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


on the


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

Marcia Miller

James O. Wilkes, U. Michigan

William J. Koros, University of Texas, Austin

E. Dendy Sloan, Jr.
Colorado School of Mines
Pablo Debenedetti
Princeton University
Dianne Dorland
University of Minnesota. Duluth
Thomas F. Edgar
University of Texas at Austin
Richard M. Felder
North Carolina State University
Bruce A. Finlayson
University of Washington
H. Scott Fogler
University of Michigan
William J. Koros
University of Texas at Austin
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
Stewart Slater
Rowan University
James E. Stice
University of Texas at Austin
Donald R. Woods
McMaster University

Sutnmer 2001

Chemical Engineering Education

Volume 35 Number 3 Summer 2001

158 Minnesota's Ed Cussler, Rutherford Aris, Alon V McCormick

162 City College of New York. Reuel Shinnar

168 Teaching Separations: Why, What, When, and How? Phillip C. Wankat
188 Teaching Biotech Manufacturing Facility Design and Regulatory
Compliance: Better Equipping Students for a Maturing Industry, David
E. Block
202 A Course in Process Simulation: Using Object-Oriented Programming
Methodologies and Java, David G. Taylor

172 An Integrated, Real-Time Computing Environment for Advanced
Process Control Development, James H. van der Lee, Donald G.
Olsen, Brent R. Young, William Y Svrcek
208 A Note on Stability Analysis Using Bode Plots, Juergen Hahn, Thomas
Edison, Thomas F Edgar

180 A Brief History of Elementary Principles of Chemical Processes,
Richard M. Felder

182 Use of Sequential and Non-Disciplinary Problems to Teach Process
Dynamics, William L. Luvben

194 Simulation of Reaction Kinetics Using Equivalent Hydrodynamic
Models: Modeling and Laboratory Experiment, Redhouane Henda
198 Low-Cost Experiments in Mass Transfer: Part 8. Absorption of Carbon
Dioxide from a Single Bubble, M.H.I. Baird, I. Nirdosh
220 Flashback and Laminar Flames: A Classroom Demonstration, Edward
S. Shanley, RonaldJ. Willey
222 Column Transport Experiments for Dissolved Pollutants and Colloids,
David R. Shonnard, Prasanna A. Deshpande

212 Some Psychological Theories in Engineering Education, E. Alpay

179, 187 Faculty Position
187 Book Review
193 Books Received
207 Letter to the Editor

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 2001 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 andfor 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.





University of Minnesota Minneapolis, MN 55455
oontide at the University of Minnesota finds a lean, polychro-
matic, Lycra-clad figure emerging from Amundson Hall and
loping off through the Twin Cities campus into the distance,
where it is swallowed up-a lone, precursory coruscation in a sea of
Midwestern mackintoshes. Is this, you wonder, an ambassador from
some place better known for the vividity of its colors than for their
sartorial harmony? No, the figure is that of IT Distinguished Professor
Edward Cussler, perhaps in training for his next marathon or prepar-
ing for a tour de part of France, or maybe getting in shape to pull his
oar to the head of the Charles River.
The "IT" in Ed's title denotes the Institute of Technology, the col-
lege that at Minnesota embraces all the natural sciences and engineer-
ing, and which gives the title "Institute of Technology Distinguished
Professor" to only a small handful of its 500 faculty who are outstand-
ing both in their teaching and their research. Indeed, Ed was already
well known for both when Ted Davis persuaded him (and his colleague
Fennell Evans) to translate from Carnegie-Mellon to Minnesota-one
of the early successes of Davis' term of headship.
Ed's undergraduate degree was from Yale (with honors) in 1961,
and his MS and PhD were taken at Wisconsin under Ed Lightfoot's
mentorship. He then had three postdoctoral years: at Wisconsin, at
Adelaide, South Australia, and in the Chemistry Department at Yale.
He was appointed at Carnegie-Mellon in 1967 and had risen to full
rank there by 1973-all this in only ten years after his first gradu-
ate degree!
Ed has always been known as an outstanding teacher, with animated
delivery, apt illustration, and engaging originality. He won the Ryan
@ Copyright ChE Division of ASEE 2001
158 Chemical Engineering Education

Undergraduate Teaching Award when he was at Carnegie-
Mellon and has won a corresponding award no less than six
times since coming to Minnesota. Of his first book, Diffu-
sion (Cambridge University Press, 1984), one reviewer wrote,
"This is an outstanding example of a text written by an ex-
pert in the field who cares as much about teaching the topic
as he does about generating the research upon which progress
in the understanding of mass transfer is based." Another sup-
porter, having Ed's teaching in mind, observed, "He has
throughout his career been something of an icono-
clast, but with a basically positive and constructive
attitude and a puckish sense of humor-three char-
acteristics that are in short supply." mi
Ed has been in demand for the special-named
lectures that are a feature of academia. He has given
the Henskie Lectures at his alma mater, the Katz
Lecture at Michigan, and the Danckwerts Lecture in
London. The last of these is easily accessible, published
under the title of "The Nature of Chemical Research" [Chem.
Eng. Sci., 53(11), 1957 (1997)]. He won the Colburn Award
of the AIChE in 1975 and ASEE's Lectureship Award in 1998.
He maintains a vantage point from which he can see the broad
trends of the profession through his chairmanship of the
American Association of Engineering Societies and his asso-
ciate editorship of the AIChE Journal.
Ed's research has been on diffusion and other modes of
mass transfer. Though the analogies with heat transfer are
illuminating, the time is long past (if indeed there ever was
one) when mass-transfer problems could be approached as
heat-transfer problems with a change of variable. Nowhere
is that more obvious than in multicomponent diffusion, the
subject of a monograph published in 1976 under that very
title. He went on to publish a more general book, Diffusion,
in 1984, of which a second edition was demanded in 1997,
Meanwhile he collaborated with Wei-Shou Hu and Paul Belter
on Bioseparations, a book published in 1988 by Wiley.
There can be no doubt of the importance of mass transfer
to the understanding of chemical processes. It is often the
rate-limiting step in a process and has to be considered both
in the absence and the presence of reaction. New configura-
tions that enhance desirable transfer, or that inhibit undesir-
able transfer, can be designed and novel situations analyzed-
Ed has been active in all these aspects. For instance, in the
so-called diffusion-solubility mechanism for a nonporous
membrane, the solute dissolves in and diffuses across the
membrane, being released on the other side to the adjacent
medium. If the membrane is porous, the transport takes place
only in the pores, but various mechanisms (e.g., Knudsen
diffusion or capillary condensation) may be involved. Ed
has shown that a nonporous polyelectrolyte membrane can
be 4000 times more permeable to ammonia than to nitro-
gen or hydrogen, a difference that is key to separations of
agricultural chemicals.

Ed has also devised membranes for drying air and for di-
rect methanol fuel cells. A barrier membrane, with imperme-
able, overlapping flakes hindering the diffusion can be re-
sponsible for a three-hundredfold reduction in flux for virtu-
ally any solute. In one dramatic series of experiments Ed veri-
fied that air could be selectively separated using a supercon-
ducting microporous membrane. This was a consequence of
capillary condensation and not, as had previously been
thought, of the Messiner effect.

Ed has excellent rapport with students at all
levels ... [A] supporter, having Ed's teaching in
nd, observed, "He has throughout his career been
something of an iconoclast, but with a basically
positive and constructive attitude and a puckish
sense of humor-three characteristics
that are in short supply."

The greater part of Ed's research has been in membrane
transport with reaction. This can selectively increase or de-
crease the transport; the former augers useful separations
while the latter indicates the possibility of barrier packaging.
The flux is increased by a reversible reaction between a sol-
ute and a mobile reagent to form a mobile complex that dif-
fuses across the membrane and dissociates at the far side.
The best-known example of this is the facilitation of oxygen
transport in blood by hemoglobin. Working independently,
Ed, Bill Ward of General Electric, and Norman Li of Exxon
have formulated the basic ideas and have gotten some strik-
ing results. In one of Ed's systems, the flux of sodium (but
not potassium) chloride can be enhanced ten-thousandfold
by reaction. In other cases, solutes can be moved from a re-
gion of low to one of high concentration. In all cases it must
be possible for all the solutes, reagents, and products involved
to diffuse, a condition that negates certain claims.
Transport across a reactive membrane is decreased when a
solute binds irreversibly to an immobile reagent. The steady-
state flux is only slightly decreased in this case, but the time
to reach steady state can lengthen a thousandfold or more, a
phenomenon that should be useful in certain packaging situ-
ations. For example, an acid that may breach the film in ten
seconds when there is no reaction may be retained for two
hours when an immobilizing reactant is used, while for oxy-
gen the breaching time can go from ten minutes to three years.
Transport into reactive films can also change the location of
reagents in unexpected ways, as when an acid dissolves cal-
cium hydroxide or hydroxyapatite and reprecipitates the min-
eral in layers. Ed has shown that this is not a Liesegang Ring
phenomenon, but comes about by the coupling of the linear
transport and nonlinear reaction.
While Ed's work is fundamental research, his consulting
keeps him ever mindful of possible applications and the de-
sign considerations that they engender. That experience is

Summer 2001

wide-ranging, and in addition to a
long-term relationship with Phar-
macia (olim Upjohn) and more casual
contact with DuPont, Exxon, Dow, et
al., it included a year (1970-71) in S
England with Unilever. Though com-
panies rarely expect much from visit-
ing academics, Unilever was well sat-
isfied with Ed's development of a
fluoridized toothpaste that was suffi-
ciently novel that it did not transgress
existing patents. He also developed a dye of notable simplic-
ity for coloring ladies' hair, its only drawback being that there
was a complete loss of hair after three applications. There is
no record of whether or not the patent rights were sold to a
wig manufacturer.
Ed first encountered the problem of texture perception when
working with Unilever. He was able to predict and verify
how texture would be perceived by tongue or fingers. Such
psychophysical measurements might be of use in a rationale
of consumer product marketing, but Ed confesses that he never
figured out how this topic could be developed further. His
is probably the only bibliography around that has such
intriguing journal titles as Cosmet.and Toil. or Percep-
tion & Psychophysics.
More successful has been his development of membrane
modules for absorption, extraction, and separation. These have
the topology of the shell-and-tube heat exchanger, but the

tubes are hollow fiber mem-
branes, typically 300gm in di-
ameter and offering virtually no
resistance to mass transfer. Be-
cause these modules can have
interfacial areas a hundredfold
greater than the areas in com-
mon chemical separators, the
modules can increase liquid
extraction rates by a factor of three hundred and absorption
rates of H,S and HCN by factors of thirty. Ed has developed
the design equations for various geometries and has shown
that the plate height can be uncommonly sensitive to some
perturbations of the fiber geometry. He has learned how to
make the equipment more robust by using gel-filled pores, a
valuable feature for racemic separations. He has eight pat-
ents on various uses to which these modules can be put. He
even developed (but did not patent) an artificial gill that suc-
cessfully allowed a beloved canine family member to stay
quite comfortable under water, breathing oxygen like a fish.
It is easy to see that the tree of Ed Cussler's research is
firmly rooted in fundamentals and ramifies into many
branches. The temptation must be resisted to part the canopy
here and there to reveal the branch of tooth decay, or of gall-
stone pathogenesis, or of the separation of protein for use in
baby food, or of breaking the bonds of ice on Minnesota roads.
Chemical Engineering Education

These topics are but a sampling from his bib-
liography of more than 180 papers and five
books. His coauthors include the twelve
Masters and thirty-six PhD students whom
he has mentored and, occasionally, a col-
league with whom he has shared a supervi-
sion or discussed a problem. Nor is it sur-
prising that his books and papers are fre-
quently cited. A 1971 paper in the AIChE
Journal on "Membranes which Pump" has
been referenced more than five hundred
times by other authors. B C
His latest book, published in February by
Cambridge University Press and coauthored
with Geoff Moggridge, is titled Chemical
Products Design. It is a product of the para-
digm shift that has been taking place over
the last decade in which Ed, both as an edu-
cator and as a professional, has played a
leading role. The book is the immediate fruit
of his sabbatical as Zeneca Fellow at the
Shell Chemical Engineering Laboratory of
Cambridge University (1999-2000).
Almost a decade before, a previous sab-
batical (1991-2), spent at the other Cam-
bridge, had been an opportunity for his con-
cerns for the profession to gestate. Elected
a Director of AIChE in 1989, he was Vice-
President in 1993 and President the follow-
ing year. In these positions, he came to feel
that the key problem the profession was fac- The Boston
ing was the change from relatively few com-
modity processes with stable and fairly pre-
dictable markets to a product-oriented output for which de-
mand would be changeable and price variable. As far as pro-
fessional education is concerned, the intellectual and techno-
logical girth is much the same, but the orientation must be
modified. With the change goes a change of expectation. No
longer can chemical engineering graduates expect to spend
their entire careers in one company, but must anticipate chang-
ing employers, probably more than once. European universi-
ties such as Cambridge, Gronigen, and Nancy have courses
in product engineering, and they are increasingly common
over here. Cussler and Moggridge will undoubtedly be a lead-
ing text in this millennial reorientation.
Ed is known far and wide for his teaching. He is conscious
of the dramatic element in memorable teaching and will not
disdain a hand puppet or a false beard if they will make a
point stick. From time to time at Minnesota, we have sur-
veyed a block of recent alumni, feeling that their undergradu-
ate emotions recollected in the tranquility of their early ex-
perience beyond the department will give their opinions a
distinctive degree of detachment. Certainly such a group is
Summer 2001


not slow to criticize, and almost everyone's
teaching comes in for a brickbat or two-
everyone except Ed. The high opinion of
him that was formed as undergraduates re-
mained uniformly favorable. One com-
mented that Ed is "especially good at pre-
senting industrially related material"; an-
other said that she "never felt I had a firm
understanding of the material (thermody-
namics) until I had been through his course";
yet another claimed that "a visitor to one of
Professor Cussler's thermodynamics lec-
tures might find a hundred engineering stu-
dents with their mouths gaping open and
their tongues pointed upwards while the in-
structor ran around the room with an imagi-
nary machete slicing make-believe sections
through the students' extended tongues,
which were serving as personal three-di-
mensional realizations of a vapor-liquid
phase equilibrium diagram far more instruc-
tive than a chalkboard drawing or a text-
book illustration alone." Small wonder that
Ed has excellent rapport with students at all
levels. It was he who had the idea of a de-
partmental recognition ceremony on the af-
ternoon of the college's auditorium-filling
official graduation, thus giving a more per-
sonal touch that the students and their par-
ents greatly appreciate.
Ed is the personification of Juvenal's mens
athon, 1987.
sana in corpore sano. He began rowing at
Yale as an undergraduate, and has rowed
ever since when his athletic timetable, the
proximity of the Cam, or the ice on the Mississippi permits.
He was in a crew that twice finished Head of the Charles.
Ed and his wife, Betsy, manifest their mutual meetness,
not only by both being excellent teachers (she of high school
English), but also in their vacations. Not for them the sleek
automobile tour or luxury cruise-rather it may be an ardu-
ous cycling trip in France or Spain. Last year they crossed
Appenines from the Adriatic to the Mediterranean.
It was the day after he ran his first marathon that I (RA)
first met Ed, and I have always regarded it as a great compli-
ment that he dragged his aching body to my lecture and man-
aged to stay awake and even to ask a question. Since then he
has run twenty-five marathons (including the Boston Mara-
thon seven times), and he still goes to seminars. Keep your
eyes open and perchance you may catch a glimpse of his re-
turn-a melange of color slipping into his telephone-kiosk
of an office, soon to emerge as the attentive and Distinguished
Professor Edward (Clark Kent) Cussler, heading for the semi-
nar room with his sandwiches and questions. C


ChE at






City College of New York e New York, NY
C ity College was founded in 1847 as the Free
Academy, the first public college in the United
States, 15 years before passage of the Morrill
Act establishing the land-grant colleges, and before wide-
spread acceptance of the principle of public higher edu-
cation. A college "where the children of laborers and
gentlemen could study together" was the dream of New
York patrician Townsend Harris, but Harris had to agree,
in the face of opposition from Columbia College, that
the Free Academy would not offer degrees in fields that
are "the sole domain of gentlemen," such as medicine,
law, and theology. The college currently occupies a 35-
acre, gargoyle-lined, Gothic and modern campus in the
historic Hamilton Heights section of upper Manhattan,
near the home of Alexander Hamilton.
Copyright ChE Division of ASEE 2001



4: 5


The bucolic City College campus contrasts with
bustling New York City.
Chemical Engineering Education

Admission was based solely on
the City of New York (as it w
CCNY, charged no tuition until 1
the vehicle for higher education f
communities, and the college w
standing student body, giving rise
of the proletariat." When opporti
tion and research began to
develop in the United States
in the 20th century, the gradu-
ates of City College found
open doors. City College has
graduated eight Nobel laure-
ates, more than any other pub-
lic-domain academic institu-
tion in the United States.
Many members of the Na-
tional Academies of Engi-
neering and Sciences and the
Institute of Medicine are its
graduates. The college ranks
fourth in the United States in
the total number of graduates
in all disciplines who have
gone on to earn the PhD, and i
economics, and psychology. It
top-level executives in American
City College always offered pi
had a professor of civil engineer
George Goethals, the builder of
student at CCNY for three years
Military Academy at West Point
designer, David Steinman, was a
A separate School of Engineeri
and the first class in chemical
1921. The class of 1926 includes
to enter the profession. The colle
engineering curriculum through
research or graduate education, 1
had industrial experience and did
haps the best known was Morri
significant role as a metallurgist
Many graduates went on to obta
leaders in the profession. The g
elude National Academy of Eng
Newhauser (1958), Institute Pr
Arnold Stancell (1958), formerly
now a Georgia Tech professor;
co-founder of Intel; Martin Sher
Grace corporate vice-president;
H.B. duPont Professor at the U
Frederick Krambeck (1963), a seni
Three more colleges were estal

merit, and the College of
'as renamed in 1866), or
976. It therefore served as
or New York's immigrant
as able to attract an out-
to the name "the Harvard
cities for graduate educa-

the growing educational needs of the City of New York,
including Hunter College for women. (At the time, only the
School of Engineering accepted women.) These four col-
leges, all of which subsequently became coeducational, later
formed the nucleus of the present City University of New
York (CUNY), which is the third largest university in the
United States and the largest urban university. Graduate

Students at any urban institution benefit
from the resources of their "city campus." City
College students, however, benefit more than most, for
their city campus is bigger and more varied than any other,
including some of the world's greatest cultural resources...
[It] is easily accessible by public transportation from any
location in New York City and nearby communities,
so students have the opportunity to live in a
variety of urban neighborhoods while
taking full advantage of New York
City's unmatched cultural life.

t ranks first in chemistry, education was introduced in CUNY in the early 1960s, and
ranks eighth in producing PhD programs in engineering were established at City Col-
business and industry. lege. Two chemical engineering graduates, Martin Sherwin
programs in technology and (the first to receive a PhD in chemical engineering from the
ing from the earliest times. college) and Frederick Krambeck are members of the Na-
the Panama Canal, was a tional Academy of Engineering.
before transferring to the In 1962, the School of Engineering moved into a new
,and the renowned bridge building which met the needs of research-oriented graduate
City College graduate. programs. The chairman of the Chemical Engineering De-
ng was established in 1917, apartment at the time was Alois X. Schmidt. While not a
engineering graduated in researcher himself, Schmidt was an excellent and insightful
d a woman, one of the first manager who spearheaded the introduction of modern chemi-
ge had a standard chemical cal engineering into the curriculum. He hired two young
the 1950s. There was no faculty members who would grow with the department, Rob-
but many faculty members ert Pfeffer, later department chair for 15 years and subse-
industrial consulting. Per- quently provost of City College, and Robert Graff, the cur-
s Kolodney, who played a rent chair. Together with Seymour Hyman (a chemical engi-
in the Manhattan Project. neering professor who was serving as the new graduate dean
in the PhD and to become for engineering who would later become the vice chancellor
graduates of this period in- of the university), Schmidt sought out established chemical
ineering members George engineers who could help develop a graduate program and
ofessor at Georgia Tech; provide guidance for the young faculty members.
a Mobil vice-president and The first senior faculty member Schmidt recruited was
Andrew Grove (1960), the Stanley Katz, then at American Cyanamid. Katz was a math-
win (1960), a former W.R. ematician by training who worked for 15 years in the chemi-
Stanley Sandler (1962), the cal industry and had written seminal papers on such chemi-
niversity of Delaware; and cal engineering topics as polymerization kinetics and the
or consultant at ExxonMobil. optimal design and operation of staged systems. He, in turn,
blished in the 1920s to meet initiated the hiring of Reuel Shinnar, and together they es-

Summer 2001

4 PhD student working inside a Class
100 clean room to identify crystals
prepared by a newly developed
templating technique.

