. Slater, "Introducing Freshmen to
Reverse Process Engineering and Design Through Investigation of the Brewing
Process," IJEE, in press
8. Bamforth, C., Beer: Tapping into the Art and Science of Brewing, Plenum Press,
New York, NY (1998)
9. "Plastic Bottle Gives Beer Market a Boost," Food Eng., 71(5), p. 25 (1999)
10. Lubliner, "7th Annual Beverage Packaging Global Design Awards," Bev. World,
119, p 84 (2000)
11. Kinghts, M., "High Hopes for Beer Bottles Enliven Packaging Conferences," Plas-
tics Tech., 46(1), p. 35 (2000)
12. Vogelpohl, Heinrich, "Beer Quality and Packaging. Part 1. Influence of Bottling
Materials on Beer Quality," Brauindustrie, 84(8), p. 462 (1999)
13. Sherwood, S., "Beer in Plastic Bottles," Bev. World, 119, p. 40 (2000)
14. Odubela, S., "Beer in Bottles with a Bounce," Waste-Age, 30(5), p. 82 (1999)
15. Bureau of Alcohol, Tobacco, and Firearms, Import/Export Branch, "International
Import Requirements for Various Countries for Beer, Wine, and Distilled Spirits,"
16. Hwang,-Suein-L, P.M., Court Allows Alcohol Levels on Beer Labels," Wall St. J.,
p. B-I, April 20 (1995)
17. ASBC Approved Methods ofAnalysis, 8th ed., ASBC, St. Paul, MN (1992)
18. Dale, C., C. West, J. Eade, M. Rito-Palmares, and A. Lyddiatt, "Studies on the
Physical and Compositional Changes in Collapsing Beer Foam," Chem. Eng. J.,
72, p. 83(1999)
19. "Do the Bubbles in a Glass of Guinness Beer Go Up or Down?" Fluent News,
8(2), Winter (2000)
20. U.S. Patent #4832968: Beverage package and a method of packaging a beverage
containing gas in solution, issued May 23, 1989
21. Fix, G., Principles of Brewing Science, Brewers Publications. Boulder, CO (1989)
22. Fix, G.J., and Fix, L.A., An Analysis of Brewing Techniques, Brewers Publica-
tions, Boulder, CO (1997)
23. Papazian, C., The New Complete Joy of Home Brewing, Avon Books, New York,
NY (1991)
24. "Distilling the Truth About Alcohol Ads, Marketing to Minorities-Special Re-
port," Bus. Society Re.:, 83, p. 12 (1992)
25. "Selling Sin to Blacks, Cigarettes, Fast Food, and Malt Liquor," Fortune, 124, p.
26. Beer,Alcohol, and Calories at (July 2001), origi-
nally from L. Hankin, Connecticut Agricultural Experimental Station and Excise
Division of the Connecticut Dept. of Revenue Services 0

Random Thoughts...



North Carolina State University Raleigh, NC 27695

Becoming a successful faculty member at a research
university is no trivial undertaking. People are not
born knowing how to prepare and deliver effective
lectures, make good use of the growing power of instruc-
tional technology, write rigorous but fair assignments and
exams, help students deal with a bewildering array of aca-
demic and personal problems, build a world-class research
program, manage research and teaching assistants, and bal-
ance the endless and often conflicting time demands imposed
by teaching, research, service, and personal life. It takes most
faculty members long years of trial and error to learn how to
do all that, and some never quite figure it out.
A new book-The Effective, Efficient Professor,"] by
Phillip Wankat-is a treasure trove of information on the strat-
egies, techniques, and tricks of the trade of successful fac-
ulty members. While the book applies to all disciplines, its
author is a well-known chemical engineering professor with
superb credentials in both education and disciplinary research,
and the writing reflects the pragmatic point-of-view of a
skilled engineer.
The book opens with a chapter that defines the dual themes
of effectiveness and efficiency and argues that one can have
both in a faculty career. The eleven chapters that follow are
grouped into four sections:
1. Time management Missions, goals, and activities; ap-
plying time management methods.
2. Effective and efficient teaching Teaching and learn-
ing; lecture-style classes; problem-oriented learning;
rapport with students and advising.
3. Effective, efficient students Undergraduates; graduate
students and graduate programs.
4. Scholarship and service Scholarship and writing; ser-
vice and administration; making changes.

The chapters begin with anecdotes about hypothetical pro-
fessors experiencing all-too-familiar problems, such as classes
full of students with glazed eyes and low motivation, heavy
pressures to churn out papers and proposals with little time
to think about what goes in them, and time-consuming ser-
vice responsibilities that offer neither tangible reward nor
personal satisfaction. The anecdotal professors return peri-
odically to illustrate the possible benefits of applying the sug-
gestions in the text (which-speaking of those service re-
sponsibilities-include pointers on when and how to say no).
Lists of practical tips are the heart of the book. Tips are
given on time management, motivating students to learn,
equipping them with good study and test-taking skills, in-
creasing their active involvement in lectures, teaching with
technology, implementing cooperative and problem-based
learning, getting the most out of office hours, minimizing
cheating, dealing with a large variety of student crises, moti-
vating and helping graduate students to finish writing their
dissertations, and many other topics. Most of the suggestions
are things I wish someone had told me when I started out in
this business. For example:

E Lecture preparation
On average, it should take two hours to prepare a new 1-

Richard M. Felder is Hoechst Celanese Pro-
fessor Emeritus of Chemical Engineering at
North Carolina State University He received
his BChE from City College of CUNY and his
PhD from Princeton. He is coauthor of the text
Elementary Principles of Chemical Processes
(Wiley 2000) and codirector of the ASEE Na-
tional Effective Teaching Institute

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

hour lecture on material you know, three at most; half-
an-hour to revise a 1-hour lecture; and roughly 30 hours
to prepare a 1-hour interactive web lecture. If it takes
you much more than that, you're probably overpreparing
(a common mistake made by new faculty members).

E Motivating students
Instructors can motivate students by learning and using
their names, communicating expectations clearly (for ex-
ample, by handing out lists of measurable learning ob-
jectives), and creating opportunities for early success.
Tests with average grades in the 50s or lower are seri-
ously demotivating, even if the grades are subsequently

E Avoiding overloaded syllabi
Have several professors independently analyze the con-
tent of a course and decide which topics they would not
include. Those topics may generally be considered op-

E[ Instructional technology
Use instructional technology only for tasks that are es-
sential to the course and cannot be done as well-if at
all-without it, and when the added cost of using it is
reasonable. If you decide to use it and you're not an ex-
pert, get help.

EI Cooperative learning
When assigning problems to groups, minimize single-
answer problems. Make the problems highly structured
early in the curriculum and more open-ended later.

EI Grading homework in large classes
Not all assignments need to be handed in, and not all
problems handed in need to be graded.

[J Test design
Design tests so that 15-20% of the material is covered
only in the lectures, 15-20% only in the readings, and
the remaining 60-70% in both. To make sure that a test
is not too long, work through it yourself and multiply
your solution time by five for freshmen, four forjuniors,
and three for graduate students.

E[ Office hours
Come out from behind the desk, keep candy on hand,
and consider requiring every student to come in once
early in the semester.

E Word-processing
Avoid perfectionism and endless revisions of documents
that are not particularly important.

El Asking for volunteers
People are more likely to say yes if you ask them indi-
vidually rather than addressing the request to a group.

[L E-mail
Assume that your messages are not private. Never send
a hot (angry) e-mail message.

E Reviewing papers for journals
Don't spend time rewriting or over-discussing what
you're reviewing. If you find a fatal error that invali-
dates the paper, stop there.

EL Business trips
Give yourself plenty of time to get to the airport-the
number one priority is the trip, not doing 15 more min-
utes of work.

When recommendations are based on research (as many are),
the sources are cited in exhaustive detail.
The book has an encyclopedic topical coverage and should
be consulted like a reference volume, not read like a novel.
Browse it to get ideas about implementing supplemental in-
struction, guided design, service learning, and inquiry learn-
ing, or to find out how contemporary learning theories (e.g.,
Lowman's model of effective teaching, the ABCF model for
student crisis management, Piaget's and Perry's models of
student development, Maslow's theory of student motivation,
and various learning style models) can be used to design ef-
fective instruction. Alternatively, just randomly open it and
start reading. I can guarantee that before you get through a
single page you'll find an idea that can help make you a bet-
ter professor.
In short, if you're looking for a psychological treatise on
the latest theories of student cognition and motivation, you
would probably do better elsewhere (and the reference list in
The Effective, Efficient Professor might be a good place to
start looking). But if you want a book that can provide an-
swers to your questions about teaching and learning that you
can start applying next Monday morning, this is the book for

1. Phillip C. Wankat. The Effective, Efficient Professor: Teaching, Schol-
arship, and Service. Boston. Allyn & Bacon ( 2002)

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

Spring 2002




University of Pennsylvania Philadelphia, PA 19104-6393

I claim no special expertise on creativity, discovery, or
innovation, and this paper does not purport to be a re-
view or scholarly treatise on any of those subjects.
Throughout my academic career, however, for practical as
well as philosophical reasons, I have strongly encouraged
my students to be creative in their experimentation, mod-
eling, analyses, problem-solving, and designs. This pa-
per describes techniques and subject matter that have
proved successful in that regard.
Scientific and technical articles in the archival literature,
even the most influential ones, rarely illuminate the creative
process itself, because the misdirections, irreproducible ob-
servations, false inferences, and discarded conjectures that
are common to most investigations go unmentioned. My pri-
mary sources of guidance for students have therefore been
the autobiographies and biographies of famous innovators,
wherein the "dirty linen" of their daily lives and their failures
as well as their successes are described. A second and per-
haps equally important set of sources has been the detailed
experiences of my own students and associates, from which I
am only once removed.
The creative process in chemical engineering differs some-
what from that in music, painting, literature, and even sci-
ence, but we can learn from the more extensive and better
documented experiences in those fields if we are careful to
keep the differences in mind. Also, we do not need to con-
ceive of ourselves as being on the same intellectual plane as
Beethoven, Rembrandt, Shakespeare, and Newton in order
to benefit from the study of their paths of creativity and dis-
covery. In that sense I have chosen four well-known person-
alities from science as primary examples.

Functioning in his manifestation as an artist, Leonardo da
Vinci in 1515, at the age of sixty-three, drew a sketch of him-

self watching the flow of a river past obstructions."1,2 In his
manifestation as an acute observer of natural phenomena,
Leonardo noted the chains of stationary vortices generated
immediately downstream from the obstructions, while in his
manifestation as a scientist he included in a descriptive cap-
tion a mechanistic explanation for this behavior. That sketch
and caption illustrate not only his universal genius, but also
the sometimes complementary roles of observation, graphi-
cal representation, and science.
Invention of the telescope in the Netherlands inspired
Galileo Galilei in 1609, when he was forty-five years old, to
construct a greatly improved one for himself. His early ob-
servations included discovery of the four largest moons of
Jupiter, the phases of Venus at different times of the year, and
the existence of sunspots. From the periodic disappearance
and reappearance of some of the latter, he inferred that the
sun rotated and estimated its rate. In an even greater intellec-
tual leap, he recognized that his observations of Jupiter and
Venus provided an irrefutable confirmation of the Coperi-
can theory of the solar system.
Issac Newton was only twenty-three years old in 1666 when
he conjectured that the same force that causes an apple to fall
to the earth might extend to the moon. Seeking an explana-
tion for the failure of the moon to fall led him, by means of
very intense and extended cerebration, to conceive of a mecha-
nistic description and explanation for all kinematic phenom-

Copyright ChE Division ofASEE 2002

Chemical Engineering Education

Stuart W. Churchil is the Carl VS. Patterson
Professor Emeritus at the University of Penn-
sylvania, where he has been since 1967. His
BSE degrees (in Chemical Engineering and
Math), MSE, and PhD were all obtained at
the University of Michigan, where he also
taught from 1950-1967. Since his formal re-
tirement in 1990, he has continued to teach
and to carry out research on turbulent flow
and heat transfer and combustion.

ena. The story of the apple may be apocryphal, but it origi-
nated from Newton himself.
In 1928, Alexander Fleming (at the age of forty-seven) was
probably not the first to observe the destruction of bacteria in
the laboratory by a contaminant, but he had enough percep-
tion and initiative to identify the agent in this instance as peni-
cillium rubrum and to successfully explore its potential as a
therapeutic agent. This led others to pursue the production of
an antibiotic drug.
The recurrent pattern in these four episodes is the recogni-
tion of anomalous behavior by a perceptive observer and the
persistent intellectual pursuit not only of an explanation, but
also of the possible consequences of that explanation. This is
the most important commonality of discovery and innova-
tion in the physical sciences and engineering. We can, how-
ever, learn many other lessons concerning the process of in-
novation from the experiences of these and other recognized
masters of the arts and sciences.

The common and salient characteristics of the great inno-
vators provide useful insight and guidance for would-be in-
novators. Some such characteristics and circumstances are
identified in the following paragraphs.
U Resilience and Self-Confidence
Most discoveries and new ideas are greeted with skepti-
cism, misunderstanding, lack of appreciation, or outright re-
jection. The writings of the great innovators reveal that they
all encountered such reactions but had sufficient self-con-
fidence to persist.
For example, the opening lines of Sonnet LV "Not marble
not the gilded monuments of princes, shall outlive this pow-
erful rhyme," demonstrates that Shakespeare knew that he
was not just another poet and playwright, and indeed was not
inferior to the royalty or the wealthy in true worth.
Beethoven's own pupil, Czerny, neither understood nor ap-
preciated the sublime music of his final period, saying,
"Beethoven's third style dates from the time when he became
gradually completely deaf.... Thence comes the dissimilarity
of style of his last three sonatas.... Thence many harmonic
roughnesses...." But Beethoven, in 1817 at the age of forty-
seven, is reported to have said of this same period, "Now I
know how to compose."
Rossini clearly understood his place in the musical hierar-
chy, saying "I know I am not Bach, but I also know I am not
The trilogy, Joseph and His Brothers, by Thomas Mann[3'
is an inspirational study of the constructive behavior of a soli-
tary genius surrounded all of his life by people whom he knew
to be intellectually and morally inferior.
When Gladstone, then Chancellor of the Exchequer, inter-
Spring 2002

rupted a description by Faraday of his work on the then-new
subject of electricity with the impatient inquiry, "But, after
all, what use is it?", the latter is reported to have responded
with, "Why sir, there is every probability that you will soon
be able to tax it."
Galileo recanted before the Inquisition in order to save his
life, but he never stopped trying to educate the leaders of the
Church and he never lost confidence in the ultimate rec-
ognition and acceptance of his findings and conclusions
by his peers in science.
Newton was perhaps more fully recognized and appreci-
ated for his scientific accomplishments in his own time than
anyone except possibly Einstein in his time, but even so he
was virtually paranoiac concerning the rejection of his find-
ings or the perceived usurpation of credit for them by others.
On the other hand, he never questioned his own intellectual
superiority or the significance of his contributions, and in-
deed finally produced his Principia'4J to remove any doubt
about that for all time.
Lord Kelvin is reported to have told an incredulous Lord
Rayleigh that as his predecessor as President of the Royal
Society he had rejected for publication the now-famous pa-
per by Josiah Willard Gibbs, "On the Equilibrium of Hetero-
geneous Substances," because the phase rule that it intro-
duced was too simple to be correct or significant. This rejec-
tion led to its publication in the obscure Transactions of the
Connecticut Academy of Arts and Sciences. But Gibbs him-
self never doubted the significance of this work, as is evident
from his subsequent submission of a reprint to virtually ev-
ery famous scientist in the world and his reciprocated corre-
spondence with many of them.
These experiences suggest that a would-be innovator must
have sufficient self-confidence and resilience to persist in the
face of skepticism and rejection.

] Patience and Refinement
Leonard Bernstein demonstrated vividly in the early tele-
vision program Omnibus that Beethoven composed his Fifth
Symphony, not in an explosion of inspiration but rather by
incessant revision and refinement.
Newton conceived of his mechanics in 1664-1666 in a great
burst of creativity, but eighteen years of incubation passed
before he was provoked by the threat of loss of priority to
publish this work. Even then, three more years of intense
mental labor were required to correct, complete, and update
these ideas for the Principia.
Seventeen years were required for the critical observation
of Fleming to be translated into the first treatment of a hu-
man patient with penicillin, and that period of time was
undoubtedly shortened by the urgency and high priority
imposed by World War II.
Subrahmanyan Chandrasekhar, whom I was privileged to

know (and whose book on creativity, Truth and Beauty: Aes-
thetics and Motivations in Science,15 has been singularly help-
ful in formulating this paper) encountered so much hostility
from his mentor Eddington and others for his theories on black
holes that he abandoned the subject for other aspects of as-
tronomy. But when he received the Nobel Prize forty-some
years later in 1983, at the age of seventy-three, it was in part
for that early, now-accepted work on cold stars.
The lesson here is that an innovator should not expect im-
mediate acceptance of his initial discovery. Rather, he should
be prepared to be patient and willing to persist, even if years
of further work in the sense of refinement and confirmation
prove necessary.
] Age and Creativity
The opinion that all important discoveries are made at a
relatively young age is widely held among mathematicians
and physicists. For example, G.H. Hardy'61 in A
Mathematician's Apology, an essay said by C.P. Snow to be
"The most beautiful statement of the creative mind ever writ-
ten or ever likely to be written," asserts that "No mathemati-
cian should ever allow himself to forget that mathematics,
more than any other art or science, is a young man's game....
Galois died at twenty-one, Abel at twenty-seven, Ramanujan
at thirty-three, Riemann at forty. There have been men who
have done great work a good deal later,... [but] I do not know
an instance of a major mathematical advance initiated by a
man past fifty.... A mathematician may still be competent
enough at sixty, but it is useless to expect him to have origi-
nal ideas." He further says, quite unkindly, of his own, far
greater protege, "The real tragedy about Ramanujan was not
his early death. It is, of course, a disaster that any great man
should die young; but a mathematician is comparatively old
at thirty, and his death may be less of a catastrophe than it
seems." For someone who criticized some of Ramanujan's
proofs for their lack of rigor, this is a strange conclusion.
What evidence is there that Galois, Abel, Ramanujan, and
Riemann would not have continued to be creative if they had
lived for a longer span?
The inclusion of ages in the preceding paragraphs and the
focus on age here has the objective of throwing light on the
possible productive span of creativity for engineers. No one
would seriously assert, in the face of overwhelming evidence
to the contrary, that creativity in painting, music, and litera-
ture is limited to the very young, but the evidence in science
is somewhat contradictory.
Newton is often cited as the prime example of a scientist
who did all of his greatest creative work while very young.
Indeed, he did first conceive of his greatest contributions in
mechanics, optics, and calculus at a very young age. But, he
greatly improved and extended this work at the age of forty-
five and demonstrated his unique mathematical acuity a de-
cade later at the age of fifty-five when provoked by a chal-
lenge concocted by Leibnitz and Johann Bernoulli. Although

Newton submitted his solution to their test problem anony-
mously, Bernoulli commented upon receiving it that "tanquam
ex ungue leonem," or, loosely, that "the lion may be recog-
nized by his paw print." Newton's celebrated hiatus from sci-
ence and mathematics at the age of thirty-three was not re-
ally due to his advancing years, but rather to his greater inter-
ests in religious history and alchemy. He subsequently wel-
comed the opportunity to leave Cambridge University and
become Warden of the Mint because of the greatly reduced
danger of his exposure and persecution as a religious heretic.
Thomas Huxley, a famous contemporary of Darwin, as-
serted that "A man of science beyond sixty does more harm
than good," even though the latter was sixty-two when he
published The Descent ofMan. Perhaps Huxley did not count
the period of reduction of ideas to print. When Lord Rayleigh,
at the age of sixty-seven and still active, was asked by his
own son to comment on this statement by Huxley, he replied,
"That may be, if he undertakes to criticize the work of younger
men, but I do not see why it need be so if he sticks to things
he is conversant with." Rayleigh's own work supports this
opinion; in a memorial lecture upon his interment in
Westminster Abbey, J.J. Thomson emphasized the uni-
formly high quality of his creative work up to his death at
the age of seventy-seven.
The span of creativity of engineers is perhaps known with
even less certainty than that of scientists and mathematicians,
but is presumably not so short as to discourage us from try-
ing to develop an innovative outlook by our students.

U Concentration and Freedom from Distraction
The power and exercise of concentration is an aspect of
creativity that is sometimes overlooked. An ability and will-
ingness to focus single-mindedly on a narrow topic for an
extended period of time has often been cited as an essential
attribute of Newton. It is probably not a coincidence that his
anni mirabiles overlapped his hiatus from Cambridge owing
to the threat of the plague. Again, when completing the math-
ematical components of Principia some years later, New-
ton went days with almost no food or sleep. An unwill-
ingness to continue to make such a commitment and the
related sacrifices with increasing age and acquired social
obligations may be an united factor in the context of the
previous subsection.
The loss of hearing and the virtual loss of human compan-
ionship by Beethoven may have been essential to his final
greatest burst of creativity.
The self-portrait of Leonardo mentioned earlier implies the
leisure to concentrate mentally on a single aspect of nature.
Although such extreme commitments as that of Newton,
such trauma as that of Beethoven, and such relative freedom
as that of Leonardo are not necessarily a prerequisite for cre-
ativity, it is not unusual for most of us lesser mortals to have
our best new ideas when we are temporarily free from the
Chemical Engineering Education

distractions of our everyday life-for example when we take
a long solitary walk, awaken in the middle of the night, or
daydream at a symphony concert.
U Interactions and Challenges
Despite the popular image of the solitary lonely genius,
interaction with one's peers as conferees or collaborators or
even as competitors, often plays an important role in innova-
tion. Again, Newton serves as a prime example. Although he
protested bitterly over his perceived harassment by Hooke,
Leibnitz, and others, had he not been provoked and challenged
by them over priorities, and had he not been urged and as-
sisted by Halley, he might never have completed or pub-
lished his work. Although Newton rarely gave any public
credit to his associates and correspondents, he tested his
ideas on them and pestered them for their own deriva-
tions and experimental data.
Mozart was certainly spurred in his own operatic composi-
tions by the competition and greater popularity of Gluck and
The implication is that innovators are apt to benefit from
interactions, challenges, and competition, and should seek
rather than avoid them.

U Fallibility
Even the greatest geniuses have proven to be fallible. For
example, Leonardo sketched symmetrical pairs of vortices
instead of the antisymmetrical ones that are now known to be
formed. Newton made countless minor errors in his zeal to
explain and model all physical phenomena. For example, he
derived an erroneous expression for the velocity of sound in
gases because of the premise that the behavior is isothermal.
Lord Kelvin estimated the age of the earth by thermal model-
ing, but was in error by several orders of magnitude (thereby
appearing to contradict the then-new theory of evolution)
because of the neglect of heating by radioactive decay, ne-
glect of the effect of pressure on the melting point of the
magma, and several other simplifications.
These examples of fallibility by truly great men illustrate
two fundamentally different sources of error. That of Leonardo
is simply one of misobservation. Those of Newton and Kelvin
were, on the other hand, the result of incomplete models; the
concept of isentropy and the existence of radioactive decay
had yet to be discovered. The latter examples provide a warn-
ing that is still valid today...predictions based on a model are
no more reliable than the model (or, in the jargon of comput-
ing-garbage in, garbage out). They also suggest a revived
opportunity for innovation when newly discovered phenom-
ena are incorporated in old models.

Acknowledgment of Error
Progress in science and engineering occurs primarily by
replacement of the old with the new and improved; that is, by
innovation. But resistance to change is deep-seated in human
Spring 2002

nature. Sometimes that resistance has religious or philosophi-
cal roots. Nietzsche has said, "Convictions are more danger-
ous foes of truth than lies." Sometimes resistance is visceral;
it is painful to have to replace knowledge acquired only after
long and arduous study. The greatest resistance to scientific
innovation, however, often comes from those whose cher-
ished contributions are thereby consigned to the dustbin
of history. The resistance may then be purely defensive
and less than objective.
Newton serves as a bad example in this respect. When his
prediction of the velocity of sound did not agree with experi-
mental measurements, he inexcusably manipulated the data
in order to produce conformity.
Acknowledgment of error by one's self as well as by one's
icons is often the first step to further innovation.

U Simplification
Considerable understanding of the most complex concepts
of science may often be achieved by means of simplifica-
tions, analogies, and rationalizations, even though their origi-
nal derivations followed a much more complex path. For ex-
ample, the proportionality of energy to mass in the most fa-
mous expression of Einstein is an obvious necessity. It fol-
lows that the proportionality constant must have the dimen-
sions of velocity squared. It is then a reasonable conjecture
that this velocity is that of light. Similarly, Planck's equation
for the spectral distribution of radiation may be recognized
as the simplest one that reduced to the previously known
asymptotes for short and long wavelengths.
It may also be inferred that complex problems in engineer-
ing, such as the behavior of an automobile engine, may be
most easily understood qualitatively and quantitatively if they
are reduced to their component parts for asymptotic condi-
tions or special cases. Skill in simplification-that is, in iden-
tifying and modeling the most important factors while elimi-
nating the secondary ones tentatively or temporarily-is a
common characteristic of successful innovators. Newton rec-
ognized the importance of three-body interactions, but real-
ized that he had no chance of solving them until he had mas-
tered two-body interactions.

U The Prepared Mind
Leonardo's experienced eye as an artist assisted him in his
scientific observations and designs.
Although Newton was relatively unschooled in mathemat-
ics and science when he came to Cambridge, part of his ge-
nius is reflected in his recognition of the need to acquire a
knowledge of these subjects extending to their very frontiers,
in his willingness to make the corresponding commitment
and effort, and of course in his accomplishment of this goal
in an incredibly short time.
Fleming was prepared for his discovery of penicillin and
for its internal application by his experiences in treating in-

fected wounds in World War I and his recognition, even then,
that bacteria could hide in the edges of a wound and thereby
resist external treatment.
Recognition of an anomaly implies knowledge of and an
expectation of somewhat simpler behavior. The explanation
of an anomaly in engineering often requires knowledge of
particular aspects of mathematics and science and/or of ex-
perimental techniques beyond those required for the origi-
nally anticipated behavior.

Finally, let us turn to teaching innovation and other forms
of creativity in the process of guiding research. Looking back
over my academic career reveals that my largely intuitive
efforts in this respect have been surprisingly successful. Over
eighty percent of my research students, both undergraduate
and graduate, have made identifiable innovations or signifi-
cant discoveries in methodology or results. These accomplish-
ments are mentioned because innovation in the sense consid-
ered herein is welcome, but not required, in doctoral work; a
contribution to knowledge may be new, meaningful, and sig-
nificant without necessarily involving innovation.
Whatever success I have enjoyed in motivating my stu-
dents may have been in large part a fortuitous consequence
of my predilection for exploratory research and of my insis-
tence on a simultaneous combination of experimental and
theoretical work. A third, more subtle factor has been a
persistent effort to convince students that they are capable of
innovation and that they can afford to take risks while within
the relatively sheltered academic environment. These
characteristics of my own work are cited because of their pos-
sible relevance to the subject at hand-not because they nec-
essarily have any special merit in the greater scheme of things.

Exploratory Research
Exploratory research is here defined as an open-ended prob-
lem for which the behavior to be determined is unknown,
perhaps even in a gross sense. A further characteristic of ex-
ploratory research is the freedom and willingness to aban-
don, at least temporarily and tentatively, the initial objective
in order to pursue the explanation of an anomaly and to specu-
late on its possible consequences. Anomalies are more likely
to be observed in open-ended problems, and students are then
more likely to be on the alert for them.
The distinction between exploratory and more-narrowly
constrained research did not arise with Leonardo, Galileo,
Newton, and Fleming and does not with most current scien-
tific research. It is, however, often an important distinction
and inhibiting factor in industrial research because of con-
siderations of time, cost, and risk, and even in academic re-
search in engineering because of the conservatism of the spon-
soring agencies and their almost exclusive favoritism to a

few currently anointed topics of the moment.
Those doing exploratory research often encounter an ob-
stacle that did not exist or was less formidable in the past.
The diversion to a new objective in midstream often requires
utilization of topics in mathematics and science beyond those
encompassed by the original objective. Doctoral students are
nowadays generally discouraged by their advisor and aca-
demic department from taking any advanced course work that
is not viewed as directly relevant to their preplanned research.
At the time of recognizing the need for such specific extended
learning, it is usually impractical to undertake the appropri-
ate formal course work even if it exists. This imposes a
serious burden of self-study that is not always pursued.
The guidance, encouragement, and patience of the advi-
sor is critical at this point.
Opportunities for Exploratory Research
Discoveries beget further discoveries. New developments
in mathematics and science suggest improvements in engi-
neering. New and improved materials, new and improved
devices, and new societal concerns provide opportunities,
motivations, and incentive for exploratory research and
thereby innovation. For example, the research of my students
has been stimulated and supported in part by concerns with
such then-current topics as nuclear weapons, nuclear reac-
tors, accidental chemical detonations, jet-engine noise, igni-
tion of solid propellants, storage and transport of cryogenic
fluids, fluid-mechanical behavior in space flight, reduction
of air pollution from combustion, incineration of toxic sub-
stances in airplanes and hospital rooms, improvement of so-
lar collectors, more efficient heating of working and living
spaces, the Strategic Defense Initiative, enhanced rates of
steam generation, controlled extrusion of Plexiglas, and the
growth of improved silicon crystals. A practical motivation
of current societal interest is usually inspiring to engineering
students because it provides a sense of relevance without nec-
essarily restricting the freedom to explore innovative approaches.

The combined improvement of computer hardware and
software has greatly impacted our ability to solve complex
models numerically. For example, the development of direct
numerical simulation has stimulated a new interest in turbu-
lence, while methods for sensitivity analysis and methods for
solving the sets of stiff differential equations that describe
free-radical chemistry have greatly abetted the modeling of
combustion. The development of lasers and spectrophotom-
eters has greatly improved our ability to make experimental
determinations of all sorts. It follows that students undertak-
ing exploratory research must be alert to and if appropriate,
master, new developments in contiguous fields. They cannot
and should not depend wholly on their advisor in this respect.

The Synergy of Experimental
and Theoretical Work
The advantage of a combination of experimental and theo-
Chemical Engineering Education

retical work was recognized by Newton who, according to
Chandrasekhar,PS] said (rather awkwardly to modern ears),
"For the best and safest method of philosophizing seems to
be, first to enquire diligently into the properties of things,
and of establishing those properties by experiments, and then
to proceed more slowly to hypotheses for the explanation of
them. For hypotheses should be subservient only in explain-
ing the properties of things, but not assumed in determining
them; unless so far as they may furnish experiments...".
In the past, unexpected behavior was most often identified
from experimental measurements, but now, because of the
increased capability for solving mathematical models numeri-
cally, previously unobserved or unrecognized behavior is of-
ten predicted. For example, in our own work, multiple sta-
tionary states in thermally stabilized combustion and a finite
time of induction for the onset of thermally generated sound
waves were both first identified from numerical solutions and
subsequently confirmed experimentally.
Students often resist a commitment to both experimental
and theoretical work because of a personal predilection, but
more often, in truth, because of their lack of experience and/
or confidence in doing one or the other. They invariably end
up most proud of their work in the resisted category. Their
opportunities and capabilities careerwise are obviously
enhanced thereby.
Guidelines for Innovation
Students are not ordinarily inspired by a detailed prescrip-
tion or discussion of how to innovate, and are either intimi-
dated or amused if told that they should emulate universally
recognized geniuses such as Leonardo, Galileo, and Newton.
On the other hand, they respond very positively to the anec-
dotal experiences described above, which emphasize the in-
fluence of everyday human factors and foibles on the lives
and work of the great ones. I do not present such material in
lecture form, but rather on an ad hoc basis when appropriate
and relevant, and then only informally during individual or
group discussions.
Establishing the Proper Environment
for Innovation
Innovation usually involves some courage and risk. In or-
der to be willing to take such risks, students must sense that
their ideas, however incomplete, unrealistic, or naive, are
welcome and will be given fair consideration. Criticism from
their peers in small informal groups, such as the weekly gath-
erings of all my research students, is more easily accepted
than from their advisor, and particularly so when it becomes
a normal procedure. Surprisingly, students who are working
on quite different topics often make very constructive and
even innovative suggestions in that format. Interaction with
other students who are clearly doing innovative work is both
encouraging and challenging.
Students should be expected to justify their new concepts
Spring 2002

or interpretations, at least after some time for incubation,
but a defensive posture on their part is to be avoided if
possible. One of the most delicate tasks of an academic
advisor is to redirect the efforts of a student from a blind
alley or unproductive path.
Presentations by my students at departmental seminars have
engendered one surprising, but perhaps significant, response.
On several occasions, other students have remarked that, be-
cause of the exploratory nature of the research and the focus
on innovation, "your students have more fun than the rest of
us." The joy and satisfaction in doing innovative work is not
to be underestimated. Such experiences may have a career-
long positive influence.
In addition to exposing their innovative work for recogni-
tion and criticism, presentations by doctoral students at pro-
fessional society meetings are of critical importance in terms
of raising their self-confidence. The implicit acceptance of
the successful performance of innovative research at the fron-
tier of their field provides a great boost in that respect at a
critical time in their career.
Association with the Immortals
New findings, either experimental or analytical, often call
for the extension, correction, or displacement of some aspect
of the work of the great scientists and engineers of the past.
At first, this is somewhat frightening. On the other hand, the
psychological rewards of success in this respect are immea-
surable. Such experiences by my students include success-
fully challenging the advice of G.K. Batchelor, disproving a
theoretical expression of Einstein, displacing results of
Rayleigh, Boussinesq, Prandtl, von Kirmin, Colburn,
Spalding, and Zel'dovich, correcting the model of Fourier
for transient conduction, and extending solutions of Birkhoff,
Debye, Schwarzschild, and Chandrasekhar.
Reviews and Rebuttals
Apart from appropriate criticisms and challenges, innova-
tive results sometimes engender an apoplectic response from
a reviewer whose work is being corrected or displaced. In
addition, physicists are sometimes enraged by the audacity
of an engineer who even attempts to correct or displace the
work of their icons. On the other hand, the famous scientists
themselves with whom we have been privileged to interact
on a personal basis, including George Uhlenbeck, S.
Chandrasekhar, Peter Debye, and John von Neumann, have
invariably welcomed and encouraged our attempts to extend
their own earlier work.

Detailed Examples
Reviews of the research of my students and associates in
the context of innovation have previously been published in
two categories: theoretically stabilized combustion'7' and heat
Continued on page 127.

a1 laboratory



In the Unit Operations Laboratory

King Saud University Riyadh 11421, Saudi Arabia

he unit operations laboratory (UOL) course is aimed
at exploring previously learned knowledge, acquir-
ing new knowledge by practice, and developing vari-
ous skills and attitudes. To successfully complete this course,
students are expected to acquire, in an integrated approach, a
variety of skills, such as experimentation, communication,
instrumentation, mathematical modeling and statistical analy-
sis, troubleshooting, startup and shutdown, safety, and main-
tenance. These and other objectives have been reviewed ex-
tensively in the literature."12] Developing and improving think-
ing skills in the UOL is another objective that has been con-
sidered recently.[3,41 Although some of these objectives have
been elaborated upon,'5-71 others (such as troubleshooting"'S
and maintenance) still need further study. The objective of
this attempt is to bridge the gap between the expectations of
industry and the education our graduates receive. Our ap-
proach is to expose students to some reality and allow them
to experience and solve real problems.
This paper shows how to develop troubleshooting skills in
the UOL. It presents the concepts of problem solving and
troubleshooting, the approaches to troubleshooting in indus-
try, and how to acquire troubleshooting skills in the UOL. A
strategy for troubleshooting is developed that is based on
understanding thinking skills, problem-solving heuristics, and
the approaches to troubleshoot in industry. A program is used

Aziz M. Abu-khalaf has educational inter-
ests that include developing new objectives
and improving the performance of laborato-
ries at the Chemical Engineering Department
at King Saud University. Research interests
include controlled release systems and cor-
rosion. He can be reached by e-mail at

\ ^______

to develop these skills in an integrated approach with other
skills. This paper emphasizes the troubleshooting part of the
program, the details of which are presented elsewhere.14,7
Students are expected to develop and improve trouble-
shooting skills by practicing, monitoring their actions,
considering feedback, and reflecting to check the effec-
tiveness of the method.

A problem is a difficulty that is viewed as a gap between
the present state and a desired state or a conflict between
what is observed and what is expected. Our goal is to over-
come the difficulty by deciding its cause, knowing what op-
erations are required to reach the desired goal, and how to
correct the situation. This is problem solving.[9-12] It requires
exercising the mind in every step of its sequence and being
skillful at both critical and creative thinking.
Troubleshooting is the ability to solve problems related to
the processes, the equipment, and the environment in order
to restore normal conditions. Typical troubles of this kind
indicate inadequate performance of equipment and processes
and the inability to meet specifications and standards. This is
usually reflected by the operating conditions and the product
quality. Restoring normal conditions usually requires that
corrective action be taken immediately, safely, and with
minimum cost.[13'14] Obviously, troubleshooting is prob-
lem solving and hence, is thinking. By understanding and
practicing thinking techniques and tools, problem solv-
ing heuristics, and how experts troubleshoot, students can
develop troubleshooting skills and become much more
efficient troubleshooters. Regular exercising in the UOL
can accomplish sharpening and upgrading troubleshoot-
ing skills.

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

Troubles in industry can be attributed to operators, equip-
ment, processes, and the environment. Sometimes these
troubles are obvious; at other times they are hidden. It is usu-
ally assumed that operators are highly proficient, knowledge-
able, and familiar with the equipment' operation and limita-
tions. Misoperation, false alarms, equipment and chemical
failure, inadequate equipment design, and process failure can
all cause troubles of this kind. Sometimes it is necessary to
deal with a process that has never worked, a process that de-
viates from the expected, or a situation that requires a change
in the capacity. The objective is to locate the problem, find
the cause of the problem, and make repairs, usually with
prompt action and quick feedback (see Table 1). A trouble-
shooter is a specialist who is called in when all other mea-
sures have failed.

Troubleshooting skills can be developed in an integrated
approach with other skills by following a program[4' that is
aimed at fostering the right attitude, acquiring and practicing
various thinking techniques and tools, and recognizing and
avoiding mental errors. This program considers the role of

Troubleshooting Steps Followed in Industry

Being aware of the problem Symptoms and deviations from
normal conditions indicate the presence of a problem. They are
determined by real observations and measurements obtained from
the control room and from the field.
Clarifying the situation by developing information-gathering
and communication procedures Organizing a multifunctional
team composed of engineering, research and development, and
plant representatives to conduct an around-the-clock information-
gathering effort; using visual aids such as videotapes, photos, and
charts; hearing personal testimony from operators; monitoring the
operation for a certain amount time to define the actual perfor-
mance and conducting empirically designed experiments to define
the expected performance; knowing the characteristics of the
chemicals being processed and the equipment being used for
processing and using the process know-how report, the process
design report, and the detailed engineering report.
Lookingfor problem causes and various remedies Considering
the use of formal diagnostic tools, e.g., comparison, exclusion,
substitution, and identification of evidence; using common sense
and engineering judgment; considering various levels of
sophisticated computer simulations or statistical design methods;
using the case-based expert system in which enough observations
are gathered until only one cause explains all symptoms.
Correcting the situation Implementing the best solution that is
expected to restore normal operations.
Checking the results for normal operation Withdrawing
samples, testing them, collecting instrument readings on a regular
basis, and checking them against standard data.

Spring 2002

the instructor and students, emphasizes inquiry-oriented and
reflective activities, and maintains continuous interaction and
immediate feedback from the instructor. Students practice
various activities in a cooperative way, learn to think for them-
selves, acquire various skills, and understand in a way that
makes content a permanent acquisition.
The program used for this purpose should provide an envi-
ronment that simulates industrial work and facilitates the in-
formation-gathering process. The following considerations,
which are related to troubleshooting, need to be emphasized:
> Providing a thorough understanding of the pro-
cesses and equipment-how they work or how they
are supposed to work. Students need to befre-
quently referred to their texts and other sources.
Developing engineering and common sense by
acquainting students with the equipment, the
processes, and the available systems. Students
should spend enough time in the laboratory working
on the available equipment; they should feel and
live in the practical environment.
Understanding and practicing safety procedures
and regulations and touring the facilities. Using a
program that is based on practicing the role of the
safety officer the safety committee members, and
performing safety assignmentst5'71 has been found to
work best.
> Facilitating access to historical data, calibration
data, maintenance and troubleshooting records,
specifications of the experiments, the flowcharts of
the equipment and processes, and the operating
manuals and testing procedures.
Emphasizing skills related to troubleshooting, such
as listening to the technicians, operators, and

One approach for tackling closed-ended problems or exer-
cises is to consider what is required from the problem, step
back to explore the required information, and apply the perti-
nent tools and techniques required to solve the problem. An-
other approach is to explore the problem and available infor-
mation first, followed by answering the questions. Open-
ended problems, on the other hand, require setting up a strat-
egy to get started, determining the direction and course of
action, monitoring progress, and choosing the best solution
among alternatives that satisfy the required goals.
Since troubleshooting is problem solving and problem solv-
ing is thinking, it would be appropriate and useful to develop
a strategy that is based on thinking techniques and tools, prob-
lem-solving heuristics, 9-1'222-25] and industrial methods of
troubleshooting. Thinking tools are related to the perception
of a situation, the related information, processing, and think-

ing errors. A heuistic is a series of steps that guides us toward
the solution by determining the starting point, the direction,
and the course of action. Troubleshooting in industry is ap-
proached by collecting data, thinking of the problem to iden-
tify causes and solutions, correcting the operations, and check-
ing the results for normal operation.
For this strategy to be a success, the student must foster the
right attitude and maintain it during the troubleshooting pro-
cess, must recognize and avoid thinking and troubleshooting
errors, and must offer feedback and reflect on each step of
the sequence. Asking questions proved to be an effective
method to explore knowledge and alternatives in order to ar-
rive at the best solution. The steps of our proposed trouble-
shooting strategy are:

Identify the problem (recognize the symptoms and
arrive at the causes by exploring the situation and
the pertinent information): feel and recognize
difficulties, gather information (symptoms, deviations,
data) and pertinent knowledge (definitions, theories,
principles); explore this information and knowledge
to reveal patterns and to determine what is missing
and what is extraneous (present in a convenient form,
analyze, ask insightful questions); collect missing
information (search and research); talk about the
situation and listen to others to know what happened,
when it happened, and how it is compared to previous
conditions; check the timing, the degree of urgency,
and flexibility of specifications; think of possible
causes and screen them using critical thinking tools
and the elimination technique.
I Set goals and strategies to generate alternative
solutions Recall or learn pertinent theory and
principles; determine if the problem should be
resolved, or just live with the situation as it is; recall
similar problems (if a solution is available, implement
it); apply pertinent thinking techniques and tools such
as: analyzing; synthesizing; seeing patterns; using
analogy; predicting using rules and laws; brainstorm-
ing; breaking methods, definitions, and assumptions;
eliminating by substitution; and restating the problem
in different ways.
Check the attitude Maintain the will, the confidence,
and the belief that the problem can be resolved; be
ready to change goals and plans; leave the problem
for a while or ask for help (depends on the timing and
urgency of the problem); be aware of and avoid
thinking errors (thinking blocks). One should consider
reviewing his/her attitude during and after each step
of the strategy.
I Choose and implement the best remedy Decide the
appropriate solution based on technical, economical,
and safety considerations; implement the solution

gradually, allowing enough time for adjustments to
take place; collect new data from tests and readings
from instruments; compare these with expected
conditions and specifications.
0 Evaluate the effectiveness of the remedy Have
normal situations been restored? Has the real
problem been resolved, or is it the symptom that has
been treated? Have the criteria and constraints been
satisfied? Does the solution cause other problems?
(Are there any side effects?)
Reflect on the procedure and the key factors What
thinking techniques and tools were applied? How was
the solution started? Was it fast or slow? What were
the crucial factors (safety, economics, etc.)? How do
you classify this problem? What do you think about
the steps you followed? What suggestions do you


The UOL provides a variety of problems: straightforward
and more difficult ones, naturally or deliberately occurring
problems, and those given in the form of assignments (e.g.,
accident investigation). Sometimes, incorrectly worded prob-
lems are given to alert students to thinking errors. Problems
occur during normal operation and during startup and shut-
down, and can be regarded generally as errors in design or
malfunction, e.g., impurities, leakage, streams off specifica-
tions, inefficient cooling or heating, etc.
The given assignments should be performed during the lab
session and within the allotted time, and could include the
1. Safety assignments such as accident investigations,
root-cause analysis, and reducing the level of
operation equipment noise.
2. Retrofit and maintenance assignments such as
modifying a batch reactor to be operated as a CSTR,
replacing corroded steel pipes with PVC pipes, and
replacing a manually operated level indicator by a
gage glass.1261
3. Startup and shutdown tasks; for example, how to mix
the reactants, which facility (mixer heater measure-
ment) to start with in order to start up the reactor,1271
and how to drain the reactor
4. Treating the errors of others. While operating the
cooling tower, leakage was detected. A loose screw
was tightened and leakage stopped for a while, but
resumed later. Thorough investigation showed that
the float was disassembled for corrosion checkup,
cleaned, painted, and reassembled. The paint
increased the weight of the float, and that was the
cause of the leakage. This group reversed the above

Chemical Engineering Education

Practical Example of Troubleshooting
A Drv Wick in the Cooling Tower

E[ Feeling the problem After starting up the cooling tower,
measurements showed that the dry-bulb and wet-bulb tempera-
tures were the same for a long time-an abnormal condition, at
least for incoming air. Occasionally, the instructor needs to direct
students' attention to this situation.
El Exploring the situation The problem should be solved quickly,
within the allotted session time. The situation was discussed with
the group that operated the tower during the last session, with the
technicians, and with the instructor, and they all confirmed that
the situation was normal during that session. Unfortunately, no
historical record was available.
El Preliminary action More measurements were collected and
compared with previous data. This was necessary for confirming
available data.
E[ Exploring the data Data followed the same trend-no change
was detected.
[I Thinking of the causes A logical move was to suspect the
source of data, so an independent measurement was taken and
the measuring equipment and the controlling mechanism were
checked, but all were found to be in order. Another suggestion
was to stop the fan and check the measurements, but this was
quickly discarded because it would interrupt the operation. The
next step was to consider the symptom itself. The students asked
the following questions: What was wrong with this situation ?
Why was it not normal to have a wet-bulb temperature that
equals the dry-bulb temperature? What is the wet-bulb
temperature and how is it actually measured? The wet-bulb
temperature depends on the dry-bulb temperature and air
humidity; therefore, if the air is saturated, the wet-bulb and the
dry-bulb temperatures will be the same. 2" This generated two
ideas: first, the air was saturated, and therefore the two
temperatures were equal. Second, students had to check the
mechanism of measuring wet-bulb temperature. The independent
measurement of the wet- and dry-bulb temperatures, combined
with hot weather, contrasted the first argument. After checking
the thermocouples, the student were able to identify the problem:
the wick was dry. Therefore, the dry-bulb and the wet-bulb
temperatures were equal.
[I The remedy The solution was obvious. The wick should have
been wetted.
E Correcting the situation Wetting the wick required adding
water to it and taking certain precautions. The students were
aware of the precautions and knew where to find them:12'1 1) the
wick should be completely wet so no dry areas of the wick are in
contact with the gas, and 2) the makeup liquid should be at the
wet-bulb temperature.
E[ Checking the results To check the effectiveness of the solution,
data were taken for a while and normal conditions were
successfully restored without any side effects.
El Reflecting on the procedure In reflection, students realized
that things should not be taken for granted, e.g., measurements
and measuring equipment should be suspect and historical
records of the equipment should be available. The steps to be
added to the troubleshooting list'" for this problem are: check the
measurement, make sure measuring equipment is working well,
check to see if air is saturated, and check whether or not the wick
is wet. The lessons pertinent to the attitudes and thinking errors
learned from this problem are: be persistent and confident, do
not jump to conclusions, and be able to break the assumptions.

Spring 2002

steps and the problem was fixed.'4"
A typical example showing the steps followed in UOL
troubleshooting is given in Table 2.


Thinking errors prevent the troubleshooter from correctly
defining the trouble or arriving at the correct solution. These
errors are known to occur in perception, information, and at-
titude. Being aware of and avoiding these types of errors
greatly improves thinking skills. It is important for students'
attention to be directed toward these errors by asking ques-
tions and reflection.
Errors in perception result from established patterns of the
brain that control attention that need to be directed in order
to change these patterns.13"i The errors include failure to see
the situation (e.g., considering the symptoms instead of the
real problem, and the opinions instead of the facts), seeing
one side only (e.g., seeing only one solution to the problem),
and considering only part of the time scale. Practical examples
include failure to challenge the assumptions (e.g., calibra-
tion curves), failure to wet the wet-bulb thermocouple before
starting up the cooling tower, draining the reactor directly
into the sewage without paying attention to the products, and
assuming that the reaction was complete.
Lacking sufficient data, using false information, or having
extra information are potential sources of errors in informa-
tion. This frequently happens in studying flooding effects
in fluidization, in startup of a reactor, and in establishing
stress strain curves.
Negative attitudes resist thinking or direct it in the wrong
direction. This happens when there is a lack of genuine curi-
osity, when others are not listened to, and when things are
taken for granted. Practical examples include ignoring
straightforward reasons and considering complicated ones,
ignoring minor things that might indicate a developing prob-
lem, and jumping to the conclusion that a particular thing is
at fault because it is a common type of failure.

Problem solving has gained much attention in the chemi-
cal engineering literature, though more toward the classroom
and less toward the laboratory. The UOL is one place to de-
velop and improve troubleshooting skills that can be devel-
oped and mastered with intensive practice and time. Students
are expected to acquire a pattern of thought and to gain the
traits of expert troubleshooters. The troubleshooting strategy
we use requires that students maintain the right attitude dur-
ing the troubleshooting process, and that they reflect on their
actions. This motivates the process and facilitates forming
the troubleshooting list,"' which will hopefully be used as
the basis for a computerized case-based system.

Acquiring problem solving, and hence troubleshooting,
skills enhances problem-based learning (PBL), which is de-
fined as any learning environment in which the problem drives
the learning process.[22J Our program provides an opportu-
nity for students in the UOL to explore previously learned
knowledge, to learn new knowledge, to acquire an orderly
pattern of thought in solving practical problems in a critical
and creative manner, and to develop the traits of skillful
practitioners. It requires continuous feedback and reflec-
tion from the students as well as involvement of the in-
structor. Therefore, troubleshooting in the UOL motivates
learning and facilitates memorization (storage and re-
trieval of information).
To assess students' performance, we use a method that con-
siders both personal and teamwork abilities and monitors the
progress of students' work both in the process (skills) and
content (text).141 This method emphasizes the achievement of
goals and considers the details in a diagnostic manner (to
reveal the strengths and weaknesses). Continuous follow-
up, immediate feedback, and reflection are essential parts
of this method.
The background of the instructor is important in the sense
that he can choose to widen the scope of troubleshooting to
include several types of problems and situations. Instructors
track the work of the students, which is not going to be graded
unless it is fully acceptable. They guide and monitor the stu-
dents in the right direction, encourage interaction and imme-
diate feedback, and provide consultation as required.
Overall, students enjoy such a program and find the UOL
an interesting and attractive place to work. They have the
opportunity to think effectively, to reenforce and emphasize
what they have learned elsewhere, to acquire and maintain
the right attitude, to appreciate the cost and effort used to
search and research the data, and to link with everyday life.
They practice this in an interactive, collaborative, and non-
competitive endeavor, with immediate feedback from the in-
structor on their performance. The instructor, however, must
be aware of the goals he wants to accomplish and rank them
according to the way they are achieved. For productive re-
sults, projects and procedures should be updated regularly.


The author thanks the reviewers of this paper and the edi-
tors of CEE for their valuable remarks.

1. Abu-khalaf, A.M., "Getting the Most Out of a Laboratory Course,"
Chem. Eng. Ed., 32(3), 184 (1998)
2. Munson-McGee, S.H., "An Introductory ChE Laboratory Incorporat-
ing EC 2000 Criteria," Chem. Eng. Ed., 34(1), 80 (2000)
3. Miller, R.L., J.F. Ely, R.M. Baldwin, and B.M. Olds, "Higher-Order
Thinking in the Unit Operations Laboratory," Chem. Eng. Ed., 32(2),
4. Abu-khalaf, A.M., "Improving Thinking Skills in the Unit Operations

Laboratory," to appear in Int. J. Eng. Ed., 17(6), p. 593 (2001)
5. Abu-khalaf, A.M., "Introducing Safety in the Chemical Engineering
Laboratory Course," Chem. Health and Safety, 8(1), 8 (2001)
6. King, J., "Incorporation Safety Into a Unit Operations Laboratory
Course," Chem. Eng. Ed., 32(3), 178 (1998)
7. Abu-khalaf, A.M., "Safety and Thinking Skills," to appear in Chemi-
cal Health and Safety, 8(6), p. 19 (2001)
8. Myers, K.J., "Troubleshooting in the Unit Operations Laboratory,"
Chem. Eng. Ed., 28(2), 120 (1994)
9. Boostrom, R., Developing Creative and Critical Thinking, National
Textbook Company (1993)
10. Fogler, H.S., and S.E. LeBlanc, Strategiesfor Creative Problem Solv-
ing, Prentice Hall (1993)
11. Sears, J.T., D.R. Woods, and R.D. Noble (editors), "Problem Solv-
ing," AIChE Symp. Series, 79(228) (1983)
12. Chorneyko, D.M., R.J. Christmas, S. Cosic, E. Dibbs, C.M. Hmielec,
L.K. Macleod, R.F Moore, S.L. Norman, R.J. Stankovich. S.C. Tyne,
L.K. Wong, and D.R. Woods, "What is Problem Solving?" Chem. Eng.
Ed., 13(3), 132 (1979)
13. Woods, D.R. (ed), "Using Troubleshooting Problems," Chem. Eng.
Ed., 14(2), 88 (1980)
14. Woods, D.R. (ed), "Using Troubleshooting Problems, Chem. Eng. Ed.,
14(3), 130 (1980)
15. Chaput, .B., "Tackle Troubleshooting With a Case-Based Expert Sys-
tem," Chem. Eng. Prog., 95(5), 57 (1999)
16. Gans, M., D. Kohan, and B. Palmer, "Systemize Troubleshooting Tech-
niques," Chem. Eng. Prog., 87(4), 25 (1991)
17. Moyers, C.G., "Don't Let Dryer Problems Put You through the
Wringer," Chem. Eng. Prog., 88(12), 34 (1992)
18. French, W.W., "Tips for Troubleshooting Pumps," Chem. Eng. Prog.,
88(6), 65 (1992)
19. Schiavello, B., "Troubleshoot Centrifugal Pumps," Chem. Eng. Prog.,
88(11), 35 (1992)
20. Ahmed, N., and A.A. Khan, "Common Telltales Can Identify Safety
Hazards," Chem. Eng. Prog., 88(7), 73 (1992)
21. Casey, R.J., and M.J. Frazer, Problem Solving in the Chemical Indus-
try, Pitman Publishing (1984)
22. Woods, D.R., The McMaster program for problem solving, available
on line at
23. Rugarcia, A., R.M. Felder, D.R. Woods, and J.E. Stice, "The Future of
Engineering Education. Part 1. A Vision for the New Century," Chem.
Eng. Ed., 34(1), 16(2000)
24. Felder, R.M., D.R. Woods, J.E. Stice, and A. Rugarcia, "The Future of
Engineering Education. Part 2. Teaching Methods that Work," Chem.
Eng. Ed., 34(1), 26 (2000)
25. Haile, J.M., "Toward Technical Understanding: Part 1. Brain Struc-
ture and Function," Chem. Eng. Ed., 31(3), 152 (1997); "Part 2. El-
ementary Levels," Chem. Eng. Ed., 31(4), 214 (1997); "Part 3. Ad-
vanced Levels," Chem. Eng. Ed., 32(1), 30 (1998)
26. Pontefract, R.A., "Don't Lose Sight of Gage Classes," Chem. Eng.
Prog., 87(4), 75 (1991)
27. Abu-khalaf, A.M., "Start-Up of a Non-Isothermal CSTR: Mathemati-
cal Modeling," Chem. Eng. Ed., 31(4), 250 (1997)
28. Felder, R.M., and R.W. Rousseau, Elementary Principles of Chemical
Processes, 3rd ed., Wiley (2000)
29. McCabe, W.L., J.C. Smith, and P. Harriot, Unit Operations of Chemi-
cal Engineering, 4th ed., McGraw-Hill (1985)
30. De Bono, E., Teaching Thinking, Temple Smith (1976)
31. Liberman, N.P., Troubleshooting Process Operations, 2nd ed., Penwell
Publishers (1985)
32. Saletan, D., Creative Troubleshooting in Chemical Process Industries,
Chapman Hall (1994)
33. Branan, C., Rules of Thumb for Chemical Engineers, 2nd ed., Gulf
Publishing (1998)
34. Goyal, O.P., "Evaluating Troubleshooting Skills," Hydrocarbon Proc.,
79(10), 100 (2000) 0
Chemical Engineering Education

Teaching Students to Be Innovative
Continued from page 121.

transfer.18' Those articles are suggested as supplements for
the present one.

For many years I conducted a seminar for doctoral students,
both my own and others, in advanced topics in fluid mechan-
ics and heat transfer. The format consisted of three assign-
ments for study, oral presentation, and written presentation,
first on some classical topic, second on some new analytical
development in the recent literature, and third on a theoreti-
cal investigation of their own of limited scope. This process
may be regarded as a three-step initiation into innovative
analysis. Many of the students in the seminar achieved a pub-
lishable result, with the same psychological benefits men-
tioned above in connection with innovation in doctoral re-
search. This seminar eventually fell victim to the unwilling-
ness of the other faculty members to tolerate such a distrac-
tion from the sponsored doctoral research of their students.
Indeed, the participants were often inspired to make a sig-
nificant, perhaps excessive, commitment of time to their
analytical investigation because of the excitement of do-
ing innovative work as compared to the more routine work
of their doctoral research.


Teaching innovation in the classroom is almost certainly
more effective within the context of a regular technical course
rather than in a special course or special designated segment
of a course on that topic. Even within the context of a regular
course, the task is more difficult than in the context of re-
search or a graduate seminar. Within the courses and topics
in chemical engineering that I have taught through the years,
speculative dimensional analysis proved to be the most ef-
fective vehicle for illustration of the process of innovation
with undergraduates (see Churchillt9l for a description of this
methodology). The development of theoretically based cor-
relating equationst"01 as well as speculative dimensional
analysis have been found to serve this role successfully
with graduate students. For both undergraduate and gradu-
ate students, the Socratic method was found to be most
effective on these topics.


The concept of innovation is highly esteemed in our cur-
rent culture, but its genesis and performance are not given
much direct attention. Furthermore, innovation is not always
welcome when it conflicts with old habits, common wisdom,
well-established practices, or deeply held convictions. In ad-
dition, innovative ideas and findings may be neglected or
Spring 2002

rejected in industry because of constraints of cost and time
and in academia because of the restrictions imposed by
their sponsorship.
It is, of course, easier to impart the science and art of engi-
neering to our students than to teach them to innovate. Dis-
covery and innovation are not programmable and are thereby
difficult to formalize, but we can stimulate innovative think-
ing by establishing an atmosphere in the classroom, confer-
ence room, and laboratory in which it is encouraged, wel-
comed, and rewarded.
The experiences of my own students indicate that innova-
tion can be fostered by proper choice of an objective and
development of the proper mindset. Exploratory research is
conducive to innovation because it implies a willingness
to take risks and to pursue a new direction when appro-
priate. Establishing confidence in their own ability to in-
novate is a first prerequisite.
The anecdotal experiences of the great innovators serve
educationally as a useful guide and source of inspiration for
students, since it is evident therefrom that they too often
experienced doubt, failure, and rejection, and only tri-
umphed by persistence.
Innovative thinking is more difficult to teach in the class-
room than in research, but it can be induced within the
context of technical subject matter and, most effectively,
by the Socratic method.
The psychological gains from innovative work may be as
important both for students and for practicing engineers as
the technical and intellectual contributions.
Despite the favorable image of innovation, it is invariably
resisted, not only by those whose contributions are dis-
placed, but also by those who are forced to discard com-
mon wisdom and relearn.