4 Doctoral student Nitin Kumar
prepares atomic force micro-
scope for observations of the
molecular surface topology of a
silicon wafer modified by a
molecular self-assembly

> Demonstration for industry of
the CCNY circulating fluid bed

tablished a pioneering research program using
applied probability and Markov processes to
solve problems in chemical engineering. The
use of population balances to model crystallizers,
coalescence processes, molecular weight distri-
butions, multiple tracer experiments, etc., was
developed and formulated in this group.
The group also worked on problems of stabil-
ity of flow of polymeric liquids. A 1969 paper
by Pfeffer, Shinnar, and their students, Goldin
and Yerushalmi, showed that the tensile stresses
developed by the presence of small amounts of
polymer can prevent a jet from breaking up un-
der the effect of surface tension. A photograph
from this classic work, which anticipates issues

in modem technologies such as jet printing, was shown as recently as a
year ago during a plenary lecture by Gareth McKinley at the Interna-
tional Congress on Rheology.
Katz died prematurely in 1971. His profound impact on chemical
engineering at City College is recalled annually through the Stanley
Katz Memorial Lecture, which is presented by a leader of the chemical
engineering profession. Shinnar subsequently moved his research em-
phasis from applied probability to the development of rigorous methods
for the design and control of chemical processes, especially chemical
reactors. He also pioneered a new methodology for the economic evalu-
ation of new processes that has been widely accepted in industry. In
1985, Shinnar was elected to the National Academy of Engineering,
becoming the first member of the department and the school to receive
this honor. The process-analysis component was augmented when Irven
Rinard joined the faculty in 1988. Rinard, a former associate of Katz,
had been in charge of reaction engineering and control at the Scientific
Design Company for more than 20 years. He continues to collaborate
with Shinnar in design and control research and has started an activity in
microreactor processes.
The department recruited Arthur Squires in the late 1960s to strengthen
the traditional engineering aspects of the program. Squires was a physi-
cal chemist who played a significant role in the isotope separation

Chemical Engineering Education

Faculty: CCNY's Department of Chemical Engineering
(Additional information can be found on the web page at
www.che.engr. ccny.cuny.edt )
Andreas Acrivos (PhD, University of Minnesota, 1954)
Rheology of concentrated suspensions
Dielectrophoresis in flowing suspensions
Dynamical systems theory and chaotic particle motions
Alexander Couzis (PhD, University of Michigan, 1992)
Polymorph selective templated crystallization
Molecularly thin organic barrier layers
Surfactant facilitated wetting of hydrophobic surfaces
Morton M. Denn (PhD, University of Minnesota, 1964)
Non-Newtonian fluid mechanics
Polymer processing and rheology
Polymer interfaces
M. Lane Gilchrist (PhD, University of California, Davis, 1996)
Bioengineering with cellular materials
Spectroscopy-guided molecular engineering
Structural studies of self-assembling proteins
Robert A. Graff (D Eng Sc, Columbia University, 1963)
Pollution prevention
Leslie L. Isaacs (PhD, Massachusetts Inst. of Tech., 1960)
Preparation and characterization of novel optical materials
Recycling of pavement materials
Application of thermo-analytic techniques in materials research
Charles Maldarelli (D Eng Sc, Columbia University, 1980)
Surfactant science and interfacial transport phenomena
Fluid mechanics and hydrodynamic stability
Nano science and engineering
Irven Rinard (Sc D, Massachusetts Inst. of Tech., 1962)
Process design methodology
Dynamic process simulation and process control
Microreaction technology
David Rumschitzki (PhD, U. of California, Berkeley, 1984)
Reaction engineering
Transport and reaction aspects of artery disease
Hydrodynamics of two-phase flow in tubes
Reuel Shinnar (D Eng Sc, Columbia University, 1957)
Advanced methods for chemical process design and control
Spinodal decomposition of binary solvent mixtures
Process economics
Systems problems in energy and the environment
Carol A. Steiner (PhD, University of Pennsylvania, 1983)
Polymeric hydrogels
Polymer/surfactant interactions
Controlled drug release
Gabriel Tardos (D Sc, Technion, Israel, 1978)
Powder technology
Granulation of powders
Granular flows
Herbert Weinstein (PhD, Case Institute of Technology, 1963)
Fluidization and multiphase flows
Multiphase chemical reactor analysis and design

Summer 2001

diffusion plant of the Manhattan Project and subse-
quently followed the leader of that group to the Hydro-
carbon Research Institute (where he hired Katz as an
assistant). Squires was instrumental in developing a
major research program in coal conversion processes,
resulting in the establishment of the Clean Fuels Insti-
tute, in which five departmental members were active.
The Institute was well funded and made significant
contributions to the technology of coal conversion.
Graff became director the Institute after Squires left in
1976, and the program continued for ten additional years
until research in clean coal technology was effectively
abandoned by the U.S. Department of Energy.
Pfeffer, in the meantime, collaborated with Sheldon
Weinbaum from the Mechanical Engineering Depart-
ment to establish a program on fluid-mechanics prob-
lems in packed beds and particulate systems. This
activity evolved in two directions. One was to rein-
force the tradition in fluidization initiated by Squires.
In 1977, this direction was strengthened by the arrival
of Herbert Weinstein, a 1956 graduate of the depart-
ment who established a very effective collaboration
with Exxon. Gabriel Tardos, who started his work
with Pfeffer in fluid-particle systems and who joined
the faculty in 1979, has become a recognized expert in
the field of powder technology. In 1992, Tardos
developed the first option in powder science and
technology integrated into a chemical engineering
undergraduate curriculum.
The other direction in which the Pfeffer/Weinbaum
collaboration evolved resulted in a very successful
research program on biomedical issues. This program
has had a strong impact on the field. Weinbaum, a
member of the National Academy of Engineering, is
now codirector of the Center for Biomedical Engi-
neering, which is a collaborative center with the Hos-
pital for Joint Diseases and the Hospital for Special
Surgery. A new PhD program in biomedical engineer-
ing was established last year at City College, and the
participants include a number of chemical engineering
faculty members and students.
David Rumschitzki (who joined the department in
1983) has been, partly in collaboration with Weinbaum,
studying the transport of water and macromolecules
such as lipoprotein cholesterol from the blood into the
walls of the vessels and the kinetics of cholesterol
binding to extracellular matrix. These events are among
the earliest in atherosclerosis, and they certainly play a
large role in vessel and graft susceptibility to disease.
He also collaborates with researchers at Exxon on the
mechanism of catalyst coking. Carol Steiner, who
joined the faculty in 1985, has been examining appli-
cations of unique polymeric hydrogels (developed in

her group) to sustained and targeted drug delivery. Lane
Gilchrist, who is currently in his second year on the
faculty, is working on the development of biomaterials
from cellular components in cooperation with the Hospi-
tal of Special Surgery.

A rare opportunity occurred in 1978 when Benjamin
(Veniamin) G. Levich was able to leave the Soviet Union.
Levich was a renowned physicist, a member of the Russian
Academy of Sci-
ences, and, most im-
portantly, a promi-
nent "Refusenik."
Robert Marshak, the
president of City Col-
lege and a prominent
physicist, moved
quickly and success-
fully to have Levich
appointed as one of
the five New York
State Albert Einstein
Professors in Sci-
ence. An interdisci-
plinary Institute of
Applied Chemical
Physics was estab-
Photo Courtesy of Departme
lished under Levich's
isednder Levich's The original Free Academy
direction. Levich's Manhattan, constructed
primary appointment
was in the Department of Chemical Engineering and he was
well known in the chemical engineering community for his
landmark monograph, "Physicochemical Hydrodynamics,"
which appeared in translation as one of the first books pub-
lished in Neal Amundson's Prentice-Hall series on chemical
engineering. (He also held appointments in physics and me-
chanical engineering.) Levich quickly established himself as
a major figure in chemical engineering in the United States
and his personal influence and that of the Institute was a major
stimulus for the college and especially for the department.

Levich passed away in 1987, and Andreas Acrivos was
induced to come to the college in 1988 to fill the Einstein
Professorship and become director of the Institute, which
was renamed the Benjamin Levich Institute for Physico-
chemical Hydrodynamics. Acrivos, a member of the Na-
tional Academy of Engineering and of the National Acad-
emy of Sciences, had previously been on the faculties at
Berkeley (1954-62) and Stanford (1962-1988). One of the
giants of modem fluid mechanics and chemical engineering
transport processes, he led the Institute and reinforced its
ties with chemical engineering until his formal retirement
earlier this year-the term "formal" indicating he is continu-
ing his influential research program on the mechanics of
particulate suspensions, as well as his active participation in

tll uJ I>
to th

departmental affairs. In fact, he took on two new PhD stu-
dents this past fall and gave an informal but well-at-
tended lecture course during the spring. Acrivos held
joint appointments as professor of physics and professor
of mechanical engineering, and he was editor of the Phys-
ics of Fluids from 1982 to 1998.

Morton Denn came to the college in 1999 and succeeded
Acrivos the following year as the Einstein Professor and as
director of the Levich Institute. A member of the National
Academy of Engi-
neering, Denn be-
gan his career at
Delaware, where
he became the
Allan P. Colburn
Professor. He sub-
sequently spent 18
years at Berkeley,
where he served as
Chemical Engi-
neering Depart-
ment chair and
head of materials
chemistry in the
Materials Sciences
Division of the
nemical Engineering, The City College of New York, CUNY D e
Lawrence Berkeley
mical laboratory in downtown
he latest standards in 1883.
tory. He is the edi-
tor of the Journal of Rheology and previously served as
editor of the AIChE Journal. His involvement with CCNY
chemical engineering goes back to his PhD dissertation on
optimization, which included an analysis of some work by
Stanley Katz. In 1986, he coauthored a book chapter with
Shinnar on coal gasification reactors. In that same year, with
his student, Bousfield, and collaborators Keunings and
Marrucci, he published the first complete analysis of the
1969 City College experiments on viscoelastic jet breakup.
His current research focuses on polymer processing and
polymer blends and interfaces. Denn's career was high-
lighted in an article in Chemical Engineering Education in
the spring of 1996. He also holds a joint appointment as
professor of physics.
The Levich Institute has seven full-time faculty members
(including Acrivos) with appointments in the Departments
of Chemical Engineering, Mechanical Engineering, and Phys-
ics, and a number of associated faculty. All Levich faculty
and students are housed in the engineering building. A com-
mon theme within the Institute is the behavior of complex
fluids and interfaces. Institute member Charles Maldarelli
joined the Chemical Engineering Department in 1980 and he
has maintained an active and internationally recognized re-
search program in interfacial fluid mechanics, which has led
Chemical Engineering Education

Doctoral student Manoj Sharma and Professor
Lane Gilchrist check the bioreactor. Microor-
ganisms grown in this reactor will be harvested
to recover self-assembling proteins used in the
construction of specialized biomaterials.

to the development of methods for the measurement of
dynamic surface tension and a theory for remobilizing fluid
interfaces retarded by the absorption of surface-active impu-
rities. This activity is part of a broad program in surface
behavior, some of which is done in collaboration with Joel
Koplik, a Levich Institute physicist interested in molecular
simulation of fluid and transport problems. Koplik also col-
laborates with Acrivos. Maldarelli and Rumschitzki, who is
an associated faculty member in the Levich Institute, also
have a collaboration on nonlinear stability issues in jets and
two-phase flows that are relevant, for example, in oil recovery.
Alexander Couzis joined the Chemical Engineering De-
partment in 1994 after several years at International Paper
and a year as an Industrial Fellow at the University of
Minnesota's Center for Interfacial Engineering. Couzis is
also an associated faculty member in the Levich Institute.
His research focuses on the absorption of surfactants at the
solid-liquid interface for the purpose of engineering materi-
als with specific surface properties. Applications include
control of wetting behavior, adhesion, novel sensor develop-
ment, and templated crystallization. Couzis and Maldarelli
have spearheaded the department's efforts in nanotechnol-
ogy and nanoscience, including a collaborative program with
Dupont, under the NSF nanotechnology initiative, to de-
velop a new molecular design approach for the control of
Summer 2001

polymorph-selective crystallization using self-assembling
surfactant molecules (SAMs) to functionalize solid surfaces
as nanotemplates for heterogenous and selective nucleation
and growth of three-dimensional crystals.
Other Levich faculty include Jimmy Feng, a mechanical
engineer who did a chemical engineering postdoc at Santa
Barbara with Gary Leal, and who is interested in non-
Newtonian fluid mechanics, and Mark Shattuck and Hernan
Makse, young physicists in the Levich mode who are initiat-
ing research programs in granular materials. Weinbaum is
also a member of the associated faculty at the Institute.
Individual faculty members have other interests in addi-
tion to the broad themes sketched above. Materials Science
and Engineering is an interdisciplinary effort at City College
and includes a major activity in photonics. Leslie Isaacs,
who joined the Chemical Engineering Department in 1974,
collaborates on materials issues with faculty in physics,
mechanical engineering, and civil engineering. Roberto Mauri
returned to Italy this year to join the chemical engineering
faculty at the University of Pisa, but he holds a position as
research professor and will spend summers at City College
continuing a collaboration with Shinnar on extraction using
spinodal decomposition of binary solvent mixtures.
Students at any urban institution benefit from the resources
of their "city campus." City College students, however, ben-
efit more than most, for their city campus is bigger and more
varied than any other, including some of the world's greatest
cultural resources. These resources include, as an incom-
plete list, the New York theaters (both Broadway and Off-
Broadway), the Lincoln Center for the Performing Arts, the
Metropolitan Museum of Art, the Museum of Modern Art
(MOMA), the American Museum of Natural History, and
the renowned Bronx Zoo and Botanical Garden complexes.
Recreational opportunities abound in all the boroughs of the
city and in the surrounding suburbs. City College is easily
accessible by public transportation from any location in the
city and nearby communities, so students have the opportu-
nity to live in a variety of urban neighborhoods while taking
full advantage of New York City's unmatched cultural life.
The hundreds of national and multinational firms head-
quartered in New York City and nearby communities consti-
tute one more resource of particular value to students. Con-
tact with businesses and governmental agencies is very ben-
eficial, and many of our students spend one summer as
interns in industrial laboratories.
Excitement and opportunities have always attracted people
to this dynamic city. The famous, the infamous, and count-
less others in the melting pot of Manhattan find a forum
here. On any given day, you can find numerous important
people making public appearances in the city. On those
same days, you need only stop by the local corner market or
a Central Park fountain to hear the orations of some slightly
less famous, but just as appealing, concerned citizens. 7



Why, What, When, and How?

Purdue University West Lafayette, IN 47907-1283

Separation costs often control the profitability of plants.
In most chemical, petroleum, petrochemical, and phar-
maceutical plants, separations account for 40% to 70%
of both capital and operating expenses." Separations are also
the basis for many businesses that involve the manufacture
of adsorption systems, distillation columns, extractors, mem-
brane equipment, etc. Finally, separations/mass transfer and
reaction engineering represent the two areas that are uniquely
"owned" by chemical engineering. Thus, every chemical en-
gineer should have a background in separations.
Unfortunately, beyond acknowledging the importance of
separations the consensus evaporates. Few curricula can cover
all of the separation techniques used commercially. Every
industry specializes in different separations-mechanical
separations such as filtration and settling; classical equilib-
rium-separations, including distillation, absorption, crystal-
lization, and extraction; newer membrane techniques includ-
ing gas permeation, reverse osmosis, ultrafiltration, and
pervaporation; and various adsorption and ion-exchange tech-
niques. And there are even newer processes, such as
supercritical fluid extraction, that are finding their way to
industrial usefulness.
Ideally, we would first decide what to teach and then de-
cide when to teach it. But often this does not occur because
of competing demands on the curriculum. We may select a

Phil Wankat received his BSChE from Purdue
and his PhD from Princeton. He is currently a
Professor of Chemical Engineering at Purdue
University. He is interested in teaching and
counseling, has won several teaching awards
at Purdue, and is Head of Interdisciplinary En-
gineering. His research interests are in the area
of separation processes, with particular empha-
sis on cyclic separations, adsorption, and pre-
parative chromatography.

time slot and then decide what can be taught there. Most
schools have chosen the junior year for a required separa-
tions course and the senior year for electives.
Finally, there is the question of how to teach separations. I
recommend an eclectic approach that includes classical
graphical and analytical methods, computer simulations, and
laboratory experience.

If most of the graduates of a program were to go into a
single industry, the needs of that industry could be used to
answer the "what to teach" question. We can classify indus-
tries by the separations they use. The petroleum industry is a
heavy user of distillation plus absorption, extraction, and sepa-
ration of two liquid phases. Petrochemicals would probably
add membrane separations and adsorption to this list. Phar-
maceutical companies are much more interested in centrifu-
gation, filtration, membrane separators, extraction, solution
crystallization, precipitation, and chromatography, while in-
organic chemical production often relies heavily on solution
crystallization and filtration. The commercial gas industry
uses cryogenic distillation, membranes, adsorption, and some
absorption, and the production of high-purity water uses dis-
tillation, membranes, and ion exchange. Fine chemicals may
use all of the above processes plus molecular distillation and
melt crystallization. Food processing would add drying and
freeze drying to the list, while environmental applications
often add various types of mechanical separators such as cy-
clones, decanters, electrostatic precipitators, magnetic sepa-
rators, and sedimentation. So, it is evident that the mix of
jobs taken by graduates of most chemical engineering pro-
grams is much too large to choose what to teach based solely
on the hiring industry.
A number of schemes have been devised to classify sepa-

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

Classification of Separations
Based on Similarities in Analysis

Classical Equilibrium-Staged Distillation, absorption, stripping.
extraction, leaching, crystallization from

Advanced Classical Extractive distillation, azeotropic distillation
Melt crystallization
Batch distillation
Rate-Based Membrane: gas permeation, pervaporation,
ultrafiltration, reverse osmosis,
nanofiltration. dialysis, electrodialysis
Molecular distillation
CSD in crystallization from solution
Adsorption: thermal and pressure swing,
desorbent and purge
Ion exchange
Chromatographic: column and simulated
moving bed
Mechanical Centrifuge, cyclone, decanter, demister,
expression, electrostatic separator, filtration,
flotation, magnetic, high-gradient magnetic
separator, sedimentation, sink-float
New Supercritical extraction
Liquid membrane

Chemical Engineering Courses in
Ideal Separations Curriculum


Mass and Energy Balances

Thermodyamics II. Equilibrium
Equilibrium-Staged Separations
Fluid Dynamics

Senior Rate Separation Processes
Process Dynamics/Control
ChE Laboratory II
ChE Elective

Possible Advanced Equilibrium-Stal
Separation Azeotropic, extractive
Electives Multicomponent batch
Supercritical extraction

Thermodynamics I

Heat/Mass Transfer
Kinetics/Reactor Design
Solids/Mechanical Separations
ChE Laboratory I

ChE Design
ChE Elective

ged Separations
and reactive distillation
distillation and extraction

Novel and Unusual Separations
Electrophoresis, isoelectric focusing, and isotachophoresis
Liquid membranes
Molecular distillation

Summer 2001

ration methods.'121 A simple classification based on similari-
ties in the theories is shown in Table 1. The classical equilib-
rium-staged processes can be operated in staged equipment
such as columns and can be analyzed with a stage-by-stage
calculation procedure, assuming each stage is in equilibrium.
The equilibrium assumption eliminates the need for a mass
transfer analysis. Of course, these systems can be operated in
different equipment such as random and structured packing,
and different analysis procedures can be used. The advanced
classical methods can also be analyzed by a stage-by-stage
calculation, but with added complexity. Extractive and
azeotropic distillation add a third component that makes ap-
plication of graphical methods difficult; melt crystallization
involves movement of solid, and batch distillation adds time
as a variable.
Rate-based separations require a mass-transfer analysis.
Membrane separations are probably the easiest of the rate-
based separations to teach since they often operate at steady-
state. Although molecular distillation is not difficult to teach,
it is usually ignored. Crystallization from solution appears in
Table 1 twice because the concentration of products can be
calculated from equilibrium considerations, but the very im-
portant crystal size distribution (CSD) requires population
balances. Adsorption, ion exchange, and chromatography are
the most difficult to teach since they are rate-based and are
usually operated as time-dependent processes.
Mechanical separations are involved with the separation
of bulk phases. Thus, the mechanisms of separation are in-
herently different than the diffusional mechanism of equilib-
rium-staged and rate processes. It has been common to cover
the mechanical separations in a fluids course, in a separate
course on solids processing, or not at all. Unfortunately, "me-
chanical separations are grossly under-represented in the typi-
cal curriculum relative to industrial practice."['6
The newer processes can be moved into the equilibrium-
staged (supercritical extraction) and rate-based (liquid mem-
branes and electrophoresis) portions of the table if educa-
tional materials are available. Only supercritical extraction,
however, seems to have enough industrial application to jus-
tify its inclusion in an undergraduate course.
An alternate approach is to pick a "typical" process in the
field being studied and to list all of the separations employed
in the process. These are then studied in their proper order.
This approach has been most commonly employed in
bioseparations.'7I8 With this approach, one covers some sepa-
rations from the mechanical, rate-based, and equilibrium sepa-
rations in Table 1. The advantage of the approach is relevance;
the disadvantage can be lack of depth.
Ideally, undergraduates would study most of the separa-
tions in Table 1 in depth. This would require three courses-
equilibrium-staged operations, rate-based separations, and
mechanical separations. I believe industry would like the re-


suiting Ideal Separations Curriculum (Table 2). Compared
to many chemical engineering curricula, this curriculum is
light in transport and design courses and may have one extra
ChE course. Unfortunately, the separation experts have been
unable to convince their faculty colleagues that separation is
more important than everything else in the curriculum!