1. Richter, J.P., ed., The Notebooks of Leonardo da Vinci, Vol. 1, Dover
Publications, New York, NY (1970)
2. Churchill, S.W., "Turbulent Flow and Convection," Adv. in Heat Trans-
fer, 34, p. 255 (2000)
3. Mann, Thomas, Joseph and His Brothers, translated by H.T. Lowe-
Porter, Alfred A. Knopf, New York, NY (1948)
4. Newton, Isaac, Principia, translated by Andrew Motte in 1729, trans-
lation revised by Florian Cajori, University of California Press, Ber-
keley, CA (1966)
5. Chandrasekhar, S., Truth and Beauty: Aesthetics and Motivations in
Science, The University of Chicago Press, Chicago, IL (1987)
6. Hardy, G.H., A Mathematician's Apology, Cambridge University Press
7. Churchill, S.W., "Chemical Kinetics, Fluid Mechanics, and Heat Trans-
fer in the Fast Lane: The Unexpurgated Story of a Long-Range Pro-
gram of Research in Combustion," Chem. Eng. Ed., 25(4), 186 (1991)
8. Churchill, S.W., "Innovation, Discovery, and Advances in Heat Trans-
fer," Thermal Sci. and Engg., 8(2), 1 (2000)
9. Churchill, S.W., "A New Approach to Teaching Dimensional Analy-
sis," Chem. Eng. Ed., 31(3), 158 (1997)
10. Churchill, S.W., "The Art of Correlation," Ind. Eng. Chem. Res., 39(6),
1850 (2000) O

2002 ASEE Annual Conference

SChemical Engineering Division Program

June 17-19, 2002 U Montreal, Quebec, Canada

Technical Sessions

Monday, June17

Session 1313 10:30 a.m. Novel Classroom Environments
Moderators: Ranil Wickramasinghe and Ann Marie Flynn
1. "An Integrated Approach to Teaching Material Balances at University of Waterloo" C. Moresoli
2. "Guilt-free Chocolate: Introducing Freshmen to Chemical Engineering" K. Hollar S. Farrell, M. Savelski
3. "Lessons with Lego Engaging Students in Chemical Engineering Courses" K. Levien, S. Rochefort
4. "Process Descriptions: An Introductory Library Research Assignment on Chemical Processes for First-Year Students" S.S. Moor
5. "Expanding Our Students' Brainpower: Idea Generation and Critical Thinking Skills" J. Jessop
6. "Project-Based Learning in Chemical Engineering" B. Marcos

Session 1613 4:30 p.m. Teaching Outside the Box
Moderator: Melanie McNeil
1. "A Modified Approach to Material and Energy Balances" D. Miller, M. Anklam, R. Artigue, M.H. Hariri, M. Misovich
2. "Don't Waste Your Breath" S. Farrell, R. Hesketh, K. Hollar M. Savelski, C.S. Slater R. Specht
3. "Learning 'Outside the Toy Box'" J. Keith
4. "Micromixing Experiments in the Undergraduate Curriculum" R. Hesketh, K. Dahm, M. Savelski
5. "The Values of a 'Hands-On' Experience in a Reactor Design Course" R. Lewis
6. "Fundamentals of Fixed Bed Adsorption Processes: Analysis of Adsorption Breakthrough and Desorption Elution Curves" J. Becnel, C. Holland,
J. Mclntyre, J. Ritter

Tuesday, June 18

Session 2213 8:30 a.m. Innovative Courses for ChE Students
Moderators: Donald Visco and Christopher Brazel
1. "Introduction to Chemical Engineering: A New Course for Freshman Students" D. Knox, B. Baltzis
2. "The Power of Pizza" W. Whiting, C. Coronella
3. "Environmental Health and Safety and Biochemical Enginering with a Chemical Engineering Foundation" M. McNeil, A. Diaz, M. Jennings
4. "A New Course in Green Chemistry and Benign Processing" D. Miller
5. "New Methods of Teaching and Learning for Industry-Based Engineering Professionals" B. Dickson, C. Grant
6. "A Course on Engineering Entrepreneurship for Chemical Engineering Seniors" R. Narayan
7. "The Effects of Physical Environment on Engineering Team Performance" E. Grulke

Session 2513 2:30 p.m. Panel: How are We Faring with EC2000?
Moderator: Joseph Shaeiwitz
1. "Before, During, and After the EC 2000 Visit" E. Grulke
2. "The EC2000 System in Chemical Engineering at Washington State University" R. Zollars
3. "EC2000 From Both Sides of the Fence" S. LeBlanc
4. "EC2000: What to Expect/What is Expected" D. Briedis
5. "Reflections on Outcomes Assessment and the ABET Accreditation Process" R. Miller

Session 2613 4:30 p.m. Assessment in Large and Small Programs
Moderator: James Newell
1. "Teaching of Thermodynamics and Fluid Mechanics using Interactive Learning Methods in Large Classes" W Dempster, C. Lee

Chemical Engineering Education

2. "Chemical Engineering and Society A Response to Constituency Concerns" R. Terry: W. V Wilding
3. "Assessment Methods for Engineering Programs Too Many Choices or Not Enough?" D. Knox
4. "A Senior Exam to Assess the Learning of Core Competencies in a Chemical Engineering Curriculum" R. Terry
5. "Student Development of Grading and Assessment Criteria" V Young
6. "Rubric Development and Inter-Rater Reliability Issues in Assessing Learning Outcomes" J. Newell, K. Dahm, H. Newell

Wednesday, June 19

Session 3213 8:30 a.m. What's in Store for the ChE Curriculum?
Moderators: David Millet; Nada Assaf-Anid
1. "Chemical Engineering: Professionally Ignored?" F* Sharifi
2. "Cooperative Engineering Education Program" W Krantz, K. Cedercreutz, A. Dardv
3. "Incorporating Biotechnology in the Chemical Engineering Curriculum" N. Assaf-Anid, H. Hollein
4. "Communication Without Borders: Collaborative Expression Within and Across Disciplines" L. Bullard
5. "Preparing International Students to Become Effective Teaching Assistants" B. Baltzis
6. "The Core Graduate Chemical Engineering Curriculum: Does It Exist?" D. Kauffinan
7. "Why Settle for an MBA?" J. Reynolds. A.M. Flynn, L. Theodore

Session 3413 12:30 p.m. The Computer, the Web, and the ChE
Moderator: David Silverstein
1. "A New Method to Calculate Phase Coexistence" D. Visco, J. Russum
2. "ChE's Teaching Introductory Computing to ChE Students A Modern Computing Course with Emphasis on Problem Solving and Programming"
D. Clough
3. "An Improved Distance Learning Environment for the Material and Energy Balances Course" D. Silverstein, G.T Lineberry
4. "Heat Transfer OnLine" W. Baratuci, A. Linse
5. "No Food Allowed The Latest Virtual Reality Laboratory Accident" J. Bell, S. Fogler
6. "Using the Modem Chemical Engineering Laboratory at a Distance" J. Henry

Session 3513 2:30 p.m. Control in the Classroom
Moderators: Stephanie Farrel and S. Scott Moor
1. "A WEB Site to Support Active Learning in Process Control" M. Thomas, E. Wood, M. Hough, S. Yip
2. "Incorporation of Process Stimulators in Teaching Separations" A. Serbezov
3. "Use of Process Stimulation and McCabe-Thiele Modeling in Teaching Distillation" K. Dahm
4. "Introducing Process Controllers Throughout the ChE Laboratory" D. Dixon, J. Puszynski
5. "Simple Modules that Illustrate Dynamic Matrix Control" C. Nippert
6. "Team Learning and Individual Accountability: A Case Study from a Senior Year Process Dynamics and Control Course" B. Baltzis
7. "The Missing Link in Process Control Education Incorporating PLC's into the ChE's Control Course" D. Clough

Session 3613 4:30 p.m. The Modern ChE Laboratory
Moderators: Kathryn Hollar and Jim Henry
1. "AIChE's New Student Chapter Competition: The Chem-E-Car Competition"
D. Dixon, M. Abraham, C. Coronella, R. Hesketh, W Rochefort, R. Zollars
2. "In Pursuit of the Perfect Potato Chip" J. Smart
3. "Integrating Team Laboratory Experiments Into a Senior Biochemical Engineering Course" C. Brazel
4. "Introducing Students to Lab Safety in Chemical Engineering: The Safety Scavenger Hunt" K. Hollar, K. Dahm, M. Harris
5. "Making 'Moder' Undergraduate Labs 'Old'" M. Cline, G. Powers
6. "The Role of Experiments in Inductive Learning" R. Hesketh. S. Farrell. C. S. Slater

Socierv-Wide ChE Division Banquet
Picnic Monday, June 17, 6 30 p m
Sunday, June 16. 6 p m. Gibby's Restaurant
Windsor Station Speaker to be announced

Meet the Board ChE Division ASEE
Breakfast Business Luncheon Annual Awards Banquet
Tuesday, June Ih, 7 -0f a.m Tuesday. June Ib Wednesday. June 1)
Wyndham Montr&al 12:30 p m. 7-00 p.m.

Spring 2002 129

T, classroom




Case Study: A Separation Processes Course

University of Newcastle Callaghan NSW 2308 Australia

Educational literature makes frequent reference to the
fact that people have many different learning styles,
and in order to tap into these different styles, a vari-
ety of teaching assessment methods can be used."' Inspired
by this literature, I decided to include written-answer ques-
tions alongside numerical problems in the assignments and
exams given to students in a third-year undergraduate sepa-
ration processes course.
The course covers topics such as gas absorption/stripping,
distillation, leaching, and extraction. When I took it as an
undergraduate, all of the assessment was based on calculat-
ing the answers to numerical problems, e.g., "How many equi-
librium stages are required to perform such-and-such a counter-
current separation?" In many introductory textbooks, the
end-of-chapter questions also consist exclusively of nu-
merical-type problems.[2-4]
The result was that this was the only style of question I
used during my first few years of teaching this course. But I
found that students often mechanically follow procedures for
constructing operating lines and stepping off stages by (for
instance) the McCabe-Thiele method without having any
physical understanding of what these lines mean. They do
not take time to read the background material in the text, look-
ing only for worked examples that will help them solve set
problems. By using written-answer questions, I hoped to force
students to understand the material at a deeper level.

A number of questions were posed at the end of each chap-
ter in the course study guide (see Table 1). Students were
referred to chapters in the textbooks where answers could be
found and were told that some of the questions would appear
Copyright ChE

Most students believed [written-answer
questions] helped them gain a deeper
understanding of the material. At the
very least, it adds some variety
and freshness to the course.

on the exams. This warning serves to motivate them to com-
plete the set readings.'15
In the assignments and exams, some additional written-
answer questions were used (see Table 2). Assignments were
performed in groups of four. The first question in Table 2,
concerning Fick's law, is based on a similar example given
by Felderr[6 for Fourier's law of heat transfer. To illustrate
Fick's law, students gave examples ranging from the diffu-
sion of the smell of toasting bread throughout a kitchen to
the interchange of oxygen and carbon dioxide in the lungs.
The second question in Table 2, on "Life Without Distilla-
tion," was designed to get students to consider the role that
distillation, and hence chemical engineering, plays in mod-

Simon Iveson completed his BChE in 1992
and his PhD in 1997, both at the University of
Queensland. Since that time he has been a
research fellow and part-time lecturer in the
Department of Chemical Engineering at the
University of Newcastle. His research interests
are in the field of particle technology, with his
focus being on the agglomeration of fine par-
ticles by the addition of liquid binders.

Division of ASEE 2002
Chemical Engineering Education

Questions Posed at the End of Each Chapter
of the Study Guide

TOPIC 1: Staged Countercurrent Gas Absorption and Stripping
Why are continuous mass transfer columns generally run
countercurrently rather than co-currently? Is there any situation
in which you would use co-current?
What are the relative advantages and disadvantages of bubble-
caps, valves, and sieve trays?
Describe the various flow conditions that can occur in tray
columns, such as flooding, weeping, entrainment, and coning.
Explain their causes and also why these conditions are undesir-
What determines the optimum solvent flowrate to use in a staged
countercurrent mass-transfer column?
Define an equilibrium stage and the Murphree stage efficiency.
Why are real stages often less efficient than ideal stages? Can a
real stage ever have greater than 100% Murphree efficiency?
Do the analytical solutions used to calculate the number of stages
required in lean-phase systems give an overestimate or an
underestimate when applied to rich-phase systems? Explain.

TOPIC 2: Distillation
What are the relative advantages and disadvantages of flash,
batch, and continuous countercurrent distillation?
Does batch distillation always give better separation than flash
distillation? Explain.
What are the required conditions for the operating lines of a
continuous countercurrent distillation column to be straight?
How realistic are these assumptions?
What determines the optimum reflux ratio at which to operate a
countercurrent distillation column?
What is the effect of preheating a feed stream before it enters a
distillation column? Does this offer any advantage? Explain.

TOPIC 3: Packed Columns for Absorption and Humidification
What are the relative advantages and disadvantages of packed
versus tray columns?
What properties are important when choosing a column packing?
What determines the optimum gas and liquid flowrates to use in
a packed column?
Is humid air denser or lighter than dry air? How is this exploited
in the design of cooling towers?
What is the difference between an adiabatic saturation line and a
wet-bulb temperature line? Why are the two similar for air-water
Why is cooling water often cooled and reused rather than using a
fresh supply of water? What problems are encountered when
water is recycled in this way, and what measures are taken to
overcome these problems?

TOPIC 4: Leaching and Extraction
What are the desirable properties for a solvent to use in liquid-
liquid extraction operations? Explain why each is beneficial.
What are the relative advantages and disadvantages of liquid-
liquid extraction versus direct distillation for the separation of
two miscible compounds?

Spring 2002

ern society. I could have simply asked them to make a list of
everyday items that require distillation at some stage in their
manufacture, but instead I asked them to speculate on how
life would be different without distillation, hoping that this
would engage their interest and imagination. Of the seven
groups, four took up the invitation to use a fictional story to
illustrate their answer (see Table 3), while the other three
groups used the more conservative approach of simply list-
ing and discussing items that would no longer be available
without distillation (see Table 4).

Many of the stories were quite imaginative and creative. In
Table 3, for instance, the group has woven into the story some
complaints about the difficulty of the previous assignment
(this may have had a cathartic benefit?) and also made hu-
morous reference to one of their professors who in lectures

Additional Questions Used in
Assignments and Exams

Question 1: Fick's Law
Describe molecular diffusion and explain Fick's law in terms
understandable to a senior high school student. Explain an
everyday phenomenon of your choice by using Fick's law.
(Maximum, 1 page)

Question 2: Life Without Distillation
Describe how daily life would be different (perhaps by using a
story?) if the distillation process had not yet been discovered. Be
creative! (Maximunm, 2 A4 pages of double-spaced, 12-point font

Question 3: Trays and Packings
Write a report on the topic "Trays and packing used in distillation
and gas absorption/stripping operations." In this report you should
Describe the different types of trays and packing commercially
available for distillation and absorption columns.
Discuss the properties of importance when choosing between
different types.
Describe the relative advantages and disadvantages of tray
versus packed columns and the factors that determine which you
should use in any given application. (Maximum, 10 pages)

Question 4: Extraction versus Distillation
Write a brief memo on the topic: "The relative advantages and
disadvantages of extraction versus distillation for separating two
miscible liquids. (Maximum, 2 pages)

Question 5: Designing a Separations Unit
You are a process engineer who has been asked to design process
plant to remove a dissolved liquid solvent from an aqueous (i.e.,
water-based) waste stream. Discuss the issues that you would
consider when choosing, designing, and sizing an appropriate
separation system. Issues you should address include (but are not
limited to) economic, safety, and space constraints. How would you
decide whether to use distillation, liquid-liquid extraction, or
stripping? You should specify what material properties and other
information you would need to know to design the system.

frequently uses the expression "There's no magic
here, people." I imagine that this group would have
had a good laugh together as they brainstormed ideas.
The only negative aspect of these creative stories
was the tendency of many groups to use the stereo-
type of engineers as heavy drinkers of alcohol. Given
the topic of distillation, however, it was perhaps in-
evitable that this would occur.
Question 3, Table 2, on trays and packing took up
an entire assignment. This question involved the stu-
dents doing research. They were encouraged to read
widely, look things up on the web, and to approach
manufacturers for information. This assignment
seemed to serve its purpose of helping students un-
derstand the pros and cons of the various types of
trays and packing.
The biggest problem in most reports was the fail-
ure of students to critically evaluate manufacturers'
claims. Many reports simply repeated verbatim from
manufacturers' sale brochures such statements as
"Packing XYZ is the best available packing for...,"
without any qualification or supporting evidence.
One additional benefit of assignments like these
is that they provide the lecturer with a useful list
of web sites to which he can refer students in sub-
sequent years.
The fourth question was fairly straightforward, sim-
ply involving summarizing the lists of pros and cons
of extraction versus distillation. These are given
in the introductory section on extraction in most
The fifth question was allocated thirty minutes
on the final exam. It was designed to test how well
the students had integrated what they had learned.
The average mark I awarded for this question was
only 43% (standard deviation 21%, number of stu-
dents 29).
Most students seemed to struggle with the open-
ended nature of the question. They often spent a great
amount of time on one specific point, such as pack-
ing types, without exploring the full range of issues
involved. In hindsight, it may have been better to give
a few of these open-ended type questions in their as-
signments in order to better prepare them for answer-
ing this type of question on the exam.

No quantitative assessment was performed to see
if use of written-answer questions enhanced student
understanding, but a qualitative evaluation was per-
formed at the end of the subject by asking students
to answer the following question:

Sample Answer (Question 2, Table 2)
From a Group Using a Creative Story

After struggling for hours on Assignment 2, Question 3, a group of discouraged
3rd year chemical engineering students were walking to the library to start
Assignment 3. While walking, one of the members said, "I wish distillation had
not been discovered." Suddenly out of the bushes on the Don Morris walk leapt
the Magical Professor, who promptly cast a spell and granted the group's wish.
Instantaneously the four members were transported to a parallel universe.
After such an experience, the group decided (as all engineers would) to go to
the bar. Upon arriving at the Godfrey Tanner Bar, wanting something a little
stronger than the usual, the group ordered a selection of spirits that included
whisky, bourbon, rum, and Malibu. Much to their disappointment, the bartender
cast them a confused look and said, "Victorian Bitter or Tooheys,t" take your
pick...and maybe a nice red for the lady."
The group, sullen with disappointment, took their beers carelessly and one
member spilled the contents of his glass all over his new $200 jacket. "This will
need dry cleaning!" he exclaimed. Another member said, "At Contact they do
dry cleaning," and so they all walked over to Contact, only to discover that this
service wasn't only unavailable, it was unheard of!
Then one of the members burst out with laughter at a fellow student lighting
his cigarette with sparks from his flint and steel. The group decided in a
unanimous vote to get out of this "bizaro world" and headed off to the car park.
Upon reaching the car park the group was confronted with a dusty plain and not
a car in sight. "Dude, where's my car?" yelled the red-header 'pwouf' boy from
chemical engineering when he discovered his Corolla wasn't there.
So the group had a long hike to the train station, with their spirits low and,
feeling like there was nowhere they knew, they just kept on going.
The train turned up on time, as usual, but there was something distinctly
different about it it was steam powered! It hurled plumes of black smoke and
soot all over the four adventurers and their beer-stained jackets.
A sudden change in the weather caused a downpour of rain. The group was
unphased as their jackets were waterproof...or so they thought. Actually, the
water soaked straight through their clothes. Professor Smith121 was also at the
train station and, overhearing their conversation, he stated, "There is no magic
here, people."
The four adventurers then had a group huddle. "What can we do next?" The
problem was, there were so many things to do, but so many things were changed
as a result of their impromptu wish. For example, Jason said, "Let's go to the
beach," but the problem with this was that sunscreen, surfboards, and
bodyboards did not exist. Kelly said, "Let's go to the movies," but once again
the movie screen, camera parts, and sound speakers need the products of
The whole group by this time was utterly distraught and had given up on the
world without distillation and its offshoots. Then the Magic Professor reap-
peared and said, "You have seen the present without distillation and it was
nothing like you could ever imagine, for you my friends are now the privileged
ones who can strive in life with the power of DISSTLEFORCE running through
your veins. Now you will uphold the law of the mighty distillation column and
never let it be ridiculed again." With two taps of his distillation wand we were
back in Newcastle.
From that day onwards, the University Distillation Inebriators Congregation
(U.D.I.C.) never allowed bad words to be said about distillation and helped
everyone in engineering to cope with life by giving free alcohol (distilled
themselves) religiously once a week at a Barby."'

The End!'41

"I Australian brands of beer
121 Name of lecturer changed to preserve anonymity.
I1' Australian slang for "barbeque"
H1 Spelling and grammatical mistakes in the original manuscript have been

Chemical Engineering Education

There seems to be no reason why

this approach could not also be used

in other courses, particularly

those involving application of

engineering principles.

Sample Answer (Question 2, Table 2) From a Group
that Did Not Use a Creative Story

Life Without Distillation?

Without the discovery of the distillation process, our modern way of
life would not exist as we know it. Many of the everyday objects.
machines, and processes we take for granted would not be around
without distillation.
If one were to visit a distillation-free world, it would be much like
stepping back in time. Cars and internal combustion engines would
not be around, making it necessary to rely on coal or wood-fired
steam engine technology. If this were the case, society would need to
have extensive public transport systems running as personal steam
engine cars would be far too impractical and polluting. If petroleum
distillates were not readily available as fuels and lubricants, transport
of goods and mail would be slow, slowing economic and technologi-
cal growth in far-away areas. Indeed, almost every aspect of modern
life would be different without the petrol/diesel engine. Even warfare
would be a totally different event-without planes, jets, tanks,
submarines, etc., wars would be fought in a totally different manner,
even if other technologically advanced weapons were available.

Many simple chemicals would not be available in their pure form
without the use of distillation. Benzene, for example, can be obtained
in large amounts from fractional distillation of crude oil. Without this
chemical, many other beneficial chemicals would not be available to
industry. TNT, for example, is based on the (methyl)benzene
molecule; without this powerful explosive, mining would be slowed
down immensely, affecting the production of many important metals
such as iron and aluminium. Benzene is also a precursor for many
beneficial pharmaceutical drugs, such as the benzodiazepine family.
Many plastics, rubbers, and other polymers could not exist without
employing the distillation process to obtain the monomers needed for
their manufacture. Without plastics, many modern-day products such
as plastic bottles, plates, food wrappings, etc., would have to be made
from other, more costly, materials such as metal and glass.
No distillation would mean that the only drinks down at the local
pub would be variations of beers and wines. Distilled spirits such as
rum and gin would not be for sale.
Many perfumes, fragrances, and esters would not be available, as
these require distillation and refluxing to separate them into a
concentrated form. Esters are used in many foods as artificial
flavorings and aromas; confectionary-like bubblegum (original
flavor) and banana Paddlepopsi" could not be made without them.
Even concentrated natural chemicals such as vanilla essence and
various essential oils would be difficult to obtain.

"I An Australian brand of ice-block/ice-cream.

Assessment in this course has involved written-answer
questions in the assignments and exams, as well as the
more traditional numerical-type problem solving. Have you
found these written-answer questions helpful in gaining an
understanding of the course? Explain.
Of 29 students in the class, a total of 20 answered this question,
with 18 responding that they found written-answer questions
helpful to their understanding. Positive comments included ref-
erence to enjoying the change from numerical-type questions,
benefiting from being forced to understand at a deeper level
than just putting numbers in equations, and having to read more
widely than they would have otherwise.
Only two students claimed that they learned better from nu-
merical-type questions. The major negative comments about
written-answer questions were
1) That they are subjective and hard to get full marks on (2
2) That they require much more time and reading to answer,
which was an overload given the course's existing heavy
workload (3 responses).
Therefore, the students themselves were overall positive about
having to do written-answer type questions as part of their as-
sessment. Most students believed it helped them gain a deeper
understanding of the material. At the very least, it adds some
variety and freshness to the course.

It is possible, without too much difficulty, to generate a list
of written-answer questions to use in assessment for a separa-
tion processes course. Students responded positively to this
addition to the traditional numerical-problem-solving type of
question. It appears to force them to read more widely and to
understand the content more deeply.
Many students also took advantage of the opportunity to write
creatively when the opportunity presented itself, something they
would otherwise rarely get the chance to do. At the very least,
this serves to create some variety and interest in the course.
There seems to be no reason why this approach could not also
be used in other courses, particularly those involving applica-
tion of engineering principles.

1. McKeachie, W.J., Teaching Tips: Strategies, Research, and Theory for
College and University Teachers, 10th ed., Houghton Mifflin, Boston,
MA (1999)
2. McCabe, Smith, and Harriot, Unit Operations of Chemical Engineering,
5th ed., McGraw-Hill (1993)
3. Treybal, R.E., Mass Transfer Operations, 3rd ed., McGraw-Hill (1981)
4. Coulson, J.M., and J.F. Richardson, Chemical Engineering. Volume 2:
Particle Technology and Separation Processes, 4th ed., Pergamon (1991)
5. Felder, R., "Beating the Numbers Game: Effective Teaching in Large
Classes." ASEE Annual Conference, Milwaukee, WI; June (1997)
6. Felder, "How About a Quick One?" Chem. Eng. Ed., 26(1), 18 (1992)
7. Treybal, R., Liquid Extraction, 2nd ed., McGraw-Hill (1963) 1

Spring 2002

[M] class and home problems

The object of this column is to enhance our readers' collections of interesting and novel prob-
lems 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 Department, Uni-
versity of Michigan, Ann Arbor, MI 48109-2136.




Tennessee Technological University Cookeville, TN

he evaluation of phase equilibria from equations of
state (EOS) is a classical problem traditionally taught
in thermodynamics courses as part of the chemical
engineering curriculum, both at the undergraduate and gradu-
ate levels. It can appear within a thermodynamics class or as
a practical example in a numerical methods course.
The solution methodology normally taught (of the several
available"J) for extracting coexistence information from an
EOS is based on solving for unknown variables in the EOS,
subject to the constraints of equilibrium. The technique is
implemented through an iterative procedurel2-31 or by using a
multidimensional root-finding algorithm.
In this work, we introduce a novel method to solve this
Jim Russum graduated from Tennessee Tech-
nological University in May of 2001 with a BS
in Chemical Engineering. He is currently en-
rolled in the graduate school of the Georgia
Institute of Technology where he is pursuing a
Ph.D. in Chemical Engineering. He is currently
conducting research in catalysis and mini-emul-

Don Visco has held the position of Assistant
Professor in Chemical Engineering at Tennes-
see Technological University since his gradua-
tion from the State University of New York at
Buffalo in 1999. His research interests are in
computational thermodynamics, bioinformatics
and impulse characterization in granular media.

problem by writing a differential form for the equilibrium
conditions that requires the numerical integration of these
coupled differential equations to trace out the coexistence
curve. For a simple example problem, we show that this
method produces reasonable results around thirty times faster
when compared to a root-finding algorithm. We also re-
visit a combined algorithm, discussed by both Asselineau,
et al.,151 and Michelsen,j61 that uses the best features of
both approaches.

Determine the coexistence curve for propane from T = 200K
to the predicted critical point. Use the Redlich-Kwong equa-
tion of state."17

Direct Method
The direct method involves writing the three equilibrium
conditions,r8' namely

TI(PI,p')= TV(PV,pv)

Pl(T' ,p)= Pv(Tv'pv)

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

where T is the temperature, P is the pressure, p is the density,
and p is the chemical potential, while the superscript "/" re-
fers to liquid and "v" refers to vapor.
According to the Gibbs phase rule'' for this one-compo-
nent, two-phase system, one independent intensive variable
must be fixed in order to establish the intensive state of the
system. Since the Redlich-Kwong equation of state is explicit
in pressure, we fix the temperature (i.e., T' = T') as per the
problem statement. From these specifications we are left with
two coupled, nonlinear equations (Eqs. 2 and 3 above) and
two unknowns (p' and p'). Here, a nonlinear root-finding
algorithm (Newton-Raphson'g9 with a forward-difference Ja-
cobian) is used to solve (root tolerance = 10s for conver-
gence) for the two unknowns at the system specifications
(fixed temperature). To reach the critical point, steps in the
temperature are taken starting from 200K. There is a finite-
sized temperature step, however, that one can take using this
solution methodology, above which this method will not con-
verge. For example, given the converged solution (liquid and
vapor density) at 200K, if we step to 200.03K (i.e., a step
size of 0.03K) and use the converged solution at 200K as the
guess into the Newton-Raphson routine for 200.03K, the
method will fail (a converged solution will not be reached).
But if we use a smaller step size (say, 0.02K), the routine
runs without problem up to the critical point. For a step
size of 0.02K, this method took about 1.5 seconds to run
on a Dell Dimension (600 MHz, 256Kb RAM, DIGITAL
Visual FORTRAN).

Integrate Mehod

Although the problem specification is for a two-phase, one-
component mixture, we will introduce the integrate method
for a two-phase, n-component mixture (keeping in mind that
what follows is applicable for three-phase or higher equilib-
rium). We do this to show the compactness of the resulting
expressions for a mixture, although the problem we are work-
ing through is for a pure component (propane).
We can write the chemical potential and pressure for each
component as

P = T',p i...,P ) (4)

dpl 4ui dT'+ Il dp
j' J^W n .P'
1 Tj= j I Tpk (k j)

dP'= d T'+ dpi
=T' =i pi T', p ,(ji)

dT( = dTV+ dpV(k (10)
B T P j=l T,p (kj)

dP" jId dT'" + dp
l p, i= p Tv,p,(ji)

If our system is at equilibrium, then the following con-
straints exist regarding the chemical potential and pressure
of the system, respectively,'"0

duj = dt' (12)
dP' = dP" (13)

We next divide each differential by an infinitesimal change
in the temperature and take these partial derivatives con-
strained to a path that satisfies the equilibrium of chemical
potential and pressure (the symbol o indicates that the de-
rivative is evaluated along that path).

aT aT

aT aT
i n ]

From the Gibbs phase rule we are able to specify n inde-
pendent variables in our system. We can fix the temperature
and the n-1 independent mole fractions of the liquid phase to
completely specify our system. Doing this, we can write the
differential equations for the liquid phase as

aIBK% (( =[JTI ),pl ()I (iJ ()
wj l+ T j I (16)
S j=1 T ,pk (kwj) (


PV =P(T',p ',...,pn)

where pi indicates a component density of species i, w
the differentials are

(5) ap' = P (d T) + in iap (17)l
) For the a pT r phase, arrie at 7)
i (6) )Tth, vjp i
(7 For the vapor phase, we arrive at


Spring 2002

a T a rT vT flapy T
a a j=1 1,, (kwj) a

SaV (aPV aT Sap (19)
aT aTV JpaBT api= taT a

where x, are the mole fractions of component i in the liquid

Substituting Eqs. (16) and (18) into

Eq. (14) and noting

T) 1 (20)
aT )Y

+ r l x
p J=l1 ) ',p (k*j) .

n aTV t (21)
V j=1+ V X Tv,pv H .
T v ap" (k j) aT )


T' aT v fapy
=1 a
l pT '=I ,pk (k j) ap

iR x p (22)
j=1 P (T',p' (kj) a

and, similarly for the pressure

Sap' P v n ap
taT' p, Tvp ) aT 0ap

xi x (23)
i= ap ',p'(ji)i a, (2

Recognizing that Eqs. (22) and (23) can be written in linear
form, we can use matrix notation to compactly represent the
system of equations in terms of a coefficient matrix (A), a
solution vector (b), and an unknown vector (x); Ax = b.

j d li mp tk' p' l T.p,(k#l) n T ,pT (k n)


x .r Cp' apTp t ap T
-j xi l p k apn

ji=1 OPi )T'tp apl tT ,p; (jkl) n T",p; (jn)

In order to implement the integrate method, a converged
starting point must be used. To do this, one would use the
direct method (or an iterative method), subject to the equilib-
rium conditions, at a specified temperature and liquid-phase
mole fraction to find the first point (i.e., the liquid phase den-
sity and the vapor phase component densities). For an equa-
tion of state that has the mole fraction, density and tempera-
ture as independent variables, all of the elements of the ma-
trix A and the solution vector b can be solved from the equa-
tion of state, either analytically or numerically. Thus, the un-
known vector x is given as x = A 'b. The n+l coupled differ-
ential equations are then numerically integrated to yield the
overall liquid phase density and the component densities of
the vapor phase. The mole fractions can be extracted from
the overall density and the individual component densities. A
step in the temperature is taken next and the process is re-
peated until the critical point is reached. A block diagram
(Figure 1) may be helpful in illustrating the technique.
For the purpose of the problem at hand, we have used a
fourth-order Runge-Kutta method""' to integrate numerically
the coupled pair of differential equations. We find that for
the same step size used in the direct method (0.02K), we
achieve reliable results for the phase densities in about the
same amount of time using the integrate method, as seen in
Figure 2 and Table 1, respectively. But, if we increase the
integration step size to 1K, the integrate method provides re-
liable results for the coexistence densities with a computa-

( aTo -, = amT"

; aT' i i, _aT
aT _( __]
Ip T I 1v

,\ a,

C aTp' a p'

TJp, aT
4 p )

tional speedup of
around 30 compared to
the direct method. At a
step size of 10K, the
integrate method starts
to fail (i.e., produce in-
accurate coexistence
densities), as seen in
Figure 2.
+ Direct Method
The deficiency in the
direct method is that
above a certain step
size in temperature, the
method will not con-
verge, while the advan-
tage is that equilibrium
is ensured upon con-
vergence (assuming the
roots are not repeated).
Mirroring this is the in-
tegrate method, whose
deficiency is that equi-
librium is not ensured
at each step (owing to

Chemical Engineering Education

CPU Time Required to Solve
the Problem for Each
Method at Various Step Sizes

Step Size CPU Time
Method (K) (sec)
Direct 0.02 1.5
Integrate 0.02 1.5
Integrate 1.0 0.05
Integrate 10.0 0.02
Integrate+Direct 10.0 0.02

E 180
> 140
5 120

200 240 280 320 360
Temperature (K)

200 240 280 320 360
Temperature (K)

the numerical integration scheme), while the advantage is that
relatively large steps in temperature can be taken. A combination
of the two methods, wherein first the integrate method predicts a
guess value for the coexistence densities at the next temperature
(as opposed to using the previously converged values), while the
direct method uses these better guesses to converge to a solution,
would seem to allow for the use of a larger step size in tempera-
ture. Such an approach has been suggested before.1[5"6"2' To this
end, a combined integrate+direct method provided equilibrium
densities in this problem for a step size of 10K with a computa-
tional speed of 75 as compared to the direct method alone.


A novel integration technique, here called integrate has been
presented to solve phase equilibrium problems using equations
of state. This method was shown to result in a computational
speedup of around 30 relative to the direct method, owing to the
larger step size the integrate method allows. Additionally, a com-
bined integrate+direct method proved most useful in using the
best features of both approaches. Future work will look at both
the integrate and a combined integrate+direct method in the so-
lution of more computationally demanding thermodynamic prob-
lems, such as tracing out liquid-liquid miscibility gaps or in de-
termining the P-T diagram for retrograde systems for compli-
cated equations of state.


The idea for this work has its roots in a special projects
course taken by DPVJ under the advisement of Professor
David A. Kofke. Helpful suggestions by Dr. Paul Mathias
are acknowledged.