S. there is the question of how to
teach separations. I recommend an
eclectic approach that includes classic
graphical and analytical methods,
computer simulations, and
laboratory experience.

The old compromise curriculum used equilibrium-staged
separations with a heavy dose of distillation as the "separa-
tions course." Mechanical separations were touched on in flu-
ids courses and in lab. Rate-processes occasionally appeared
in design courses, but the students really did not know how
to design them. Dual-level electives might be available on
advanced classical processes, rate-based separations, and sol-
ids processing. Details for an equilibrium-staged course and
a rate-based separations course are given by Wankat, et al. [S
At a time when almost all chemical engineering graduates
went into the petroleum and chemical process industries, this
compromise made sense. "Distillation is used to make 90-
95% of all separations in the chemical process industry."'1'
Since the ideal separations curriculum is unlikely to hap-
pen, what is an appropriate compromise for current times
when chemical engineers work in such a wide variety of in-
dustries? I suggest the following: the "separations course"
should cover the classical equilibrium separations in both
staged and packed columns (-80%) and membrane separa-
tions (-20%). These topics fit together because the separa-
tions tend to be complementary in process plants and the peda-
gogical difficulty is about the same. A modest amount of
mechanical separations (filtration and sedimentation) should
be included in the fluids course, and laboratory and design
classes should include a variety of separations from the three
major categories in Table 1. Dual-level electives should be
available in both rate-based separations and solids process-
ing for those students who want more depth.

Some chemical engineering curricula teach separations in
the sophomore year, some in the junior year, and some in the
senior year. Equilibrium-staged courses have an advantage
in that they can be taught at any time after the basic mass and
energy balance course. They can be taught before mass trans-

fer if this will help balance the curriculum. An interesting
experiment and book combined equilibrium-staged separa-
tions processes with mass and energy balances in a sopho-
more course.191 This book was not widely adopted, perhaps
because it required changing well-established curricula. A
more common place for equilibrium-staged separations is in
the junior year after thermodynamics and either
before or at the same time as mass transfer.
A few programs have put the main separations
course in the senior year, but it is more common
cal to have laboratory and design courses that require
separations as a prerequisite in the senior year. The
senior year is the most common time for technical
We also teach separations to graduate students.
This used to be the course where students learned
computer calculations for multicomponent distil-
lation, absorption, and extraction, but with the advent of pow-
erful simulators these topics have been moved into the un-
dergraduate curriculum. Graduate school is the ideal time to
teach adsorption and chromatography, which are conceptu-
ally more difficult than most of the other separations. Ideally,
schools would make an advanced separations course a core
requirement in the graduate program. This would be particu-
larly appropriate in a nonthesis masters programs since it
mainly educates students for industrial careers. One possible
core course would start with an overview of equilibrium and
rates separations and then focus on either adsorption and
chromatography or mechanical separations. An alternative
is a course focusing on modern analyses of advanced dis-
tillation topics, including azeotropic, extractive, and re-
active distillation.1101 A third, more general, alternative is
an advanced mass transport course with applications in
separations and other areas.

How to teach separations may be the most interesting ques-
tion of all, and of course, it interacts with the why, what, and
when. The classical method of teaching separations was to
use graphical methods, including McCabe-Thiele, Ponchon-
Savarit, and triangular diagrams. Lectures were comple-
mented by laboratory operation of distillation, adsorption,
extraction, and evaporation equipment, sometimes of fairly
large scale. As computers became readily available, however,
the graphical approaches were supplemented with assign-
ments to write FORTRAN code for distillation columns.
Graphical methods have the advantage of helping students
visualize the separation, but they no longer represent the
modern practice of chemical engineering.
Modern chemical engineering practice to design and simu-
late equilibrium-staged separations to a large extent involves
using commercial process simulators such as AspenPlus,

Chemical Engineering Education


ChemCad, Hysys, and Prosim. To be prepared for commer-
cial practice, students need experience simulating and design-
ing equilibrium-staged separations using a commercial simu-
lator. Although the simulators have differences, the one that
is used is probably not important (the companies that sell
simulation packages may disagree). Unfortunately, students
often treat simulators as black-boxes and tend to believe the
results they obtain without further checking. Thus, fundamen-
tals and hand calculations (graphical or with a calculator)
should still be required. It is also useful to require students to
repeat a simulation with different equilibrium correlations.
Trying to explain the differences in their results will con-
vince many students that the choice of VLE or LLE corre-
lations is critically important. Suddenly, thermodynam-
ics is relevant!
I believe that graphical methods still have an important place
in the curriculum since they foster visualization and serve as
a thinking tool. Practicing engineers commonly use McCabe-
Thiele diagrams to understand or help debug simulation re-
sults. Modern tools such as simulators or spreadsheets'11I can
be used to draw accurate graphical solutions and thus remove
the tedium associated with graphical solutions. Laboratory
experience is still necessary since simulators need data, and
simulations do not always match reality.
Best-practice principles for teaching should, of course, be
employed.'21 Introductory courses need to be structured to
lead from inductive to deductive reasoning. Start with simple,
specific examples and build to more complex cases, then gen-
eralize and develop an abstract understanding of the analy-
sis. Finally, deductively show how other separations can be
designed using similar techniques (see Haile'3' for a com-
plete description of this procedure). Reviews of previously
studied techniques can be done deductively. Be sure students
actively process material.
Simulators can be incorporated into lecture courses by
scheduling a computer lab that meets approximately every
other week. Essentially no class time needs to be spent train-
ing students to use steady-state simulators. With a good set
of instructions and help from the laboratory teaching assis-
tants, the students can become proficient with the separation
parts of the simulator while they solve problems. The same
simulator should be used throughout the curriculum. Students
will then obtain more practice solving separation problems
in their laboratory and capstone design courses.
Simulators are beginning to be used extensively in indus-
try for the design of rate-based separations. The adsorption
and chromatography simulators are powerful, but quite com-
plicated. In a few years these simulation packages should be
used in graduate courses on rate-based separations.

The questions in the title have been answered as follows:


Because separations have overwhelming economic signifi-
cance and they are at the core of chemical engineering.


All students should study the classical equilibrium-staged
processes and receive an introduction to a rate-based process
such as membrane separators. Required courses or electives
should be available in rate-based separations and mechanical
separations. Different separation experiments should be avail-
able in laboratory, and separation should be an integral part
of senior design projects. Graduate students should study
adsorption and chromatography or mechanical separations.


The required course should be taught when it fits into the
curriculum, which is often the junior year. Electives are nor-
mally taught in the senior year. A graduate core course in
separations is recommended.


Lecture courses should be integrated with a computer labo-
ratory for practice with a modem process simulator. Intro-
ductory courses should follow an inductive pattern. Both
graphical and analytical methods should be included. The
courses) in separations should be reinforced with sepa-
rations laboratory experiments and design projects that
include separations.

1. Humphrey, J.L., and G.E. Keller, II, Separation Process Technology,
McGraw-Hill, New York, NY (1997)
2. Keller. 11. G.E.. Separations: New Directions for an Old Field, AIChE
Monograph Series. 83, 17 (1987)
3. King, CJ., Separation Processes, 2nd ed., McGraw-Hill, New York,
NY (1980)
4. Wankat P.C., Rate-ControlledSeparations, Kluwer, Amsterdam (1990)
5. Wankat, P.C.. R.P. Hesketh, K.H. Schulz, and C.S. Slater, "Separa-
tions: What to Teach Undergraduates," Chem. Eng. Ed., 28, 12 (1994)
6. Bryan, Paul, private communication, April (2000)
7. Belter, P.A.. E.L. Cussler, and W.-S Hu, Bioseparations: Downstream
Processing for Biotechnology, Wiley, New York, NY (1988)
8. Blanch, H.W., and D.S. Clark, Biochemical Engineering, Marcel
Dekker. New York, NY (1997)
9. Luyben, W.L., and L.A. Wenzel, Chemical Process Analysis: Mass
and Energy Balances, Prentice-Hall, Englewood Cliffs, NJ (1988)
10. Doherty, M.F., and M.F. Malone, Conceptual Design of Distillation
Systems, McGraw-Hill, New York, NY (2001)
11. Bums, M.A., and J.C. Sung, "Design of Separation Units Using Spread-
sheets," Chem. Eng. Ed., 30, 62 (1996)
12. Wankat, P.C., and F.S. Oreovicz, Teaching Engineering, McGraw-Hill,
New York, NY (1993) Available free at>
13. Haile, J.M., "Toward Technical Understanding: Part 4. A General Hi-
erarchy Based on the Evolution of Cognition," Chem. Eng. Edi., 34,
48 (2000) 1

Summer 2001

S SOclassroom



For Advanced Process Control Development

University of Calgary Calgary, Alberta, Canada T2N IN4

oday's process control field is such that control tech-
niques that were considered advanced even ten to
twenty years ago are now becoming commonplace."
Model predictive control (MPC) in all it's incarnations is a
good example-today there are well over two thousand MPC
controllers reported to be in operation industrially.12' Despite
this abundance of MPC technology, however, commercial
simulation software packages have been slow to incorporate
MPC algorithms. Even when they are included, the algorithms
are prescribed and the software does not allow for
customization of the algorithm(s) by users such as process
engineers. This can be attributed to the fact that there are
many MPC algorithms and it would take large develop-
ment teams to incorporate them all; but even if this were
possible, it would not be particularly useful for the test-
ing of a new algorithm.
This limitation must be accepted unless you decide to pro-
gram your own code to simulate your own process and con-
trol algorithm, using a programming language such as C++
or Visual Basic for Applications (VBA). This approach is
time-consuming, however, and is typically attempted only by
process engineers with prior experience in such an exercise.
This lack of both the flexibility of commercial packages
and the experience or education required to build one's own
simulator and control algorithm can cause process engineers
to steer clear of MPC, even though it may provide the solu-
tion they are looking for and despite its relative abundance
and growing acceptance in a number of industries. This bar-
rier to understanding and implementation also exists for many
other related advanced process control (APC) technologies
that are not as widespread as MPC.
From an educational perspective, this barrier to implemen-
tation has also largely prevented the facile inclusion of MPC

Hysys Excel Matlab

Figure 1. Software communication pathways.
and other APC technologies even in senior, advanced under-
graduate process control technical electives, e.g., as advo-
cated by Edgar.'341 Even some excellent graduate courses
emphasize the fundamentals and steer clear of APC,'5'
leaving those students who wish to enter the field of pro-
cess control devoid of practical experience with MPC and
APC algorithmns.
This paper presents a simulation environment, composed
of three readily available commercial software packages, that
allows for quick and easy development of custom APC
schemes on a wide variety of processes. This effectively re-
moves the implementation barriers described above and
allows the advanced undergraduate and graduate student
an opportunity to study and implement APC schemes of

James van der Lee is a full-time graduate student in Chemical and Pe-
troleum Engineering at the University of Calgary He received his BS
(1999) in chemical engineering from the University of Calgary. His re-
search focuses on the model predictive control of a pilot amine absorp-
tion/stripping plant.
Don Olsen is a part-time graduate student in Chemical and Petroleum
Engineering at the University of Calgary and an applications engineer
with Hyprotech Ltd. in Calgary He received his BS in chemical engineer-
ing from the University of British Columbia, and his research focuses on
model predictive control within the framework of a vinyl acetate process
Brent Young is Associate Professor of Chemical and Petroleum Engi-
neering at the University of Calgary. He received his BE (1986) and his
PhD (1993) degrees in chemical and process engineering from the Uni-
versity of Canterbury in New Zealand. His teaching and research inter-
ests center on process control and design.
William Svrcek is Professor of Chemical and Petroleum Engineering at
the University of Calgary He received his BSc (1962) and his PhD (1967)
degrees in chemical engineering from the University of Alberta. His teach-
ing and research interests center on process control and design.

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

varying complexity, thus allowing for a level of under-
standing of APC that only a "learning-by-doing" ap-
proach can provide.

The methodology behind this project is similar to that
used by an ever-increasing number of chemical process
control authors and educators in that it uses commercial
software in order to perform tasks that, while important,
would to a certain extent impede specific control educa-
tion objectives. There are several successful examples of
this approach'671 that quite effectively use Matlab to handle
modeling and solution methods to clearly demonstrate a
variety of fundamental control concepts. But given the
nature of the problem of being able to overcome the imple-
mentation issues associated with APC for educational pur-
poses in a way that would provide easy implementation in
advanced technical elective or graduate classrooms, the
following characteristics were determined to be important:
The need for an interactive dynamic simulation
environment capable of using a wide variety of
process models.
A means for interacting with this environment so
that custom algorithms can be implemented.
An ability to be able to see what steps are
occurring as they happen, while the simulation is
Tools that contain many of the standard opera-
tions required for APC applications.
Any software packages) that meets these requirements
has the potential to remove the implementation barrier. We
found that Hyprotech's Hysys, Microsoft's Excel, and
Mathwork's Matlab could be configured in such a manner
that all these requirements were met. Excel is the industry
standard for spreadsheets, and as a result many programs
include provisions for two-way communication with Ex-
cel. Both Hysys and Matlab contain such links, and it is
possible to use both links simultaneously in order to ex-
ploit the strengths of all three programs for use in APC
applications. Figure 1 illustrates the basic communication
pathway when the three programs are linked.
The benefits of this type of system are rigorous steady
state and dynamic plant simulation (e.g., Hysys), a large
library of functions useful for APC applications (e.g.,
Matlab), and powerful data handling and visualization tools
(e.g., Excel) in software packages that are already familiar
to many chemical engineers. The steps involved in creat-
ing a simulation using these three programs in conjunc-
tion with each other are:
A dynamic simulation case of the process required.
It is necessary to make note of the values that will

need to be read from and written to the process for the
algorithm to work effectively.
Add the necessary read/write variable on the variable page
of the 'data book.' Create a new 'Process Data Table' (PDT)
by pressing add on the PDT page in the 'data book,' and
then add the desired variables by checking the showbox on
the PDTpage. Then press the view button, add 'tag names,'
and select the 'access mode'(read, write, or read/write)for
each variable. It should be noted that the order in which the
show boxes are selected is the order in which the variables
appear in the 'PDT set up'page (accessed by pressing the
view button). Because of this property, it has been our
experience that it is easiest to group read, write, and read/
write variables and add them to the PDT in blocks, making
note of the order in which each variable is added.
Save the updated simulation case and make note of itsfile
name and location.
Start Excel and open the Visual Basic for Applications (VBA)
Editor Make sure the Excel link, Matlab Automation Server
Type Library, and Hysys Type Library references are selected
by selecting References in the Tools menu list in the VBA
editor This ensures that VBA will recognize the Matlab and
Hysys object types.
Insert a 'Module' into VBA and add the code as the numeri-
cal order of the following steps indicate.
Step 1. Variable declaration (see Figure 2).

Option Explicit 'Ensures that undefined variables are not allowed in code
'Objects required to bind Hysys/Process Data Table(PDT) to Excel/VBA
Public HyApp As HYSYS.Application
Public SimCase As SimulationCase
Public dt As HYSYS.DataTable
'Variables that are used in communicating information from the PDT to Excel/VBA
Public Wtags As Variant
Public WvaluesO As Double
Public Rtags As Variant
Public Rvalues As Variant
Public RWtags As Variant
Public RWvalues As Variant
'Special Flags used for error checking, in regards to establishment of
Public dtValid As Boolean
Public simCaseValid As Boolean
Public hyAppValid As Boolean
'counter variable used to keep track of when the control algorithm should execute
Public i As Integer
'Excel worksheet Initialization
Public xlSheet As Excel.Worksheet
'Variables used in Registering of HYSYS/Excel/VBA Interface
Public notifyeventl as EventSink

'used to control the integrator from Excel/VBA
Public integrator As Variant

Figure 2. Step 1 Variable declaration.

Summer 2001

Figure 3. Step 2 Main function. Llean up coae ror Notly vent
Dim result As Boolean
result = dt.RemoveNotifyEventSink(notifyevent I)
Set notifyeventl = Nothing

Sub Init() dt.EndTransfer
On Error Go To InitError Set dt = Nothing
'Regular HYSYS case Initialisation End If
Set HyApp = CreateObject("HYSYS.Application")
HyApp.Visible = True If simCaseValid = True Then
hyAppValid= True SimCase.Close
Set SimCase = GetObject("c:\your Hysyscase.hsc") Set SimCase = Nothing
SimCase.Visible = True End If
simCaseValid = True If hyAppValid = True Then
'initialises the active Excel worksheet HyApp.Quit
Set xlSheet = Sheets("Sheetl") Set HyApp = Nothing
'the following space could be used to initialised other VBA variables, End
'transfer data to or from Matlab or perform initialisation using Matlab functions
i = 0 'sets itegrator step counter to zero Exit Sub
Cleanup Error:
Exit Sub MsgBox "Error in Quit" & ErrDescription
Init Error: End Sub
MsgBox "Error initializing" & ErrDescription Figure 5. Step 4.
End Sub

Sub BindData Table( 'Dispatch Interface
On Error Go To Bind Error Dim instanceName As Variant
Dim result as Boolean
'Bind datatable to object Private Sub ClassInitialize(
Set dt = SimCase.DataTables.Item(0) 'Initiation Steps related to notify event interface if needed are placed
'Bind object to EBSink class 'here
Set notifyeventl = New EventSink End Sub
'register instance with datatable
result = dt.AddNotifyEventSink("LBSink". notifyeventl) Public Function AdviseEvento
'Enable Data Transfer 'increments time counter with every solver event
dtValid = True i = i +
'flashes counter to Excel worksheet
Wtags = dt.WriteTags 'object linked to write only variables in PDT xlSheet.Range("b22") = i
Rtags = dt.ReadTags 'object linked to read only variables in PDT 'after a predetermined # of integrator steps the contents of the if
RWtags = dt.ReadWriteTags 'object linked to read/write variables in PDT 'statement are implemented

Exit Sub If i = Switch Then
Bind Error: APCAlgorithm
MsgBox "Error in Binding Data Table" & Err.Description i =
Cleanup End If
End Sub End Function
Figure 4. Step 3. Figure 6. Event Sink Class Module.
Figure 4. Step 3. Figure 6. Event Sink Class Module.