1. Fotouh, K., and K. Shukla, Chem. Eng. Sci., 51, p. 3763 (1996)
2. Elliott, J.R., and C.T. Lira, Introduction to Chemical Engineering Thermo-
dynamics, Prentice Hall PTR, Upper Saddle River, NJ (1999)
3. Sandler, S.I., Chemical and Engineering Thermodynamics, 3rd ed., John
Wiley & Sons, Inc. (1999)
4. Smith, J.M., H.C. Van Ness, and M.M. Abbott, Introduction to Chemical
Engineering Thermodynamics, 5th ed., McGraw-Hill Company, New York,
NY (1996)
5. Asselineau, L., G. Bogdanic, and J. Vidal, "A Versatile Algorithm for Cal-
culating Vapor-Liquid Equilibria," Fluid Phase Equil., 3, p. 273 (1979)
6. Michelsen, M.L., "Calculation of Phase Envelopes and Critical Points for
Multi-Component Mixtures," Fluid Phase Equil., 4, 1 (1980)
7. Redlich, 0., and J.N.S. Kwong, Chem. Rev., 44, p. 233 (1949)
8. Prausnitz, J.M., R.N. Lichtenthaler, and E.G. de Azevedo, Molecular Ther-
modynamics ofFluid-Phase Equilibria, 2nd ed., Prentice Hall PTR, Upper
Saddle River, NJ (1999)
9. Press, W.H., S.A. Teukolsky, W.T. Vetterling, and B.P. Flannery, Numeri-
cal Recipes in FORTRAN: The Art of Scientific Computing, 2nd ed., Cam-
bridge University Press, New York, NY (1992)
10. Modell, M., and R.C. Reid, Thermodynamics and Its Applications, 2nd
ed., Prentice-Hall, Englewood Cliffs, NJ (1983)
11. Hoffman, J.D., Numerical Methods for Engineers and Scientists, 2nd ed.,
Marcel Dekker, Inc., New York, NY (2001)
12. Heidemann, R.A., "Computation of High Pressure Phase Equilibria," Fluid
Phase Equil., 14, p. 55 (1983) 0

A Figure 1.
A flowchart
the integrate

Figure 2. >
The satu-
rated liquid
volume (top)
and satu-
rated vapor
from the
compared to
that pre-
dicted from
the integrate
method at
various step

Spring 2002

MR% laboratory



Rowan University Glassboro, NJ 08028

he fluidized bed polymer coating process is a unique
experiment that can have a large impact on student
learning and retention. It was first developed for a
National Science Foundation Novel Process Workshop"''21 and
is a highly visual experiment in chemical engineering pro-
cesses and experimentation. In addition, the coating process
is environmentally benign because it has essentially no vola-
tile emissions. The object of the experiment is to place a pro-
tective coating on a metal object by first heating it in a hot-
air stream and then dipping it into a fluidized bed of thermo-
plastic powder. The powder is contained within a clear plas-
tic cylinder (acrylic) which allows the students to see and
feel the fluidization. At the end of the experiment, students
are able to take home a metal object of their choosing, cov-
ered with a brightly colored polymer.
The experiment can be used throughout the engineering
curriculum. For recruitment at the precollege and freshman
level, the fluid motion of the gas and the brightly colored
particles attracts the attention of everyone in the laboratory.
The brightly colored powders contained within the clear plas-
tic walls of the fluidized bed gives this experiment the pro-
fessional look of an executive desk amusement. Prospective
students and freshmen also are given the opportunity to feel
the water-like quality of the bed using a rod or a ruler. This is
done by asking the prospective student to move the ruler
around in a slumped bed (bed without air flow) and then in a
fully fluidized bed. The resulting look of amazement is one
of those uplifting moments in a professor's life!
Freshmen use the fluidized bed as an example of the engi-
neering measurements of flowrate, temperature, pressure, and
coating thickness. They design an experiment to determine
the desired coating thickness by varying the dipping time and
temperature of the object. Simple Excel plots are produced
from their experiments. Sophomores measure pressure drop
through the distributor plate to determine the relationship
between flowrate and pressure drop. In an advanced fluids
class, the fluidization regimes can be identified from a pres-
sure-drop-vs.-flowrate plot. For transport phenomena, the

combined heat and mass transport of the coating process can
be examined. Many other advanced experiments can be con-
ducted using this apparatus if pressure transducers are placed
in the bed. These measurements allow the students to deter-
mine the bubble size and frequency,'31 the minimum fluidiza-
tion,141 and to characterize the gas-solid flow regimes.15,61
This experiment is compact and cost-effective; the cost of
fabricating the equipment is about $830. The colored poly-
mer powder makes the experiment enjoyable to watch and to
collect data. Student feedback has been extremely positive.

The business of polymers is a major component of the pro-
cess industry and represents a significant area of opportunity
for the chemical engineering profession. The field encom-
passes many technologies, ranging from polymerization
processes used for chemical production of materials to
fabrication processes needed to transform the materials
into usable products.
The use of polymeric material continues to expand. Ad-
vanced polymers are being developed for use in emerging
areas of technology such as medical devices, smart packag-
ing systems, fuel cells, and electronic device fabrication.
Conventional plastics find extensive use as a material of con-
struction for many products common in daily life. Their low

Robert Hesketh received his BS from the University of Illinois and his
PhD from the University of Delaware. His research is in the areas of reac-
tion engineering, novel separations and green engineering.
C. Stewart Slater is Professor and Chair of Chemical Engineering at
Rowan University. He received his BS, MS, and PhD from Rutgers Uni-
versity. His teaching and research interests are in separation and purifi-
cation technology, laboratory development, and novel processes for
Stephanie Farrell received her BS from the University of Pennsylvania,
her MS from Stevens Institute of Technology, and her PhD from New
Jersey Institute of Technology. She has research expertise in the field of
drug delivery and controlled release.
Michael Carney received his BS from Rutgers and is currently a gradu-
ate student at Rowan University He is currently the Pilot Plant Manager
for Johnson Matthey in West Deptford, NJ.

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

weight, resistance to weather and wear, and economical pro-
duction, make them attractive alternatives to glass, metal, and
wood for use in products ranging from food and beverage
containers to recreational equipment to automobile compo-
nents to building materials.
Coating processes fall in the area of polymer fabrication
technologies along with molding, extrusion, casting, form-
ing, and calendaring. In parts that must be constructed of metal
for structural reasons, a plastic coating may be applied for
decorative and/or functional purposes such as electrical in-
sulation, corrosion protection, and abrasion resistance.170'l
Fluidized bed coating is a commercially important process
in many industrial fields. The main uses of fluidized bed coat-
ing techniques are in the pharmaceutical industry for con-
trolled-release coatings on drugs and for microencapsulating
drug components;'1"12' protective coatings for glass contain-
ers and other components;"31 coating of particles as small as
50%m;"141 fluidized-bed electrostatic coating;"5' chemical va-
por deposition in a fluidized bed for metal-coated
microspheres for inertial confinement fusion targets;"6171 and
fluidized bed coating of aluminum."]8 Additional examples
of fluidized bed coating of materials are available (a recent
search in chemical abstracts yielded over 800 references).
The major advantage of powder coating a substrate is that
these processes use no solvents and thus provide an environ-
mentally friendly alternative to older techniques such as dip-
ping, brushing, and spraying."9'20 Fluidized bed coating is a

Industrial Applications of Fluidized Beds

Polymeric Materials

* Gas-phase polymerization of polyethylene
* Production of silicon for the semiconduc-
tor industry

Biochemical Cultivation of microorganisms for the food
and pharmaceutical industries

Chemical Synthesis

Petroleum Processing

* Phthalic anhydride
* Fischer-Tropsch synthesis of hydrocarbons
* Acrylonitrile, maleic anhydride, activated
carbon, calcination, roasting of sulfide
ores, chlorination, reduction
* Fluid catalytic cracking (FCC) for
production of gasoline from oil
* Coal gasification
* Thermal cracking of naphtha petroleum
fractions to produce ethylene and
* Fluid coking

Combustion Coal combustion
Solid waste incineration
Steam raising

Physical Operations

novel process that offers the advantages of efficient use of
materials (near 100%), the ability to coat irregular shapes,
high coating rates, simple and inexpensive equipment require-
ments, process automation, and smooth and continuous coat-
ing applications.

Fluidization finds application in many important industrial
processes. Examples of fluidization are given in Table 1. In
fluidization, a gas or liquid is passed through a bed of solid
particles that is supported on a perforated or porous plate. In
fluidized bed coating, air is passed through a bed of polymer
particles. These particles become fluidized when the frictional
force of the air on the particles equals or exceeds the weight
of the bed. The minimum velocity required for fluidization
can be determined by equating the weight per unit area of the
bed to the pressure drop of air through the bed. At this mini-
mum fluidization velocity, all of the powder particles become
suspended and the bed exhibits liquid-like behavior. This can
easily be observed through the clear plastic wall of a labora-
tory fluidized bed. As shown in Figure 1, at gas flowrates
less than the fluidization velocity, the bed is a fixed bed and
there is no movement of particles. At flowrates above mini-
mum fluidization, this bed of Geldart class A particles'2" first
expands and then at higher velocities, bubbles appear.
For a given system, minimum fluidization velocity can be
determined from a pressure drop vs. air velocity diagram as
shown in the flow regimes figure. The value of the minimum
fluidization velocity can be compared with correlations pre-
sented by Wen and Yu'22' to get an effective particle size of
the powder mixture

-Fixed or Packed Bed Fluidized Bed-

Air Flowrate

Fixed Bed FLOWRATE Bubbling
Minimum Fluidization

Figure 1. Fluidization regimes.

* Coating metal objects
* Drying of solids
* Adsorption of solvents

Spring 2002

S 2dpg(pps-pg)g
umf- (.3.7)2+0.040 O -33.7
dpPg 9 3




superficial velocity at minimum fluidization, m/s
gas viscosity, kg/(ms)
gas density, kg/m3
particle density, kg/m3
9.81 m/s2
particle diameter, (m)

This paper does not intend to review the theory and equa-
tions governing the fluidized bed process. Instead, the reader
is referred to the many texts on the subject for additional in-
As air flow is increased above the minimum fluidization
velocity, the bed may exhibit behaviors ranging from smooth
fluidizaiton to bubbling fluidization to fast fluidization and
pneumatic transport, in which the particles are transported
by the air stream. Smooth fluidization is desirable for opti-
mal performance in the powder-coating process.
The liquid-like nature of the fluidized powder bed allows a
metal object to be easily dipped into it. The metal object is
preheated to a temperature above the melting point of the
polymer prior to being dipped. As the hot metal object is
dipped into the bed, the polymer particles contact and melt
on the hot surface. Additional heat is transferred from the
object through the initial polymer layer to additional layers
of polymer. After the coated metal object is removed from
the bed, it is then allowed to cool. In many of the dipping
operations, the outer layer of polymer powder has not com-
pletely melted on the object. To give a smooth texture to the
surface coating the object may be reheated, allowing this outer
layer of particles to melt and become incorporated with the coat-
ing. For a given object, the thickness of the coating is depen-
dent on two process variables: preheat temperature of the ob-
ject and the amount of time it is submersed in the powder bed.

The purpose of this experiment is to introduce students to
basic measurements of temperature, pressure, flowrate, and
film thickness using a fluidized bed coating unit. By con-
ducting this experiment, they will also be introduced to the
chemical enigneering operation of fluidization. The experi-
ment is broken into two parts. The first part is a demonstra-
tion of the basic fluidization regimes. Students operate a labo-
ratory fluidized bed and take measurements to generate a clas-
sical pressure-drop-vs.-flowrate diagram to determine the
minimum fluidization flowrate for the system. During this
part of the exercise, students get a chance to observe the be-
havior of the fluidized bed over a wide range of air flowrates.
In the second part of the experiment, the participants are
charged with conducting coating trials to determine process

conditions (preheat temperature and dip time) necessary to
achieve a specified coating thickness on sample objects.
This experiment is designed to be a cost-effective bench-
scale experiment that can be easily integrated into lower-level
courses like other experiments we have developed at Rowan
University.'26-291 Our goal is to excite lower-division under-
graduate students in chemical engineering while at the same
time imparting some of the process aspects of emerging tech-
nologies. We recognize that larger pilot-scale processes in
fluidization or polymer processing have a role in senior
courses and unit operations laboratories and are presented in
the literature. The use of the experiment described in this paper
will have the added effect of generating excitement in stu-
dents for these subsequent experiences.

The following is the laboratory experiment developed for
a freshman engineering laboratory course that we call the
Freshman Engineering Clinic.'30-311 The experimental system
is shown in Figure 2. A team of students (typically three to
four) conducts the experiment, and it can be easily conducted
in a 3-hour lab period with minimal assistance. The experi-
ment was developed with the support of the National Sci-
ence Foundation for a workshop on teaching novel processes
in the curriculum (see Figure 3).
Freshman Laboratory Objectives
Using a calibration curve, convert the rotameter read-
ings in mm to a flowrate in mL/min
Measure the temperature of an object using a bare wire
Measure the pressure of the inlet air stream using a
Bourdon gauge.
Measure the pressure difference across a fixed and a

Figure 2. Fluidized bed apparatus.
Chemical Engineering Education

- Substrate

- Manometer


... Rotameter


fludized bed using a liquid filled manometer.
Estimate the thickness of a polymer coating from know-
ing the surface area of an object and the masses of the
coated and uncoated objects.
Determine the optimum temperature, dipping time, and
fluidization regime to obtain an average coating of 0.025
Explain the effect of temperature and dipping time on
the coating thickness of
an object.
Safety Considerations .
Specific hazards of this lab .*,
include the heating of metal
objects to very high tempera-
tures. Students are asked to
wear appropriate gloves and '
to use tongs when possible
when handling these hot ob-
jects. This is especially true
ivith the use of the heat guns.
Safety goggles or glasses are
required since there is a pos-
sibility of fine powder, hot
objects, or line breakages en-
tering the eyes of the partici-
pants. The polymer powder
used in the coating process is
very fine and will produce Figure 3. Fluidized bed ex
dust. Loading of the fluid bed Workshop on Novel
column should be done to Robert Ybarra, C. St
and Robert I
minimize exposure of the par-
ticles. A cloth or Kimwipe
should be secured over the top of the column when oper-
ating at high air-flow rates where entrainment of the pow-
der can occur.
Relevant Data
The coating material is functionalized polyethylene copoly-
mer-based powder (Polyarmor Powder-PFS Thermoplastic
Powder Coatings, Big Spring, TX). It has a particle density
of p =0.934 g/cm3. The particle size distribution is not given
by the vendor, but a recommended literature value for the
average polyethylene particles size is 75 p m.i110 The polymer
melting point is 1050C (221F). The polymers can be ob-
tained in a variety of interesting colors, such as safety yel-
low, safety green, safety red, safety orange, and light blue.
The use of the various colors and mixing of them adds visual
fun to this experiment. The fluidized bed is made of Plexiglas
and the coated substrates are easily seen in a laboratory. The
metal substrate used is a steel washer (Hillman #270067)
1/2-inch nominal size. The exact dimensions are OD 1.376
in (3.495 cm), ID 0.563 in (1.430 cm), and thickness 0.117 in
(0.297 cm). The substrate surface area is calculated to be 3.19
Spring 2002


in2 (20.6 cm2). Any size washer or metal may be coated, and
trial runs are suggested to optimize the experiments. We rec-
ommend that the students do not dip their team member's car
keys or residence hall keys into the fluidized bed!
Required Equipment
The fluidized bed can be fabricated from clear plastic
(acrylic) tubing and sheets. The clear plastic tube is glued to
a flat sheet flange and a rubber gasket material is used to seal
the distributor plate to the unit. The distributor plate is a poly-
ethylene porous sheet
manufactured specifically
for heat treating fluidized
beds. This plate can be
Obtained from POREX
Technologies. The drop
mechanism for the metal
S( samples was fabricated by
bending stainless steel
tubing into a U shape and
running a thin metal cable
through the center of the
tubing. An attachment de-
vice is placed at one end
to hook a wire loop to it
and the other end has an
adjustable stop. The wire
is weighted using washers
to obtain a fast drop into
ment being conducted for NSF the fluidized bed. The re-
cesses. Drs. Carolyn Lee, maining components
t Slater, Dave Kauffman, shown in Figure 2 are
th (lefunctio of rir shown in Figure 2 are
th (eft to rit) standard laboratory units
given in Table 2.
Experimental Procedure
Part 1: Investigation of Fluidization Regimes
The first experiment in the freshman laboratory is to have
the students investigate the flow regimes of the fluidized bed.
In this experiment they identify the equipment and identify
the point of incipient fluidization. They are asked to place a
ruler into the fluidized bed and feel the difference between a
slumped bed (no air flow) and a fully fluidized bed. Students
always marvel at the fluid-like behavior of particles. The next
step is to obtain a fluidization curve of bed pressure drop as a
function of air flowrate shown in Figure 4. In addition, in
advanced courses students can make a plot of bed height as a
function of air velocity to determine the Geldart particle clas-
sification of the powder.21 In this experiment the freshmen
use several measurement devices: air pressure gauge, rota-
meter, ruler, and a U-tube manometer. This helps fulfill one
of the objectives of our Freshman Clinic, which is to intro-
duce process measurements. This experiment also serves a
role of reinforcing graph preparation from experimental data
using spreadsheet tools such as Microsoft Excel.

At the end of the laboratory the students Determination of Minimum Fluidization
1. Submit afludization chart (graph). This includes bed
pressure and height vs. flowrate.
2. Show the value of the minimum fluidization velocity that is Z 3.0 *
determined on the graph (Figure 4). 2.5
3. Submit a laboratory notebook yellow sheets containing data 2.0
and a sample calculation of the flowrate. 1.5
4. Submit sample calculation of step 2 in the next experiment.
Experiment ,'0 0.5
Part 2: Polymer Coating 0.0
0.OOE+00 1.00E-04 2.00E-04 3.00E-04 4.00E-04 5.00E-04
The next part of the experiment is where students coat the Flowrate mix's )
metal samples. They are told that a metal part is to be used in Figure 4. Detn of m u
Figure 4. Determination of minimum fluidization.
an application where the rate of heat transfer through the part
is critical. This part will be exposed to a corrosive environ- TABLE 2
ment and a coating is required for protection. Increasing the Parts List for Fluidized Bed
coating thickness increases corrosion protection, but decreases
the heat transfer rate. Initial calculations indicate that a coat- Fluidized bed unit: fabricated by Pemm Corp., Chelsea Industrial Park, $140
ing thickness of 0.02500.001 inches (0.06350.0025 cm) Brockway Road, Wappingers Falls, NY 12590. Phone 914-831-5828;
will maximize corrosion protection while allowing for an (includes polkyethylene porous distributor plate, fine sheet 15-45
adequate heat transfer rate. g m, 0.25"X42"X44", part no. 4902 from POREX Technologies,
Using this problem statement, the students conduct a se- Rotameter, Cole-Parmer, N044-40, SS float, 41,512 mL/min max of air $220
ries of pilot runs in the fluidized bed coating system to deter- Heat gun, Wagner TurboCool H13000, PN 0503835, 1200W $139
mine values for the process variables (pre-heat temperature Handheld thermometer, Omega 26K $ 99
Handheld thermometer, Omega HH-26K $ 99
and dip time) that will produce the desired coating thickness.
Ring stand $ 47
To examine the behavior of the coating process, they con-
duct runs of constant temperature and constant time. They Polyarmor powder (3-lb sample)-PFS Thermoplastic Powder Coat- $ 40
Sr o ings, 3400 W. 7th, Big Spring, TX 79720;
are given a range of temperatures that start below the melting or e-mail; Telephone 800-753-5263
point of the polymer (1050C, 221 F) and extend it to 4500F Three-prong extension clamps (two) each $ 20
(2320C). The dip time ranges between 2 and 10 seconds. Tilt stand (for thermometer) $
The students determine an average coating thickness from Stopwatch $ 12
the formula mc = pAt, where mc is the mass of the polymer U-tube manometer 0.24" H,O, McMaster-Carr, #3985K25 $ 40
coating, p is the density of the coating, A is the area, and t is
Castaloy clamp regular holder $ 10
the desired thickness. The mass of the coating is determined
by difference using the electronic balance. A wire is attached Plastic tubing or air lines 25
to the sample and placed on the hook to dip and remove the Max O xplorer toploaing 20$1,500
sample from the fluidized bed. Using the heat gun, the sample
is heated to a temperature approximately 10-15F above the Metal samples and disposable hanging wire
desired temperature. The sample is then dropped into the flu- *Not included in total experiment cost.
idized bed for a given time, then removed. After the sample
has cooled, the wire is removed and the coated sample is weighed
using the electronic balance. To give the sample a more attrac- 0.040 -
tive finish, it can be reheated to obtain a smooth finish. 0.035
An example of the student data is shown in Figure 5. They 0.030
find that the coating thickness can be increased by increasing 0.025 oo
either the coating temperature and/or the time. Many students E 0.020 0 -
also find out that if they use a temperature near or below the o0.015 -'-
melting point of the polymer, the polymer particles do not 0.010o *
coat the metal object! 0.005
At the end of this laboratory the students 0.00o
Submit a summary graph of data from the coating experiment. 0 5 10 15
Submit a summary paragraph on the effect of temperature and Dipping Time (s)
dipping time on the coating thickness. Figure 5. Example of student coating thickness results.
142 Chemical Engineering Education

Predict, based on the data, the time and temperature required
for a coating thickness of 0.05 inches.
Submit laboratory notebook yellow sheets containing data and
sample calculations (showing units).


This laboratory is both meaningful and fun. Students prac-
tice principles of measurement and learn about fluidization,
coating, and environmental principles. They also have fun
coating objects of their own choosing. In addition to the stan-
dard samples, students have coated keys and flashlights made
in an earlier lab. Student feedback on the fluidized bed ex-
periment has been extremely positive. Representative com-
ments are "one of the best experiments in the Freshman
Clinic," "this got me excited about process measurements
and chemical engineering," and "I like the visual coating part."
One of the course evaluation questions, Was the experiment
interesting and information, got a 4.71/5.00 score, making it
the second highest rated of all freshman laboratory experiments.


The experiment helps in student recruitment and retention
and provides a focal point of laboratory demonstrations to
pre-college students. Beyond the visual nature of the experi-
ment are some key engineering objectives that students ac-
complish. Some of the process measurements performed in-
clude: use of a rotameter to measure air flowrate, measuring
the temperature using a bare wire thermocouple, measuring
the pressure of air stream using a Bourdon gauge, and mea-
suring the pressure difference across a fixed and fluidized
bed using a liquid filled manometer. Students also do some
problem solving by estimating the thickness of a polymer
coating from knowing the surface area of an object and the
masses of the coated and uncoated objects and determining
the optimum temperature, dipping time, and fluidization re-
gime to obtain the desired coating thickness. Finally, they ex-
plain the effect of temperature and dipping time on the coating
thickness of an object. We believe that this experiment is both
motivational and an excellent learning experience.

Support for the laboratory development activity described
in this paper is provided for by a grant (DUE-9752789) from
the National Science Foundation through the Division for
Undergraduate Education.

1. Slater, C.S., and R.P. Hesketh, "A Hands-On Workshop on Novel Pro-
cess Engineering," Proc. Conf. ASEE, Session 1526 (2000)
2. Hesketh, R.P., and C.S. Slater, "Novel Processing Workshop," Proc.
Conf. ASEE, Session 1526 (1999)
3. Ramayya, A.V., S.P. Venkateshan, and A.K. Kolar, "Estimation of
Bubble Parameters from Differential Pressure Measurements in Gas-
Fluidized Beds," Powder Tech., 87, 113 (1996)
4. Wilkinson, Derek, "Determination of Minimum Fluidization Velocity
Spring 2002

by Pressure Fluctuation Measurement," Can. J. Chem. Eng., 73(4),
5. Bai, D., E. Shibuya. N. Nakahawa, and K. Kato, "Characterization of
Gas Fluidization Regimes Using Pressure Fluctuations," Powder Tech.,
87, 105 (1996)
6. Yerushalmi, J., and N.T. Cankurt, "Further Studies of the Regimes of
Fluidization," Powder Tech., 24, 187 (1979)
7. Crabb, C., "Powder Coatings Find Cures," Chem. Eng., 108(2), 54
8. Handbook of Plastics, Elastomers, and Composites,3rd Ed., Charles
A. Harper, Ed. McGraw-Hill, New York, p. 638 (1996)
9. Gaynor, J., "Fluidized Bed Coating," Chem. Eng. Prog., 56(7), 75
10. Rodriguez, F, Principles of Polyner Systems, Hemisphere Publishing
Corporation, Washington, p. 431 (1982)
11. Wan, Lucy S.C., and W.F. Lai, "Fluidized Bed Coating of Particles: A
Review," Chin. Pharm. J., 47(3), 185 (1995)
12. Li, Shun Por, Chana R. Kowarski, Kenneth M. Feld, and Wayne M.
Grim, "Recent Advances in Microencapsulation Technology and Equip-
ments," Drug. Dev. Ind. Phann., 14(2-3), 353 (1988)
13. Hinz, Paul, "Powder Coatings on Glass," Fachforum Schichten Glas:
Herstell., Eigenschaften, MessPruefnethoden, Anwend., 155,157(1999)
14. Gaupp, W., "Powder Coating," Recents Prog. Genie Prod., 13(63),
201 (1999)
15. Bellemare, David J., "Web Coatings by Powder Deposition Technol-
ogy, J. Coated Fabr, 27(Oct), 84 (1997)
16. Meyer, S.F., "Metallic Coating of Microspheres, J. Vac. Sci. Tech.,
18(3), 1198(1981)
17. Rouyu, Hong, and Li Hongzhong, "Progress in Coating Ultrafine Par-
ticles Using Chemical Vapor Deposition in Fluidized Bed Reactors,"
Prog. Nat. Sci., 6(3) 269 (1996)
18. Landrock, Arthur H., "Coating of Aluminum with Plastics by the Flu-
idized Bed and Electrostatic Powder Techniques, PLASTEC Note, 18,
p. 18(1968)
19. Reisch, M.S., "Paints and Coatings," Chem. Eng. News, 18 Oct, p. 22
20. Narkis, M., and N. Rosenzweig, eds., Polymer Powder Technology,
Wiley, Chichester (1995)
21. Kunii, D., and 0. Levenspiel, Fluidization Engineering, 2nd ed.,
Butterworth-Heinemann, Newton, MA (1991)
22. Wen, C.Y., and Y.H. Yu, AIChE J., 12, 610 (1966)
23. Yates, J.G., Fundamentals of Fluidized-Bed Chemical Processes,
Butterworths, London (1983)
24. McCabe, W.L., J.C. Smith, and P Harriott, Unit Operations ofChemi-
cal Engineering, McGraw-Hill, p. 165 (1993)
25. Fan, Liang-Shih, and Arvind Varma (eds.), Principles of Gas-Solid
Flows, Cambridge University Press (1997)
26. Hesketh, R.P., and C.S. Slater, "Innovative and Economical Bench-
Scale Process Engineering Experiments," IJEE, 16(4), 327 (2000)
27. Hesketh, R.P., D. Bosak, and L. Kline, "Automotive Catalytic Reac-
tion Engineering Experiment," Chem. Eng. Ed., 34(3), 240 (2000)
28. Hesketh, R.P., and C.S. Slater, "Cost Effective Experiments in Chemi-
cal Engineering Core Courses," Proc. Conf. ASEE, Session 1613(1999)
29. Hesketh, R.P, and C.S. Slater, "Demonstration of Chemical Engineer-
ing Principles to a Multi-Disciplinary Engineering Audience," Proc.
Conf ASEE, Session 2513 (1997)
30. Jahan, K., A.J. Marchese, R.P. Hesketh, C.S. Slater, J.L. Schmalzel,
T.R. Chandrupatla, and R.A. Dusseau, "Engineering Measurements in
the Freshman Engineering Clinic at Rowan University," Proc. Conf
ASEE, Session 1326 (1998)
31. Hesketh, R.P, and C.S. Slater, "Using a Cogeneration Facility to Il-
lustrate Engineering Practice to Lower Level Students," Chem. Eng.
Ed., 33(4), 316 (1999)
32. Hesketh, R.P, J. Jahan, A.K. Marchese, C.S. Slater, J.L. Schmalzel,
T.R. Chandrupatla, and R.A. Dusseau, "Multidisciplinary Experimen-
tal Experiences in the Freshman Engineering Clinic at Rowan Univer-
sity," Proc. Conf ASEE, Session 2326 (1997) 0

MR]^= laboratory




Clarkson University Potsdam, NY 13699-5705

Traditional undergraduate chemical engineering labo-
ratory courses place emphasis on familiarizing stu-
dents with the equipment of unit operations, chemi-
cal reaction engineering, and process control. Few environ-
mentally oriented experiments, such as wastewater treatment,
are available to students. The experimental project described
in this paper is concerned with the application of electrochemi-
cal engineering principles to wastewater treatment. Waste-
water containing toxic heavy metal ions is generated in large
quantities from the microelectronics, metal finishing, min-
ing, and photographic industries. Metal Waste sources and
their characteristics are listed in several publication.1'21
Wastewater may be freed of toxic metal ions by electrode-
positing them in metallic form. An example of this was the
use of a silver tower electrolyzerE31 to recover silver from spent
photographic fixing solutions. Zhou and Chin14'5] described
an eletrolytic process for simultaneously recovering heavy
metal ions from wastewater at the cathode and destruction of
cyanide ions at the anode. Khristoskova and Lazavou'6] stud-
ied an electrolytic process to remove hexavalent chromium
from wastewater. Shifrin, et al.,'7] discussed the advantage of
adding RuO2 to a TiO2 anode to treat wastewater. Hertwig, et
al.,E81 and Tison'1g used a rotary drum electrode to recover
copper. Fleishmann, et al.,[E ] used a fluidized bed as the cath-
ode to plate copper onto metal particles. Bennion and
Newman1]" used a flow-through porous-electrode to remove
copper ions from dilute solutions. Robertson and
Dossenbach[121 developed a gas-sparging cell to improve mass
transfer in an electrolytic cell for wastewater treatment. A
general review of electrochemical removal of metals from
aqueous solutions was given by Kuhn,E13' and a review of elec-
trochemical cell design for metal recovery was given by
Robertson, et al.114] The basic electrochemistry and engineer-
ing principles were discussed by Weininger,11[5 and the envi-
ronmental and economic factors of the electrolytic metal re-
covery process were discussed by O'Keefe and Ettel.'"6

The objectives of this project in a chemical engineering
laboratory course are
To demonstrate to students that toxic heavy metal ions in
wastewater can be reduced by electrodeposition of the metals
at a porous cathode
To enhance students' experience of chemical reaction
engineering by determining the reaction rate constant of the
electrodeposition reaction, its activation energy, and the
effect of mass transfer on the rate constant
To improve students' economic consciousness by evaluating
the current efficiency and energy consumption of the process
Students perform the electrolysis experiment to reduce toxic
metal ions from an initial concentration of 50-250 parts per
million (ppm) to a low level acceptable for discharge as re-
quired by the U.S. Environmental Protection Agency's (EPA)
regulations. By measuring the concentration changes of the
metal ion at various controlled anode-to-cathode cell volt-
ages, temperatures, and water recirculation rates, students
determine the reaction rate constant, its activation energy, and
the effect of mass transport of metal ions on the rate of elec-
trodeposition reaction. By measuring the amount of metal
recovered and the total electric charges used in the electroly-
sis, students calculate the current efficiency and energy re-
quirement of the process and compare the results to those
reported in the literature.

Der-Tau Chin is Professor of Chemical
Engineering at Clarkson University He has
more than 30 years teaching and research
experience in the areas of corrosion and
electrochemical engineering. He is a Fel-
low of the Electrochemical Society and prior
to joining Clarkson, he was a senior re-
search engineer in the Electrochemistry
Department of General Motors Research
Laboratories. He received his PhD from the
University of Pennsylvnia in 1969.

Copyright ChE Division ofASEE 2002

Chemical Engineering Education

In this project, a flow-through electrochemical cell operat-
ing in a recirculation mode in a solution tank is used for re-
covering copper from wastewater containing 50-250 ppm of
cupric (Cu'2) ion and 0.05 M Na,SO4. The cathodic and an-
odic reactions in the cell are



Cu2++2e =Cu

H20=1/202+2H+ +2e

The cupric ion concentration in the wastewater is to be re-
duced by electrodepositing copper on a porous carbon cath-
ode to less than 1.0 ppm to satisfy EPA regulations for dis-
charging copper ions to waste streams."[71
The copper electrodeposition reaction is a first-order reac-
tion with respect to cupric ion concentration in the wastewa-
ter. By assuming a uniform concentration throughout the so-
lution tank and electrolytic cell, the concentration change of
cupric ion at a given set of controlled anode-to-cathode cell
voltage, temperature, and wastewater recirculation rate, can
be described by


with a=--

where C is the concentration of cupric ion in wastewater at
time t; Co is the initial concentration of cupric ion; k is a first-
order reaction rate constant for the deposition of cupric ion at
the cathode; A is the total cathode area; and Vsoi is the volume
of wastewater in the solution tank.
Since the total surface area of the porous cathode is not
easy to determine, one may consider the product, ka, as the
effective volumetric reaction rate constant of the copper elec-
trodeposition reaction. According to Eq. (3), a plot of In C
vs. t would yield a straight line, and the value of ka can be
evaluated from the slope of the straight line. The values of ka
depend on the wastewater temperature, the electrode poten-
tial of the cathode, and the water recirculation rate. The cath-
ode potential can be changed by varying the anode-to-cath-
ode cell voltage in the experiment.
The activation energy, Eac. is evaluated by measuring the
volumetric reaction rate constant, ka, at various solution tem-
peratures at a constant cell voltage and a wastewater recircu-
lation rate according to the Arrhenius equation

ka= kaexp- Eact (4)

where T is the absolute temperature, R is the universal gas
constant, and ka is the frequency factor. According to Eq.
(4), a plot of the logarithm of volumetric reaction rate con-
stant, ka, versus the reciprocal of absolute temperature (1/T)
would result in a straight line, and the activation energy can
be evaluated from the slope of the straight line.
Spring 2002

The instantaneous current efficiency of copper electrodepo-
sition reaction at a given electrolysis time can be evaluated
by calculating the rate of change of cupric ion concentration,
dC/dt, from the experimental C vs. t curve and by comparing
the value to the cell current, I, at the same electrolysis time,
according to Faraday's law

Vsol ,
Current Efficiency(%)= I xl 00 (5)
where F is the Faraday constant (96,500 C/equiv) and n is the
number of electrons transferred in the cathode deposition re-
action (2 equiv/mol). The average current efficiency is ob-
tained by comparing the mass of copper recovered at the end
of a run to the total charge passed during the run according to

Average Current Efficiency(%)= x100 (6)
M fldt

where 9 is the total electrolysis time in seconds, W is the
mass of copper deposited at the cathode, and M is the atomic
mass of copper (63.5 g/mol).
The energy consumption per kilogram of copper removed
from wastewater is calculated by integrating the experimen-
tal cell current and voltage curve with respect to the time
according to
f IEceldt
Energy(kWh/kg-metal)= 0 (7)

where Ec1 is the anode-to-cathode voltage and W is the mass
of copper recovered in kilograms.