Chemical Engineering Education

Sub Main()
'lines of code in italic type are used for error handling
'and will allow easier debugging and to ensure the HYSYS/
'Excel/VBA is terminated in the event of an error
On Error GoTo mainerror
'Binds Excel/VBA to the PDT
BindData Table

Form.Show 'Shows Form
Exit Sub
main error:
MsgBox "Error in Main" & Err.Description
End Sub

SubAPC Algorithm()
On Error GoTo APC_AlgorithmError

'Place Advanced Process Control Algorithm here

Rvalues = dt.GetValues(Rtags) 'reads data from read tags in PDT
dt.SetValues Wtags, Wvalues'writes to write tags in PDT

Exit Sub

MsgBox "Error in APCAlgorithm" & ErrDescription
End Sub

Sub Cleanup(
On Error GoTo CleanupError
'This procedure terminates the HYSYS/Excel/VBA
If dtValid = True then

Public Sub QuitBttn_Click()
On Error GoTo QuitError
'allows user to terminate transfer using VBA GUI
Exit Sub
MsgBox "Error in Quit" & E: Description
End Sub
Public Sub IntegratorStartClick()
On Error GoTo StartError
'allows user to start Hysys integrator using VBA GUI
Set integrator = SimCase.Solver.Integrator
integrator.IsRunning = True
Exit Sub
MsgBox "Error in Start" & Err.Description
End Sub
Public Sub IntegratorStop_Click()
On Error GoTo StopError
'allows user to stop Hysys integrator using VBA GUI
Set integrator = SimCase.Solver.integrator
integrator.IsRunning = False
Exit Sub
MsgBox "Error in Stop" & ErrDescription
End Sub
Figure 7. MPC control form.

out I
Quit '


: Stop: :

Figure 8. Visual Basic button code.

Summary of the Excel-Matlab Link Commands

Matlab Function Syntax
mlevalstring"Matlab command"

mlputmatrix mlvar
"worksheetRange "

mlgetmatrix mlvar

mlputvar mlvar, VBvar

Mlgetvar mlvar. VBvar

Performs the string of Matlab enclosed in
the quotes
Copies the matrix defined by 'worksheet-
Range' in the Excel worksheet to Matlab
variable 'mlvar'
Copies the contents of the Matlab variable
'mlvar'to the Excel worksheet. 'work-
sheetcell'represents the location of the
upper left-hand cell of the matrix.
Copies the contents of the Visual Basic
variable 'VBvar'to the Matlab variable
Copies the contents of the Matlab variable
'mlvar' to the Visual Basic variable

* Step 2.
Main function (see Figure 3). This function calls func-
tions that initialize the Hysys-Excel link, binds the PDT
to Excel variables, and causes the 'Form' GUI to be

* Step 3.
Functions that initialize the Hysys Excel/VBA link and
bind the PDT to VBA variables (see Figure 4).
* Step 4.
Functions that contain the APC algorithm and terminate
the Hysys-Excel link (see Figure 5).
* Step 5.
Insert a 'Class Module,' change its name to 'Event Sink,'
and add the following code (see Figure 6).
* Step 6.
Insert and create a 'Form' of similar structure to that in
Figure 7 and insert the code in Figure 8 for the appropri-
ate buttons.
The previous steps result in the basic structure of a Hysys-
to-Excel link that will recognize when the simulation case un-
dergoes a solver event, which may be either the steady-state
solver updating the solution for a change in operating condi-
tions or when the dynamics solver completes a time step.
At this point it would be possible to fill areas as indicated
in the code in Figure 8 with an APC algorithm, using VBA
and Excel alone, and then run an APC-enabled simulation
case. But this would typically involve writing a substantial
amount of code for routine matrix manipulation procedures
and data handling, etc., effectively overshadowing the APC
algorithms if the user is inexperienced. This is where the
"power" of the Excel-Matlab link is most apparent. By al-
lowing direct access to all of Matlab's functions and the abil-
ity to read and write values to both the Excel worksheet and
VBA variables through function calls (summarized in Table
1) in the VBA code, the majority of the student's time can be
spent developing and testing various APC algorithms. In fact,
Matlab's toolbox functions, such as those from the MPC
toolbox could also be used in this environment if one
wished to implement Matlab's algorithms for APC on a
case-study plant.
The following case study is an example of how to fill in
some of the blanks in the code above to obtain a useful algo-
rithm. It also is provided to give examples of PDT format,
Matlab calls via the Excel link, and the associated VBA code.
The example details one way of implementing a series of
pseudo random-binary sequences (PRBS) used in the
identification of a distillation column. The distillation is one
column of the Dimethyl Ether production described in Turton,
et al. '8 The column separates a stream primarily composed
of methanol (25-40 wt%) and water (75-60 wt%) ranging in

Summer 2001

.. .~a. ~ ~ *.1. T jJ ~ .:AL..

suran8 D C 9flD
162 924

Wnugrao s t, 5.e (SI

~ L~LeUiut9ajnc lIjscrordsra Ituuicya Ibuvalli In~nhwUe 18mm sDg


300 8 8 4 151
e nalharep S unioorder ighvIs lwceris nom lvl firm stepOinn
c 300 8 0847 5847 6347 10

comp OP comp I
588 1800 delay bolen 1a and 2nd prbs
tmpOP 1empi
8347 1200

deisjbeb ptl s md dy lsd i
6D 7200



A Figure 9. Format of Excel worksheet
used in PRBS example.

0 Figure 11. Additions/modifications to
the variable declarations and init()
for the PRBS example.

flow from 6000 to 12000 lb/h. The distillation
is performed at 35-40 psia and produces high-
purity water of lower than 220 ppm methanol
in a column of 17 theoretical stages. The best
conventional PI control configuration was
determined to be LVt91 using reflux (L) to con-
trol for top-tray temperature and boil-up (V)
via reboiler duty to control for bottoms metha-
nol composition in water.

Figure 9 shows the Excel workbook that is
used to hold the data that is necessary to con-
figure the individual PRBS signals, the delay
between the two signals, the starting "b" value
that allows the identification process to start
at any desired pont, the means to display the
link status, and the progress of the number of
steps since the last data exchange. Figure 10
shows what the PDT should look like for this
case. Figures 11, 12, and 13 show the neces-
sary additions/modification to the VBA code.

Figure 14 shows a typical result when the
above additions/modifications are made and
the simulation is run.

Although the environment is extremely flex-
ible and provides easy set up, these benefits
would be negated if this environment proved
to dramatically slow the speed of the integra-
tor. The performance of the simulation can be
measured by comparing the real-time factor
(RTF) [which is defined by (simulated time
interval)/(actual time required to compute
simulated time interval)] of a simulation us-

'The following changes should be made to the variable declarations
'remove variables associated with read tags and modify the Wvalues as follows
Public Wvalues (2) As Double
'add the following variables
Public b As Integer 'variable that allows identification to start at given point
Public comptime As Double variable that holds number of integrator steps for time in comp PRBS
Public temptime As Double variable that holds number of integrator steps for time in temp PRBS

Public delay As Double variable that holds number of integrator steps for time delay between PRBS signals

'The following should replace Sub init()
Sub Init()
On Error GoTo InitError
'Regular Initialization
Set HyApp = Create Object("HYSYS.Application")
HyApp.Visible = True
hyAppValid = True
Set SimCase = GetObject("c:\yourdirectories\yourfilename")
SimCase. Visible = True
simCaseValid = True
Set xlSheet = Sheets("Sheetl")
initializationn of counter and starting variables
b= xlSheet.Range ("b2")
inputs necessary values into Matlab
mlputmatrix "compchar", xlSheet.Range ("a6:f6")
mlputmatrix "tempchar", xlSheet.Range("a8:f8")
mlputmatrix "delaytime", xlSheet.Range ("j6")
mlputmatrix "integerstep", xlSheet.Range ("e2")
calculates PRBS signals, and number of integration steps to for delay and time to next move(in terms of i)
mlevalstring "compdist= idinput(compchar(l),'prbs',[compchar(2) l],[compchar(4) compchar(3)])"
mlevalstring tempdist= idinput(tempchar(l),'prbs',[tempchar(2) l].[tempchar(4) tempchar(3)])
mlevalstring" comptime= round((60*compchar(6)/integerstep))"
mlevalstring temptime= round((60*tempchar(6)/integerstep))"
mlevalstring delayi = round((60*delaytime/integerstep))"
'the following prints the above results to screen
mlgetmatrix "comptime", "h6"
Matlabrequest 'forces data transfer to occur between vba/Excel and matlab
mlgetmatrix "temptime", "h8"
mlgetmatrix delayyi, "k6"
Exit Sub
MsgBox "Error initializing" & Err.Description
End Sub

Figure 11. Additions/modification to the variable declarations and init()
for the PRBS example.

Chemical Engineering Education



Q ~____.

EnMe 7

TsnsW 7

Object Variable Value Units Tag Access Mode

Reflux Rate flow controller I1
Reboiler Duty Flow Controller

5688 Ib/hr Reflux Rate [Temp Control MV)
63.47 % Reboiler Duty (Comp Control MV)

I View DataBook...

'The following procedures need to be added to the module
Sub WriteOPcompO
On Error GoTo WriteOPcomperror
'Procedure that writes values to PDT when prbs signal is on temperature is activated
'the following transfer Matlab data to Excel worksheet
mlevalstring Compp = compdist(a)"
mlgetmatrix Compp". "g6" 'calculated prbs value for temp disturbance
mlevalstring "temp = tempchar(5)"
mlgetmatrix "temp", "g8" 'nominal value for temp
"The following transfers data to PDT
Wvalues(O) = xlSheet.Range("g6") 'input to comp
Values (1) = xlSheet.Range("g8") 'input to temp
dt.SetValue Wtags. Values
Exit Sub
Write OPcomperror:
MsgBox "Error in WriteOPcomp & Err.Description
End Sub
Sub Write OPtempO
On Error GoTo Write OPtemperror
Procedure that writes values to PDT when PRBS sigal is on temperature is activated
the following transfer Matlab data to Excel worksheet
mlevalstring Compp = compchar(5)"
mlevalstring "temp = tempdist((a-compchart( )-1))"
mlgetmatrix "temp", "g8" 'calculated prbs value for temp disturbance
'The following transfers data to PDT
Wvalues(O) = xlSheet.Range ("g6") 'input to comp
Values (1) = xlSheet.Range("g8") 'input to temp
dt.SetValues Wtags. Values
Exit Sub
Write OPtemperror
MsgBox "Error in Write OPtemp" & ErrDescription
End Sub
Sub WriteOPnom)
On Error GoTo Write OPnomerror
'Procedure that write nominal values to PDT
'the following transfer Matlab data to Excel worksheet
mlevalstring Compp = compchar(5)"
mlgetmatrix Compp", "g6"
mlevalstring temp = tempchar(5)"
mlgetmatrix "temp", "g8"
'The following transfers data to PDT
Values (0) = xlSheet.Range ("g6")
Values (1) = xlSheet.Range ("g8")
dt.SetValues Wtags. Values
Exit Sub
Write OPnomerror
MsgBox "Error in WriteOPnom" & Err Description
End Sub

Summer 2001

A Figure 10. PDTfor the PRBS example.

4 Figure 12. Additions to the VBA module
for the PRBS example.

V Figure 13. Additions/modifications to
AdviseEventf) for the PRBS example.

'Replace the contents of public function events with the following
Public Function AdviseEvent()

i = i + 1 'i is a counter which counts the number of solver steps that occurred
xlSheet.Range ("b22") = i 'flashes i to the worksheet
xlSheet.Range ("c2") 'flashes b to the worksheet
mlputmatrix "a". Cells (2.3) 'allows Matlab to see the progress in the identification
'The contents of this if statement are responsible for the first prbs sequence
If i = xlSheet.Range("h6") And b <= xlSheet.Range ("a6") Then
End If
'The contents of this if statement are responsible for the delay between sequences
Ifi = xlSheet.Range("k6") And b = (xlSheet.Range("a6") + 1) Then
Write OPnom
End If
'The contents of this statement are responsible for second prbs sequence which follows the
If i = xlSheet.Range("h8") and b>=(xlSheet.Range("a6")+2) And b<(xlSheet.Range("a6") +
xlSheet.Range("a8") + 2) Then
Write OPtemp
End if
'Stop integrator when identification sequence ends
If b = (xlSheet.Range("a6") + xlSheet.Range ("x8") + 2) Then
Set integrator = SimCase.Solver.lntegrator
integrator.lsRunning = False
End If
End function

Figure 14.
Typical result using the
environment for the PRBS
(The x-axis shows
simulation time in hours,
minutes, and seconds,
and the y-axis shows the
trends of various
process variables.

Figure 15.
Response to a
bottoms methanol
composition set point
change using PID
(The x-axis shows
simulation in
hours, minutes, and
seconds, and the
y-axis shows the trends
of various

Figure 16.
Response to a bottoms
methanol composition
set point change
using a linear 2x2
DMC algorithm
developed using the
(The x-axis shows
simulation in
hours, minutes, and
seconds, and the
y-axis shows the
trends of
process variables.)

6 23.2- ... ..

Enn nn

,,, lf,,. n. nnn n n nnnn n .


JU Uu U LUuu uuu_ UU Uli L .

6so904838 5s51 6545 .3s 4.6 5 60 201263

564 042072 56 7480 2


Chemical Engineering Education

ing the environment to one that does not. Figure 15 shows
the system response using PID controllers to control for
bottom's composition and top-tray temperature. Figure 16
shows the same distillation process with control carried out
using a linear 2x2 Dynamic Matrix Controller (DMC)"'O"
that has been implemented using the link described in this
paper. It can be seen that the RTF for the 2x2 DMC controller
case is comparable to that of the PID controller case, and
also gives better performance in terms of controller move-
ment and oscillation around the set point.

The development of an integrated, real-time computing
environment for advanced process control development and
education using Hysys, Excel, and Matlab linked with each
other has been outlined in this article. The methodology used
to develop the environment was detailed in order to enable
the reader to substantially reduce the learning curve involved
in developing the communication structure itself, thus allow-
ing a means to focus the attention onto a large variety ofAPC
algorithms. The potential of the environment has been dem-
onstrated using an example that implements a PRBS iden-
tification sequence on a methanol water distillation col-
umn simulation.

1. Ramaker, B.L.. H.K. Lau, and E. Hernandez, "Control Technology

Challenges for the Future." in Proc. CPC V J.C. Kantor, C.E. Garcia,
and B. Carnahan. eds. AIChE Snyp. Series No. 316, 93, 1 (1997)
2. Qin. S.J., and T.A. Badgwell. "An Overview of Industrial Model Pre-
dictive Control Technology," in Proc. CPC V J.C. Kantor, C.E. Garcia,
and B. Carnahan, eds, AIChE Symp. Series No 316, 93, 232 (1997)
3. Edgar, T.F.. "Process Control Education: Past, Present, and Future,"
in "Chemical Engineering Education: Curricula for the Future," Proc.
Ilndo-US Seminar: D. Ramkrishna. P.B. Deshpande, R. Kumar, and
M.M. Sharma. eds. Bangalore, India, pp. 117 (1998)
4. Edgar. T.F. "Process Control Education in the Year 2000: A Roundtable
Discussion," Chem. Eng. Ed., 24(3), 72 (1990)
5. Rhinehart, R.R., S. Natarajan, and J.J. Anderson, "A Course in Pro-
cess Dynamics and Control: An Experience to Bridge the Gap Be-
tween Theory and Industrial Practice," Chem. Eng. Ed., 29(4) 218
6. Doyle, III, F.J., E.P. Gatzke, and R.S. Parker, "Practical Case Studies
for Undergraduate Process Dynamics and Control Using Process Con-
trol Modules." Comp Appl. in Eng. Ed.. 6, 181 (1998)
7. Doyle III. F.J. V. Venkatasubramanian, and T.A. Kendi, "Purdue Con-
trol Modules: A Flexible Set of Software Modules for an Undergradu-
ate Process Dynamics and Control Lboratory," Comp. Apps. Eng. Ed.,
4(3), 179 (1996)
8. Turton. R., R.C. Bailie. W.B. Whiting, and J.A. Shaeiwitz, Analysis,
Synthesis, and Design of Chemical Processes, Prentice-Hall, Upper
Saddle River, NJ (1998)
9. Shinskey, F.G., Distillation Control for Productivity and Energy Con-
servation. 2nd ed., McGraw Hill, New York, NY (1984)
10. Cutler. C.R.. and B.L. Ramaker, "Dynamic Matrix Control: A Com-
puter Control Algorithm," AIChE National Meeting, Houston, TX
11. Cutler, C.R., and B.L. Ramaker, "Dynamic Matrix Control: A Com-
puter Control Algorithm," in Proc. Joint Automatic Cont. Conf., paper
WP5-B (1980) 1

King Fahd University of Petroleum & Minerals

Department of Chemical Engineering


Applications and/or nominations are invited for the position of Saudi
Basic Industries Corporation (SABIC) Chair of Pollution Control in
Chemical Process Industries in the Department of Chemical Engineering
at King Fahd University of Petroleum & Minerals (KFUPM), Dhahran,
Saudi Arabia. The department has 23 full-time faculty members. 600 stu-
dents and offers undergraduate degree programs in Applied and Chemi-
cal Engineering Science, and graduate degree programs.
The candidate must have an earned doctorate in Chemical Engineer-
ing, hold full professorial rank, and have achieved an outstanding reputa-
tion in the field of Pollution Control. The candidate will be responsible
for developing expertise in this area within the region through his occu-
pancy of the Chair. A strong commitment to the development and mainte-
nance of high quality education and research are required. The candidate
should have a demonstrated track record in teaching, research, and pro-
fessional activities. Industrial experience would be an asset.
Specifically, the Chair holder will be expected to provide leadership in
both the areas of academia, through undergraduate and graduate course
development, and in research through development of a research labora-
tory as a center of excellence in Pollution Control in the region. He will
be responsible for teaching undergraduate and graduate courses, teaching



professional-development courses, consulting and conducting research
of direct interest to SABIC. and supervision of graduate students.
King Fahd University of Petroleum & Minerals is a leading techni-
cal university in the Middle East. The university has six colleges with
five engineering disciplines in the College of Engineering. English is
the medium of instruction. SABIC is the foremost non-oil company in
the Middle East and one of the world's fastest growing industrial con-
cerns, producing chemicals polymers, metals, and fertilizers.
The position is for a 3-year term, with an attractive salary (tax-free)
and benefit package. Funds for equipment, conference travel and re-
search assistance are available.
Applicants should send curriculum vitae to:

Dean of Faculty and Personnel Affairs
KFUPM Box 5005, DEPT SABIC-201
Dhahran 31261, Saudi Arabia

Fax: 966-3-860-2429

Sunmer 2001 17

Random Thoughts...




North Carolina State University Raleigh, NC 27695

Elementary Principles of Chemical Processes has been other reasons
used as the introductory chemical engineering course text at would first ki
nearly 150 American universities and many others elsewhere ourselves, an
since it first appeared in 1978. The first two editions of the
"red book" can still be found on the shelves of thousands of ere is a c
chemical engineers, mainly (as many alumni have told me) as Ican reca
because of the conversion factors on the inside front cover.
The red and blue edition came out in 1999 and may come to 1972
occupy the shelves of many future engineers, since the con- 1
version factors are still there. 1973
I am occasionally asked about how the book came to be, s
how long it took to write, who did which parts, and why any-
one in his or her right mind would write an undergraduate t
textbook with the faculty reward system being what it is. I t
thought the responses to these inquiries might interest some r
CEE readers, so here they are. c
I joined the N.C. State University faculty in July of 1969 s
and Ron Rousseau joined in August. In 1972, Alan Lesure of c
John Wiley & Sons invited Ron to write a stoichiometry book. a
Ron thought it sounded like a fine idea and said "sure," and u
when soon afterwards he asked me if I'd like to come in on 1
it, I thought it might be fun and said "sure." Although I was
slightly older and much wiser than Ron and had that critical
extra month of academic experience, neither of us had a clue
about writing textbooks. By the time we realized the magni-
tude of what we had agreed to do, we had invested far too
much time and effort to back out.
The first edition of EPCP made its appearance in the spring
of 1978 and was an almost instant success. Its success is not
something we could have anticipated when we were writing
it, however; in fact, we doubted that we'd ever finish it, among
Copyright ChE Division of ASEE 2001

because it seemed almost certain that one of us
11 the other. Fortunately we managed to restrain
d we're still good friends 23 years later.
hronology of the book's development, as closely
l it.

Work begins on A First Course in Chemical
Engineering (working title).
We write an outline and a sample chapter and
end it to Wiley for review. The reviews are
nixed. The chemical engineering editor at the
ime, Thurman Poston ("Post"), encourages us
o continue, but says that we'll need to get
nore chapters reviewed before he can issue a
contract. We continue writing. In the fall the
students in CHE 205 get drafts of the first few
hapters as their course text. They discover
bout 25 mistakes per page of text and an
uncountable number of glitches in the prob-
ems. The course ends before we ever get to

Chemical Engineering Education

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

energy balances. (I hope those students
eventually learned them somewhere.)

1974 We send about five chapter manuscripts off for
review and get back two good reviews and one
declaring that the book has few redeeming
features and no future. Post says that he needs
to see still more chapters before deciding. We
reply that perhaps it's time for us to open up
discussions with McGraw-Hill. We get a
contract by return mail.
One of the anonymous reviews is so insightful
and well written that we persuade Post to hire
the reviewer to critique the rest of the manu-
script for us. The reviewer turns out to be a
pair of professors from Iowa State-John
Stevens and Dick Seagrave. We owe a great
deal of the book's success to the comments we
got from those two, even though we still
bitterly resent the fact that no matter how
many chapter-end problems we wrote they
always wanted more.

1975-1977 We continue to write chapter drafts, class-test
them, find glitches, write new drafts, class-test
them, and so on, until in the spring of 1977 we
declare the manuscript perfect and send it off
to Wiley. Much of 1977 is spent proofreading
galleys and page proofs and writing a solution
manual. We are extremely careful about the
proofreading and are sure the book will be
Before we send the manuscript off, we set out
to select a permanent title. Neither of us has
any brilliant ideas, so what we do is write the
words Basic, Fundamental, Elementary,
Introduction, Principles, Elements, Methods,
Chemical, Engineering, Stoichiometry, and
Processes on pieces of paper and shuffle them
into every combination that makes sense.
When we get to Elementary Principles of
Chemical Processes we look at each other,
shrug, say "Good as anything else, I guess" or
words to that effect, and make that the title.
1978 The first printing appears. In the fall we use
the book in its published form for the first time
and offer the students 250 for each previously
undiscovered error they come up with in the
unlikely event that any survived our proofing.
We stop counting after 200.

1979-1980 By the third printing we are offering $1 for
each new mistake and are not getting many
takers. By its third year the text has taken over
most of the market.
1981-2001 The text continues to enjoy success. A
Spanish translation, the second edition (1986),
a Chinese translation, an International edition,
and the third edition with bundled CD-ROM
(1999) all appear. We live happily ever after.

As I mentioned before, I am often asked who did what.
Here's the answer. Problems containing weird characters with
unpronounceable names are mostly mine. All case studies in
the first two editions and two out of three in the third edition
are Ron's. The rest of the book bounced back and forth be-
tween us so much that it's impossible to ascribe any of it to
either author, except that any remaining mistakes are Ron's.
I am also occasionally asked how I account for the book's
success. Of course it's anybody's guess, but I believe a large
part of it is that we wrote the text for students and not to
impress potential adopters, and both our colleagues and their
students appreciated it. We emphasized physical and chemi-
cal phenomena and minimized abstract mathematical formal-
ism-there would be time enough for that later in the cur-
riculum. We provided one or more examples to illustrate ev-
ery problem-solving procedure, included self-tests so the stu-
dents could make sure they grasped the main ideas in every
section, and wrote close to 700 chapter-end problems and
three comprehensive case studies that applied the text mate-
rial to real process systems. We even took a page from Oc-
tave Levenspiel's playbook and included some humor.
To be fair, I must note that not all of our colleagues wel-
comed our approach, and the text has on occasion been la-
beled insufficiently rigorous. My favorite criticism was one I
overheard in an AIChE meeting hospitality suite some years
ago, when one professor snorted to another that he found
Felder and Rousseau "sophomoric"-an interesting way to
insult a sophomore textbook. I was tempted to thank him but
instead just filed the comment away for possible future use
and continued threading my way to the bar.
Finally, is it worth it to spend the years it takes to write a
textbook? Unquestionably, it is. Nothing compares with the
satisfaction Ron and I feel when we see students carrying our
book, or when they write us from campuses all over the world
telling us how our splendidly written and inspirational text
has changed their lives (usually followed by a request for a
copy of the solution manual). The nickel-per-hour return we
get in royalties is just icing on the cake. 0

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

Summer 2001

W class and home problems




Lehigh University Bethlehem, PA 18015

This paper illustrates two useful pedagogical techniques
for motivating and teaching students that can be eas-
ily applied to teaching process dynamics. The two
basic ideas are: 1) use situations that are not chemical engi-
neering and 2) use different versions of the same problem
sequentially throughout the duration of the course.
The first helps to motivate students because they can see
that the basic principles of developing dynamic mathemati-
cal models have wide application in many aspects of life.
The second provides the "creative redundancy" that is needed
to really understand a subject.
One example of this approach is presented here. There are
four similar problems that have slightly different mathemati-
cal models and/or boundary conditions:
E The 1805 Battle of Trafalgar (Version 1)
E The Battle of Trafalgar (Version 2)
The Battle of the North Atlantic (1940)
U The 2200 battle between the Federation fleet of starships,
led by Captain Kirk, and the evil Klingon fleet
The originator in chemical engineering of the idea of mo-
tivating students by using non-chemical engineer-ing ex-
amples was Octave Levenspiel. In his pioneering textbook,
Chemical Reaction Engineering,' I he presented a number of

problems that were outside the chemical engineering field.
As a graduate student studying this book, I found these prob-
lems very refreshing. The typical textbook back in those days
(and still true for many books today) was dry as dust. The
language was very stiff and formal. The use of the first per-
son was unheard of, as was any attempt to inject humor. All
the material was straight-line chemical engineering.
Levenspiel changed all that and produced a very "user-
friendly" book. His "reactor design" problems included the
Battle of Trafalgar, Snake-Eyes Magoo betting habits, inves-
tigation of the missing operator by Sherlock Holmes and Dr.
Watson, etc. These problems were a great help in letting stu-
dents understand that the basic principles could be applied to
a wide spectrum of life situations. In my own writing over the

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

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, University of
Michigan, Ann Arbor, MI 48109-2136.

William L. Luyben earned degrees in chemi-
cal engineering from Penn State (BS, 1955) and
Delaware (PhD, 1963). His industrial experience
includes four years with Exxon, four years with
DuPont, and three decades of consulting with
chemical and petroleum companies. He has
taught at Lehigh University since 1967 and has
participated in the development of several in-
novative undergraduate courses, from the in-
troductory course in mass and energy balances
through the capstone senior design course and
an interdisciplinary controls laboratory

last forty years, I have tried to follow Levenspiel's example by
using problems drawn from such diverse areas as farming, whis-
key making, mechanical and aerospace systems, etc.
In the following sections, I will show how the "Battle of
Trafalgar" problem can be extended to teach students the prin-
ciples of dynamic mathematical modeling and the use of
Laplace transforms.


The second idea has grown out of over three decades of un-
dergraduate teaching. The approach is to assign a homework
problem and go over its solution in class. Then in the first ex-
amination give a similar problem that is a slight extension of the
first, and in the second examination given another similar problem
that adds different features to the mathematical model that must be
derived and solved. By the time the students get to the end of the
course, they have figured out that there will be a similar problem
on the final examination, and they all know how to solve it.
The principle behind this approach is "creative redundancy."
How many times have you heard the remark "I really didn't un-
derstand thermodynamics until I took the third thermo course."
Repetition is a fundamental approach to learning. The idea is to
make sure it is not boring.
This paper presents one example of this sequential, non-disci-
plinary problem approach. The problems are published in Luyben
and Luyben.12i


This problem is assigned early in the course after the funda-


5 er ch V(t)

Number 20
of Ships
NN. English NWt

Time (ki)

Figure 1. Battle of Tafalgar No. 1.

mental of dynamic modeling have been reviewed and il-
lustrated with several chemical engineering processes.

* Problem Statement

Solve the following problem, which is part of a problem
given in Levenspiel's Chemical Reaction Engineering,
using Laplace transform techniques. Find an analytical
expression for the number of Nelson's ships, N, and the
number of Villeneuve's ships, V0 as functions of time.
The great naval battle, to be known to history as the
Battle of Trafalgar (1805), was soon to be joined.
Admiral Villeneuve proudly surveyed his powerful fleet of
33 ships stately sailing in single file in the light breeze.
The British fleet under Lord Nelson was now in sight, 27
ships strong. Estimating that it would still be two hours
before the battle, Villeneuve popped open another bottle
of burgundy and point-by-point reviewed his carefully
thought-out battle strategy As was the custom of naval
battles at that time, the two fleets would sail in single file
parallel to each other and in the same direction, firing
their cannons madly. Now, by long experience in battles
of this kind, it was a well-known fact that the rate of
destruction of a fleet was proportional to thefirepower of
the opposing fleet. Considering his ships to be on a par
one-for-one, with the British, Villeneuve was confident of
victory. Looking at his sundial, Villeneuve sighed and
cursed the light wind; he'd never get it over with in time
for his favorite television western. "Oh, well," he sighed,
"C'est la vie." He could see the headlines next morn-
ing--"British Fleet annihilated, Villeneuve's losses
are...." Villeneuve stopped short. How many ships would
he lose? He called over his chief bottle-cork popper,
Monsieur Dubois, and asked this question. What answer
did he get?

Problem Solution

The mathematical model of this problem consists of two
linear ordinary differential equations:

dN k

dV -kN (2)

with the initial conditions N10 =27 and V0 = 33, and k
is a constant equal to the rate of destruction per ship. This
rate constant is analogous to the rate constant in a chemi-
cal reaction. Note that the actual variables in this problem
are discrete (integers), but we are approximating the sys-
tem with continuous variables to keep the mathematics
simple. For reasonably large values (>10), this approxi-
mation is probably fairly accurate.
Laplace transforming gives the two algebraic equations

Summer 2001

sN() 27 = -kV() (3)

sV() 33 = -kN(s) (4)

where s is the Laplace transform variable. Combining these
equations gives expressions for N(s) and V(s), which can be
inverted back into the time domain to obtain N,) and V(.

27s 33k
(s) S2 k2
33s 27k
V(s) 2 k2
N(t) = 30ekt 3ekt

V -3ekt +30e-kt

Note that there are positive eigenvalues. This does not mean
that variables become infinite because the solution is limited
to finite values of time. The battle ends at tF when the number
of Nelson's ships goes to zero, N(tF) = 0,

N(F) = = 30e-ktF 3ektF
Solving for t, gives
en 10
tF 2k

The number of Villeneuve's ships left is

V(tF) = 3ek(cn 0)/2k -30e-k(fn l)/2k = 18.95

Therefore, Villeneuve has lost 33-19 = 14 ships. Figure 1
shows the dynamic changes in the number of ships in each
This problem is assigned and its solution is discussed care-
fully in class before the first examination.


Now we modify the problem by breaking it into two differ-
ent battles. The model is the same, but there are two time
periods with different initial conditions. Note that some of
the characters are personalized to increase the interest level
of the class (Bethany Steadman was the student who asked to
review the original homework problem).

* Problem Statement

While Adminral Villeneuve was doing his calculations about
the outcome of the Battle of Trafalgar Admiral Nelson was
also doing some thinking. His fleet was outnumbered 33 to
27, so it didn't take a rocket scientist to predict the outcome
of the battle if the normal battle plan was followed (the
opposing fleets sailing parallel to each other). So Admiral

Nelson turned for help to his trusty young Lt. Steadman,
who fortunately was an innovative Lehigh graduate in
chemical engineering (class of 1796). Steadman opened up
her textbook on Laplace transforms and did some back-of-
the-envelope calculations to evaluate alternative battle
After several minutes of brainstorming and calculations
(she had her PC on board, so she could use MATLAB to aid
in the numerical calculations), Lt. Steadman devised the
following plan: The British fleet would split the French
fleet, taking on 17 ships first and then attacking the other 16
French ships with the remaining British ships. Admiral
Nelson approved the plan, and the battle began.
Solve quantitatively for the dynamic changes in the number
of British and French ships as functions of time during the
battle. Assume the rate of destruction of a fleet is propor-
tional to thefirepower of the opposing fleet and that the
ships are on a par with each other in firepower

* Problem Solution

(9) The ordinary differential equations are exactly the same as
in the homework problem, but the initial conditions are dif-
ferent. Generalizing the solution for arbitrary initial numbers
(10) of ships in each fleet, let N,_o) = No and V),_)=Vo. The solu-
tion is

N(t) = N kt

- 2 ekt

( V + N (V N )
V^t)= ej-kt + ekt

During the first battle when Nelson takes on half the French
fleet, the initial conditions are Nt= = 27 and V o = 17. The
end of this initial battle occurs at tp, when V,(t = 0, or

Figure 2. Battle of Tafalgar No. 2.

Chemical Engineering Education

of Ships


* Problem Statement

Vl(t)= 0= 0 e-ktF + 0ektF

t n (4.4) 0.7408 / k
2 k

and the remaining number of British ships is

NF) 44 )-0.7408 +( 10e0.7408 21 (16)

Now the second phase of the battle begins with the initial
conditions N,= = 21 and V=O = 16. The dynamic changes in
the number of French ships is

V 37 e-kt + -5 ekt (17)
(o) = 2(f 2

The time it takes Nelson to completely demolish the French
fleet is

t n (37/5) 1.0007 (18)
tF2 = 2k k 8)
The remaining number of British ships is

N(t) ) -1.)el- 10007 = 13.5 (19)

Figure 2 shows the dynamic changes in the number of ships
in each fleet during the two phases of the battle.
So Nelson and the British fleet win the day (with a little
help from a Lehigh chemical engineer)!


On the next test the problem is modified to include genera-
tion terms in the differential equations in addition to the deple-
tion terms.

Figure 3. Battle of the North Atlantic.

Summer 2001

The 1940 Battle of the North Atlantic is about to begin. The
German submarine fleet consists of 200 U-boats at the
beginning of the battle. The British destroyer fleet, under
the command of Admiral Steadman (a direct descendant of
the intelligence officer responsible for the British victory at
the Battle of Trafalgar), consists of 150 ships at the
beginning of the battle. The rate of destruction of subma-
rines by destroyers is equal to the rate of destruction of
destroyers by submarines: 0.25 ships/week/ship.
Germany is launching two new submarines per week and
adding them to its fleet. President Roosevelt is trying to
decide how many new destroyers per week must be sent to
the British fleet under the Lend-Lease Program in order to
win the battle. Admiral Steadman claims she needs 15 ships
added to her fleet per week to defeat the U-boat fleet. The
Secretary of the Navy, William Gustus, claims she only
needs 5 ships per week. Who is correct?

* Problem Solution

The dynamic mathematical model describing the number
of destroyers, D and the number of U-boats, U ,, is

= -kU+ PD (20)
dt -kD + Pu (21)

where k is the rate of destruction (0.25 ships destroyed per
week for each ship in the opposing fleet), PD and Pu are the
weekly rate of addition of destroyers and U-boats to the fleets,
and time, t, is in weeks. The initial conditions are D=01 = Do =
150 and Uo) = Uo = 200. Laplace transforming and combin-
ing gives
Uo2 + (Pu kD)s kPD (22)
"() s(s + k)(s k)

Inverting to the time domain gives
U(t) =

PD +[k(Uo+Do)-(P +PD)le-kt +k(U -Do)+(P -PD) kt
k 2k 2k

If destroyers are added at a rate PD = 15, the number of U-
boats goes to zero at t, = 16.5 weeks, and the number of re-
maining destroyers is D = 71.8. If, however, destroyers are
added at a slightly reduced rate PD = 14, the number of de-
stroyers goes to zero at t, = 11.2 weeks, and the number of
remaining U-boats is U = 81.2. Figure 3 shows the dynamic
changes in the number of vessels in each fleet for the two
cases. Thus, Admiral Steadman's claim that 15 ships are
needed per week is correct, and the Secretary of the Navy's


claim of 5 ships per week is a gross underestimate.


The final sequential problem moves into future star wars.
The battle is between Captain Kirk's fleet of starships and
the evil Klingon fleet in the year 2200. Now there are two
types of starships: some have better defensive shields and
some have more firepower than the Klingon ships. The math-
ematical model now has three ordinary differential equations
with different coefficients in the destruction-rate term.

* Problem Statement

Captain James Kirk is in command of a fleet of 16 starships
of the Enterprise class. A Klingon fleet of 20 ships has been
spotted approaching. The legendary Lt. Spock has recently
retired, so Captain Kirk turns to his new intelligence officer,
Lt. Steadman (Lehigh Class of 2196 in chemical engineer-
ing) for a prediction of the outcome of the upcoming battle.
Steadman has been working with the new engineering
officers in the fleet, Lt. Moquin and Lt. Walsh, who have
replaced the retired Lt. Scott. These innovative officers have
been able to increase the firepower of half of the vessels in
Kirk's fleet by a factor of two over the firepower of the
Klingon vessels, which all have the same firepower The
firepower of the rest of Kirk's fleet is on a par with that of
the Klingons. But these officers have also been able to
improve the defensive shields on the second half of the fleet.
The more effective shields reduce by 50% the destruction
rate of these vessels by the Klingon firepower.
Thus, there are two classes of starships: eight vessels are
Class E,, with increased firepower and eight vessels are
Class E2, with improved defensive shields. Assume that half
of the Klingon fleet is firing at each class at any point in

Calculate who wins the battle and how many vessels of each
type survive.

* Problem Solution

The dynamic model of the system is

dEt -k(K) (24)

dE2 _2k 5K (25)

S-kE2 -2kE, (26)

with the initial conditions E, ,o, = 8, E2(t=O)= 8, and K,-, = 20.
Laplace transforming and combining give

K 20s 24k 20.73 0.7335 (27)
(s) -2 _5k2 s + 1.118k s-1.118k (

14 Kling s K(t)
of Ships 1
s____ Class 2 Starshios E

Figure 4. Star Trek Battle.

8 -kK() /2
) (s)
1(s) S


8 kK(s) / 4

Inverting to the time domain gives

K(t) = 20.73e-l1.18kt -0.7335el11 8kt

El(t)= 9.271e-1.1I8kt +0.328e. 118kt -1.599
E2(t)= 4.636e-1.118kt -0.164el118kt +3.528

The end of the battle occurs when K(tF)=0. Solving the
first Eq. 30 yields tF = 1.494/k. The number of surviving
starships is El(t=494/k) = 1.889 and E2(t=1494/k) = 3.528. So it is
better to be on a ship with better shields than on a ship with
more firepower in this matchup. Figure 4 shows the dynamic
changes in the number of vessels in each fleet.

This sequential non-chemical engineering problem illus-
trates the basic ideas of the teaching methods proposed in
this paper. Students respond when you show them how they
can apply the fundamental principles they are learning about
chemical engineering processes to many other real-life situa-
tions. "Variations-on-a-theme" problems help students learn
the basic principles of dynamic modeling in a variety of situ-
ations. They learn how to think and how to derive models
instead of trying to find a formula in a book.

1. Levenspiel, Octave, Chemical Reaction Engineering, John Wiley &
Sons, New York, NY (1962)
2. Luyben and Luyben, Essentials of Process Control, McGraw Hill Book
Company, New York, NY, Problems 7.23, 7.24, 7.28, and 7.29 (1997)

Chemical Engineering Education

S book review

The Dynamics of Fluidized Particles
By Roy Jackson
Cambridge University Press, Cambridge, CB2, 2RU, UK
(2000); 339 pages, $69.95 (2000)

Reviewed by L.T. Fan
Kansas State University

In the author's own words in the preface, "This book ad-
dresses the motion of systems of solid particles immersed in
a fluid that may be a liquid or a gas..." This motion is indeed
ubiquitous in natural as well as man-made environments.
Moreover, this motion attributable to the interaction between
the two environments has been accelerating in both frequency
and magnitude. Often, it is detrimental to the ecosystem that
is shared by humans, animals, and plants.
In fact, it has frequently brought about catastrophic events.
Some of the well-publicized instances are large-scale land-
slides due to the loss of vegetation on hillsides around Los
Angeles from excessive housing development and the mam-
moth floods around the Yangtze River basin caused by the
serious erosion accompanied by the flow of silt into the river
from the over-cultivation of farmland and extensive defores-
tation. On the other hand, the motion of systems of particles,
if induced under tightly controlled and well-managed condi-
tions, can create immeasurable benefits to our daily lives.
For example, it is involved in a wide variety of processes
yielding the products essential to our very existence, e.g. food-
stuffs, fuels, and cement.
Obviously, common principles underlie the motion of sys-
tems of particles found in sundry situations. Naturally, it is
highly desirable that such principles be learned, explored,
and understood in a unified manner. As far as this reviewer
knows, a concise volume containing a systematic and com-
prehensive account of these principles has been totally lack-
ing. Anyone who would even simply browse through the book
will readily realize it has definitely filled this void.
At this juncture, the book may be the only suitable text-
book on the market for a one-semester graduate course on
fluidization. Certainly, it would be highly recommended as a
supplementary textbook or reference in different courses such
as transport phenomena, multiphase flows, and chemical-re-
action engineering. In the author's own words again, "...The
book is intended as an introduction to this field for graduate
students and others entering it for the first time, but by draw-
ing together widely scattered material..." For the convenience
of those who would be interested in a quick glimpse of the
book's contents, the chapters are listed below:
Summer 2001

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

Department of Materials and Process Engineering
School of Science and Technology
The University of Waikato

Applications are invited from suitably qualified people
for a continuing (tenure track) academic position. Duties
consist of teaching at undergraduate and graduate levels
and conducting innovative research. Applicants should
have a first degree in engineering and either a PhD or
extensive industrial R&D experience in chemical or me-
chanical engineering or related areas. Preference will be
given to candidates with teaching and research interests
in process design. The ability to forge close links with
industry would be an advantage.
The Department currently specializes in advanced ma-
terials (metallurgy, composites) and processing, biochemi-
cal engineering, process engineering, innovation manage-
ment and environmental technology.
Enquiries to Associate Professor Janis Swan, Chair-
person of the Department of Materials and Process Engi-
neering, telephone 07 838 4049; fax 07 838 4835; email
Applications close on Thursday, 30 August 2001.
Further information and an application form are avail-
able from: or the Hu-
man Resource Management Division, The University of
Waikato, Private Bag 3105, Hamilton, phone (07) 838
4003, fax (07) 856 0135, email .
The University is committed to providing equal op-
portunities for all.
E herea ana Te Whare Wananga ki te kaupapa kia
whakaratohia te mea angitii rite ki nga tangata katoa.