Apparatus The experimental apparatus is shown schemati-
cally in Figure 1. It consists of a solution tank, and electro-
lytic cell, a recirculation pump, a control valve, a rotameter,
a connecting pipe, a direct current (DC) power supply, an
immersion heater, and a glass thermistor probe. The solution
tank is a 16-gallon rectangular polypropylene container with
quarter-inch wall thickness. It is equipped with a spigot at
the bottom of the front wall for draining wastewater at the
end of the experiment. The electrolytic cell is a commercial
electrochemical reactor in the form of two concentric cylin-
ders, as shown in Figure 2. The inner cylinder is a ruthenium
oxide-coated titanium mesh anode (commercially called "di-
mensionally stable anode," or DSA) of 2 13/16" outside di-
ameter, 12" long, and 1/16" thick. The outer cylinder is a
porous carbon felt cathode that was pre-coated with a thin
layer of copper. The cathode dimensions are 4 3/4" outside
diameter, 3/8" thick, and 10 3/4" long.

The anode and cathode are placed in a cylindrical polyvinyl
chloride (PVC) casing of 12 1/2" in length and 5 1/2" outside
diameter, as shown in Figure 3. Wastewater containing heavy
metal ions is pumped by the recirculation pump to one end of
the electrolytic cell and flows through the annular space be-
tween the anode and the cathode where the heavy metal ions
are recovered by electrodeposition at the porous cathode sur-
face. The wastewater exits at the other end of the concentric
cell and is continuously recirculated by the pump between the
solution tank and the electrolytic cell through a 3/4" PVC pipe,
as shown in Figure 1.
The solution flow rate through the electrolytic cell is read on
the rotameter and can be adjusted at the control valve to a maxi-
mum of 15 gallons per minute. Two insulated electric cables
connect the anode and the cathode to the DC power supply.
During the electrolysis, the anode-to-cathode cell voltage is set
at the power supply, and the cell current is read on a built-in
ammeter of the power supply. The immersion heater is an elec-
tric heater encased in a stainless steel sheath and Teflon coat-
ing. A power controller is used to adjust the heater output from
0 to 1.0 kW. The heater and the glass thermister probe are con-
nected to a temperature controller that maintains the wastewa-
ter temperature at a constant value for the runs above the room
temperature. Due to the stability of the construction materials,
the maximum wastewater temperature that could be operated
on this equipment was 500C.
Experimental Procedures 1 .Calibration of Copper Ion Selec-
tive Electrode The concentration of cupric ion in wastewater
during the run is measured with a cupric ion selective electrode.
The potential of this electrode with respect to a built-in refer-
ence electrode is first calibrated with the following standard
solutions: 1000 ppm Cu2+ + 0.05 M Na2SO4; 100 ppm Cu2+ +
0.05 M Na2SO4; 10 ppm Cu2+ + 0.05 M Na2SO4; 1 ppm Cu2 +
0.05 M Na2SO4; 0.1 ppm Cu2+ + 0.05 M Na2SO4. The calibration
procedures involve insertion of the cupric ion selective electrode
in 50 mL of a standard solution, addition of 1 mL of an ionic
strength adjuster (5 M NaNO3), and reading the electrode poten-
tials with a digital pH/mV meter. A sample calibration curve show-
ing the potential of cupric ion selective electrode as a function of
cupric ion concentration at 250C is shown in Figure 4.

2. Wastewater Treatment Experiment A simulated waste-
water containing 50-250 ppm Cu2+ and 0.05 M Na2SO4 is pre-
pared in the solution tank by adding an appropriate amount of
cupric sulfate and sodium sulfate salts into 53 L of tap water.
The recirculation pump is turned on for 10 to 15 minutes, until
the solution is well mixed. The initial cupric ion concentration
and solution pH is measured with the cupric ion selective elec-
trode and a combination pH electrode on the pH/mV meter. The
wastewater flow rate (10-60 L/min) through the electrolytic cell
is adjusted at the control valve to a desired level, and a constant
water temperature (25-500C) is set at the temperature control-
ler. Electrolysis is started by applying a constant anode-to-cath-
ode cell voltage (3-6 V) from the DC power supply. The cell

current, cupric ion concentration, and solution pH are mea-
sured at specified time intervals. The run continues until the
cupric ion concentration in the wastewater drops to less than
1.0 ppm to permit discharge of wastewater into the drain sys-
tem. Electric power to the DC power supply, the recircula-
tion pump, and the immersion heater are then turned off. The
solution tank is cleaned and rinsed with fresh tap water. The
electrolytic cell is removed from the solution tank and disas-
sembled for visual inspection of the cell components.
System Cost The costs of the electrolytic cell, the equip-
ment, and the construction materials are listed in Table 1.
The total capital cost (excluding the labor for assembly and
testing) of the experimental system was $6,340 in 1998 U.S.
dollars. The system cost could be reduced to less than $4,000
if a spare pH/mV meter, a pH electrode, and a temperature
controller were available in the laboratory. The material cost
to run an experiment is estimated to be $18.63, as shown in

Figure 1. Schematic arrangement of solution tank, elec-
trolytic cell, recirculation pump, control valve, rotame-
ter, DC power supply, immersion heater, thermister probe,
pipe connection, and direction of solution flow.

Figure 2. Schematic of electrolytic cell.

Figure 3. Photo of the electrolytic cell showing the ar-
rangement of Ti-mesh anode, Cu-precoated porous cath-
ode, and PVC casing prior to assembly.
Chemical Engineering Education

Table 1. The replacement cylindrical cathode can be reused
in several experimental runs. With proper cleaning and rins-
ing after each experiment, we used a single cathode cartridge
six to ten times without affecting the quality of data collec-

80 --------------------------

eo Temperature: 25 C


E 20
.0 0
2 -0

110 100 1000 10000 I
Cupric Ion Concentration (ppm)

Figure 4. Calibration curve showing potential versus
cupric ion concentration of a cupric ion selective elec-
trode at 250C.

Capital Cost of Experimental System
and the Materials Cost for One Experimental Run
in 1998 U.S. Dollars

Item Cost
Electrolytic cell with 3 replacement cathode
cartridges and DC power supply $2,000
16-gallon polypropylene solution tank with cover 160
SPolypropylene recirculation pump with motor 360
Rotameter 70
S* Teflon-coated immersion heater, power controller,
S thermister probe, and temperature controller 1,600
U Combination copper ion selective electrode with
filling solution kit 450
pH/mV meter and combination pH electrode with
filling solution kit and electrode holder 1,200
Other construction materials (plastics, hardware,
connecting pipies, pipe fittings, etc.) 500
TOTAL $6,340

Replacement cathode cartridge ($55 per cartridge,
pro-rated for approximate useful life of six
Experimental runs) $ 9.17
0.025 kg of CuSO, 5HO (technical grade,
S$189.06 per 3-kg bottle) 1.58
0.4 kg of anhydrous Na,SO4 (technical grade,
U $52.38 per 3-kg bottle) 6.98
I 20 mL of 5M NaNO, solution ($45 per 1000-mL
bottle) 0.90
TOTAL $18.63

Spring 2002

The project has been a part of a senior-level chemical engi-
neering undergraduate laboratory course at Clarkson University
since 1998. The course consists of a weekly lab period of 6 hours,
and the project was assigned to a group of 3 to 4 students as a
three-week mini-thesis project. Each student group was able to
complete one experimental run in one 6-hour lab period. Prior to
the first-week lab period, the students were asked to submit a
preliminary lab report describing the equipment setup, experi-
mental procedures, theoretical background for data analysis, and
safety precautions. The instructor then held a pre-lab meeting
with the students and assigned specific objectives to the group.
Since each group was able to carry out only three experimental
runs during the three-week period, the objectives were confined
to one of the following scenarios:
To examine the effect of anode-to-cathode cell voltage on
the reaction rate constant, ka, by conducting the experi-
ment at three cell voltages while holding wastewater
temperature and recirculation rate constant
To determine the activation energy of the copper elec-
trodeposition reaction by conducting the experiment at
three temperatures while holding the cell voltage and
wastewater recirculation rate constant
To examine the effect of mass transfer on the reaction rate
constant, ka, by conducting the experiment at three flow
rates while holding the cell voltage and solution tempera-
ture constant
In the subsequent two weeks, the student group was asked to
make an oral report to the instructor prior to the second and third
lab period. The oral report consisted of presenting the previous
week's experimental results, data analysis, comparison with
known literature results, discussion of the difficulties encoun-
tered in the previous lab period, and modification of experimen-
tal procedures, if necessary. The safety precautions were also
reviewed in the oral report. One week after completion of the
third week's experiment, the group submitted a final project re-
port to the instructor, summarizing all experimental results, theory,
and data analysis

This section presents the typical experimental results collected
by the students. Figure 5 shows a Ti-mesh anode, an unused cath-
ode, and a cathode after the electrolysis experiment. The black
ruthenium oxide coated Ti-anode was inert and stable in the
present electrolytic system, and the electrolysis did not change
its color. The unused porous carbon cathode was dark in color,
although it was pre-coated with a thin-layer Cu from the com-
mercial supplier (the dark color was caused by oxidation of the
thin copper coating by air during storage and shipping). After the
experiment, the color of the porous cathode changed to bright
metallic copper, indicating that a significant quantity of copper
was recovered from the wastewater by electrolysis.

Figure 6 is a semi-logarithmic plot of cupric ion concentration ver-
sus electrolysis time for three controlled anode-to-cathode cell volt-
ages of 3.5 V, 4.5 V, and 5.5 V at 280C and a wastewater recirculation
rate of 38 L/min. In all the runs, the students were able to reduce the
cupric ion concentration in the wastewater from an initial value of 60-
250 ppm to less than 0.5 ppm, permitting discharge of the wastewater
to the waste stream at the end of each experimental run. The concentra-
tion of cupric ion decreased logarithmically with the time as described
by Eq. (3), and the rate of concentration change increased with increasing
cell voltage. Using regression analysis, the effective volumetric reaction
rate constant, ka, could be calculated from the slope of the linear lines.
Reaction Rate Constant Figure 7 is a plot of ka versus cell voltage
for a series of runs at 280C and 38 L/min of wastewater recirculation
rate. The value of ka varied linearly with the cell voltage from 0.0003 s-1
at 3 V to 0.0006 s-' at 6 V, as described by Ohm's law. This indicates
that ohmic resistance of the wastewater played an important role in the
copper electrodeposition reaction.
The pH of the wastewater decreased from an initial value of 6-7 to a
final value around 3 at the end of the run. This increase in acidity was
caused by the anode reaction in Eq. (2), where H20 molecules were
decomposed to 02 gas and Ht ions. The rate of pH decrease was fast at
the beginning of an experimental run and slowed down with increasing
electrolysis time.
Current Efficiency and Energy Requirement The main side reaction at
the cathode was the reduction of H20 molecules to H2 gas and OH ions.

2H20+2e- =H2+20H- (8)
The OH- ions generated from the above reaction neutralized the H+
ions produced at the anode, and thus decreased the rate of pH change at
large electrolysis time. Figure 8 shows the instantaneous cathode cur-
rent efficiency for the copper electrodeposition reaction as a function
of electrolysis time for two controlled cell voltages at 280C and 38 L/
min of wastewater recirculation rate. The instantaneous current effi-
ciency was calculated from the rate of cupric ion concentration change
in the wastewater using Eq. (5). The rate of copper electrodeposition

Anode Un sed Cathode
I Oa ode after
b, 1. ElectrolMpi

Figure 5. Photograph showing a Ti-mesh anode, an
unused cathode, and a bright metallic copper-colored
cathode after the electrolysis experiment.

decreased with decreasing concentration of cupric ions.
At the beginning of the electrolysis, the cupric ion con-
centration in the wastewater was high (50-250 ppm)
and the current efficiency was nearly 100%. As the
electrolysis proceeded, the cupric ion concentration was
reduced, and the current efficiency for the copper
deposition reaction decreased logarithmically with
increasing electrolysis time. At the end of an ex-
perimental run, the instantaneous current efficiency
was typically less than 1%.
The average current efficiency can be calculated by
integrating the instantaneous current efficiency curve
in Figure 8 with respect to the time and dividing the
result with the total electrolysis time of an experimen-
tal run. It can also be calculated using Eq. (6) from the
total amount of copper removed from the wastewater

W=Vsol(Co -Cfinal) (9)
where V, is the volume of wastewater in the solution

1 o Temperature: 28 C
E Flow Rate: 38 L/min

0 o
[] o

0 3 4 .5 V
0 4.S6V 0
0.1 5.5V o

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 2000 22000
Time (s)

Figure 6. Cupric ion concentration versus elec-
trolysis time for three controlled anode-to-cathode
cell voltages at 280C and 38 L/min of wastewater
recirculation rate.




3.0 3.5 4.0 4.5
Cell Voltage (V)

5.0 5.5

Figure 7. Effective volumetric reaction rate con-
stant, ka, of copper electrodeposition reaction
versus cell voltage at 280C and a solution flow rate
of 38 L/min.
Chemical Engineering Education

Temperature: 28 C
Flow Rate: 38 L/min


0 0O

tank and C. and C ina are the initial and final cupric
ion concentrations in the experimental run. The two
methods provided an opportunity for the students to
check the accuracy of mass balance in their experi-
mental measurement. The cell current increased with
increasing the controlled cell voltage. For a given
experimental run, the current was nearly constant. It
typically exhibited a 2-5% decrease throughout the
entire run. Consequently, the product of average cur-
rent and experimental duration was used as the total
charge in Eq. (6) for calculation of the average cath-
ode current efficiency.
Table 2 summarizes the results of average current
efficiency for three-controlled cell voltages of 3 V, 4
V, and 5 V at 280C and 38 L/min of solution recircu-
lation rate. The table also lists the average cell cur-
rent and the initial and final cupric ion concentra-
tions in each run. All the experiments were able to
reduce cupric ion concentration from 210-220 ppm
to less than 0.5 ppm in the wastewater. Although Fig-

0 5000 10000 15000
Time (s)

Temperature: 28 C
Flow Rate: 38 L/min


5 V 3 V

Figure 8. Instantaneous cathode current effi-
ciency for copper electrodeposition reaction for
two controlled cell-voltages at 280C and 38 L/min
of solution recirculation rate.

000310 0.00315 000320 0,00325 0.00330 0.00335
1/T (K"1)
Figure 9. Semi-logarithmic Arrhenius plot of ka
versus 1/T for copper electrode deposition reaction
at the porous cathode.
Spring 2002

ure 8 seems to show that the instantaneous current efficiency for copper
elctrodeposition decreased with increasing cell voltage, the average
current efficiency remained approximately constant at around 25%. This
is because the cell current was larger and the total electrolysis time was
shorter at a higher cell voltage.
The consumption of electric energy per kilogram of copper recov-
ered from the wastewater by electrolysis was calculated by integrating
the product of cell current and cell voltage with respect to the time, as
shown in Eq. (7). The results are listed in the last column of Table 2 for
the three controlled cell-voltage runs at 28'C and 38 L/min of wastewa-
ter recirculation rate. Although the average current efficiency for cop-
per electrodeposition was independent of the cell voltage, the electric
energy consumption increased with increasing cell voltage. The elec-
tric energy per kilogram of copper recovered varied from 11 kWh at 3
V to 17 kWh at 5 V. These values agreed with the work of Zhou and
Chin,E'" who reported a value of 13-30 kWh/kg-Cu for the electrolytic
treatment of a wastewater using a rotating barrel plater.
Activation Energy Some groups carried out the electrolytic wastewa-
ter treatment at several temperatures with constant cell voltage and so-
lution recirculation rate. Figure 9 is an Arrhenius plot of ka versus re-
ciprocal of absolute temperature (1/T) for ten experimental runs over
the temperature range of 28-500C at a constant cell voltage of 5 V and
38 L/min recirculation rate. The data exhibited a linear relationship be-
tween ln(ka) and 1/T, as suggested by Eq. (4). The activation energy,
E ,,, as calculated from the slope of the straight line, was 22 kJ/mol.
Mass TransferAspects ofElectrodeposition Reaction For electrodepo-
sition of metal from a dilute solution, the rate-controlling step is gener-
ally the transport of metal ions from the bulk solution to the cathode
surface. Increasing the solution velocity near the cathode enhances the
mass transfer rate. Figure 10 shows the values of ka over a range of
water recirculation rate of 10-60 L/min at a common cell voltage of 5.5
V and a temperature of 330C. The results indicate that ka increased with
increasing recirculation rate, and thus the velocity of wastewater pass-
ing the porous cathode. The linear relationship between ka and the re-
circulation rate of the log-log plot implies a strong mass transfer influ-
ence on the rate of copper electrodeposition reaction in the present sys-
tem. But if the copper electrodeposition reaction was completely con-
trolled by mass transfer, the reaction rate constant would be expected to
be independent of the anode-to-cathode cell voltage. The fact that ka
Continued on page 155.

Summary of Experimental Results for
Three Controlled-Cell Voltage Runs at 280C and
38 L/min Solution Recirculation Rate

Controlled Average
Cell Cell
Volt (V) Current (A)

Initial Cu2 Final Cu2* Average Electric
Concen. Concen. Current Energy
(ppm) (ppm) Efficiency (%) (

0.2 23 11.3
0.3 26 13.2
0.5 24 17.7

20000 25000



M^]^= laboratory



University Rovira i Virgili Tarragona, Spain

n the coming century, chemical engineers will face many
new challenges. The needs of the chemical industry are
progressively moving from process-oriented engineering
to product-based engineering, and the new environment re-
quires that chemical engineers address a broader body of
knowledge and collaborate with other specialists.'" Hence,
industry expects to hire graduates capable of applying their
understanding without further training, of finding creative
solutions, and of communicating the outcomes. Technical
competence is no longer sufficient if it is not combined with
non-technical abilities such as problem solving, management,
leadership, teamwork, decision making, and ethical respon-
sibility.[2' This has been recognized by the Accreditation
Board for Engineering and Technology (ABET)'3' which
has specified that engineers should demonstrate not only
a broad scientific base but also a set of skills linked to
social capabilities.
As a result, the paradigm of engineering is shifting from
hard engineering to soft engineering, although technical as-
pects are still the core. This shift involves dealing with issues
such as more efficient teaching methodologies, different learn-
ing styles, new learning materials, and the revision of course
syllabi, which must evolve to fit the new paradigm of educa-
tion by switching the emphasis from instructor-based teach-
ing to student-centered learning.'41
Since real problems do not recognize disciplinary bound-
aries, the unit operations laboratory could easily be a suitable
place for a holistic approach to chemical engineering.5"141 In
addition to the classical understanding of unit operations, a
professionally oriented chemical engineering laboratory could
provide creative and critical thinking, the ability to design
experiments, and the capacity to analyze data and draw rea-
sonable conclusions. Simultaneously, the laboratory should
incorporate aspects that are necessary to achieve a global
education of the chemical engineer, such as safety and envi-
ronmental concerns, commercial relevance, troubleshooting,

and design of procedures. A similar laboratory with struc-
tured experiments was recently proposed.151
In response to these expectations, the School of Chemical
Engineering of the Rovira i Virgili University (URV) has a
laboratory that addresses soft skills and requires rigorous
understanding of the basic operations. The course is based
on a constructivist approach, and students learn by forming
their own interpretation of open-ended experiments. The
instructor's role is to guide the students and prevent mis-
conceptions, rather than to transmit formal knowledge to
passive students.

The chemical engineering degree at URV takes five years
to complete. Each course is divided into two fifteen-week
semesters. The courses are run using a credit system in which
one credit is equivalent to ten hours of lectures. The com-
plete degree requires students to obtain 405 credits.
The unit operations laboratory is a nine-credit course given
during the second semester of the third year. By this time the
students have taken the basic subjects, several fundamental

Laureano Jimdnez is Associate Professor in the Chemical Engineering
Department at the URV He holds a BCh and PhD in Chemistry from the
University of Barcelona and has eight years experience in laboratory teach-
ing. His research interests are process synthesis, process modeling, pro-
cess simulation, and teaching methodologies.
Josep Font completed both his BCh and PhD in Chemistry at the Univer-
sity of Barcelona. At present he is Associate Professor in the Department
of Chemical Engineering, URV His research interests are chemical reac-
tion engineering, wastewater treatment, and membrane processes.
Josep Bonet completed both his BCh and PhD in Physics at the Univer-
sity of Barcelona. At present, he is Associate Professor in the Department
of Chemical Engineering at URV His research interests are polymer mod-
eling and molecular simulations.
Xavier Farriol completed both his BCh and PhD in Chemistry at the Uni-
versity of Barcelona. Atpresent, he is Professorin the Department of Chemi-
cal Engineering, URV His research interests are wood science and lignin

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

The laboratory ... simulates a professional environment in which students must
design experimental procedures to meet customers' demands. [It]also addresses other no-less-
important topics such as safety, legal regulations, economy, troubleshooting, and the environment.

laboratories, and a few other major subjects, so they can in-
terpret the basic concepts underlying unit operations. The
course is devoted to classical unit operations (except heat
transfer) and includes water and wastewater treatment.
The planning and execution of the experimental work and
the subsequent interpretation and presentation of the results
is the essence of the course. The students (approximately 60
per year) are organized into teams, usually of three to four
members. Random
teams are preferable
since this promotes a
mixture of learning Insights of unit operation
styles and the develop-
Selectkey variables
ment of interpersonalky
skills. The instructional Planning experiments
objectives are Confidenceand
-' ~~confidence and ^ ^^r

* To test real equip- consistency of results
ment, manage pos- Write the report Tea
sible upsets, and
solve operational Literature search
To design proce-
dures for start-up,
steady-state opera- Practical aspects
tion, and shutdown Analyta methods
Analytical methods
To identify key Deal
variables during soft
normal operations
To search for, con- Figure 1. Team roles and
sult, and interpret
technical documents
To process data and check the mass and heat balances,
physical properties, thermodynamics, transport phenom-
ena, and chemical reaction
To develop decision-making criteria depending on prod-
uct specifications, environmental constraints, legal regu-
lations, safety, and economic reasons
To consider the importance of errors in the validation of
the results obtained
To formulate hypotheses and simplifications to facilitate
the analysis and modeling of unit operations
To optimize the operating conditions according to the
experimental results
To present effective oral and written results and conclu-
The laboratory course was devised around the following
Spring 2002



The students have access to the laboratory for one three-
hour session, four days per week throughout the semes-
They are supervised by one faculty member and one as-
sistant lecturer
The experimental equipment in the laboratory is divided
into three different blocks
Students spend four days completing the classical unit
operations experi-
ments (distillation,
absorption, liquid-
Report editing liquid extraction, and
a set of reactors);
Project management three days for the
water treatment
Leader modules (reverse os-
Plantiming mosis and ion ex-
changer); one day for
working Scheduling the wastewater
treatment plants
Super (flocculation-sedi-
mentation, aerobic-
activated sludge, and
Calculations anaerobic fluidized
Error analysis bed). Each group
Data collection must perform two ex-
,rdand periments from the
cations first block, one from
the second, and the
anization in the laboratory whole of the third
whole of the third
Thus, the maximum involvement of the students in experi-
mental work is forty-two hours. They spend the rest of the
time planning, analyzing results, and reporting.
The methodology is to imitate a professional environment
in which decisions have to be taken, responsibilities as-
sumed, mutual confidence experienced, and tasks pro-
grammed and distributed. The tasks are randomly assigned
to each team. During the course, each team member must
perform, at least once, the role of coordinator, operator,
and analyzer (see Figure 1).

Once the problems are assigned, students are provided with
a simple scheme of the equipment. In turn, we expect them
to find all the relevant information needed to design and con-
duct the experiments and analyze the data obtained. Instruc-


tors merely act as supervisors and assist teams if they fall
into a dead-end situation or if potential safety risks are de-
tected. As the laboratory slogan says, "Good judgment comes
from experience. Experience comes from bad judgment."
The course progresses through a six-step procedure that
must be satisfactorily completed:
1) Experiment preparation and preliminary report
2) Two sessions of experimental work (just one for
block 3, with no possible extension)
3) Intermediate report and new planning tutoring
4) Two additional experimental sessions (only one for
block 2)
5) Technical report
6) Oral presentation and question-and-answer session
This schedule enables better monitoring of the performance
and evolution of students and also provides continuous feed-
back. The preliminary report must contain practical proce-
dures (start-up, routines for steady-state operation, and shut-
down protocols). After students have been tutored, the plan-
ning is accepted on the basis of the experimental aspects,
time management, and analytical methods. Therefore, fac-
ulty efforts are especially important at the beginning of the
course to prevent inaccurate procedures. As students become
increasingly familiar with the methodology, instructors fo-
cus their attention by posing challenging questions that en-
courage creative thinking. Once the plan has been accepted,
the team must reserve the equipment for two laboratory ses-
sions. All experimental data, tasks, and incidents must be de-
tailed in the laboratory notebook for reproducibility. The note-
book is checked periodically and graded at the end of the course.
After the first part of the experimentation, each team must
check the coherency of the data (mass and energy balance)
and draw preliminary conclusions. At this point, a progress
report must be written to compare the results with the model
predictions and to discuss the goals reached. Frequently,
partial results induce changes in the subsequent experi-
mental plan and students are forced to make decision cri-
teria for themselves.
At the end of the second period, the teams must deliver a
technical report of each experiment. The final report is ex-
pected to contain all the valuable information needed to jus-
tify the conclusions in a concise and clear manner. In any
case, students are asked to report the confidence interval of
the results and to provide explanations for the behavior ob-
served and any possible deviations from the theoretical
models. Usually, the results and discussion are presented
in the same section. Finally, the students must propose a
solution in a few lines.

The evaluation is mainly based on the oral presentation of

the final report for each assignment. Each member of the team
is required to give an oral presentation of one randomly cho-
sen assignment. As all members of the team are fully account-
able for all the assignments, they do not know in advance
which one they have to present. After the presentation, the
student is questioned about the statements (from the report
and/or the presentation), the procedures and the conclusions,
as well as any unclear parts of the discussion. Three faculty
members judge the quality of the presentation and the over-
all knowledge of the problem.
Table 1 shows the grading scheme used to obtain the stu-
dents' final qualification. As can be seen, the final grade de-
pends on their knowledge of unit operations, their perfor-
mance in the laboratory, and their personal skills. There is a
good balance (45% versus 55%) between personal skills and
collective skills, but an individual factor assigned to each stu-
dent can increase or decrease the final grade by 10%. With
this factor, we attempt to account for the greater or lesser
involvement of a particular student in the group performance.
This involvement is easily detected in the group's daily work.
The examination at the end of each experiment permits stu-
dents to learn from their own experience and mistakes. On
average, 12% of the students fail at the first attempt and after
additional work just 5% do not qualify. It should be pointed
out that the pre-laboratory and intermediate reports are a cru-
cial part of the learning procedure, so they are used basically
to collect information about course dynamics and as a first-
hand source of feedback.

Nine different experimental set-ups are currently available
in the laboratory. The present structure of the laboratory and
the assignments are the result of the evolution toward design
of a zero-waste disposal laboratory. Table 2 shows a list of
the problems addressed during a typical course. Usually, 3 to
4 new problems are posed every course (i.e., the Murphree
efficiency at the distillation column or the HETP at the ex-

Grading Scheme

Course Component Grade Allocation
Individual Assignment +10%
Methodology: planning, methods, group dynamics 10%
Reproducibility: experiment description in notebook 10%
*Final Report
Editing: structure, distribution, composition, numbering, visual impact 5%
Readability: composition, grammar, conciseness, neatness 5%
Results: goodness of data, proper discussion, appropriate solution 20%
* Oral Presentation
Editing: visual impact, relevant slides, content 5%
Performance: preparation, timing, tone, contact with audience 5%
Question session: fundamentals, experimental, results, evaluation 40%

Chemical Engineering Education

traction column), so the assignment for each experimental
set-up is not unique. All the problems are depicted as poten-
tial real-life cases that are not necessarily limited to the chemi-
cal process industry. This encourages multidisciplinarity and
forces students to the knowledge of related areas such as en-
vironmental engineering.
The laboratory is based on open-ended problems. In con-
trast, in the classroom courses on unit operations, students
solve close-ended problems in a tight environment. In the
laboratory, however, the students face situations in which they

Problem Statement for Each Experiment

Block 1
1. Distillation A client asks for the best economic conditions for operating
a continuous distillation column. The column is fed with an ethanol-water
mixture containing 60% w/w of ethanol and a flowrate up to 25 L/h. The
product composition must achieve 90% of the azeotropic. The reboiler
and pre-heater power are 2 and 0.3 kW, respectively. The feed cost is
$0.5/L and the product is sold at $2/L. The power cost is 0.1 $/kWh.
2. Absorption A customer has to decrease the ammonia content in a waste
air stream (3.2 m3/h) from 15% v/v to less than 1%. An absorption tower
is available where the ammonia could be absorbed with water. The
availability of water is limited.
3. Liquid-Liquid Extraction An industrialist needs to purify 5 L/h of a
binary mixture (45:55 w/w) containing MIBK and acetic acid (HAc). the
MIBK recovered must retain a maximum of 2% HAc. Two technologies,
liquid-liquid extraction with water and conventional distillation, have to
be checked. Operating conditions must be optimized.
4. Reactors A small industry produces an aqueous stream contaminated
with ethyl acetate (20 g/L). The acetate content must be reduced to 3 g/L
or less before disposal. Hydrolysis using sodium hydroxide (NaOH) is
proposed as treatment. Unlimited 0.2 mol/L NaOH solution is available at
low cost. The customer asks for the reactor type and the operating
conditions to comply with legislation.
Block 2
5. Reverse Osmosis In an isolated farm, a saline source is used (1.25 gNcl/
L) to purify water. The farm needs 100 L/h of water with conductivity up
to 50 pS/cm. A second-hand reverse osmosis module is available without
technical characteristics. Operating conditions must be set so that water is
produced with minimum energy expenditure.
6. Ion Exchanger A laboratory received a new piece of equipment. The
product specification is 150 pS/cm. The conductivity of the crude water,
which is freely available, is close to 1 mS/cm. The water production cycle
must attempt to maximize the pure water yield.
Block 3
7. Flocculation/Sedimentation The acidic effluent of a galvanic plant
must be treated. The plant generates 5 m'/h of water with 1000 ppm of
copper. A preliminary design and scale-up of the treatment plant must be
made using the data collected from the 100 L/h physical-chemical
treatment plant.
8. Sewage Treatment Plant The Mayor of a city on the Mediterranean
coast (with a population of 100,000) is aware that the urban sewage is
more refractory than expected and cannot be biologically treated. The
sludge plant needs to be re-engineered, so preliminary scale-up from
laboratory data (2 L/h) must be carried out.
9. Denitrification Plant A modem farmer has implemented a sophisticated
hydroponics system, but the purged water (5 m3/day) does not comply
with environmental law. Biological denitrification is proposed as
treatment. Experimental data can be retrieved from a 0.1 L/h lab-scale
equipment. Scale-up must be done.