1. The Mathematical Modeling of Fluidized Suspensions
2. Equations of Motion
3. Fluidization and Defluidization
4. Stability of the Uniformly Fluidized State
5. Bubbles and Other Structures in Fluidized Beds
6. Riser Flow
7. Standpipe Flow
This reviewer is confident that the book will be a
"Schlichting" of particle dynamics if it is continually updated
and expanded for years to come. O





Better Equipping Students for a

Maturing Industry

University of California Davis, CA 95616

Over the past fifty years, the field of biochemical en-
gineering has evolved tremendously, from early work
on fermentation of antibiotics and other small or-
ganic molecules to the more recent production of recombi-
nant proteins as human and animal therapeutics.
The curriculum for training biochemical engineers has also
evolved. Early courses in fermentation used textbooks such
as Biochemical Engineering"'' and covered enzymatic reac-
tions, cell growth kinetics (e.g., Monod kinetics), microbial
fermentor design, oxygen transfer, and sterilization with a
limited analysis of purification of biomolecules. Other simi-
lar courses, using books such as Biochemical Engineering
Fundamentals[2] and more recently Bioprocess Engineering:
Basic Concepts["1 and Biochemical Engineering[41 followed.
Even more recently, courses geared toward purification of
molecules produced by fermentation have been added to bio-
chemical engineering curricula, using such texts as
Bioseparations: Downstream Processing for Biotechnology. [5]
These courses have typically focused on the theoretical back-
ground of the unit operations, however. Teaching the de-
sign of specific equipment has normally been left for the
mainstream chemical engineering course on equipment
design, using texts such as Plant Design and Economics
for Chemical Engineers. 61
The first commercial recombinant protein production (i.e.,
recombinant human insulin by Eli Lilly) began in 1982. Since
then, the number of therapeutic recombinant protein prod-
ucts on the market has grown steadily to over seventy in 2000,

with at least 360 more currently in development.171 This has
resulted in a relatively mature industry with very specific train-
ing needs. In particular, the industry uses specialized unit
operations such as fermentation and chromatography, as well
as unique critical utilities. In addition, sanitary design is es-
sential to the successful production of biologics. Regulatory
issues such as current Good Manufacturing Practices (cGMP)
and validation are also critical, not only to manufacturing
biopharmaceuticals, but also to all facets of process develop-
ment and process engineering in this industry.
For these reasons, in 1997 we initiated a course on Biotech
Manufacturing Facility Design and Regulatory Compliance
at the University of California, Davis, that covers this mate-
rial and which is now required of all senior-level biochemi-
cal engineering majors.
The course covers material necessary in the design and
operation of a facility producing biological products, with
emphasis on recombinant proteins. It introduces the students

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

David E. Block is an Assistant Professor in
the Department of Viticulture and Enology
and the Department of Chemical Engineer-
ing and Materials Science at the University
of California, Davis. He holds a BSE and PhD
in Chemical Engineering from the University
of Pennsylvania and the University of Min-
nesota, respectively. Prior to joining the uni-
versity, he worked in the Biopharama-
ceuticals Department at Hoffmann-La Roche
in Nutley, New Jersey.

to concepts such as aseptic processing and Good Manufac-
turing Practices (GMP). After introducing these general de-
sign and regulatory issues, each area of a biotech facility is
discussed following process flow. In the discussion, general
equipment-design issues are included, such as piping and
pumpsizing, and pressure-vessel design. The second focus
of the course is compliance with cur-
rent GMP, the laws governing pro-
duction of human and veterinary
drugs, as enforced by the Food and
Drug Administration (FDA). Concepts Cour
such as validation masterplanning,
validation protocols, protocol execu- 1 Overview of
tion, and change control as they apply 2 Overview of
to both equipment and process quali- 3 Aseptic Proc
fiction are discussed. Ideas for suc- 4 Piping and II
cessful project management and facil- 5 Inoculum Pr
ity startup are introduced throughout 6 Media Prepa
the course. 7 Pumps/pipe
This course is typically taken con- 8 Fermentatiol
currently with a traditional course in 9 Basic desig
fermentation and a course in 10 Oxygen train
biopurification. In the following quar- 11 Sterilizatior
ter, students take a bioprocess engi- 12 Primary Rec
neering laboratory course and biotech- 13 Centrifuges
related senior design project to com- 14 Cell Disrup
plete their area of emphasis. 15 Buffer Prepa
16 Purification-
GOALS 17 Ultrafiltrati
There are five goals for this course. 18 Viral inacti
By the completion of the ten weeks of 19 Water Systel
instruction, students are expected to 20 Heat excha
21 CIP Systems
Be familiar with bioprocess 22 Biowaste Tn
equipment, processflow, and 23 Basic Contre
interactions between facility Control and
areas 24 GMP Reviev
Understand aseptic processing 25 Validation 1
and the idea of sanitary design 26 Protocols
27 Change Con
Be able to specify critical 28 Batch Recor
28 Batch Recor
equipment parameters and
understand design issues

Understand the importance of GMP in manufactur-
Understand and be able to write validation docu-

All course lectures, demonstrations, and projects are geared
toward these goals. In general, we do not expect a student to
complete the course and immediately to be able to perform a
detailed design of a piece of equipment. Mentoring by expe-
rienced engineers will always be a necessary step in training.
The information that the students receive in this course, how-
Summer 2001

ever, will allow them to identify equipment as they walk
through a biotech manufacturing facility, to have a good idea
of the critical parameters for each piece of equipment so they
can more readily participate in operation and design of the
equipment, and to understand the intense regulatory environ-
ment in which they will work.

The general outline for the course in
1 shown in Table 1. After some introduc-
tine tory material, the course outline follows
the process flow through a production
ch Facilities facility, including the fermentation, pu-
,liance Issues rification, and utility areas. The last sec-
/Materials tion of the course focuses on cGMP, in-
nentation Drawings cluding validation and other regulatory
tion documentation. The book used as a text
n for the course is Bioprocess Engineer-
es/filters ing: Systems, Equipment, and Facili-
ties, 81 which is a compilation of chap-
pressure vessels ters on various parts of a biotech facil-
ity written by industrial practitioners.

Introductory Material
Since this is a new area of study for
most of the students, a fairly large
omatography amount of time (approximately six fifty-
afiltration minute lectures) is spent on introductory
/removal material to give them a strong common
background. First there is an introduc-
tion to the bioprocess industry and its
various segments, which produce every-
nt: HVAC thing from bulk chemicals to recombi-
nant proteins. Reasons for the orders-
acquisition of-magnitude difference in product cost
from one end of the spectrum to the
Plan other are discussed as a way of get-
ting students to focus on what is dif-

e Ou

n and
, filter

Il Syst



ferent about producing pharmaceuti-
cals. Next is an introduction to regu-
latory compliance, including a history
of the FDA in this country, a brief
overview of cGMP and validation, and an idea of the ethi-
cal and legal ramifications of noncompliance.
After these overviews, the course focuses on the ideas of
sanitary design and aseptic processing that are critical to
biopharmaceutical production. Introductory information is
given to the students on the two facets of sanitary design:
materials of construction (e.g., grades of stainless steel and
welding considerations), and design (e.g., pipe sloping,
low-point drains, avoidance of dead legs, cleanability of
various types of valves and flange connections, and in-
strument mounting in tanks).

Finally, piping and instrumentation diagrams (P&IDs) are
introduced, using actual examples. The types of information
contained in these drawings and how to use this information
for equipment design and operation are stressed in both class
exercises and homework problems (see the class example in
Figure 1). Aside from their importance in facility design,
P&IDs are a convenient means of introducing and dis-
cussing each of the pieces of equipment or systems in the
remainder of the course.

Process Areas and Design Details
After completing the introductory material, the next major
series of lectures covers design of equipment in each of the
process areas in the order of process flow, as shown in the
block diagram of a generic facility in Figure 2.
This section of the course begins with the inoculum and
media preparation areas of the facility. For inoculum prepa-
ration, specialized equipment used for microbial and mam-
malian cell systems is discussed, such as laminar flow hoods,
incubating shakers, and CO, incubators. General guidelines
are given for inoculum sizes and stages used in industry. For
media preparation, a description of various types of media

prep systems (e.g., ones using sterile filtration or continuous
heat sterilization) is given, along with a detailed discussion
of all of the components. Since this is the first time in the
course that we come across valves and pumps, a survey is
completed of the various types that are available and which
ones are typically used for biotech facilities. Next, the pro-
cess for sizing pipes for turbulent flow with reasonable
pressure drops, as well as sizing pumps using a total me-
chanical energy balance, desired flow rate, and pump
curves is explained.
For the fermentation area, we discuss a simplified P&ID
for a generic microbial fermentor. Then, as with the media
prep area, each component part is considered. This includes
a discussion of the sizing and the various aspect ratios of
fermentors, of the design of pressure vessels and ASME code,
and of the control systems typical for fermentors (e.g., pH,
dissolved oxygen, foam, aeration, agitation, temperature).
Special lectures are given on oxygen transfer in fermentors
(and how it affects design) and on designing for sterilization.
Differences between microbial fermentors and mammalian
cell bioreactors are then highlighted.
Primary recovery (i.e., separation of cells from the cell-

Figure 1. Example of a piping and instrumentation diagram (P&ID) used in the course. This P&ID for
a mammalian cell culture media prep area was used for homework and exam purposes.
Each of the student teams generates a similar diagram for different pieces of
equipment as part of the quarter-long group project.
Chemical Engineering Education

free broth) is covered next. This is mainly a description of
tangential flow filtration and centrifugation, the two main
means of cell separation currently used in the industry. Again,
the important design parameters are stressed, as well as the
scale-up issues for each unit operation. Factors affecting the
choice between these unit operations are also discussed.
For purification, the discussion centers primarily on chro-
matography since it is currently the main means of purifying
therapeutic recombinant proteins. After discussing the four
most common types of chromatography (i.e., ion exchange,
hydrophobic interaction, affinity, and gel filtration) and their
specific uses, the focus turns again to the equipment. This
includes a description of the components of chromatography
skids, including specialized valves and sensors. Scale-up and
chromatography column sizing is also covered for the vari-
ous types of columns. Ultrafiltration and diafiltration are then
discussed. Since the equipment for these processes is essen-
tially the same as for the primary recovery tangential flow
filtration, emphasis is placed on the specific use of the equip-
ment for product concentration or buffer exchange.
Finally, we cover the utilities critical for biotech manufac-
turing facility operation. While these utilities commonly make
up one-third of a facility, they are largely ignored by tradi-
tional biochemical engineering courses. USP purified-water
production systems are described, as well as Water-for-In-
jection (WFI) and Clean Steam stills that are fed by the puri-
fied-water systems. In discussing WFI systems, heat ex-
changer design and sizing is covered.

Next, the concept of Clean-in-Place systems, or CIP, is in-
troduced. Cleaning cycles and necessary flow rates and tem-
peratures are discussed. Heating, ventilation, and air condi-
tioning (HVAC) is then covered briefly-only in regard to
what it controls in the facility, namely temperature, humid-
ity, room cleanliness (and classification), and pressure. Dif-
ferential room pressure and HEPA filtration are used in these
facilities as a means of maintaining a clean-room environ-
ment. Control-system alternatives such as programmable
logic controllers (PLCs) and distributed control systems
(DCS) are introduced.

Regulatory Compliance
Aside from the introductory material mentioned above, the
last section of the course covers regulatory compliance ac-
cording to cGMP. The bulk of this time is spent in discussion
of Validation. Validation is defined as the process of estab-
lishing documented evidence that provides a high degree of
assurance that a specific process will consistently produce a
product meeting its predetermined specifications and qual-
ity attributes.[9' For Validation, facility masterplans are dis-
cussed briefly prior to discussing the details of a valida-
tion test plan or protocol.
All of the typical activities associated with Installation
Qualification (IQ), Operational Qualification (OQ), and Per-
formance Qualification (PQ) are covered in a fair amount of
detail. The emphasis here is that testing should be performed
to assure that the equipment does what it is designed for: 1)

Buffer Purification

Clean USP t- WFp Biowaste
Steam Water

Facility Block Process Control and Monitoring Automation
Diagram __
Figure 2. Block diagram for a generic biotech manufacturing facility. The course
developed generally follows the process flow as an outline.

Summer 2001

reproducibly and 2) for good technical
reasons. Testing for the sake of testing,
and creating paperwork, is discouraged. [Studi

the in*
In order to reinforce the lecture mate-
rial and achieve the course goals, several throL
class activities are carried out during the
course. For instance, when discussing COU
sanitary fittings and design, various types
of valves (e.g., ball, butterfly, diaphragm, practi
and gate), welding samples (e.g., hand
and machine welded), flanges or other maste
unions (e.g., threaded joints and sanitary
clamps), and objects of various materi- subject
als (e.g., different grades of stainless
steel, brass, glass, and various types of allows
flexible tubing) are brought into the class.
Students work in small teams to deter-
mine what the various objects are and rapid
whether or not they would be considered r
cleanable in a biopharmaceutical facility, to indt
Aside from being interesting for the stu-
dents, this activity gives us the opportu- th e
nity to stress the limitations to design
caused by the idea of cleanability and to be
point out the difference between "clean"
and cleanablee." pr
Another class activity takes place dur-
ing the introduction of P&IDs when the soC
students are again arbitrarily split up into
small groups and are given a list of ques-
tions to answer about a P&ID that they
have not yet reviewed extensively. They are given a limited
time to answer all the questions, forcing them to work to-
gether as an actual team (which will be the norm when they
enter an industrial environment).
Another major part of the course (30% of the final grade)
is a group project. The premise of the project is that the stu-
dents comprise a central engineering group in a company
where a new production facility is entering a detailed design
phase. Each group of three students is given one piece of
equipment or system in the facility for which they must cre-
ate a P&ID, write a detailed design specification that could
go out for bid, and then write a validation protocol or test
plan for the same piece of equipment. This project gives the
students an opportunity to use the lecture material from the
class as well as to think about a particular unit operation in
great detail. It also usually involves interaction with actual
equipment vendors to get ideas or to answer questions about















Chemical Engineering Education

the functionality or design of their equip-
ment. This aspect of the course has
ts] see proven to be quite successful. Former
students have specifically pointed to
Sa tion these unique activities as the reason they
were offered jobs in the industry.
nated Since no actual textbook exists for this
topic, homework problems have to be
h thiS created each year from personal knowl-
edge or limited literature. Problems gen-
c as erally involve sizing the relevant equip-
ment or designing portions of a facility
1, and based on design principles covered in
lecture. Exams closely follow the form
Of the of homework.

em] to AND TOURS
To familiarize the students further with
the type of equipment discussed in class
nsition and to make the course more "hands on,"
an extra hour each week is set aside for
ry and demonstrations. They consist of either a
vendor or the instructor demonstrating
._to a piece of equipment, and they follow
the flow of the course material, roughly
e corresponding to the material that is be-
ing covered that week in lecture. For
,tive example, during the introductory phase
of the course, a representative of Arc
Machines comes in to demonstrate or-
bital machine welding, the type of weld-
ing predominantly used in piping sys-
tems in biotech facilities.
At another time, a representative of ITT PureFlo Valves
shows the students a large collection of sanitary diaphragm
valves made especially for the biotech industry; full-scale
sanitary centrifugal and positive displacement pumps are
demonstrated (including assembling/disassembling the pump
heads to examine the mechanism of pumping, a discussion
of mechanical seals, priming capability, and cavitation) at
the pilot winery at UC Davis.
Small autoclavable fermentors have also been used for dem-
onstrations of associated control systems and mixing (with
various impeller types and placement, with and without
baffles). Tours of local fermentation process development
facilities have been used to demonstrate larger sterilize-in-
place fermentors. Millipore has brought in both cartridge and
small tangential flow filtration units for demonstration.
Amersham Pharmacia Biotech covers different types of chro-
matography resins and brings in pilot-scale process columns.

Finally, programmable logic controllers are demonstrated in-
house, both for their functionality and ladder-logic program-
ming. These demonstrations provide a valuable experience
for the students that pictures and diagrams alone cannot equal.
In addition to the weekly demonstration periods, one or
two plant tours have been scheduled each year to local biotech
manufacturing sites. This experience has been extremely use-
ful to the students in integrating the knowledge they have
gained from the remainder of the course material.

To assure that the students taking this course remain mar-
ketable in the wider chemical engineering job market, many
of the equipment design issues of a mainstream equipment
design course are still taught (e.g., pipes, pumps, pressure
vessels, heat exchangers)-but the equipment is discussed in
the context of a biotech facility. Unit operations and ideas
unique to the biotech/pharmaceutical industry, such as fer-
mentation, chromatography, clean utilities, sanitary design,
and cGMP, are also stressed throughout the course.
The addition of instruction in reading P&IDs, demonstra-
tions of actual equipment used in processing, and tours of
production facilities have been particularly useful and would
likely be beneficial in a mainstream chemical engineering
equipment design course as well. Students completing this
course have received simultaneous job offers from biotech
and chemical companies, so it appears that the flexibility of
this approach is evident to potential employers.
Teaching a course in this subject does require some knowl-
edge of the industry. This knowledge could be gained from
previous industrial experience, extensive reading, collabora-
tive teaching with personnel from local companies, or an in-
dustrial sabbatical.

This new course on biotech manufacturing facility design
and regulatory compliance is a unique experience for chemi-
cal engineering students who are planning to join the biotech
or biopharmaceutical industry after graduation. The recep-
tion from students and industry alike has been overwhelm-
ingly favorable. Both groups see the information dissemi-
nated through this course as practical, and mastery of the sub-
ject matter allows the students to make a more rapid transi-
tion to industry and thereby to become productive sooner.

The author would like to acknowledge the support and en-
couragement of Professors Karen McDonald, Alan Jackman,
and Subhash Risbud in the Department of Chemical Engi-
neering and Materials Science at the University of Califor-
nia, Davis, during the development of this new course and its

Summer 2001

transition to a permanent, required course. Acknowledgment
is also extended to the various companies that sent represen-
tatives to talk to students over the last four years (Digital
Welding Systems, ITT PureFlo Valves, Millipore, Amersham
Pharmacia Biotech, Ingold Mettler Toledo, and New
Brunswick Scientific) and the companies who allowed us
to visit their sites, including Genentech, Chiron, and
AgraQuest. The author would also like to thank Professor
Michael Delwiche at UC Davis for the demonstration of
PLCs in his laboratory.

1. Aiba. S., A.E. Humphrey, and N.F. Millis, Biochemical Engineering,
2nd ed., Academic Press, New York, NY (1973)
2. Bailey, J.E., and D.F. Ollis, Biochemical Engineering Fundamentals,
2nd ed., McGraw-Hill, New York, NY (1986)
3. Shuler, M.L., and F. Kargi, Bioprocess Engineering: Basic Concepts,
Prentice Hall, Englewood Cliffs, NJ (1992)
4. Blanch, H.W., and D.S. Clark, Biochemical Engineering, Marcel
Dekker. New York, NY (1996)
5. Belter, P.A., E.L. Cussler. and W.-S. Hu, Bioseparations: Downstream
Processing for Biotechnology, John Wiley & Sons, New York, NY
6. Peters, M.S., and K.D. Timmerhaus, Plant Design and Economicsfor
Chemical Engineers, 3rd ed., McGraw-Hill Book Company, New York,
NY (1980)
8. Lyderson, B.K., N.A. D'Elia, and K.L. Nelson, eds, Bioprocess Engi-
neering: Systems, Equipment, and Facilities, John Wiley & Sons, New
York, NY (1994)
9. FDA, "Guideline on General Principles of Process Validation" (1987)

S books received

Principles of Condensed Matter Physics, by P.M. Chaikin and T.C. Lubensky;
Cambridge University Press, 40 West 20th Street, New York, NY 10011-
4211; 694 pages, $69.95 (hardback), $47.95 (paperback) (2000)
An Introduction to Nonlinear Chemical Dynamics: Oscillations, Waves,
Patterns. and Chaos, by Irving R. Epstein and John A. Pojman; Oxford
University Press, 198 Madison Avenue, New York NY 10016; 392 pages,
Polymers in Sensors: Theory and Practice, edited by NaimAkmal and Arthur
M. Usmani; ACS Symposium Series 690; Oxford University Press, 198
Madison Avenue, New York NY 10016; 302 pages (1998)
Energy Storage Systemsfor Electronics, Edited by Tetsuya Osaka and Madhav
Datta; Gordon and Breach Science Publishers; 579 pages, $145 (2000)
Oxford Dictionary of Biochemistry and Molecular Biology, edited by A.D.
Smith, S.P. Datta, and G. Howard; Oxford University Press, 198 Madison
Ave., New York, NY 10016; 738 pages, $60 (1997)
Fundamentals of Industrial Catalytic Processes, by R.J. Farrauto and C.H.
Bartholomew; Blackie Academic & Professional; 754 pages $149 (1997)
Fluid Dynamics and Transport of Droplets and Sprays, by William A.
Sirignano; Cambridge University Press, 40 West 20th Street, New York, NY
10011-4211; 311 pages, $80 (1999)
Flow Measurement Handbook: Industrial Designs, Operating Principles,
Performance, and Applications, by Roger C. Baker; Cambridge University
Press, 40 West 20th Street, New York, NY 10011-4211; 524 pages, $110

r, M. laboratory




Modeling and Laboratory Experiment

King Saud University Riyadh 11421, Saudi Arabia

Laboratory work is an important segment of chemical
engineering instruction that cannot be overempha-
sized. It reinforces the theoretical foundation of
courses laid out during lectures, and it is a way to help stu-
dents develop vital experimental skills (e.g., planning, ob-
servation, analysis, communication, etc.) to succeed in their
future careers.",21 Learning skills through "doing," however,
is not always an easy task because of budgetary constraints
and the high costs incurred by equipment maintenance and
safety requirements. These shortcomings can be countered
by resorting to the simulation tool. That is, expensive and/or
hazardous experiments can be simulated by simple analo-
gous physical concepts.
An interesting example offered by this approach is the anal-
ogy that exists between the hydrodynamics of nonviscous
liquids and chemical reaction kinetics. The idea is that the
mathematical models describing the velocity as a function of
the hydrostatic height of free-falling liquids in ducts, with
appropriate shapes, are similar in form to models expressing
the rate of simple chemical reactions as a function of species
This paper describes a safe and inexpensive experiment to
simulate the kinetics of simple chemical reactions by analo-
gous hydrodynamic models. Among the advantages of such
simulation work is the possibility of studying the rate of
chemical reactions in a hazard-free environment necessary
for an experimenter, to avoid unnecessary generation of haz-
ardous waste, and to lower the cost of experimental runs that
involve large quantities of chemicals. The proposed approach
is not intended to be a substitute for real kinetic measure-
ments, but the breadth of the analogy is valuable from the
pedagogical point of view, and the data analysis is an ex-
cellent exercise for a student taking a class in chemical
reaction kinetics.

Figure 1 shows a cylindrical duct within which a fluid flows.
The duct holds some volume of a nonviscous liquid, which
drains through a small area or orifice. The hydrostatic height,
H, of the liquid decreases with time as the volume of the
liquid in the system decreases; the flow is clearly unsteady.
From the macroscopic expressions of the energy and mass
balances within the boundaries of an isothermal system, as-
suming frictional losses to be negligible and the liquid den-
sity, p, to be constant, it is easy to establish the following
relationship (Torricelli's law) between the velocity of the liq-
uid, vo, at the outlet of the cylinder, the velocity, v, of the
falling liquid at the liquid-free surface, and the hydrostatic
height, H,'3'
v2 -v2=2gH (1)
The macroscopic mass balance within the boundaries of the
system states that

-pA dt =pvoA0=pvA=q (2)

where A, Ao, and q are the cylinder cross-sectional area, out-
let orifice area, and mass flow rate of the falling liquid, re-
spectively. If the outlet orifice area is much smaller than the
area of the liquid-free surface, the velocity of the liquid in
Redhouane Henda is an Assistant Professor
at King Saud University, Saudi Arabia. He re-
ceived his MSc and PhD degrees in chemical
engineering from the Ecole Nationale
T Superieure d'lngenieurs de Genie Chimique,
Institute National Polytechnique de Toulouse
B J (France). His interests include diffusion-convec-
tion-reaction systems, solid coatings, and ap-
plication of microwave frequencies to chemical

Copyright ChE Division of ASEE 2001
Chemical Engineering Education

the duct is a small fraction of the velocity of the liquid leav-
ing the duct. Under this condition, the term v2 can be ne-
glected in Eq. (1). This assumption holds as long as the liq-
uid-free surface is far from the outlet (i.e., for values of z
large enough).
A substitution of v obtained from Eq. (2) into Eq. (1) leads
to the final differential form of Eq. (1) as

dt- -H(t)2 (3)

K= (4)

Equation (3) expresses the dependence of the velocity of the
falling liquid on the hydrostatic height and has a similar math-
ematical form to the equation of mass balance for a reactive
system in a closed well-stirred reactor wherein a simple re-
action of order one-half takes place.[4'" In Eq. (3), the hydro-
static height mimics a species concentration in the equation
giving the rate of a chemical reaction. On the other hand,
noticing that the velocity, -dH/dt, of the falling liquid decays
with decreasing hydrostatic height, H, with time is a valu-
able exercise from the practical and instructional viewpoints.
It is exactly the same as the variation of the rate of a simple
chemical reaction of order one-half with concentration. Fi-
nally, by judiciously choosing the geometry of the duct, it is
possible to simulate the kinetics of simple chemical reactions
of various orders.
For simplification, let us consider a duct with the geometry
illustrated in Figure 2. The hydrostatic height, H, and width,

Flow out

Figure 1. Schematic of the physics of a cylindrical duct.
Summer 2001

x, of the liquid inside the duct are assumed to be related by a
simple equation of the form
H = ax" (5)

where n can be an integer or a fractional number, and a is a
If the duct has a constant depth, D, and using Eq. (5), the
expression for the area of the falling liquid-free surface be-

A=Dx= D(HIn (6)

Using the same approach described previously, the term v2 is
neglected in Eq. (1) provided that An is much smaller than A.
By substituting A obtained from Eq. (6) into Eq. (1), the dif-
ferential form of Eq. (1) takes the form

dH(t) n-2
dt =K'H(t)2n (7)

A an
K '= V2g (8)

The analogy between the hydrodynamics of draining a duct
and the kinetics of simple chemical reactions is now estab-
lished. Let us examine some simple cases.
Model of order 0 A nonviscous liquid falling through
a duct whose lateral boundary (see Figure 2) is
parabola-like (n = 2) simulates a chemical reaction
or order zero.

Figure 2. Schematic of an arbitrary-shaped duct.

free surface



Model of order 1/2 This is the situation for which the param-
eter n must be very large. The lateral boundary of the duct
is parallel to z-axis (n >> 1, x = const). The draining of a
cylindrical duct is a special case simulating a chemical
reaction of order one-half.
Model of order 3/2 By equating the exponent in the hydro-
static height in Eq. (7) to 3/2, the boundary of the duct takes
the form of a hyperbola, with n = -1.
Model of order -1/2 A chemical reaction of this order can be
simulated by an analogous hydrodynamic model for which
n = 1. The lateral boundary of the duct for this case is linear
in form.
Similarly, further simple chemical reactions can be simulated using
the hydrodynamic analogy described above.

The equipment required for this experiment is rudimentary and in-
expensive. It consists of ducts that were made from Plexiglass wherein
water flows and leaves through an orifice. Duct surfaces have been
tailored after a precise large-format drawing of the function defining
the shape of the duct as expressed by Eq. (5). Pieces of Plexiglass
were then carefully cut according to the drawing, and the resulting
surfaces were assembled using chloroform as a gluing agent. Four ducts
with shapes as described in the previous section were designed so as
to simulate the rates of four simple chemical reactions. The suggested
experimental procedure is:
1. Fill the duct with a nonviscous liquid (e.g., water) until the zero
reading of the scale.
2. Make sure all air bubbles in the duct are eliminated.
3. Refill the duct with the liquid until the zero reading after
elimination of air bubbles.
4. Open the orifice and start reading the height of the liquid as a
function of time using a stopwatch. Either a time or liquid
height increment can be fixed during the experiment.
5. Take frequent readings and tabulate the results as time vs.
6. Repeat steps 1 through 5 a number of times for a statistical
analysis of the results.

The results obtained for models -l/2(n=1), 1/2(n=oo), and 3/2(n=-1)
considered in this study are shown in Figures 3a,b,c, respectively. All
figures depict values of log(-dH/dt) as a function of log(H). Data rela-
tive to the order (n-2)/2n and constant K' for a given model can be
estimated from the corresponding curve slope and intercept, respec-
tively. These model parameters are summarized in Table 1. The ex-

Figure 3.
Log (-dH/dt) as a function of log (H) for
(a) model -1/2, (b) model 1/2, and (c) model 3/2.
Different markers correspond to different experimental runs.

Chemical Engineering Education

perimental results, at least those relative to the parameter (n-
2)/2n, are in good accordance with theoretical predictions,
discussed earlier, within experimental errors. The latter are
mainly attributed to inaccuracies in readings done by human
operators. For this reason, automation of the experimental
procedure is desirable as it allows for flexibility in data col-
lection and it yields more precise data.
One point worthy of notice in Figure 3 is related to the
distributions of experimental data depicted in Figures 3a and
3b. A careful analysis of the data indicates that the values of
-dH/dt are not sensitive to the hydrostatic height, H, within
many (between 2 and 4) successive readings for the models
with n=l and n=o. These results are attributed to two fac-
tors: first, in order to have a good approximation of -dH/dt
by -AH/At, a small height increment AH has been chosen so
that within the time elapsing between two successive read-
ings, the velocity of the liquid has not changed significantly;
second, both sets of data (in Figures 3a and 3b) correspond to
models with "weak" orders, i.e., -1/2 and 1/2, respectively.
The parameters of the different models were also estimated
by an optimization technique, namely the simplex method,"'6
using the commercial software MATLAB 5.3. The computed
values are in good agreement with those estimated from the

Parameters of Hydrodynamic
Models 1/2, 3/2, and -1/2
(G)-graphical estimation: (M)-MATLAB results

Model (n-2)/2n (G) (n-2)/2n (M)

1/2 0.529 0.537 8.71 10-2 8.34 10'2
3/2 1.535 1.525 4.67 104 4.15 104
-1/2 -0.541 -0.535 14.79 13.59

240 A


s A
1 40

0 100 200 300 400 500 600 700 800
Time, t, s

Figure 4. Height of liquid-free surface as a
function of time.
Summer 2001

plots (Figure 3) and are shown in Table 1.
Figure 4 depicts the hydrostatic height, H, of the liquid-
free surface as a function of draining time, t, for model 0
with n = 2. It can be seen that the height, H, varies linearly
with time so that the liquid velocity, -dH/dt, which is the op-
posite of the slope of the curve H vs. t shown in Figure 4, of
the freely falling liquid has a constant value of -0.3 mm/s-
and is independent of H.

I have reported on a novel experiment for the simulation of
the kinetics of simple chemical reactions by analogous hy-
drodynamic models. The experiment is safe and was set up
with normal laboratory glassware and fittings. It exposes stu-
dents to the power of simulation and to the tools of model
parameter identification. The analogy between the hydrody-
namics of non-viscous liquids and chemical reaction kinetics
can be extended to any reaction order provided the duct shape
is adequately chosen. The obtained experimental data for the
different models considered in this study are in good agree-
ment with the theory and numerical calculations.

Thanks are due to Mr. A.R. Killani for helping in setting up
the equipment and to two undergraduate students for the ex-
perimental results and model parameter estimation.

A cross-sectional area for flow in duct, m2
A cross-sectional area for flow leaving duct, m-
a constant, m1-n
D depth of duct, m
g constant of gravitation, m/s2
H height of liquid in duct, m
n number
q mass flow rate of liquid, kg/s
t time, s
v liquid velocity, m/s
vo liquid velocity at the outlet, m/s
x width of liquid, m
x,y,z Cartesian system of coordinates
K constant, m/2/s
K' constant, mn+2/2n/s
p liquid density, kg/m3

1. Lauterbach, J., S. White, Z. Liu, G.M. Bonder, and W.N. Delgass,
Chem. Eng. Ed., 31, 260 (1997)
2. Mackenzie, J.G., R.M. Allen. W.B. Earl, and I.A. Gilmour, Chem. Eng.
Ed., 33, 150 (1999)
3. Middleman, S., An Introduction to Fluid Dynamics: Principles of
Analysis and Design, John Wiley and Sons, New York, NY (1998)
4. Levenspiel, O., Chemical Reaction Engineering, 3rd ed., John Wiley
and Sons, New York, NY (1999)
5. Froment, G.F., and K.B. Bischoff, Chemical Reactor Analysis and
Design, 2nd ed., John Wiley and Sons, New York, NY (1990)
6. Oberg, T.G., and S.N. Deming, Chem. Eng. Prog., 96(4), 53 (2000) 0

iK' (G)

K' (M)

MlR%= laboratory



Part 8. Absorption of Carbon Dioxide from a Single Bubble

McMaster University Hamilton, Ontario, Canada L8S 4L7

Efficient gas-liquid contact for mass transfer purposes
can be obtained by creating a dispersion of gas
bubbles, as (for example) in a bubble column reactor,
a stirred-tank reactor, or a bubble-cap plate distillation col-
umn.111 An earlier paper in this series described an experi-
ment to measure mass transfer in a bubble column.[2] For a
more basic approach, an experiment can be done with a single
gas bubble of known size that allows for a more precise com-
parison of the results with predictions based on transport phe-
nomena and fluid mechanics.
From the point of view of an inexpensive undergraduate
laboratory experiment, there are some problems associated
with measurements on a freely rising gas bubble. The termi-
nal velocity of a freely rising bubble is typically about 20
cm/s, so a tall column must be constructed to measure mass
transfer over a reasonable period of rise time. Careful ac-
count must also be taken of mass transfer end effects at the
points of bubble release and collection. The experimentation
and analysis require more time than is available in a typical
undergraduate laboratory course.
If some way can be found to observe the mass transfer from
a bubble while it is held stationary in a downflowing stream
of liquid, the experiment is considerably simplified. This pa-
per describes such an experiment in which a cylindrical bubble
of carbon dioxide, which is moderately soluble in water, is
held stationary in a downflow of water in a vertical tube. The
mass transfer rate is calculated from the measured rate at
which the bubble volume decreases with time.

It has been found that good results are obtained with inter-
nal tube diameters in the range of 7 to 10 mm. Only a moder-
ate flow of water is needed to hold the gas bubble station-

* Address: Lakehead University, Thunder Bay, Ontario, Canada P7B 5E1

ary, and it takes up an elongated cylindrical shape as in-
dicated in Figure 1.
Rao and Baird'3' have studied the relationship between the
downward critical superficial velocity of the liquid phase (uc)
for bubble stabilization and the tube diameter and other prop-
erties of the liquid. Data for several different tube diameters
and liquids can be correlated in terms of the Froude number
and the E6tvos number

Fr05 = 0.163 (n(Eo)- 0.222


Fr = c

and Eo=4r2p

Equation (1) has been found to hold for 3.9 independent of the length of the bubble, provided this ex-
ceeds about one tube diameter. Table 1 shows the four differ-
ent values of tube radius r that have been studied in the ex-
periment at McMaster University, with the corresponding
values of u estimated from Eq. (1).
If a single vertical tube is used, the superficial liquid ve-
locity, u, must be controlled precisely at the appropriate value
of u from Eq. (1) in order to hold the bubble stationary, re-
quiring constant attention and small flow adjustments with a
needle valve. In order to avoid this limitation and to allow
four different tube radii to be studied, a glass tube has been
made up of four 12-cm long sections A to D (see Table 1),
with the radius decreasing with height as shown in Figure

Malcolm Baird received his PhD in chemical engineering from Cambridge
University in 1960. After some industrial experience and a post-doctoral
fellowship at the University of Edinburgh, he came to McMaster University.
His research interests are liquid-liquid extraction, oscillatory fluid flows, and
hydrodynamic modeling of metallurgical processes.
Inder Nirdosh received his BSc and MSc in chemical engineering from
Panjab University (India) and his PhD from Birmingham University (United
Kingdom). He joined Lakehead University in 1981, and his research inter-
ests are in the fields of mineral processing and electrochemical engineering.

Copyright ChE Division of ASEE 2001
Chemical Engineering Education

2. Water is fed down through the tube at a measured flow
rate as shown.
In this type of multisectioned tube, a cylindrical gas bubble
tends to stabilize in one of the four sections, depending on
the flow rate of water supplied. For example, for any flow
rate of water between 0.896 and 1.80 mL/s, a cylindrical
bubble will tend to rise to the top of section B since the water
flow is less than 1.80 mL/s; but it cannot rise into section A
since the water flow rate exceeds 0.896 mL/s. Thus it can be
seen that a bubble can be held stationary in any section by
setting the flow rate at any required value between certain
limits, without the experimenter having to worry about the
effect of any slight drift in the water flow rate upon the bubble

The rate of mass transfer is given by the well-known equa-

m'= kAAc (2)

The area A is taken to be that of the curved cylindrical sur-
face enclosing the bubble; the nose and tail areas are ne-

velocity, us


Flow Q

S--8 L

-\ L

- I -

vel. =Q/r r2

Figure 1. Cylindrical gas bubble, showing some of the
symbols used.

Tube Radii, Critical Liquid Velocities, and Water Flow
Rates for Bubble Stabilization

Internal tube radius, mm
Eotvos number for water
Velocity u from Eq. (1), mm/s
Flow rate, 7r'u mL/s

3.51 3.98
6.708 8.642
23.14 36.08
0.896 1.80



glected. Typically, these areas are no more than 5% of the
total area. From Figure 1

A = 2 7t(r 8)L (3)
The concentration driving force, Ac, is taken to be the solu-
bility c* of carbon dioxide in water, assuming that the
downflowing water is free of dissolved gas.
Since carbon dioxide has a low solubility in water, and pure
carbon dioxide gas is used, it can be assumed that the mass
transfer rate is liquid phase diffusion-controlled.
If t is the contact time of the moving liquid film at the cylin-
drical bubble surface, the well-known Higbie penetration
model'4 provides an expression for the mass transfer coeffi-

k = 2 D (4)

where D is the molecular diffusivity of carbon dioxide in water.

While the gas bubble is stationary in the tube, the water flows
downward past it as a moving film. If the surface velocity of
the film is u, then

L=L (5)
Assuming that the film is laminar and that the film thickness,
8, is much less than r (typically 8 / r = 0.05), the simple ex-

To sink '

Figure 2. Experimental arrangement with
multisectioned tube.