Spring 2002

encounter build-up equipment, but no precise step-by-step
guidelines. Thus, they need to understand the principles of
unit operations, since mathematical models are of no use for
a rapid qualitative interpretation of how each variable influ-
ences the unit operation performance. For instance, the
reboiler power or the top condenser duty are seldom set in
distillation design. On the contrary, they are usually calcu-
lated using the reflux ratio required for a particular separa-
tion. It is noteworthy that the main disturbance for students
is that the reboiler power is fixed in a real distillation col-
umn, so that its capacity and the reflux are constrained.
One of the characteristics of this laboratory is the mini-
mum waste production, where students experience aspects
covered by several elective subjects (i.e., environmental en-
gineering). At the moment, waste production has been re-
duced almost completely at no significant additional cost. In
fact, the only waste that cannot be treated in situ is the pre-
cipitate from the flocculation-sedimentation plant (copper
hydroxide and calcium sulfate), which is sent to a qualified
waste-treatment company. The other wastes are either reused
or treated. For instance, the water-ethanol mixture used in
distillation is reused throughout the course in the same equip-
ment. Extraction requires more complex treatment. Refined
methyl-isobutyl ketone (MIBK) is directly reused. After ex-
traction, however, the acetic acid and water mixture still con-
tains a certain amount of MIBK, which is recovered by dis-
tillation. Moreover, since distillation of the acetic acid-water
mixture is difficult, the MIBK-free mixture is reused as a
feed for the activated sludge plant. Notice that the principal
environmental impact produced by the laboratory is due to
the life-cycle impact of the electrical power used.

When problems are assigned, students who are not famil-
iar with problem-solving schemes often miss the point, and
continuous assessment is required. Hence, instructors act more
as counselors, redirecting students efforts, than as formal
teachers of structured knowledge. Once students realize that
there is no single solution or approach to each problem, they
connect up the disparate pieces as a whole and develop their
problem-solving skills exponentially.
The next critical point is when students prepare the inter-
mediate report. They tend to make a list of results and do not
estimate errors, check the robustness of the experiment, or
explain deviations. Discrepancies and unexpected results are
ways of identifying and correcting mistakes. The final re-
ports are generally well-structured and carefully edited, and
above all, the discussion of the results explains the de-
pendence with the process variables and makes compari-
sons with model predictions.
At the end of the course, students are required to anony-
mously answer a feedback questionnaire. Table 3 summa-
rizes the answers from the last course, which are similar to

those of previous years. In general terms, the responses show
that the course is well-accepted and they are particularly fa-
vorable for those statements about what students perceive
they had learned/improved. This demonstrates that the edu-
cational objectives of the course were mostly attained.

It should be pointed out that students feel much more com-
fortable with this kind of teaching, although they demand
more supervision. We should focus our efforts on providing
students with training based on creative thinking, critical cri-
teria, and problem-solving skills rather than providing them
with a better understanding of unit operations, which they
are capable of learning for themselves. Overall, the students'
main objection was the amount of time they had to devote to
the course, which was greater than the time scheduled. We
should point out, however, that the extra time was spent on
planning, data analysis, and reporting, since self-motivation
often led students to go beyond the requirements of the course.

Many favorable comments have been received during the
four years that the laboratory course has been running, which
encourages us to continue pioneering the application of new
educational methodologies in Spain. In the near future, this
course will be part of an even more ambitious one-project-
per-year strategy to stress holistic education in the chemical
engineering undergraduate program.


We have designed a unit operations laboratory course with
the main objective of providing chemical engineering under-
graduate students with creative thinking skills, criteria for
decision making, problem-solving and communication skills,
and the capability to monitor and operate unit operations. The
laboratory, therefore, simulates a professional environment
in which students must design experimental procedures to
meet the customers' demands. The course also addresses other
no-less-important topics such as safety, legal regulations,
economy, troubleshooting, and the environment.
Faculty members act as mere advisors, so students are not
subjected to passive teaching. Student skills are developed
through open-ended problems and by posing Socratic ques-
tions that enhance critical thinking. Obviously, we do not
expect students to magically develop their entire individual
potential within this laboratory, but as the course advances,
most of the students become capable of designing experi-
ments, analyzing results, and suggesting solutions. Simulta-
neously, they improve their self-confidence and learn to make
attractive presentations. Faculty members must provide mo-
tivation when students fail and continuous assessment is
needed if students are to make headway. The laboratory pro-
cedure (preliminary report, two-day experiments, intermedi-
ate report, and two additional experimental days) forces stu-
dents to adopt a very useful stop-and-go procedure.
The benefits of the course largely make up for the tremen-

Results of the Feedback Questionnaire
(Class Size, 60 students: Score, 0-strongly disagree, 5=strongly agree)

Ave. St.
Question Dev.
1. After the informative session I understood the course methodology 3.8 1.0
2. The laboratory schedule was well-programmed and coordinated 3.1 1.2
3. The problems matched my academic background 3.6 1.0
4. The laboratory exigency fitted my previous formation 3.5 1.1
5. Facilities and infrastructure were suitable 3.1 1.1
6. The duration of the course was appropriate 2.8 1.4
7. In this course, I improved
A. My basic knowledge of the unit operations 4.0 1.2
B. My management and organizational abilities 3.4 1.0
C. My report-writing skills 3.1 1.1
D. My oral-presentation skills 3.6 1.0
E. My documents/information-search skills 3.4 1.0
8. All holistic aspects were taken into account in the final grading 3.8 2.3
9. The team performed reasonably well 4.7 1.9
10. I prefer this stye of teaching to a pre-set lab methodology 3.4 1.5
11. Instructors were always available 3.8 1.0
12. Instructors made sure that the experimental objectives were clear 2.9 1.7
13. Instructors supervised the team performance sufficiently 2.9 1.9

dous effort required. The driving force for all of us is the same
as for the students-the excitement of learning by doing.

I. Wintermantel, W., "Process and Product Engineering: Achievements,
Present and Future Challenges," Chem. Eng. Sci., 54, 1601 (1999)
2. Nguyen. D.Q., "The Essential Skills and Attributes of an Engineer: A
Comparative Study of Academics, Industry Personnel, and Engineer-
ing Students," Global J. Eng. Ed., 2, 65 (1998)
3. "Criteria forAccrediting Engineering Programs," Accreditation Board
for Engineering and Technology (ABET) , Bal-
timore, MD (1999)
4. Hurst, K.D., "A New Paradigm for Engineering Education," Proc.
ASEE/IEEE Frontiers in Education (1995)
5. Giralt, E, M. Medir, H. Thier, and F.X. Grau, "A Holistic Approach to
ChE Education. Part 1. Professional and Issue-Oriented Approach,"
Chem. Eng. Ed., 28, 122 (1994)
6. Giralt, F., A. Fabregat, X. Farriol, F.X. Grau, J. Giralt, and M. Medir,
"A Holistic Approach to ChE Education. Part 2. Approach at the Intro-
ductory Level," Chem. Eng. Ed., 28, 204 (1994)
7. Miller, R.L., J.F. Ely, R.M. Baldwin, and B.M. Olds, "Higher-Order
Thinking in the Unit Operations Laboratory," Chem. Eng. Ed., 32,
146 (1998)
8. Davies, W.A., and T.A.G. Langrish, "Putting Commercial Relevance
into the Unit Operations Laboratory," Chem. Eng. Ed., 29, 40 (1995)
9. McCallum, C.L., and L.A. Est6vez, "Introducing Process-Design Ele-
ments in the Unit Operations Lab," Chem. Eng. Ed., 33, 66 (1999)
10. Middelberg, A.P.J., "Laboratory Projects. Should Students Do Them
or Design Them?" Chem. Eng. Ed., 29, 34 (1995)
11. Marrero, T.R., and W.J. Burkett, "Introducing Industrial Practice in
the Unit Operations Lab," Chem. Eng. Ed., 28, 128 (1994)
12. Myers, K.J., "Troubleshooting in the Unit Operations Laboratory,"
Chem. Eng. Ed., 28, 120 (1994)
13. Abu-Khalaf, A.M., "Getting the Most Out of a Laboratory Course,"
Chem. Eng. Ed., 32, 184 (1998)
14. King, J.A., "Incorporating Safety into a Unit Operations Laboratory
Course," Chem. Eng. Ed., 32, 178 (1998)
15. Munson-McGee, S.H., "An Introductory ChE Laboratory Incorporat-
ing EC2000 Criteria," Chem. Eng. Ed., 34, 80 (2000) 0
Chemical Engineering Education

Metal Recovery from Wastewater
Continued from page 149.

increased with increasing both cell voltage and water-flow rate sug-
gests that the copper electrodeposition reaction in the present dilute
electrolytic system was under a mixed control of mass transfer and
ohmic resistance of the wastewater. Increasing temperature increased
the diffusivity of cupric ion and electric conductivity of wastewater,
and thus the rate of electrodeposition of copper at the cathode.

This experimental project is relatively safe. No toxic or harmful
chemicals are used, and it is carried out at the ambient temperature
to 50'C. Wastewater in this temperature range would not scorch
the operators in the event of accidental contact. Although the elec-
trolysis releases 0, and H, gases from the solution tank, their quan-
tities are small and do not pose a fire hazard. The equipment should
be placed in a well-vented area, however, and students are advised
not to light any matches near the experimental area. To avoid elec-
tric shocks, students are advised to dry their hands before operating
power switches. They should also wear safety goggles to protect
against accident spill of wastewater samples or standard solutions
during the measurement of cupric ion concentration.
Although the present experiment was assigned to students as a
three-week mini-thesis project, it can also be used as a short ex-
perimental project to be completed in one 6-hour lab period. In this
case, the experimental objectives will be confined to a determina-
tion of the reaction rate constant for a given set of cell voltage,
water recirculation rate, and temperature. To enhance students' eco-
nomic consciousness, they should also perform the current effi-
ciency and energy requirement calculations and compare the re-
sults with the literature values.


An experimental project to enhance students' experience in elec-
trochemical reaction engineering and wastewater treatment has been
developed for use in a senior-level undergraduate chemical engi-
neering laboratory course. The project involves electrolysis experi-
ments to reduce cupric ions in wastewater by electrodepositing them
in metallic form at a porous cathode. By measuring the concentra-
tion changes of cupric ion as a function of electrolysis time at vari-
ous controlled cell voltages, temperatures, and water-flow rates,
students determine the reaction rate constant, its activation energy,
and the role of mass transfer in the electrodeposition reaction. They
also calculate the current efficiency and energy requirement for
recovery of copper from the wastewater. The project is also suit-
able for use as a short experiment in a single lab period, or it can be
assigned to students as a mini-thesis project to be completed in a
period of three to four weeks.

The work described in this paper was supported by the U.S. Na-
tional Science Foundation under a grant DUE9650068.
Spring 2002

o.oor Temperature: 33

Figure 10. Log-log plot of the effective volumetric


at Cell Voltage: 5.5 V
.l ... Temperature: 33 C

10 20 30 40 50 60
Flow Rate (L/min)

Figure 10. Log-log plot of the effective volumetric
reaction rate constant, ka, as a function of
wastewater flow rate through the ellectroytic cell
at a cell voltage of 5.5 V and 330C.

1. Palmer, S.A.K.. M.A. Breton, T.J. Nunno, and D.M. Sullivan, Metal/Cya-
nide Containing Waste Treatment Technologies, Noyes Data Corporation,
Park Ridge, NJ; p. 10-100(1988)
2. Reed, A.K., J.F. Shea, T.L. Tewksbury, R.H. Cherry, and G.R. Smithson,
"An Investigation of Techniques for Removal of Cyanide from Electro-
plating Wastes," Report for Project No. 12010EIE, Battelle Columbus
Laboratory, Columbus, OH (1971)
3. Hickman, K., W. Weyerts, and O.E. Goehler, "Recovery of Silver," Ind.
Eng. Chten., 25, 202 (1933)
4. Zhou, C.-D., and D.-T. Chin, "Copper Recovery and Cyanide Destruction
with a Plating Barrel Cathode and a Packed-Bed Anode," Plat. and Sur-
face Fin.. 80(6), 69 (1993)
5. Zhou, C.-D., and D.-T. Chin, "Continuous Electrolytic Treatment of Com-
plex Metal Cyanides with a Rotating Barrel Plater as the Cathode and a
Packed-Bed as the Anode," Plat. and Surf. Fin., 81(6), 70 (1994)
6. Khristoskova, S., and D. Lazavou, "Electrochemical Purification of
Chromium(VI) in Waste Water," Nauchni Tr Plovdivski Univ., 22, 153
7. Shifrin, S.M., VS. Varygin, V.K. Golyshev, I.G. Krasnobrod, and E.V.
Khosid, "Selection of Anode Material for Electrochemical Treatment of
Waste Water," Zh. Prikl. Krim., 52(7), 1648 (1979)
8. Hertwig, K., H. Bergmann, and E Nieber, "The Rotating Cylinder Cath-
ode: A Novel Electrochemical Reactor for Electrochemical Effluent Treat-
ment," Galvanotech.. 83, 1696 (1992)
9. Tison, R.P., "Copper Recovery Using a Tumbled-Bed Electrochemical
Reactor," J. Electrochem. Soc., 128, 317 (1981)
10. Fleishmann, M., J.W. Oldfield, and L. Timakoon, "Electrochemical Re-
moval of Copper Ions by Use of a Fluidized Bed Electrode," J. Appl.
Electrochem., 1, 103 (1971)
11 Bennion, D.N., and J. Newman, "Electrochemical Removal of Copper
Ions from Very Dilute Solutions," J. Appl. Electrochem., 2, 113 (1972)
12. Robertson, P.M., and 0. Dossenbach, "Stirring by Gas Introduction and
Its Application in the Electroplating Industry," Oberflaeche-Surf., 22(9),
13. Kuhn, A.T., and R.W. Houghton, "The Electrochemical Treatment of
Aqueous Effluent Streams," in Electrochemistry of Cleaner Environments,
edited by J. O'M. Bockris, p. 98, Plenum Press, New York, NY (1972).
14. Robertson, P.M., J. Leudolph, and H. Mauret, "Improvements in Rinse
Water Treatment by Electrolysis," Plat. andSurf Fin., 70(10), 48 (1983)
15. Weininger, J.L., "Electrochemical Recovery of Metals from Waste Wa-
ter," AIChE Svmp. Series, 79, 179 (1983)
16. O'Keefee, T.J., and V.A. Ettel, "The Electrolytic Recovery of Metals from
Aqueous Solution," Electrochem. Soc. Symp. Proc., PV87-7, 103 (1987)
17. U.S. Environmental Protection Agency Regulations, "Electroplating Point
Source Category," Guideline 40, CFR 413 (1992) O

MM a laboratory


Determination of Liquid Diffusion Coefficients

University ofAlicante Alicante, Spain

Diffusion is the process by which matter is transported
from one part of a system to another as a result of
random molecular motions. The rate of many
mass-transfer operations is determined by molecular dif-
fusion, so finding it is important for predicting rates of
mass transfer. Diffusion in liquids and solids is a slow
process, with diffusivity in liquids about 8 x 10-6 m/s and
in solids about 2 x 10-9 m/s.'01
There are many accurate methods available to measure dif-
fusion coefficients, explained in detail in the literature,"l-31 but
when we are designing a laboratory experiment for students
to measure diffusion, accuracy is not as important as is visual
insight into the phenomena. This paper describes a simple
experimental method to determine diffusion coefficients in
liquids that works well for laboratory classes.
While the theoretical background of diffusion in liquids is
described in detail in many textbooks," -3 an introductory treat-
ment is given here as an immediate reference. For diffusion
in unsteady state and without chemical reaction, Fick's first
law of diffusion is

aCA V2
=DABV2CA (1)

where DAB is the diffusion coefficient for A in a stationary
liquid B, CA is the concentration of A at time t and position
(x,y,z), and V is the gradient operator. Solutions of the diffu-
sion equation can be obtained for a variety of initial and
boundary conditions.1' We are going to consider diffusion in
one direction (z) in a system where the diffusing substance
(A) and the stationary substance (B) occupy an infinite re-

Francisco Rulz Bevld is Professor and Head of the Chemical Engineer-
ing Department at the Alicante University (Spain). He received his PhD
from Valencia University (Spain). He conducts research in phase equilibria
and in holographic interferometry applied to mass transfer.
Marfa del Mar Olaya L6pez received her BS in chemistry and her PhD in
chemical engineering. She is currently Assistant Professor at the Univer-
sity of Alicante (Spain). Her research interests include phase equilibria and
polymer structure, properties, and processing.

gion, and the initial state is
t=0 for ZCA=CAo and for Z>O-CA=O
This can be, for example, a column of clear water resting on
a column of A solution at t=0 (see Figure 1). After a time t,
those molecules ofA close to the interface have diffused across
the column of water. The concentration of A is dependent on
the position and time and can be calculated by

CA(z,t)=-CA erfc z (2)
2 2 ABt
where erfc(x) is the complementary error function.
Considering that erfc(O) = 1, it is clear from Eq. (2) that
CA 2-CAo

a) t-0

Clear water Co ]1

0 C,.J2 C C

SuJlI ion ofA A

b) t


0 cj C,.,

Figure 1. Unsteady-state diffusion experiment
in liquids:
a) at time t=0 and b) at time t=t.

Copyright ChE Division of ASEE 2002

Chemical Engineering Education

Figure 2. Qualitative concentration profiles
for different times for nonstationary diffusion
experiments as predicted by Eq. (2).

~-- -- u -a
(a) (b)
Figure 3. Experimental method to obtain
diffusion coefficients: a) before discharge,
and b) after concentrated solution is dis-

Figure 4. Initial situation in the diffusion cell:
dimensions of the cell, sample extraction points,
and initial limits of the solutions.
Spring 2002

at z = 0 for all t > 0, even though the volumes of both solutions are not
equal. This is because this model considers that both columns are infinite.
Therefore, this model will be adequate only for short experiments where
the concentrations at the top and the bottom of the diffusion cell do not
change with time. Figure 2 shows a qualitative representation with con-
centration curves as a function of position and time that can be obtained
using Eq. (2).
If the column of clear water is changed to a diluted solution of A, Eq. (2)

CA(Z,t)=CA,m (AO,M -CAO,m)erfc 2 (3)
2 2 DABt

where CAOM and CAOm are the concentrations of the solute A in the more
concentrated and diluted solutions, respectively.

The apparatus that we propose using to obtain diffusion coefficients in
liquids is shown in Figure 3. It consists of
1 A poly methyl methacrylate (PMMA) diffusion cell where four syringes
have been incorporated at positions where the samples will be taken out of
the system.
> A funnel or tank with the A solution that will be discharged slowly to the
bottom of the diffusion cell (which previously contains clear water or a
diluted solution) through a pipe with a capillary tube in the extreme.
The experimental method consists of the following steps:
1) Clear water or diluted solution is placed inside the diffusion cell.
2) The funnel is filled with the more-concentrated solution of A. The solution
should reach the extreme of the capillary tube and no bubbles should be
3) This capillary tube is placed inside the diffusion cell, on the bottom of the
vessel under the column of water (or the more-diluted solution).
4) The key of the funnel or tank is opened and the solution starts going out to
the vessel, raising the column of water very slowly as a piston. The
objective of the discharge process is to raise the water column (with a lower
density) to have the system prepared in the initial conditions to begin the
diffusion experiment. This process should be done carefully in order to
avoid mixture of both solutions and to maintain a sharp interface between
the solutions.
5) The previous process is finished when the two columns reach the desired
volume. The key in the funnel is closed. In Figure 4 the initial situation in
the diffusion cell is shown with dimensions and positions for sampling, in
accordance with our laboratory experiment.
6) The system is maintained in this way (avoiding movement) until the
experiment is finished (we consider two or three weeks).
7) Samples from each syringe are taken out every two or three days. The
sample volume (5 ml) is negligible compared to the initial volume.
8) Samples are analyzed, using an adequate technique for the solute used, to
obtain concentrations.
9) When a set of concentrations as a function of position and time is obtained,
Eq. (2) or (3) is used to calculate the experimental diffusion coefficient DAB.
This calculation should be done by optimization. A spreadsheet can be
prepared to perform nonlinear regression with DAB as the parameter that
should be optimized to minimize the objective function


for the diffusion experiment with ethylene glycol are two
(4) aqueous solutions of ethylene glycol, 5 and 20% mass.

O.F.= (CA(exp) C A(cal))

where CA(exp) is the experimental concentration of A obtained ir
each position of the diffusion cell, and CA(ca is the concentrate(
calculated using Eq. (2) or (3).
Two experiments are proposed using the same apparatus:
Diffusion of the electrolyte CuSO4
> Diffusion of ethylene glycol
The first of these experiments was selected because student
see how the color, concentrated in the bottom of the bottle at
slowly spreads through the diffusion cell. Concentrations of C
as a function of position and time are obtained using UV-spec
copy (Jenway 6300 spectrophotometer). The wavelength sel
was 800 nm. Due to the small volume of sample avail;
microcuvettes were used. Students prepare standards with kr
concentrations of CuSO, to obtain the calibration curve.
The experiment with ethylene glycol was chosen to intro
gas chromatography as an analytical method to obtain the cor
tration as a function of position and time. The apparatus
Shimadzu GC-14A with an AOC-14 Auto Injector and an
tronic integrator C-R64 Chromatopac. The column was a 2 rr
8" Chromosorb 102. The column temperature was 180'C, anm
tection was carried out by thermal conductivity (TCD) if v
and ethylene glycol were to be analyzed to check mass balance
by flame ionization detector (FID) if only ethylene glycol was
lyzed. The helium flow rate was 30 cm3/min. The internal
dard method was applied for the quantitative analysis, using t
nol as the standard.


Diffusion of CuSO. In Table 1 the concentrations of CuSO,
function of position and time obtained in the diffusion experii
for CuSO4 are presented. Initial conditions were a column of i
water resting on a column of an aqueous solution of CuSO4
Figure 4). These experimental results are used to obtain the d
sion coefficient for CuSO4

Dcuso =4.9xl0-6cm2/s

Figure 5 shows experimental points and calculated curves with this
diffusion coefficient. Diffusion coefficients are strongly concen-
tration-dependent; therefore, comparisons should be done for simi-
lar values of initial concentration. In the literature we found a pa-
per14j where aqueous diffusion coefficients for CuSO4 are deter-
mined for different concentrations using a diaphragm cell technique
(see Table 2). The initial concentration that we have used in the
experiment is 60 g/L. Therefore, according to these authors, a value
between 4.86 x 106 cm2/s and 4.95 x 10-6 cm2/s should be obtained
for the concentration that has been used in the laboratory, which is
consistent with our result.
Diffusion of ethylene glvcol The concentrations obtained for the
ethylene glycol experiment are shown in Table 3. Initial conditions

For this experiment, the ini-
tial interface (z=0) was placed
at 13.5 cm high in the diffu-
sion cell. These experimental
results are used to obtain the
diffusion coefficient for eth-
ylene glycol by optimization
using Eq. (3). The diffusion

Concentration of
CuSO4 as a Function
of Position and Time
(z=0 at the initial
interface between
the two liquid
Initial concentrations:
Concentrated solution
is 60 g/L and diluted
solution is clear water

time z Conc.
(h) (cm) CuSO4 (g/L)

46 8 0.00
6 0.00
4 0.00
2 1.96
70 8 0.00
6 0.00
4 0.00
2 3.86
95 8 0.00
6 0.00
4 1.02
2 8.13
120 8 0.00
6 0.03
4 1.58
2 9.33
143 8 0.06
6 0.22
4 2.15
2 10.2
174 8 0.15
6 0.24
4 2.78
2 11.2
193 8 0.48
6 0.52
4 3.25
2 12.0
215 8 0.50
6 1.04
4 4.65
2 13.5

Diffusion coeffi-
cients at 250C for
copper sulfate in
water obtained by
the diaphragm cell

Cone. 106 DAB
CuSO4 (g/L) (cm2/s)
0 8.50
16 5.64
32 5.37
48 5.23
56 4.95
64 4.86
96 4.45
128 4.24
159 4.07
191 3.95
224 (saturation) 3.83

Concentration of
ethylene glycol (EG) as
a function of position
and time
(z=0 at the initial
interface between the
two liquid columns).
Initial concentrations:
concentrated solution is
20% mass and diluted
solution is 5% mass

time z Cone. EG
(h) (cm) (% mass)

21 6.5
40 6.5
71 6.5
158 6.5
192 6.5


Chemical Engineering Education

coefficient obtained is

Dethyleneglycol = 1110-6cm2 /s (6)

Figure 6 shows the experimental points and calculated
curves with this diffusion coefficient. Fernmndez-Sempere,

coefficients at
250C for ethylene
glycol (EG) in
water obtained
using holo-
graphic interfer-

WEG 106 D,,
(% mass) (cm2/s)
5.0 11.35
22.5 9.33
40.0 7.82
62.7 5.59
99.2 2.70

et al., 5I determined the diffusion co-
efficient for aqueous solutions of
ethylene glycol at different initial
concentrations and 250C using the
holographic interferometric tech-
nique. Table 4 shows the results pub-
lished by these authors. The result
that our students obtained in the
laboratory using a much more rudi-
mentary or basic technique is con-
sistent with the literature data.
The reproducibility of the ob-
tained diffusion coefficients is dif-
ficult to evaluate, but we can give
an approximate value of 10% when
results obtained using a good labo-
ratory practice are considered.

Figure 5. Experimental and calculated (-) concentrations
profiles at different times obtainedfor the experiment with
Copper Sulfate (curves calculated for the experimental
value D=4.9 x 10'6 cm2/s and Eq. (2).
Diffusion for ethylene glycol in water

*t=21 h
5 t=40 h
4 a t=71 h
3 ot=158 h
3 o0 t=192 h

5 6 7 8 9 10 11 12 13
EG (Amass)

Figure 6. Experimental and calculated (-) concentrations
profiles at different times obtained for the experiment with
ethylene glycol (curves calculated for the experimental
value D=11 x 106 cm'/s and Eq. (3).
Spring 2002

We have described a simple laboratory experiment to intro-
duce diffusion in liquids to students. We designed a diffusion
cell (made of poly methyl methacrylate) to obtain diffusion co-
efficients in liquids; Equations (2) and (3) can be used to obtain
the experimental diffusion coefficients for CuSO4 and ethylene
glycol, respectively, when the concentration as a function of
position and time is previously obtained in diffusion experiments
using the diffusion cell and the methodology proposed; and to
validate the diffusion coefficients obtained in this study, com-
parisons with the values previously determined by other authors
using high accuracy techniques have been made. The results
show that it is possible to obtain good results for the diffusion
coefficients using the methodology proposed in this paper.


The experiments described in this paper are integrated in a
group of transport phenomena laboratory classes for second-
year students. The objective is to provide students with clear in-
sight to the phenomena explained in the classroom by using simple
experiments. One drawback is that several weeks are required, but
most of the students feel that this laboratory helps them under-
stand diffusion phenomena and how a diffusion coefficient can be
evaluated from data and by using a model. The experiments on
unsteady-state diffusion in liquids are completed with another in
steady state (vapor diffusion), described by Nirdosh, et al.'1]


CA concentration of A as a function of position and time
D diffusion coefficient for A in a stationary liquid B
erfc(x) complementary error function
O.F. objective function
t time
z direction for diffusion
o initial
M concentrated solution
m diluted solution
exp experimental
cal calculated
infinite dilution

1. Crank, J., The Mathematics of Diffusion, 2nd ed., Oxford University Press (1975)
2. Tyrrel, H.J.V., and K.R. Harris, Diffusion in Liquids: A Theoretical and Experi-
mental Study, Butterworth & Co. Ltd. (1984)
3. Cussler, E.L., Diffusion: Mass Transfer in Fluid Systems, 2nd ed., Cambridge
University Press (1997)
4. Emanuel, A., and D.R. Olander, "Diffusion Coefficients of Copper Sulfate in Wa-
ter and Water in n-Butyl Alcohol," Chem. and Eng. Data., 8(1), 31 (1963)
5. Fernlndez-Sempere, J., F. Ruiz-Bevia, J. Colom-Valiente, and F. MAs-P6rez, "De-
termination of Diffusion Coefficients of Glycols," J. Chem. Eng. Data, 41, 47
6. Nirdosh, I., L.J. Garred, and M.H.I. Baird, "Low-Cost Mass Transfer Experiments:
Part 6. Determination of Vapour Diffusion Coefficient," Chem. Eng. Ed.. 34(2),
158(2000) 0

Diffusion for CuSO4 in water

8 II |t-=46h
m t=70 h
v 6 t_" h

S4 ot-143 h
x t=174h
2 t=-193 h
=-215 h

0 5 10 15 20 25 30
CuSO4 (g/n)

e, classroom




Louisiana Tech University Ruston, LA 71272

Students in the undergraduate heat transfer class seem
to become more excited about the subject when they
begin solving realistic problems that somehow con-
nect to their experience. While many of these problems can
be solved using the approximations of one-dimensional sym-
metry, a large body of interesting and relevant problems must
be tackled with two-dimensional (2-D) methods. This paper
describes a simple method for solving these problems using
any of a number of spreadsheet programs, such as Microsoft
Excel, Corel Quattro Pro, Lotus 1-2-3, etc. We have success-
fully used this method in junior-level heat transfer at Louisi-
ana Tech University for the past two years.

Various approaches are available for solving 2-D problems.
Analytical solutions to engineering problems are highly de-
sirable due to the elegant connection that becomes visible
between physical and mathematical principles. For a few
simple geometries, methods such as separation of variables'
can be applied, or solutions to characteristic differential equa-
tions may be available,[2] but they cover only a small fraction
of the possible problems.
Graphical methods[3,41 have been used for many years to
produce solutions for situations requiring qualitative or ap-
proximate answers. Information about these methods is avail-
able in several textbooks, but graphical techniques may be
perceived as excessively approximate compared to the nu-
merical methods that are so accessible today. Although no
statistics to this effect are known, the sense is that the graphi-
cal methods are seldom taught.
The explosion of the information age has provided ready
access for engineers and students to high-powered desktop

The method can .., be extended by solving
problems with time dependence (transient
problems), problems with geometries that
would benefit from rectangular rather
than square elements, geometries with
edges at oblique angles, and even
three-dimensional problems.

machines that are suited for numerical solutions to heat trans-
port and other engineering problems. While commercial soft-
ware such as ANSYS, PDEase, FlexPDE, etc., can tackle two-
and even three-dimensional problems, extremely useful 2-D
solutions using the Finite Difference Method (FDM) can be
easily obtained by students or engineers with an ordinary
spreadsheet. Furthermore, the process of setting up the prob-
lem, including formulating the boundary conditions, laying
out the geometry statement, determining the convergence
conditions, etc., reinforces the understanding of heat trans-
fer principles by the student. Obtaining the solution by
this process also promotes understanding of how com-
mercial solvers work.

Ronald S. Besser has been Associate Pro-
fessor of Chemical Engineering at the Louisi-
ana Tech University Institute for Micro-manu-
facturing since 1999. He holds a BS in chemi-
cal engineering from U.C. Berkeley, and an MS
and PhD in materials science and engineering
from Stanford University. His research and de-
velopment interests are in chemical-MEMS and
sensors, thin-film materials, plasma deposition
and etching, sub-micron processing, device
physics, and characterization.

@ Copyright ChE Division of ASEE 2002

Chemical Engineering Education

The value of convenient spreadsheet programs for solving
a variety of chemical engineering problems has been previ-
ously demonstrated in various sources.1' While the descrip-
tion here applies solely to heat transfer, a nearly identical
approach can be used to solve problems of mass transfer, fluid
flow, electric current flow, mechanical stress, etc., because
of analogous mathematical descriptions.

The FDM starts by taking the system under study and di-
viding it into a large (but "finite!") number of rectangular
elements. Each element is assumed to be isothermal, i.e., the
entire element exists at a single temperature. At the center of
each element is a "node" or "mesh point" with a unique iden-
tifier based on its position in the nodall network" or "mesh."
Integer subscripts (m,n) relate to position on an x-y axis sys-
tem with a discrete value range. Figure 1 displays this setup.
In order to solve for the temperatures in the system, we
need temperature derivatives for insertion into the heat equa-
tion. Consider the temperature Tm of an arbitrary element
(m,n) as shown in Figure 1. The first derivatives (in x and y)
are written by assuming linear variation of temperature be-
tween node points. Since second derivatives are just first de-
rivatives of first derivatives,


aT aT
ax m+1l ax a_1
mn x m,n
2 2

Tm+l,n -Tm,n Tm.n -m-l,n
Ax Ax Tm+l.n -2Tm.n +m- (1)
Ax (Ax)2

Figure 1. Mesh representing the body or
system under study. T n is the uniform
temperature of the
shaded element.
Spring 2002

Similarly, for the vertical direction

02T Tm.n+l -2Tm,n +Tm,n- (2)
y2 Lmn (Ay)2
We know the three-dimensional heat equation that relates
conductive fluxes and heat generation to the time rate of
change of the temperature of a system as

Sa T)+} k( T) +- ,T)+ PCPT (3
_ak + ak + =k (3)
ax ax ay 3y az z at

We can apply some simplifying assumptions that neverthe-
less hardly reduce the usefulness of the equation by impos-
ing steady-state conditions, no generation, and a thermal con-
ductivity that is temperature independent. The result is
LaPlace's equation, which, when we make the additional as-
sumption of 2-D symmetry, i.e., T is constant in z, is

a2T a2T
+ =0 (4)
ax2 ay2

Inserting the second derivatives (Eqs. 1 and 2) into Eq. (4),
taking the case of a square mesh (i.e., Ax = Ay, assumed
throughout the rest of this article), and doing some algebra to
solve for Tm.n, we get

Tmn m+.n +Tm-ln +Tm,n+l +Tm,n- (5)
In other words, the temperature of interest of a location within
the "bulk" of the system, and not at a boundary, is just the
average of the four temperatures surrounding it.

We can apply this result immediately to solve a real prob-
lem. Consider a metal plate 0.9 m x 0.9 m in size that has its
edges held at constant temperatures, as shown in Figure 2.
We ask, what is the temperature field that develops in the

298 K

( 4


273 K

Figure 2. Metal plate 0.9 m x 0.9 m
in size with edges held at constant

plate once steady-state conditions are attained?
We set up a simple spreadsheet, as shown in Fig-
ure 3, with a cell representing the nodal tempera-
ture of each 0.1 m x 0.1 m element in the plate.
The nodes are at the center of each element, ex-
cept at the boundaries. The boundary nodes sit at
the edge of their elements, and the elements are
half the size of the nodes in the center portion
(note that the corner elements are one-fourth the
size of the central elements).
The spreadsheet is set up by first turning off
any limitations on circular references. In Excel,
this is done by going to Options and selecting
the Calculation tab, then choosing manual cal-
culation. The number of iterations and conver-

gence criteria are also set there. These are important steps, as
without them, Excel will return errors when copying the nodal
equations, leading to untold frustration.
Now the perimeter temperatures can be input as constants.
Then the nodal equations are entered at the interior points.
The equation for cell B9, for example, is

=(A9+B8+C9+B10)/4 (6)
Once typed into B8, the equation can then simply be copied
and pasted to the rest of the interior cells.
After setting up the equations, hitting F9 causes the spread-
sheet to calculate a number of times set by the calculation
limit or the convergence limit ("maximum change" as labeled
in the Excel Calculation option) entered previously. Repeated
presses of F9 guarantee that the solution converges before
reaching the number of iterations limit.
This problem converges almost immediately (in 80 itera-
tions) using a maximum change criterion of 0.001. Reducing
the size of this convergence limit will result in a higher preci-
sion solution at the cost of increased CPU time. The accu-
racy of the solution depends on how closely the mesh ap-
proximates the actual geometry. In general, accuracy improves
as the node spacing decreases. Accuracy can be checked by
halving the node spacing and recalculating a solution. The
calculation has reached its highest accuracy if the two solu-
tions are found to be essentially the same. If substantial dif-
ference exists, the process of reducing node spacing and re-
calculating is continued until the difference diminishes.
The solution, shown in Figure 4, was also graphed using
the surface plot option in Excel. The plot gives an excellent
view of what is going on with the plate's temperature field.
Higher spatial resolution could be obtained by setting a smaller
increment size, resulting in a larger number of cells. Though
execution time trades off with resolution, even highly resolved
arrays iterate quickly with a current-model PC. Moreover,
the linear nature of the equations being evaluated tends to
prevent the occurrence of computational instabilities that
sometimes appear with iterative methods.

M il4 Pi l, E'dgis vf i ,'nian. T-mpBrai.,a
del. 0 1 m,

1- a D 1' 1) r r ,1 0 D ]-I
at it 2t A2 aB ; 2t a
2i's I] '' uA 0 CL I:I Ci 'O At '0 J J '3
S273 CID CI ,u 00 a O GE D uu or u ,.1P
S273 0 0 D| 'u0 r 0E 0 r') iu A 73
2773 'l0 O0 O1 00 CA0 O1 A 11C 37'3
27 A0 1 I:i Ar0 lA (I' I Au C'i] AD 3w3
ST 00 I i 0 00 0 00 C3'
272i '30 'a AI -' c0 all Q :o '2E' 313
272 7"3 2i" *'3 2-_ 3 2

Figure 3. Spreadsheet set up to solve for the temperature distribution
of the plate in Figure 2. The interior nodes consist of formulae (as
shown for B9), while the perimeter values are constants, as shown.

The above example is especially easy because of the con-
stant boundary conditions (BCs). Changes in heat transfer
mode (e.g., convection or radiation from a solid) or a change
in material (transition to a region of different thermal con-
ductivity) necessitate more complicated equations in the
boundary cells. Several textbooks list these boundary condi-
tions for various cases.6,7'1 The ability to derive arbitrary BC
equations,[8',9 however, gives one the confidence to attack a
variety of problems.
The basic approach to deriving a BC equation is to per-
form a heat balance on the boundary element of interest. Since
we have assumed the absence of generation, this amounts to

Iqin =0 (7)
where the q,, are rates of heat transfer from adjoining cells.

.1 :'B X'Cs 3s5 B1 d %' 3li 2W % M Ln
I: 2' E ?676 A16 301JL jluj 31! 6; iK 1 a 7-
*-3 2 %7 1. 3 30, 3 ; Li 30Sj 10 ;3.) -"j, 373
3 N ; 3M" 6 3. T 3 ` iE. 373
r3 ?7868 5 1 I IH: Z %E 301 .i A t9 a 373
2'3 7" A" W, e ..U6 3021 3 i .11 1 32 .. .' 1i
r 7. 3 I3 3I73 8 iT 2B6 5 30 i It'9 i 6" 371
3-7 :.6 77P 1 15 2 3 3 1 _q I 3 jil 12'.j rI
'J '3 271 3 :-j 7i 7 ? :2j 731

Figure 4. Solution of slab problem after 80 iterations.
Chemical Engineering Education

A surprising array of complicated problems can be solved using this
method that cannot be directly solved with analytical methods.

This can be illustrated by example. Consider convection above
a horizontal surface, as illustrated in Figure 5. Based on the
figure and Eq. (7),

q +q2 +q3+qc =0 (8)
Some students may see a comfortable analogy between this
equation and Kirchoff's current law in electric circuits, i.e.,
the currents entering a circuit node must sum to zero. Now,
using Fourier's Law for the conduction rates,
q =-kA =-kA (9)
dx Ax
and the standard expression for the convection transport rate
q, =hAAT (10)
we have

k(A (lm) (Tm- Tmn .+k(Ax)(1m)(Tmn-Tm,n)
2 Ax Ax

k(A)(Tlm ) + mhAx(lm)(T Tm,)=0 (11)
S2 )Ax

For this equation, we have assumed an arbitrary depth of
the system of 1 m. In the 2-D symmetry that we have adopted,
all properties of the system, including its structure and tem-
perature, are constant with depth (i.e., the z direction). The
arbitrary choice of 1 m simplifies the arithmetic and permits
calculated secondary quantities to be considered on a per-
meter-of-depth basis. The flux "faces" at the left and right
ends of the cell are only a half-width tall since the element
sits at an edge. The Fourier Law terms are cast to have a
positive sign by listing the T terms in the order of exterior
temperature minus interior temperature with respect to the
element being analyzed.
After canceling and solving for the node temperature, we

Tm.-n *Tm.n Tm.1,n

qi q3

Figure 5. Schematic for deriving
boundary equation for the case of
convection above a horizontal

Spring 2002

Tmn h Tm+ln +2Tm,nI +Tm _,n + AxToi (12)
I k )
2 2(+ kAx

This equation is inserted into the corresponding spreadsheet
cell. References to the convective heat transfer coefficient
(h), thermal conductivity (k), element width (Ax), and fluid
temperature (T) can be made by giving these variables names,
making it easier to transcribe and debug equations.

The effects of heat generation may be included by adding a
term to the heat balance of an element. This applies to bal-
ances done on both edge cells and interior cells. Consider
first the edge cell with convection from its top surface that
we analyzed above. With the generation term, Eq. (7) be-
gqin +qV=O (13)
where 4 is the volumetric rate of heat generation (W/m3)
that is considered to be uniform within the element. The vol-
ume of the element is given by V. Now Eq. (13) becomes

ql q2+q3 +q +4(Ax)2(m)=0 (14)

After making the substitutions we made above and applying
some algebra, the nodal temperature is found to be
T =

1 2h (Ax)2
I h. Tm+i,n +2T,.n_- +T-,_,, + AxT, +q +
2 2+A k k


Similarly, applying this approach to an interior element yields
an equation analogous to Eq. (5)

Tm+I,n +Tm-l,n +Tm,n+l +Tm,n-1 + (Ak)2
Tm,n 4 (16)

By assigning different generation rates to different areas
within the system, problems involving non-uniform genera-
tion can be attacked. Problems of this type are in general very
difficult to solve. Because of its relative simplicity, this method
can help one set up and solve many situations involving non-
uniform generation and thereby build intuition about this class
of problem.


Fourier's law written for 2-D permits the determination of
the heat rate at an arbitrary point given knowledge of the
spatial temperature distribution. The heat rate can be treated
as a vector quantity and solved at an arbitrary direction by
this means, but for the purposes of junior-level heat transfer,
it is sufficient to determine fluxes that are parallel to the x-
and y-axes.
Figure 6 shows an interior element where conduction is
occurring. We know the heat transfer rate from element A to
element B from Fourier's law by

q A-B -kA AT = kAx(lm) TA TB (17)
Ax Ax
By now applying this formula along a vertical or horizontal
length of several elements, and adding up the heat rate con-
tributions of all these elements, we can calculate the total
rate across any plane in the system. Not only is this capabil-
ity useful for determining heat transfer rates for specific cases,
but it also permits one to check the self-consistency of any
solution by making sure that the heat balance around any
boundary is correct.
The perimeter heat fluxes of the metal slab above were cal-
culated as shown in the spreadsheet of Figure 7. Here the
heat rates from each side of the body have been determined
by first calculating the rates on a cell-by-cell basis and then
adding them up. As we see from the figure, the heat rates
from all the sides, when summed, add to zero-as they should
for a body in steady-state without generation.


We further illustrate the method by solving a more practi-
cal, real-life problem taken from our heat transfer course. The
problem shows the utility of this method, as solutions by other
routes would only come with much difficulty. The problem
is stated this way:

A large, horizontalfiberglass slab serving as a
floor is heated by hot air passing through ducts
buried in it, as shown in the cross-section in Figure
8, where S = 160 mm. The square ducts are centered
in the fiberglass, which is exposed to the ambient
above and insulating earth below. For the case with
the top surface and duct surfaces at 25 and 85 C,
respectively, calculate the heat rate from each duct,
per unit length of duct. The thermal conductivity of
the fiberglass is 2.5 W/(mK).

The spreadsheet solution to this problem is shown in Fig-
ure 9, with key equations spelled out for clarity. An area in

the upper left allows input parameters (whose cells are named
as variables) to be changed so that the problem can be solved
for a variety of cases. The solution exploits the inherent sym-
metry of the problem by solving only a segment of the sys-
tem, yielding the heat rate for a half-duct, which is then
doubled to arrive at an answer.
A detail of the spreadsheet shown is the labeling of the

Figure 6. Heat transfer rate from an
internal element A into another
internal element B.

Figure 7. Determination of heat rates from all sides of the
metal slab pictured in Figure 2. The sum of the rates
equals zero, verifying the solution of the
temperature distribution.

Figure 8. Fiberglass slab with embedded ducts for
heating by air flow. The top surface of the slab is
a room floor, while the bottom surface
lies against insulating earth.
Chemical Engineering Education

-* H a' .. ... .. f.... .. ... :
SW 1 6 ... ... .. ? 1 ff i ai
s a. I *,r 1 ,
i ~ i 5 Si ni I5' r l b *'i- Ui C- -
-" :** E: 1- r r j r -
r- .' a. akn 5 a. 3. fl,. a j-
K ._* a, I. i s n- c: .-

b *k i-i. -. L- B1-l l -^ I_ liJY p * a. 'i- ij
'. 5. as 3 : 1.; C .'c
IK*h -3 ________r_______
13 1' 1 i |r
Is M... U;-..u q>.* .q.~ ja jii. o.. a

lii S.313. "I3.39
ot~ ll nal,

- -IAB C OHE r| I 1 J I | |I L I M I N

Tempi 85 C
Temp2 25 C
k 2.5 W/m/K
delx 0 02 m

row distance
0 0.000
1 0.020
2 0.040
3 0.060
4 0.080
5 0.100
6 0.120
7 0.140
8 0.160
9 0.180
10 0.200
11 0.220
12 .240

=delx*A8 stanc


e 0
n 0

qlop 14.1176 29.3319



25 25 25
367328 381 81 40.9816
474689 49.9 2 54.9760
56.4739 59.2 97 64.3467
635962 66.1 00 70.4785
69.0259 71 1 55 74.5820
730662 747141 773932
76.0101 77.2925 793415
78.1081 79.1044 80.6829
79.5498 80.3338 81.5730
80.4758 81.1082 821039
80.9863 81.5193 82.3358
81.1486 81.6468 82.3999

002 0.04 0.06
1 2 3


25 25 25
45.7821 52.5292 54 3348
646176 85 85
72.7124 85
76.8854 85
79 3507 85
809353 85
81.9974 85
82.7129 85
83.1712 85
83.3988 85 85
83.3200 84.2642 84.6516
83.2813 84.0853 84.5314


32203 39.9539 51.9553

K0F8-F3) 07189
20 2865
[__=k_(ll _l_ IC 7178
=k(1157-H15) 71

25 25
54.8099 54 9050
85 85


85 85
84.8108 8 8543
84.7372 84.7958

0.1 012 0.14 016
5 6 7 8

68.8231 733370 74.5249 37,3812

qtop total

766630 75.4751 37.6188

qduct tot
0871 0.473 0073 422.236

Q for one duct is 2 422 2362 or 844 4723 W
per meter of duct length.

nodal rows and columns and their locations in vertical and
horizontal directions. These labels help one to keep track of
the geometry of the problem.

The lower portion of the sheet (row 24 down) shows the
calculation of heat rates at nodes along the floor or top sur-
face ("qtop") and along the duct perimeter ("qduct"). The
close agreement of these heat rates serves as a check on the
answer and reinforces the notion that input and output rates
must be equal in a steady-state problem.


We have described here a computer-aided, finite-difference
approach to solving 2-D heat transfer problems in the under-
graduate curriculum. A surprising array of complicated prob-
lems can be solved using this method that cannot be directly
solved with analytical methods. Once learned, the method is
applicable to a number of other course situations. For ex-
ample, our students in unit operations laboratory have solved
for heat transfer characteristics of critical but odd-shaped
components in heat transfer equipment under study. The
method can also be extended by solving problems with time
dependence (transient problems), problems with geometries
that would benefit from rectangular rather than square ele-
ments, geometries with edges at oblique angles, and even
three-dimensional problems.

Spring 2002

The author thanks the students of Chemical Engineering
313 for their interaction with and feedback on the methods
described here. He also thanks Dr. James Palmer and Mr.
Rohit Ghan for reviewing the manuscript.

1. Boyce, W.E., and R.C. DiPrima, Elementary Differential Equations
and Boundary Value Problems, 2nd ed. John Wiley and Sons, New
York, NY: p. 423 (1969)
2. Jaeger, J.C., and H.S. Carslaw, Conduction of Heat in Solids, 2nd ed.,
Oxford University Press, Oxford, England (1986)
3. Welty, J.R., C.E. Wicks, R.E. Wilson, and G.L. Rorrer, Fundamentals
ofMomentum, Heat, and Mass Transfer, 4th ed., John Wiley and Sons,
New York, NY; p. 246 (2001)
4. Incropera, F.P., and D.P. DeWitt, Fundamentals of Heat and Mass
Transfer; 4th ed., John Wiley and Sons, New York, NY; p. 167 (1996)
5. See, for example, B.S. Gottfried, Spreadsheet Tools for Engineers,
Excel 97 Version, WCB/McGraw-Hill, Boston (1998); also see vari-
ous articles in Chemical Engineering Education, e.g., M. Misovich
and K. Biasca, "The Power of Spreadsheets in a Mass and Energy
Balances Course," Chem. Eng. Ed., 25(1), 46 (1991) and P.E. Savage,
"Spreadsheets for Thermodynamics Instruction," Chem. Eng. Ed.,
29(4), 262 (1995)
6. Reference 4, page 178
7. Mills, A.F., Heat and Mass Transfer, Irwin, Chicago, IL, p. 131, 197
8. Chapra, S.C., and R.P. Canale. Numerical Methods for Engineers, 3rd
ed., McGraw-Hill Book Company, Boston, MA; p. 820 (1998)
9. Arpaci, V.S., S.-H. Kao, and A. Selamet, Introduction to Heat Trans-
fer; Prentice-Hall, Upper Saddle River, NJ; p. 194 (1999) J

Figure 9.

844 W

MIN laboratory



Johns Hopkins University Baltimore, MD 21218

Chemical engineering students typically take a unit op-
erations laboratory course and a capstone design
course in their senior year. Experiments in the unit
operations laboratory might focus on a practical design prob-
lem, while the design course is often based on a case study
drawn from the experience of a practicing engineer. These
courses are intended to confront students with problems that
are not well-defined; problems that will be encountered in
the complex "real world" of chemical engineering practice.
The capstone design course, in particular, requires integra-
tion of basic science and engineering facts, theory, and quan-
titative problem-solving skills learned in the earlier years of
the curriculum. The scope of these "real world" problems is,
in most cases, limited by the traditional classroom or labora-
tory settings in which they are taught. Moreover, laboratory
experiences that complement classroom instruction in other
courses are the exception rather than the rule. A major con-
straint to developing new, innovative teaching laboratories
that offer more practical engineering experiences to under-
graduates is that operating and maintaining them is expen-
sive in terms of manpower, space, and equipment.
In this paper, we describe a virtual laboratory now being
created in the Department of Chemical Engineering at Johns
Hopkins that features real-time dynamic simulations of each
experiment in our undergraduate unit operations laboratory
course. The overall goal of this virtual laboratory project is
to enable our students to experience a broad range of design,
scale-up (and even start-up) problems, normal and unusual
operating conditions, and safety issues going well beyond

Michael Paulaitis is Chair of the Department of Chemical Engineering at
Johns Hopkins University He holds degrees from Princeton University
(BS, 1968), Stanford University (MS, 1970), and the University of Illinois
(PhD, 1976). His research interests include chemical engineering ther-
modynamics, statistical/molecular thermodynamics, and molecular simu-
Patrick Fleming is Principal Research Scientist in the Department of
Chemical Engineering at Johns Hopkins University. His BA (1969) is from
Kalamazoo College and his PhD (1975) is from the University of Michi-
gan. His research interests include simulation of protein interactions and
solvation and protein folding.
Copyright ChE Division of ASEE 2002

the scope of experiences that could ever be contemplated in a
traditional chemical engineering unit operations laboratory.

We have so far built, parameterized, and in-house tested a
batch distillation column simulation model as the first ex-
periment in the virtual laboratory. The simulation and pro-
cess control software, provided by GSE Systems, Inc.,E" is
that used in the chemical, pharmaceutical, energy, process,
and manufacturing industries (see Figure 1).
The software comprises a D/3 distributed control system
(D/3 DCST) with SimSuite Pro" simulation software and runs
on PC-level computers. The D/3 distributed control software
permits display and control of any process communicating
via analog or digital signal. SimSuite Pro is stand-alone soft-
ware that is used to build, test, and run process simulation
models dynamically and in real time. Additional Active X'"-
based software donated by GSE enables display of variable
output, control panel, and process graphics inside a web
browser. With these software modules we are able to control
the pilot-scale distillation column now used in our chemical
engineering unit operations laboratory, collect data from the
column during an experiment, or simulate it in real time.
The actual column consists of six identical bubble cap trays,
each fitted with a liquid sampling port and a copper-constan-
tan thermocouple to measure the temperature of the entering
vapor. The reboiler is a 20-liter vessel with variable electric
heating. The total condenser at the top of the column is sup-
plied with cooling water at 250C and 30 psig. Reflux to the
column is controlled by an electrically operated solenoid
valve. During normal operations, this valve is in the "open"
position when energized, and the liquid condensate is col-
lected as product. In the event of electrical failure, grav-
ity causes the valve to close such that all liquid conden-
sate is returned to the column. The reflux ratio is con-
trolled by a timer on the solenoid that periodically opens
and closes the valve.
The distillation column simulation model is a computer

Chemical Engineering Education

program with graphic displays on an instructor's workstation and on an
operator's console. It performs stagewise vapor-liquid equilibrium calcu-
lations based on tray pressure, mixture composition and enthalpy, and va-
por-liquid traffic in the column. The model considers tray holdup, external
heat transfer, and column geometry. It calculates in real time the tempera-
ture, vapor and liquid flow rates, enthalpies, and compositions for each
tray. During equilibrium staged operations, the column is modeled by solv-





*-- -

i.nceqer aa;t
ineager niea'Btrne
ra1.4 radxundist
real"4 x f(.az.:a fi -, */.-nax'.).z2 axat'l
zeal' 4 ox.K oraoa s).yyy naxota tr; azr r;ia
do nin=l rt x --t
d. : o 2
if (dist Ie rad us) then

and do


Figure 1. Configuration of process control and simulation modules for
the distillation column experiment. The operator's station communicates
with the simulator by means of a process control module that translates
input/output signals to the appropriate format and display. The actual
distillation column (or any other unit operations equipment) may also be
controlled from the same operator's station by using a similar active dis-
play that is connected to the input/output addresses of components on
the real equipment. The instructor's station may also control the simula-
tion model or initiate malfunctions during the experiment.

*G Wi ," _

M i-,.
I dl'TG .,I -I -
II L__,_,__ I I-_

.l u.' -

Figure 2. Active computer display seen by students during simulated
operation of the distillation column.
Spring 2002

ing simultaneously component material and en-
ergy balances, and vapor-liquid equilibrium rela-
tionships using Wilson equation activity coeffi-
cients to account for thermodynamic non-ideali-
ties in the liquid phase.
The simulation model was validated using tem-
perature and composition data collected at each
tray for ethanol-water mixtures in the actual dis-
tillation column. During transient regimes, it is
modeled by allowing the stage efficiencies, Wil-
son equation parameters, and column heat losses
to be functions of time. These functions were de-
termined by tuning the model using data collected
for ethanol-water mixtures in the actual column
during transient operations. All the measurements
were made by a team of Hopkins undergraduate
chemical engineers as a summer project.
Flooding in the column, loss of cooling water,
failed valves, or other malfunctions can be simu-
lated from the instructor's station of the virtual
laboratory. During development, the station is used
to tune the running model, and during a virtual
distillation experiment, it is used to set initial con-
ditions, control model execution, and insert inter-
ventions, such as malfunctions. The system is ca-
pable of maintaining up to 100 different initial
conditions: values for all process variables, remote
functions, and malfunctions.
The simulation model exists not only as a com-
puter program running in the background but is
also experienced by the students as an on-screen
graphic image depicting a distillation apparatus.
Stream flows, temperatures, liquid compositions,
and volumes are displayed in real time (see Fig-
ure 2). Operation of the distillation process is con-
trolled by students from the screen by mouse-click-
ing on the various components. Students enter their
choice of operating parameters such as heating
current or valve openings or closings. The same
graphic image may be used to monitor and con-
trol the actual distillation column. In principal, the
student need not know if the virtual or actual dis-
tillation process is connected to the display, although
in practice the virtual process provides more on-
screen information such as tray compositions.
Data on many process variables is continuously
collected and may be viewed in real time (see Fig-
ure 3) or downloaded to a spreadsheet program
for later analysis.

The distillation simulation model was used for
the first time during the 2000/01 academic year in

two chemical engineering courses: Chemical Engineering
Process Design and Separations Processes. The latter is ajun-
ior-year chemical engineering core course taught in the Spring
semester. Chemical Engineering Process Design is taken by
seniors in the same semester. In the previous (Fall) semester the
seniors normally have taken Unit Operations laboratory. Batch
distillation is a required experiment in this laboratory course,
and the students are required to design and then carry out ex-
periments to characterize the actual batch distillation column.
The context and motivation for the experiment is to train
personnel with limited technical backgrounds to separate sev-
eral liquid solvent mix-
tures using the actual pilot- -- *
scale batch distillation col- IT "I ~ I 1 I12j l
umn to accomplish the
task. The first objective is
to decide what data needs 12. |
to be collected to accu-
rately characterize the col-
umn. As a team, the stu- ...
dents identify tray efficien-
cies as the key information
they need and carry out
McCabe-Thiele analyses
to get this information .ooo
from the measured tray
temperatures and compo- M TI--S TRAY TEMP SIM
sitions. The students are T-03S TRAY 3 TEMP SIM
also required to select a TIr--STRAY 5 TEP SiM
suitable "model" system "-
(ethanol/water is selected Figure 3. Real-time display of
in the end), specify the op- lated operation of the distillation
rating conditions for the in temperature in all trays sign
distillation of this mixture,
determine startup and shutdown procedures, etc. They also
consider safety, cost, and environmental impact.

Once the actual distillation column has been characterized,
the students move on to another experiment in the unit op-
erations laboratory. The stated overall objective to train un-
skilled personnel to use the equipment to separate other sol-
vent mixtures is never met or even considered beyond moti-
vating the original experiments. With the ability to carry out
essentially an unlimited variety of separations using the vir-
tual distillation column, it is now straightforward to ask the
students to run the virtual column to complete the overall
objective of the experiment.

We implemented this "mini-design" project in Chemical
Engineering Process Design using a problem-based learning
approach.'12 Teams of senior chemical engineers who were
now familiar with operating the actual distillation column
were given the assignment of repeating the experiment using
the simulation model. Each team of three seniors was intro-
duced to the distillation simulation model during an initial

two-hour session in which they operated the simulation ac-
cording to prescribed instructions contained in a tutorial de-
veloped for the mini-design project of one of the authors (PJF).
During the second session, the students operated the vir-
tual column independently. They charged the column pot with
an ethanol/water mixture, brought the column to equilibra-
tion, and gathered data on the temperature and composition
of each tray. For the third session, each team was given a
virtual column with a different binary mixture, listed in Table
1. Their task was to determine the optimal operating param-
eters for performing a single batch distillation enrichment of
the mixture and to prepare
a training program and

In c

manual to teach inexperi-
enced operators how to
carry out the specific sepa-
ration. Although guide-
lines on size and format
were provided, each team
could decide what the best
content of such a training
manual would be.

S. ...The juniors in the Sepa-
rations Processes course
served as the untrained
personnel in this project.
82.421600 14:07:02 07/11/201 HisTrmnd The timing of their in-
73.603210 14:07:02 o071 12001 HisTrend
72J.9807 14:070 12001HiTrd volvement was coordi-
72.053436 14:07:02 07/1112001 HisTrend
69.390259 14:07.02 07/11/2001 HIsTrend nated such that they were
61451916 14 Oi 02 011112001 ,lrSrT*d
6 14 0 S! 0 00 r nated such that they were
-"-- ll 0-I- just completing course
temperature data during simu- work on the fundamentals
column. Note the abrupt changes of distillation. Teams of
g a recharging of the reboiler. three juniors were paired

with the senior teams for
this training session. After a short oral presentation by the
seniors, the junior operators carried out the batch distillation
of their particular binary mixture according to the training
manual written by the senior team. After the juniors became
familiar with the column operation, the seniors were encour-
aged to introduce various malfunctions and the juniors were
prompted to analyze the problem from displayed process vari-
ables and perform corrective actions in real time. One typical
malfunction was a failure of the cooling water flow. This fail-
ure could be announced by an alarm (if desired by the in-
structor) or observed on the graphic display as a change in
the flow rate of cooling water or an increase in the tempera-
ture of the cooling water return. The students usually come
to the conclusion that the best corrective procedure is to turn
off the heating coils immediately. Malfunctions were also in-
troduced inadvertently. For example, one team accidentally
opened the drain valve on the reboiler and quickly drained
all of a noxious mixture out of the pot! The versatility of
the virtual laboratory allowed us to explore the safety
ramifications of this virtual accident, thus providing an
Chemical Engineering Education


............ ................. ..... ..... ..

educational opportunity that never would have occurred
in the actual laboratory experiment.
One of the great advantages of the virtual column is the
ability to speed up process clock time by as much as a factor
of 10. Thus, instead of waiting hours for the column to reach
steady state from a cold start, the start-up can be shortened to
minutes. This accelerated process time encourages the stu-
dents to carry out "what if" experiments that could never be
contemplated with the actual distillation column.

Each team (seniors and juniors sepa-
rately) were required to give oral pre-
sentations of their results and to pre-
pare written reports. The seniors were
present for the junior team presenta-
tions to ask questions. In both oral pre-
sentations, the students were asked to
describe what they learned during the
training session from the perspective
of either trainers or trainees. The com-
bination of problem solving, group in-
teraction and delegation of teaching/
learning responsibility uses many of
the elements described by Wankat for
efficient effective teaching.[3'
In their written evaluations of the

It is important to point out that computer simulations such
as those described here are not meant to supplant existing
real laboratory experiments, but rather to complement them.
As stated previously,'5' neither the virtual nor the real labora-
tory format is "better"-rather, they have different roles in
the curriculum. We have used the virtual experiment as a tool
to integrate efficient, effective teaching elementst3' into our
curriculum. In designing the virtual laboratory experiments, we
recognize a key responsibility is to ensure that students con-

Binary Mixtures for
Distillation Experiments


Methylene Chloride
/Ethylene Chloride

( )0


project, the seniors were unanimous in their opinion that the
opportunity to separate different mixtures and to explore vari-
ous operating conditions in the virtual column, which they
could not do in the unit operations laboratory, allowed them
to gain a deeper understanding of distillation. The juniors, in
their written evaluations, indicated that the project provided
them with a perspective of distillation that went well beyond
the fundamentals they learned in the classroom.
To evaluate the continuing impact of this junior-level ex-
posure to virtual distillation experiment, we gave a written
survey to the juniors after they had taken the senior unit op-
erations laboratory course and performed the actual distilla-
tion experiment. This survey was designed in the format of
the Student Assessment of Learning Gains instrument avail-
able at the National Institute for Science Education.'4' We
asked the students to rank a list of classroom aspects on a
five-category scale from "No Help" (1) to "Very Much Help"
(5). The questions addressed the value of: a) working in
groups; b) hands-on control of simulation; c) the written lab
manual; d) verbal instruction by seniors; e) final presentation
report; f) impact of virtual lab on real lab. The average scores
for the first five aspects (a through e) of the virtual lab were
between 2.8 and 3.0 (a little to moderate help); the average
score for the last question on the impact of the virtual lab was
3.4 (moderate to much help). Verbal comments from the stu-
dents that correlate with this high level of impact indicated
that their virtual experience instilled more confidence in per-
forming the actual experiment.
Spring 2002


Extremely Flammable
Extremely Flammable
Carcinogenic, Azeotrope
Extremely Flammable
Flammable. Azeotrope
Flammable, Extremely Toxic

tinue to have actual hands-on experience
with an actual process or device.

A major goal for the coming year is
to make the virtual distillation experi-
ment web-accessible and to create and
evaluate virtual laboratory experiments
in partnership with several other uni-
versities. We also plan to introduce the
virtual laboratory in a local high school.
We intend to design and evaluate sev-
eral specific experiments and educa-
tional modules around the virtual batch
distillation column simulation and use
them as templates for a much larger
effort to develop a variety of virtual ex-

periments maintained by faculty at other universities. Web-
accessibility of virtual experiments has the potential to give
chemical engineering students at many educational institu-
tions access to specialized virtual experiments that can
complement existing laboratory curricula. Our "mini-design"
project demonstrated the important element vertical integra-
tion to other courses in the undergraduate chemical engineer-
ing curriculum.

The authors gratefully acknowledge support from the Special
Grant Program in the Chemical Sciences, the Camille and Henry
Dreyfus Foundation, and computer software, hardware, and train-
ing from GSE Systems, Inc., Columbia, Maryland 21045. The ef-
forts of undergraduate students Robin Cohen, Elizabeth Cham-
bers, Gabe Farkas, and Austin Lin are also greatly appreciated.
We thank Professor Joseph Katz for help with the actual distil-
lation column operation.

1. GSE Systems, Inc., is a leading global provider of
dynamic simulation and process control systems to the chemical, phar-
maceutical, energy, process, and manufacturing industries
2. Woods, D.R., Problem-Based Learning: How to Gain the Most from
Problem-Based Learning, Hamilton, Ontario, Canada (1994)
3. Wankat, P.C., "Effective, Efficient Teaching," Chem. Eng. Ed., 35(2),
p. 92 (2001)
5. White, S.R., and G.M. Bodner, "Evaluation of Computer-Simulation
Experiments in a Senior-Level Capstone ChE Course," Chem. Eng.
Ed., 35(1), p. 34 (1999) C

e, classroom
, J5oo~_




University of Waterloo Waterloo, Ontario, Canada N2L 3G1

any undergraduate courses in chemical engineer-
ing are essentially deterministic where numerical
values of variables and parameters are often speci-
fied to several digits past the decimal point. Randomness,
probability, and statistical concepts usually get much less at-
tention. But, random-process oriented thinking is becoming
increasingly important in a world where uncertainties, ambi-
guities, misrepresentation, misinformation, and partial infor-
mation are increasingly unavoidable. The quip that every-
thing carries its own probability distribution"1 is much less
facetious than it sounds.
Every undergraduate chemical engineer knows about
McCabe-Thiele diagrams, for instance-but how many real-
ize that the equilibrium curve may represent only single mea-
sured set of data? The true equilibrium curve would be in-
side a band of (repeated) measurements, and it is anybody's
guess as to what extent the number of theoretical plates/stages
obtained from a diagram based on a single measurement set
is reliable. Similar questions can be asked in chemical reac-
tion engineering (e.g., how reliable is a rate constant ob-
tained from a single set of concentration versus time mea-
surements), in transport phenomena and, in fact, through-
out our entire discipline.
A major purpose of a course in applied probability and sta-
tistics should be to sensitize students to such questions and to
provide them with at least basic means to deal with them.
One could, indeed, claim that a single course in this subject
is insufficient for a modern undergraduate engineering cur-
riculum of any kind, and that probability-oriented approaches
should be given appreciably more room than currently avail-
able in undergraduate engineering education.
At Waterloo, two such courses are given in the second
year-one to chemical engineering students and the other to

environmental/chemical engineering students. It is perhaps
arguable that the second year is too soon for such a course,
but we are convinced that our students should be exposed to
the subject matter at a relatively early stage of their career.
The difference between the two courses is primarily in illus-
trative and homework exercise problems, and the same text-
book by Walpole, et al., 2] (with its earlier editions in the past)
has proven itself to be quite adequate in spite of its limited
repertoire of problems specific to chemical and environmen-
tal engineering.

The cooperative scheme of alternating school terms and
work terms at Waterloo results in three terms per academic
year, with slightly varying lengths, allowing 33 to 36 fifty-
minute lectures and 11 to 12 weekly tutorials per term. As
shown in Table 1, the fraction of time spent on the probabil-
ity and the statistics sections is roughly equal. The tutorial
periods are devoted to representative numerical problems
arising from material presented in the lectures. Weekly home-
work assignments contain typically 5 to 7 nontrivial prob-
lems taken from the assigned text, from a variety of other
texts, and from data sources available in the technical jour-
nals (e.g., Chemical Engineering Progress). Students are ex-

Thomas Z. Fahidy is Professor of Chemical
Engineering at the University of Waterloo
(Canada). He obtained his BSc and MSc de-
grees at Queen's University (Kingston,
Ontario) and his PhD at the University of Illi-
nois, Urbana-Champaign. His major research
and teaching interests are in applied electro-
chemistry, electrochemical engineering, ap-
plied engineering mathematics, and applied
probability and statistics. He can be reached
at .

@ Copyright ChE Division of ASEE 2002

Chemical Engineering Education

amined by a two-hour-long midterm test given, usually, in
the sixth or the seventh week of the term, and by a three-
hour-long final test; notes, books, and calculating devices
are usually allowed.

In the first part of the course, discussion of fundamental
concepts and theorems of probability theory, is followed by
applications to specific problems of interest to chemical and
environmental engineering. There is little emphasis on find-
ing, for example, the probability P(=7/15) of taking randomly
one black ball and two green balls from an urn containing
two black balls and eight green balls, even if problems of this
kind appear in most textbooks. The three following examples
represent probability problems much closer to the heart of
the chemical and environmental engineer.

Example 1
Processing of Sludge Water by Water Purification Plants
A municipality operates two sludge water processing plants. The amount of sludge
water that one plant can process in one day has an exponential distribution with a
mean of t tons for each plant. Assuming that the operation of one plant has no
influence on the other plant, what is the probability that one of the two plants
would process more than l tons during a randomly chosen day?
First, we compute the probability that any of the two plants processes more than l
tons by

Course Syllabus

1 Basic concepts of probability; counting rules; permutations and
2 Basics of probability distribution theory; expectation and variance
3 Binomial and Poisson distribution
4 Uniform continuous, normal, and exponential distribution
5 Lognormal and Weibull distribution; normal approximations to
binomial and Poisson distributions; sample mean and sample
variance; central limit theorem
6 Confidence intervals for means, difference in means, and
7 Type I and Type II error; the p-value concept; tests of hypothesis
for means
8 Variance-based tests of hypothesis; chi-square and F-distribu-
tions; contingency tables; goodness of fit
9 Simple regression analysis and associated confidence intervals
and hypothesis testing; simple nonlinear regression
10 Multiple linear regression
11 Analysis of variance; one-way classification for comparison of
12 Tests for homogeneity; multiple-range test
13 Review

Spring 2002

P{X>}t= f(I/p)exp(-x/l)dx=exp(-1)=0.37

Then the probability that one of the two plants processes more than the mean
number of tons per day is given by the binomial distribution

b(2; 1, 0.37) = 2(0.37)(1 0.37) = 0.466 (2)
The answer is independent of the numerical value of the mean.

Due to their recently recognized importance in science and
engineering, the Weibull and the lognormal distributions are
now regular course topics. A typical classroom (or tutorial)
illustration is given by Example 2.

Example 2
Application of the Lognormal Distribution to Metal Structure
This is a modified version of Supplementary Exercise 4.92 in Scheaffer
and McClave.'1' The grains comprising a certain type of aluminum have a
lognormal distribution of their mass with parameters a = 0.03 and P =
0.04 g.
a) Find the mean and variance of the grain mass-the mean (or expecta-
tion) is given by

E[X]= exp(a + i2 / 2) 1.0313 g

and the variance is given by

V[X]= exp(2 + 2 )[exp(2) l] = 0.0017 g

b) Find the probability that a randomly chosen grain has a lower mass
than the mean mass-the probability of finding a lognormal variable
between limits of a and b is given by

P{a X < b} = FN{[,n(b)-a]/3}-FN{[n(a)-x/1]} (3)


P{X E(X)} = FN[n E(X)- a] / P} = FN(0.02)= 0.508

from standardized normal distribution tables; FN is the standardized
cumulative normal probability distribution of the grain mass.
c) Find the probability that a randomly chosen grain has a mass between
the mean mass and twice the mean mass-from Eq. (3)

P{1.0313 X<2.0626}=FN (17.35)-FN (0.02)=1-0.508=0.492

Because of the time limitations, the beta, gamma, hyper-
geometric, and multinomial distributions are not treated in
the course, although some reference may be made to them
during tutorial periods. We feel that the probability sections
in class provide an adequate foundation for students' self-
directed study of other probability distributions.
Statistics are often viewed with suspicion because its results
can willfully be distorted or misrepresented (how to lie with
statistics?). The intent of Example 3 is to deal with this point.

Example 3
How Reliable is a Highly Touted (Hypothetical) TB-Testing Device?
A tuberculosis (TB) testing device is reported to be 99% effective if a
(randomly) tested person is infected by TB and 95% reliable if the person
is not infected with TB. If a person taken randomly from a group where it
is known that 5% is infected by TB, what is the probability that the per-
son is healthy, although the test indicates infection?
This is a standard problem in conditional probability for the application
of Bayes' rule. We have four events to consider: 1) event A, the device
does not indicate TB; 2) event A', the device does indicate TB; 3) event
B, the person is not infected by TB; and 4) event B', the person is infected
by TB. Bayes' rule may then be expressed as

P(BIA')=[P(A'IB)P(B)]/[P(A'IB)P(B)+ P(A'I B')P(B')] (5)

The terms with vertical bars denote conditional probabilities, e.g., P(BIA')
is the probability of the person being healthy when it is known that the
device indicated infection; P(A'IB) is the probability that the device indi-
cates infection, although the person tested is healthy, etc. We have P(AIB)
= 0.95, P(AIB') = 0.01, P(A'IB) = 0.05, P(A'IB') = 0.99, P(B)= 0.95, and
P(B') = 0.05. Eq. (5) yields the disturbingly high value of P(BIA') = 0.49;
the device is prone to be incorrect in about fifty out of one hundred times.
Even if P(AIB) were as high as 99%, Eq. (5) would still yield a not-too-
comfortable probability of about 16%. If this device were announced by
stating only that it is 99% effective, it would be a misrepresentation, if
not false advertising. Equally thought-provoking conclusions can be
reached by solving Problem 2.40[] and review exercise Problem I121 on
page 49.

A number of traditional topics of applied statistics shown
in Table 1 can be covered in a more-or-less "cookbook recipe"
fashion, once the underpinning principles of formulae and
the proper choice of tests have thoroughly been explained.
Due to its highly important position in contemporary engi-
neering, regression analysis is treated as a major theme. Stu-
dents are repeatedly warned against its indiscriminate use,
however, as well as against an overconfident reliance on the
widely used and often abused R2 statistic (the square of the
correlation coefficient, also known as the coefficient of de-
termination). Examples 4 and 5 illustrate these cardinal points.

Example 4
Data Linearization for Reg,

session Analysis

Given the data in Table 2 for an irreversible isothermal first order type
decomposition reaction, estimate the regression value of the rate coeffi-
The traditional handling of such a problem is to first linearize (unwit-
tingly) the integrated form of the differential rate equation dY/dt = kY,
then perform linear regression on

fnYi=-kti i=1...,N (6

where N is the size of the data set. The regression value of the rate coef-
ficient obtained in this manner, k = 0.00376 min' is statistically biased,
as seen from the following general argument.
Let Q = f(q) denote a transformation of random variable q to random
variable Q. Then, the variances of the two random variables are related

and for f(q) = Inq, we can write specifically

CO =0q exp(-2Q)

Observations in
Example 4

Time Fraction of
t(min) species A
unconverted, Y

5 0.985
10 0.966
15 0.971
20 0.966
25 0.947
30 0.947
40 0.814
50 0.861
60 0.832
70 0.670

and setting it to zero by

An important point here is that the vari-
ance of Q is a variable from observation
to observation. Statistical bias arises,
therefore, in applying Eq. (5) unless it is
known that dQ/dq = 1 everywhere along
the observation set; this is a seldom-met
condition in practice. To circumvent bias,
the so-called weighted least squares
method, using weighting factors
w=l/so, was discussed in the chemi-
cal engineering literature about half a cen-
tury ago.1' It is not used widely due to its
unwieldiness ( sQ, the sample variance of
Q, has to be computed usually for each
observation). It is more inviting to use
nonlinear regression as a more powerful
In applying nonlinear regression, first we
obtain the least squares relationship by
differentiating the sum of the error squares

(SSE) / k= a 1 [Yi -exp(-kti)]2 /ak=0 (9

The numerical value of k is obtained by using a root-finding procedure
applied to the resulting expression

f(k)=X Yl ti [Yi exp(-kti)-exp(-2kt )]=0 (10

In our case, the familiar Newton's rule (chosen arbitrarily; other root-
finding procedures would also do)

kn+, =kn +f(kn )/f'(kn) (11
yields k = 0.00348 min-' [f(k) = 10-; f'(k) = 104]. The fact that the error
variance of the linearization procedure, s2 = 0.113 is about 2.5 times the
value of s 2 = 0.047 of the nonlinear procedure (s.2 is the sample variance
of the experimental error), is in support of nonlinear regression in this
case. In general, model assumptions about error (structure) should influ-
ence the decision about the choice when either linear, weighted, or non-
linear regression could be used.

Example 5
Data Fitting and Goodness of Fit
) Given the data in Table 3, analyze the linear least squares expression
proposed to link the mass of certain vehicles to their mean fuel consump-
Chemical Engineering Education

tion rate, with

Y=a+bx; a=-0.8696; b=8.5164 (12)
obtained via conventional techniques of regression analysis.
Students are warned against accepting the relatively large R' = 0.95 value
as a "proof' of a good linear relationship between mean consumption
rate and mass. They are urged to carry out an appropriate analysis of
confidence intervals for the true (population) parameters of the linear re-
gression. Using the computed values of

Yx2 =18.3258; s, =0.3232; sc =0.6840
the confidence intervals are established as

-0.8696-0.8995tr truee <-0.8696+0.8996tcr (13)

8.5164-0.6644tcr where tCr is the critical value of the T-distribution at a significance level
a, evaluated at a/2, with (N-2) degrees of freedom. The confidence
intervals shown in Table 4 for selected values of a are wide, even at
a =0.2 (i.e., at an 80% level of confidence; conventional statisticians
would not assign importance to confidence levels below 90%).
The weakness of a purely R -based decision is further demonstrated by
hypothesis testing on assumed values of the true (population) correlation
coefficient p. Applying the Fisher transformation z = (1/2) In [(l+r)/(l-
r)] = 2.178 with mean (1/2) In [(+ p)/(1-P)] and standard deviation 1/
/(N-3), and if we postulate p = 0.995, we obtain the standardized normal
variate z = (2.178-2.995)/0.378 = -2.161. Since FN = (-2.161) =0.016, it
follows that the p = 0.995 postulate can be rejected at the significant
level (a =0.05), but not at the highly significant level ( a =0.01) accord-
ing to the rules of classical hypothesis testing. Without the evaluation of
the confidence intervals shown above, a purely correlation-based analy-
sis would induce a much stronger "faith" in Eq. (12) than it actually de-
serves. Students are reminded that the widespread practice of considering
only R2 values is statistically insufficient.

Mean Fuel Consumption Rates and Mass of Certain
Vehiclest5'6l in Example 5*
(MFCR = mean fuel consumption rate)

Vehicle mass MFCR
Vehicle type x (metric ton) y (liter/100 km)
AMC Concord 1.54 12.93
Chevy Caprice 1.72 13.87
Ford Country Squire Wagon 1.86 15.28
Chevette 1.00 7.76
Toyota Corona 1.18 8.46
Ford Mustang Ghia 1.31 10.82
Mazda GLC 0.91 6.82
AMC Sprint 1.22 8.46
Buick Century 1.54 11.52
VW Rabbit 0.86 7.29
*Original data in Imperial units were converted to SI units by the author

Spring 2002

The discussion of significance/confidence levels logically
leads to the concept of the P-value, which may be regarded
as a modem alternative to the a = 0.05 and a = 0.01 test of
hypothesis scenario of traditional statistics. Among many
sources dealing with this subjects, the textbooks of Doherty,7l
Milton and Arnold,'8' Walpole, et al.,12' and Scheaffer and
McClave'3' provide a good background. The P-value of a test
of hypothesis is the smallest level of significance (i.e., the
highest level of confidence) at which the null hypothesis
would be rejected due to the numerical value of the test sta-
tistic. Put alternatively, the P-value, known also as the at-
tained significance level,'3 may be regarded as the Type I
error a if the test statistic is considered to be the critical
value (the Type I error is the error committed when the hy-
pothesized value of a true statistical parameter is rejected in
favor of an alternative value). The level of significance would
thus appear "floating," but Miller, Freund, and Johnsonl'' rec-
ommend that a level of significance should be specified prior
to the performance of the test. Arnoldi"0 argues that the P-
value provides "...a more continuous response to the data..."
than a black-and-white accept/reject decision on the null hy-
pothesis, but on the other hand " is difficult to give a theory
of statistics on P-values..." even if they are "...often useful in
applied statistics...." A particularly convincing case for P-
values is made in an earlier book by Huntsberger and
Billingsley."" Example 6 illustrates a typical demonstration
of the P-value concept to students.

Example 6
Applying the P-Value Concept to Decision Concerning Variances
The problem is based on Example 10.16, Section 10.9, of Walpole.l21
A manufacturer claims that the standard deviation of the lifetime of its
car batteries in 0.9 year. A random sample of ten batteries yields a stan-
dard deviation of 1.2 years. Can it be assumed that the true standard de-
viation is larger than 0.9 year?
The conventional way of treating this problem is to compute the chi-
square variate

x2 =(N-l)s /O2 =(9)(1.2)2 /(0.9)2 =16 (15)

Confidence Intervals of the True Regression Param-
eters in Example 5.
(Degree of Freedom is d.f = 10-2 = 8)

L.L U.L. L.L. U.L.
a t( / 2) fora,,, for a,, for b, for b,

0.2 1.397 -2.1261 0.3870 7.5882 9.4446
0.1 1.860 -2.5427 0.8035 7.2806 9.7522
0.05 2.306 -2.9438 1.2046 6.9843 10.0485
0.01 3.355 -3.8874 2.1482 6.2873 10.7455

and to compare it to the critical value of X =16.92 at the significant level
a =0.05 with degree of freedom d.f. = N-1=9. Since 16 is less than 16.92,
the null hypothesis that there is no change in the true standard deviation
of battery lifetime is not rejected.
Now we apply the P-value approach by examining the variation of the
numerical value of chi-square with the level of significance at d.f. = 9.

Illustration for
Example 6
The Critical X
versus a
Relationship at
a Xcr
0.3 10.66
0.2 12.24
0.1 14.68
0.05 16.92
0.025 19.02
0.01 21.67

Plotting the values shown in Table 5,
we find that X2 becomes critical at
a = 0.065. This is also the P-value. It
follows that the size of the Type I error
associated with rejecting the null hy-
pothesis in favor of the alternative hy-
pothesis a > 0.9 is about 6.5%.
Is this error appreciably larger than
an error of 5%? The answer depends
on the nature of the industrial product.
If the answer is "no" in the case of these
batteries, then rejecting the a = 0.9 hy-
pothesis may be a justifiable decision.
Going beyond the original problem,
assume that in a subsequent test, fif-
teen batteries yield a standard devia-
tion of 1.116 years. Equation (15)
yields X2 = 21.53 and the related P-

value is about 0.014.
Is a 1.4% large Type I error pragmatically different from an error of
1%. Most likely not; so the hypothesis of c = 0.9 could be rejected even
at the highly significant level, for practical purposes. It is recognized, of
course, that a traditional statistician would not reject the a = 0.9 hypoth-
esis after the first test, and reject it only at the significant level ( a =0.05)
after the second test.

An additional advantage of the P-value concept is in esti-
mating the magnitude of the Type I error when the test statis-
tic is appreciably lower than the critical value associated with
a = 0.05, or appreciably higher than the critical value asso-
ciated with a = 0.01.

Consider, for example, two equal-sized sets of data with
eight elements in each set from two different populations of
the same species. The individual sample standard deviations
are 0.25 and 0.05. The task is to test the null hypothesis that
the two population variances are equal. The ratio of the two
sample variances is an F-distribution variate and the numeri-
cal value is f = (0.25/0.05)2 = 25. The degrees of freedom
being 7 and 7, we find a P-value of about 0.0001 upon some
numerical manipulation.

It follows that, if the null hypothesis is true, the probability
that f 25 would occur by chance is at most 0.01%; thus
the fact that f = 25 did occur forces us to reject the null

Example 7
The Type II Error in Hypothesis Testing: When Does It Matter?
The Type II error 3 is the error committed when a null hypothesis is not
rejected, although it is false. It seems to be a more difficult concept for
students than its Type I cousin. The lecture devoted to the Type II error
emphasizes the fact that although this error is just as important in prin-
ciple as the Type I error, its practical importance may be small, if not
negligible, in certain instances. To illustrate this point, consider a catalyst
manufactured by a certain firm with a mean life (before regeneration) of
300 days of continuous operation and standard deviation of 24 days. The
manufacturer claims to have recently developed a new process that makes
this catalyst with a larger mean life, but with a negligible change in the
standard deviation. A test of 70 catalyst samples from the new process is
planned, with an anticipated sample mean xm.
Computing the normal variate z = (/70)(xm-300)/24 = 0.3486 x 104.58,
we can establish for a specified Type I error the value of the Type II error
for specified values of xm (the size of the Type I error has to be a-priori
specified in order to determine the size of the Type II error for an as-
sumed value of x ). The procedure is described in detail by Problems 7-
26 and 7-27 in Lindley and Scott."'4 Assume that a 0.05 Type I error is
chosen; since FN (1.645) = 0.05, z = 1.645 and xm = 304.7 days. Thus any
sample mean larger than 304.7 will induce our willingness to accept the
manufacturer's claim. Now we ask the question, however, "What is the
probability of not rejecting the competitor's claim with xm = 304.7 if the
true new life is a variable?" To obtain the Type II error, we set z = /70
(304.7 9)/24 = 106.22 0.3486 I9, then specify various values of the
true mean, compute z, and determine from standardized normal distribu-
tion tables the related values of P. The procedure is summarized in Table
6, which also includes the case where we pre-specify a Type I error of
0.01. If the true mean is well below the xm value determined by the choice
of the Type I error, it is essentially guaranteed that the old mean will be
considered true, since there is no justification in the claim of a longer-
lasting catalyst. Conversely, if the true mean is considerably higher than
xm, it is highly unlikely that the old mean would be retained. The shift
toward higher P values with an increase in a is due to the shift of xm

Illustration for Example 7
Variation of Type II Error P with True (new) Mean

True mean Type I error a = 0.01
g (days) x,= 304.7 (days)
z [



Type I error a =0.05
x = 306.7 (days)
z B



Chemical Engineering Education

toward higher values, indicating that the acceptance of xm versus .l > xm
would be more tempting (since Ixm ti I is smaller, as a increases, for a
set value of .).
A discipline- and distribution-free argument can be made as follows. Let
the false null hypothesis be that the mean of a random population is 10.00,
whereas the true alternative hypothesis is that the mean is 9.95 (in consis-
tent units). The Type II error could be almost as high as unity, but the
(wrong) acceptance of 10.00 instead of 9.95 may have little importance
(would the mean price of an industrial product costing nine dollars and
ninety-five cents per gallon instead of ten dollars per gallon upset a cus-
tomer wishing to buy only half a gallon?). It would be incorrect, of course,
to draw the conclusion that Type II errors are secondary to Type I errors.
Table V on page 610 of Devore and Peck""3 makes this point very clear.

Time limitations allow only a brief treatment of traditional
one-way classification problems and basic tests of homo-
geneity when at least one of the means is indicated to be
different from the rest, or the sample sizes are unequal.
Duncan's multiple-range test usually concludes the se-
quence of lectures.
The tutorial period, along with the last homework assign-
ment, provides numerical opportunities for students to famil-
iarize themselves with the structure of ANOVA tables. Use
of statistical computer software (e.g., SAS, Minitab,
Mathematica, Polymath 5, etc.) is accepted only with stu-
dents' own reasoned evaluation of computer printouts. Their
attention is routinely drawn to the usual futility of eight-digit
decimal "accuracy" of sum of squares, mean squares, param-
eter estimates, standard error of estimates, etc., when physi-
cal measurements might be only one or two decimal accu-
rate. Similar caveats are raised with respect to computer-based
multiple regression analysis, where printouts may be carry-
ing R2 values with six-digit decimals. The use of computer
software is encouraged, but the choice (including the use
of statistical calculators) is left to the students' discre-
tion. A typical example for the handling of regression
analysis is given in the Appendix.

In addition to the textbook chosen for the course, students
are encouraged to consult the wealth of excellent books on
probability and statistics. Beyond the usual statistical tabula-
tions carried by such books, students are introduced (e.g.,
during tutorials) to the well-known Cambridge tables,[141 and
mention is made of similar material supplied by Kokoska
and Newison1151 and Powell."61 The books by Crow, Davis,
and Maxfield,'17' Bruning and Kintz,E"81 and Freund and Wil-
liams[l91 may be recommended as references. The gourmet
student can find much delight in the impressive works of
Guenther,'201 Snedecor and Cochran,t211 and Neter, Wasserman,
and Kubner.'22] The CRC tables231 are one of the most com-

prehensive of the kind, but most of its contents are beyond
the scope of the course as well as students' financial means.
Spiegel's compendium of problems in probability and statis-
tics,'241 along with brief theory, is a useful and moderately
priced study aid.

In general, difficulties are experienced by many students
in absorbing the plethora of new concepts, and the workload
is considered to be high. This impression is probably due to
the sudden transition from the familiar territory of black-and-
white answers to black-and-white problems to the unfamiliar
(and perhaps strange) domain of chance, odds, inexactitudes,
and the absence of engraved-in-stone solutions.
Tutorials are popular, and instructors' efforts are well-rec-
ognized (in the student evaluation of the course given in the
Spring 2000 term, 60% of respondents put the value of tuto-
rials in the top two score categories, and the author's dedica-
tion to the course was rewarded by 88% of the responses in
these score categories).

In a somewhat iconoclastic monograph,1251 Cheremisinoff
advocates an early treatment of confidence limits (Chapter
2), data scatter, and the use of control charts (Chapter 3) when
teaching statistics tuned to practical engineering. Spectral
density analysis, conventionally left to courses dealing with
stochastic processes and signal processing, is also included
(Chapter 4). This is a thought-provoking scheme, especially
for a chemical engineering curriculum, given that employ-
ment opportunities in areas traditionally reserved for other
engineering disciplines may well require chemical engineer-
ing graduates to increase their (limited) knowledge of prob-
ability and statistical methods beyond the conventionally
taught concepts and techniques.
The current trend to compress the treatment of probability
distributions and classical tests of various hypotheses, in fa-
vor of a deeper discussion of regression analysis, design of
experiments, and extended applications of the analysis of
variance is illustrated by the recent texts of Montgomery, et
al.,'261 and Kinney.l127 This trend reflects an increasing em-
phasis in engineering practice on data handling and its in-
creasing level of sophistication supported by modern com-
puter techniques. The author's choice of Montgomery's text
for the Winter 2002, and possibly Spring 2002, term course
is an exploration in this direction, but a somewhat similar
shift in the newest (and heavier) edition of Walpole, et
al.,[281 is duly noted.

The author's interest in the subject area has been strongly
influenced by his departmental colleagues, Distinguished
Professor Emeritus Park M. Reilly and Professor Tom Duever.

Spring 2002


1. Reilly, P.M., private communication
2. Walpole, R.E., R.H. Myers, and S.L. Myers, Probability and Statistics
for Engineers and Scientists, 6th ed., Prentice Hall, Englewood Cliffs,
NJ (1998)
3. Scheaffer, R.L., and J.T. McClave, Probability and Statisticsfor Engi-
neers, 2nd ed., Duxbury Press, Boston, MA (1986)
4. Mickley, H.S., T.K. Sherwood, and C.E. Reed, Applied Mathematics
in Chemical Engineering, 2nd ed., McGraw-Hill, New York, NY (1957)
5. Henderson, H.V., and P.F Velleman, "Building Multiple Regression
Models Interactively," Biometrics, 37, 391 (1981)
6. Hogg, R.V., and J. Ledolter, Engineering Statistics, Macmillan, New
York, NY (1987)
7. Dougherty. E.R., Probability and Statistics for Engineering: Comput-
ing and Physical Sciences, Prentice Hall, Englewood Cliffs, NJ (1990)
8. Milton, J.S., and J.C. Arnold, Probability and Statistics in the Engi-
neering and Computing Sciences, McGraw-Hill, New York, NY (1986)
9. Miller, R.J., J.E. Freund, and R. Johnson, Probability and Statistics
for Engineers, 4th ed., Prentice Hall, Englewood Cliffs, NJ (1990)
10. Arnold, S.F., Mathematical Statistics, Prentice Hall, Englewood Cliffs,
NJ (1990)
11. Huntsberger, D.V., and P. Billingsley, Elements of Statistical Infer-
ence, Allyson and Beacon, Boston, MA (1981)
12. Walpole. R.E. Introduction to Statistics, Macmillan, New York, NY
13. Devore, J., and R. Peck, Statistics: The Exploration and Analysis of
Data, 3rd ed., Duxbury Press, Pacific Grove, CA (1997)
14. Lindley, D.V., and W.F Scott, New Cambridge Statistical Tables, 2nd
ed., Cambridge University Press, Cambridge, UK (1984)
15. Kokoska, S., and C. Nevison, Statistical Tables and Formulae,
Springer-Verlag, New York, NY (1989)
16. Powell, FC., Statistical Tables for the Social, Biological, and Physi-
cal Sciences, Cambridge University Press, Cambridge, UK (1982)
17. Crow, E.L., EA. Davis, and M.W. Maxfield, Statistics Manual, Do-
ver, New York, NY (1960)
18. Bruning, J.L., and B.L. Kintz, Computational Handbook of Statistics,
Scott, Foresman and Co., Glenview, IL (1977)
19. Freund, J.E., and FJ. Williams, Dictionary/Outline ofBasic Statistics,
Dover, New York, NY (1966)
20. Guenther, W.C., Analysis of Variance, Prentice Hall, Englewood Cliffs,
NJ (1964)
21. Snedecor, G.W., and W.G. Cochran, Statistical Methods, 8th ed., Iowa
State University Press, Ames, IA (1989)
22. Neter, J.W., W. Wasserman, and M.H. Kubner, Applied Linear Statis-
tical Models, R.D. Irwin Inc., Homewood, IL (1990)
23. Beyer, W.H., ed., CRC Standard Probability and Statistics Tables and
Formulae, CRC Press, Boca Raton, FL (1991)
24. Spiegel, M.R., Probability and Statistics, Schaum's Outline Series,
McGraw-Hill, New York, NY (1975)
25. Cheremisinoff, N.P., Practical Statistics for Engineers and Scientists,
Technomic Pub. Co., Inc., Lancaster, PA (1987)
26. Montgomery, D.C. G.C. Runger, and N.F. Hubele, Engineering Sta-
tistics, 2nd ed., Wiley and Sons, New York, NY (2001)
27. Kinney, J.J., Statistics for Science and Engineering, Addison-Wesley,
New York, NY (2002)
28. Walpole, R.E., R.H. Myers, S.L. Myers, and K. Ye, Probability and
Statistics for Engineers and Scientists, 7th ed., Prentice Hall, Upper
Saddle River, NJ (2002) O


Typical Handling of a Regression Analysis Problem
via Computer Software

Given: six catalysts containing 12, 10, 14, 11, 12, and 9
coded platinum mass units X, The corresponding coded
effectiveness indicators Y are 18, 17, 23, 19, 20, and 15,
respectively. The regression section of Polymath 5 yields
parameter values shown for the simple linear regression y
= a + bx in the tabulation below

Intersection (a)
Slope (b)
95% confidence for aO
95% confidence for al
Coefficient of determination
RMS (root mean square) of deviations
Error variance

Numerical value

The 95% confidence interval for the true (population) re-
gression parameters can readily be computed as
1.91304 7.93568 < a < 1.91304 + 7.93568,
i.e., [-6.02264; +9.84872],

1.47826 0.69334 < P < 1.47826 + 0.69334,
i.e., [0.78492; + 2.1716]
It is instructive to consider the size of these intervals against
the "comfortably" high correlation coefficient of 0.947!
Other statistical packages will deliver the ANOVA table
shown below, which provides the numerical value of the
F-statistic to test the null hypothesis that Y and X are not
related linearly (i.e., P = 0).


Sum of Degree of Mean
Squares Freedom Square


1 33.505
4 0.957


The critical values of F (1,4), as shown below, indicate
that the P-value is about 0.0042, hence the hypothesis of
no linear relationship between X and Y can be rejected at
an approximately 99.6% confidence (this is not a surpris-
ing result, since a b = 1.48 is appreciably larger than zero).
Significance level 0.05 0.01 0.005 0.001

Critical F (1,4)

7.71 21.20 31.33 74.14

Any hypothesis concerning the true slope (e.g., p = 1.5,
or 1.0, or 2.0) can be tested conveniently by using the T-
distribution, as described routinely in the textbook litera-
ture. 0

Chemical Engineering Education


Deadline is June 1, 2002

Full Text