Summer 2001

pressioni5' for the surface velocity of a liquid film flowing
down a vertical flat plate
us = 0.3054 3 (6)
pIr2 )(6)
can be used. It can also be shown that the film thickness is

8= 3IQ 3 (7)
When Eqs. (3) through (7) are substituted into Eq. (2), a
theoretical expression is obtained for the rate of mass trans-
fer in terms of the flow rate of water, the tube radius r, the
bubble length L, and the system properties. For many repeated
calculations, writing a small computer program is recom-
mended. It is important to note that the theoretical approach
depends on several assumptions, including neglect of the nose
and tail contributions, the validity of the Higbie penetration
model, and the simplification 8< It is also assumed that the contact time, T, (usually less than
0.5 s) is much shorter than the time, t, over which the experi-
ment is carried out.

The procedure is to start the liquid flow at the desired value
(Figure 2) and then inject a controlled amount of carbon di-
oxide from the cylinder by means of the toggle valve as shown.
A needle valve is placed just upstream of the injection toggle
valve and the gas pressure upstream of the needle valve is
regulated at a low gauge pressure, preferably about 10 kPa or
1.5 psig, to avoid excess gas injection. The gas injection tech-
nique requires some practice and dexterity, but if a mistake is
made (e.g., injection of too much gas), the system can be
freed of bubbles by momentarily increasing the water flow
or by suction, using the vacuum purge as shown in Figure 2. It
is important to ensure that no air remains in the injected carbon
dioxide bubble, as this would lead to slower mass transfer.
Once a bubble of suitable length (typically 10 cm) has been
injected and stabilized in one of the tube sections, its length
will be seen to gradually shrink. This is due to dissolution of
carbon dioxide in the water; the rate of shrinkage provides a
direct measurement of the mass transfer rate. A millimeter
scale mounted behind the vertical glass tube enables the
bubble length L to be measured as a function of time. The
best arrangement is to have one student calling out the values
of L and another keeping a note of the values of L and time;
alternatively, a close-up video camera can provide a timed
record of the bubble shrinkage.

The linear rate of shrinkage of the bubble can be related to
the volumetric rate of shrinkage and to the rate of mass trans-
fer by

dV -(-2dL ( RT
dt --(r) d= MP (8)
The term R is the universal gas constant, and M denotes the
molecular weight of carbon dioxide (= 44 kg/kmol).
The theoretical mass transfer rate can be expressed using Eqs.
(2) to (5), and hence

_(_2 dE ('DLu RT>
_(r -_8)2 dL= 4 c*(r-8) DLu-- 2 ) (9)
dt ( c MP) (9)

Note that this expression is in the form
dL=KL0.5 (10)
which can be integrated to give
L0.5 = L05 Kt (11)
0 2
where Lo is the bubble length initially measured (t=0). Thus,
a plot of L1/2 versus time should be linear with a negative
slope. The slope should be predictable from theory, accord-
ing to which

(Du 0.5
4c* --
SI ) T) (12)
K (r-8) (12)

where us and 6 are given by Eqs. (6) and (7), respectively.

Some typical data from a student report161 are shown as a
plot of L"5 versus time in Figure 3. The linearity of the plot
provides a qualitative support for the penetration model.14'
The slopes of these plots, -K/2, are determined by linear
regression and the experimental values of K can thus be cal-
culated. The corresponding experimental values of K for the
four typical cases are compared with the theoretical values
(Eq. 12) in Table 2.
It can be seen that the experimental values of the mass trans-
fer rate constant are about 30% below the theoretical values.
Students should be encouraged to carefully examine the vari-
ous assumptions in the theoretical treatment. The assump-
tion that 6< already been mentioned. Another simplifying assumption is
made following Eq. (3), namely that the driving force for
mass transfer is the gas solubility. But a material balance for
carbon dioxide indicates that in a typical case the exit liquid
concentration is as much as 10% of the saturation value. This
does not invalidate the theoretical model, because it can be
shown that the dissolved carbon dioxide is present in a thin
boundary layer near the surface, rather than being uniformly
distributed. This concentration profile is accounted for in the
derivation of Eq. (4) from the unsteady diffusion equation."7
The surface velocity of the falling film (Eq. 6) is calcu-
lated on the assumption that the shear stress at the gas-liquid
Chemical Engineering Education

interface is negligible. But there is evidence from the litera-
ture'81 that surface tension gradients can cause a deceleration
of the liquid surface and the formation of ripples at the rear
of a cylindrical gas bubble. This has the effect of reducing
the average velocity, therefore increasing the surface contact
time and significantly reducing the mass transfer rate. Van
Heuven and Beeki81 reported mass transfer rates about 30%
below theoretical predictions and attributed the reduction to
the surface deceleration effect.

It can be concluded from this experiment that transport
phenomena are helpful in understanding the basic mecha-
nism of mass transfer and obtaining an estimate of mass trans-
fer rates from first principles. The derived equation for the
mass transfer rate, however, is subject to simplifying assump-
tions, and therefore the estimate is only approximate.
The experiment is simple and inexpensive to construct and
is much less elaborate than the rig used by Van Heuven and
Beek,181 which gave similar results. The most costly items are

I5 10 15 20 25 30
Time, t, s
Figure 3. Typical data on bubble shrinkage.!'
0 Tube radius 3.51 mm, waterflow 0.352 mL/s
[ Tube radius 3.98 mm, water flow 1.68 mL/s
V Tube radius 4.41 mm, water flow 2.69 mL/s
A Tube radius 5.05 mm, waterflow 3.96 mL/s

Typical Observed and Calculated Values'6] of Mass
Transfer Rate Constant

Tube Flow
Radius Rate Values of K, m"S.s' Ratio
Sect. (mm) (mL/s) Observed Calculated Kb/K,
A 3.51 0.352 0.00741 0.01051 0.705
B 3.98 1.68 0.00964 0.01537 0.608
C 4.41 2.69 0.01097 0.01563 0.702
D 5.05 3.96 0.01089 0.01478 0.737

Summer 2001

the rotameter and the needle valve. Some manual dexterity is
required on the part of the students, but experience has shown
that with a little practice the technique can be made to work.

The authors are grateful to the Natural Sciences and Engi-
neering Research Council of Canada for partial financial help
in the preparation of this paper.

A surface area, m2
c* solubility of CO, in water, kg/m3
D molecular diffusivity, m2/s
Eo Eotvos number
Fr Froude number
g gravitational acceleration, m/s2
k mass transfer coefficient, m/s
K rate constant, mo 5/s
L bubble length, m
L0 initial bubble length, m
m' mass transfer rate, kg/s
M molecular weight, kg/kmol
P gas pressure, Pa
Q liquid flow rate, m3/s
r tube internal radius, m
R universal gas constant, J/(kmol.K)
t time, s
T temperature, K
u critical flow velocity, m/s
u bubble surface velocity, m/s
V bubble volume, m3
Greek symbols
6 liquid film thickness, m
.t liquid viscosity, Pa.s
p liquid density, kg/m3
a surface tension, N/m
T contact time, s
obs observed
calc calculated

1. Treybal, R.E., Mass Transfer Operations, 3rd ed.. McGraw Hill, New
York, NY: Ch. 6(1980)
2. Nirdosh, I., L.J. Garred, and M.H.I. Baird, "Low-Cost Experiments in
Mass Transfer: Part 3. Mass Transfer in a Bubble Column," Chem.
Eng. Ed., 32, 138 (1998)
3. Rao, N.V.R., and M.H.I. Baird, "Continuous Measurement of Surface
and Interfacial Tension by a Stationary Slug Method," Can. J. Chem.
Eng., 61, 581 (1983)
4. Higbie, R., "The Rate of Absorption of a Pure Gas Into a Still Liquid
During Short Periods of Exposure," Trans. AIChE, 31, 365 (1935)
5. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport Phenomena,
John Wiley and Sons, Inc., New York, NY: p. 37 (1960)
6. Chong, C., Chemical Engineering 3L2 Report, McMaster University,
February 2 (1999)
7. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport Phenomena,
John Wiley and Sons, Inc., New York, NY: p. 537 (1960)
8. Van Heuven, J.W., and W.J. Beek, "Gas Absorption in Narrow Gas
Lifts," Chem. Eng. Sci., 18, 377 (1963) O

e, curriculum





University of Ottawa Ottawa, Ontario, Canada

During their fourth year of undergraduate study, chemi-
cal engineering students at the University of Ottawa
complete a course entitled "Computer-Aided Design
in Chemical Engineering." This course is intended to pro-
vide students with some knowledge of process simulation and
its use in the design of chemical unit operations.
The computer-aided design course has undergone several
transformations since first being introduced into the Univer-
sity of Ottawa's chemical engineering undergraduate curricu-
lum. Initially, the course centered on the application of soft-
ware designed by one of the faculty members. The original
intent of the software, however, was as a research tool rather
than a pedagogical one, with the end result that the students
using the package learned nothing about the construction of
process simulators and gained little in the way of additional
engineering skills. It was then decided that the course should
employ a commercial software package-in this case FIDAPO,
a computational fluid dynamics package produced by Fluent,
Inc. While this had the advantage of giving students experience
in using a commercial product, it provided little insight into the
design and construction of process simulators.

David G. Taylor is Associate Professor of
Chemical Engineering at the University of Ot-
tawa. He received his BASc in engineering sci-
ence at the University of Toronto and his PhD in
chemical engineering from the University of Brit-
ish Columbia. His research focuses on process
modeling and computer simulation using object-
oriented approaches. Dr. Taylor is also keenly
interested in distributed and computer-mediated
learning. He can be reached via email at

In 1999, the computer-aided design course was redesigned
with an entirely new approach and philosophy in mind: the
revised course required students to design, develop, and test
their own simulators of various unit operations using object-
oriented programming. This article discusses the rationale
behind the changes, provides details regarding the revised
course structure, and presents students' reactions to the
course's new focus.

Chemical engineering students at the University of Ottawa
undertake a programming course in procedural language such
as C in their first year. Hence, prior to the computer-aided
design course, these students receive no training in object-
oriented programming. The question then arises, why intro-
duce object-oriented programming into this course?
It so happens that all object-oriented programming lan-
guages share four characteristics that make them particularly
well suited to process simulation: abstraction, encapsulation,
inheritance, and polymorphism. Abstraction is key to the pro-
cess of object-oriented design. As Booch"' describes it, "An
abstraction denotes the essential characteristics of an object
that distinguish it from all other kinds of objects and thus
provide crisply-defined conceptual boundaries, relative to the
perspective of the viewer." In other words, the abstractions
are conceptual models of the physical entities we seek to de-
scribe programmatically. These abstractions are represented
in object-oriented programs by independent units of code re-
ferred to as classes. The process of defining classes to repre-
sent the system is therefore intimately linked to the modeling
process itself.

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

Encapsulation, also known as data hiding,21 refers to the
way in which object-oriented programs combine data and
behavior to assure the greatest degree of code integrity. By
controlling exposure of code through public and private in-
terfaces, encapsulation simplifies program debugging and
assists the designer in building more robust applications.
Inheritance is a mechanism in object-oriented programming
languages that allows one to take an existing class, referred to
as the parent class, and with comparatively little effort, extend
it to produce a second class, called the child class, that describes
a new and distinct component. Through inheritance one can
program from the general to the specific. For example, one may
design code to describe a generic heat exchanger class which is
later extended to represent more specific types, such as counter-
current or co-current exchangers. All of this is done with a mini-
mum of additional programming and without altering the par-
ent class in any way, which, in turn, opens opportunities for
creating highly flexible and reusable code.
Finally, polymorphism, as defined by Kafura,'3' is "the abil-
ity to manipulate objects (i.e., instances of classes created
during execution) of distinct classes using only knowledge
of their common properties without regard to their exact
In practical terms, polymorphism is a mechanism whereby
object-oriented languages can determine the particular ex-
ecution of code at run time rather than during compilation, as
occurs with procedural languages. This allows "execution on
the fly," in many instances eliminates the need for switch or
other logical control structures, and provides a high degree
of interactivity between the user and the simulation during

These concepts are best illustrated through a simple example.
Consider the heat exchanger shown in Figure 1. Many design
problems require us to estimate the outlet temperature of one of
the two fluids associated with the exchanger, which in turn de-
pends on the specific heat exchanger used. Standard heat ex-
changer configurations, however, can be analyzed using the NTU
method,'4' which relates the inlet and outlet temperatures of the

Figure 1. Schematic diagram of a heat exchanger.

In 1999, the computer-aided
design course was redesigned with an
entirely new approach and philosophy in
mind: the revised course required students
to design, develop, and test their own
simulators of various unit operations
using object-oriented programming.

two fluids to the capacity of these fluids (Chot and Cco.d) and the
effectiveness factor, e:

C hot THot.out Ccold (TCold.out TColdin
=T (1)
C in i( mi Hotin
where Cn is the minimum of Chot and Ccold'
The effectiveness factor depends on the type of heat ex-
changer in operation, the exchanger's heat transfer charac-
teristics, (i.e., its heat transfer coefficient, surface area prod-
uct, UA), and the capacity rates of the fluids. For a counter-
current heat exchanger, for example, we have

1-ex [At 1 Cmin
1cxpf UA (f Cmin 1

C m UA Cmax
- C exp C (I C
maX m max
'- epI- '- 11

while for a co-current heat exchanger the effectiveness factor
is given by the following:

-ex UA (, Cmin
-exPI 1+ C
Smm max

In both cases, C refers to the maximum of Chot and Cold'
Since the effectiveness factor depends only on the fluid-ca-
pacity rates and exchanger heat-transfer characteristics, it can
be determined without knowing the fluid tem-
peratures. One can then evaluate any of the four
fluid stream temperatures by setting the other
three, calculating the effectiveness factor for the
given exchanger and subsequently solving the
1 first equation.

To apply abstraction to the problem, we con-
sider the nature of the device itself. Here, we
focus on the noun "heat exchanger" in order to
consider those features of the physical sys-
tem that are relevant to our analysis. The
closer this conceptual model is to the origi-
nal system, the better will be our object-ori-
ented representation. For simplicity's sake,

Summer 2001

however, we will only consider the unit operation in the
context of the NTU analysis.
It turns out that this problem is ideally suited to the appli-
cation of inheritance, since there are numerous forms of heat
exchangers that can be analyzed using the NTU method. We
therefore begin by defining a generic
class that will serve as a template for the more specific de-
scriptions. We then define various child classes, such as a
class and a
class that represent specific forms of the generic heat ex-
changer (see Figure 2).
Through encapsulation, our
class will contain both properties (through its instance vari-
ables) and behavior (via its methods). We decide which in-
stance variables and methods to include by considering those
properties and behaviors common to the NTU analysis of
any heat exchanger. For example, the UA product is com-
mon to all heat exchangers, so we will include it as an in-
stance variable in our parent class. Likewise, all heat exchang-
ers will rearrange Eq. (1) to calculate one of the stream tem-
peratures-hence we include in the parent class a method
such as
public double getHotOutletTemp
(double THotIn, double TColdIn,
double CHot, double CCold)
double CMin;
if (CHot else CCold=CMin;
return getEFactor (CHot,CCold)
Note that this method makes use of another method:
The latter method calculates the appropriate effectiveness
factor using the capacity rates passed to it. The details of this
method, however, will not be defined in the parent class. In-
stead, each child class, e.g.,
will redefine this method, using its appropriate definition of
the effectiveness factor. Meanwhile, the child class will in-
herit the parent's
method and its UA product and, as a result, will not need to

SCounterCurrentHeatExchanger CoCurrentHeatExchanger

Figure 2. Inheritance structure between various heat
exchanger classes. Child classes
(CounterCurrentHeatExchanger and
CocurrentHeatExchanger) derive properties and
behavior from the parent class (HeatExchanger).

include code for these elements. In other words, to create a
new heat exchanger class, we need only write code that rede-
fines the
method for the particular exchanger. The rest of the function-
ality is inherited from the parent class and needs not be ex-
plicitly defined.
Now suppose we want to design a simulator in which the
user can select, during the simulation, the type of heat ex-
changer to be included in the calculations. Using a proce-
dural language, we would have to include a switch statement
or other logical structure within the code identifying the spe-
cific heat exchanger to be used.
Object-oriented programming and polymorphism make the
process far simpler for us. In this case, we create a simulator
class containing a method
that includes the generic
in its parameter list. This method will carry out the necessary
heat transfer calculations using the object-oriented represen-
tation of the heat exchanger. During run time, the user can
select the particular heat exchanger to be used in the calcula-
tions. The program then creates an instance of this particular
heat exchanger class, e.g., a
object and passes it to
Although this method is unaware of the type of
it has received, it simply expects a generic
and it will still use the form of the
method appropriate for the specific heat exchanger object
passed to it, without the need for any switch statements or

Chemical Engineering Education

logic controls. It is polymorphism that makes this possible.
Now suppose that we decide to extend our list of available
heat exchangers. In this case, we need only create the appro-
priate child class, defining the new
method there, and adding it to the list of exchangers avail-
able to the user during run time. When the user selects this
new exchanger type, it is passed to the simulator's
method. No alteration or recompilation is required by
Once again, this extensibility and flexibility in the simulator
is possible because of polymorphism.

When introducing object-oriented programming into a
course, one is faced with the further decision of which lan-
guage to use. There are a number of languages from which to
choose, perhaps the most popular being C++; however, Java
was selected for the computer-aided design course for the
following reasons:
Java is easier to learn than C++
because Java and C++ share a similar syntax, upon
mastering Java it is relatively straightforward to
migrate to C++, should one wish
given the current trend toward delivering information
and services via the Internet, it makes sense to exploit
a language that is "Internet-ready," namely, Java

In designing the content and structure of the computer-aided
design course, two objectives were kept in mind: to provide
students with a solid grounding in object-oriented design and
programming, and to provide them with experience in de-
signing, building, and employing process simulators. In or-
der to achieve the first objective, the course content centered
on object-oriented design and the Java language. Students
learned about classes, objects, and the principles of object-
oriented design. They were taught the language syntax, as well
as Java's means for handling errors at run time and the use of
streams (Java's mechanisms for input and output). They were
also given instruction in the building of graphical user inter-
faces using Java.
Meanwhile, matters surrounding the design and construction
of process simulators were introduced via numerous in-class
examples, through assignments and by means of a major project
in which small groups (four students or less) designed and con-
structed simulators of a particular chemical unit operation. As
part of the project, each group produced two interim reports, a
final report, and made an oral presentation to the class. The
Summer 2001

students were also required to write a final exam.
When the revised version of the computer-aided design
course was first introduced in 1999, it centered largely on
weekly lectures when the students would receive formal pre-
sentations of the course material along with in-class demon-
strations. The lectures also exploited collaborative learning5'
on a regular basis. In this case, the class was divided into
small groups and each group was routinely given the same
problem to work on. After 10 to 15 minutes, one of the groups
was randomly selected to present its findings. In this way the
in-class group problem sessions served to identify areas of
confusion for the students and to act as a springboard for
further discussions.
In addition to the regular lectures, four hours of operational
tutorials were also provided in one of the university's computer
labs. During these tutorials, students received small program-
ming assignments to work on while the professor circulated
through the classroom to assist where needed. A web site was
also built from which students could obtain lecture material,
download assignments and solutions, view course announce-
ments, and retrieve their marks.

All lecture courses taught at the University of Ottawa un-
dergo student evaluation near the end of term. The evaluations
include a set of standard questions seeking feedback re-
garding the course content and its delivery by the profes-
sor, as well as space for students to provide their anony-
mous written comments.
Presented here are the students' responses to four of these
questions that focus on course content and structure. The first
question asked the students to evaluate the course overall, rank-
ing it from very poor to excellent. The second question asked
for the students' perspective regarding organization of the course.
The third question asked the students to rank how well the course
provided feedback of their progress throughout the term, while
the fourth question asked for their impressions of the course
workload relative to others in the program.
The students' responses to these questions during the 1999
course evaluation are given in Figure 3 (next page). As the fig-
ure shows, students were, for the most part, satisfied with the
course overall, including its organization and feedback mecha-
nisms. Based on the students' comments, it is also clear that
they enjoyed the general structure of the lectures, as evidenced
by the typical response:

- "I really enjoyed the overall organization of this course;
multimedia presentations (PowerPoint slides, in-class
running programs), in-class problems, etc."
It was clear from both the questionnaire and from the written
comments, however, that the students found the workload in
the course much heavier than in their other courses. As one
student stated on a comment sheet,

- "There is too much material to cover The course load
needs to be smaller in order to allow students to
actually learn OOP especially since most of us only
know Fortran."

Students also expressed dismay with the course textbook, but
showed considerable support for the tutorial sessions:

- "Thanks for the extra tutorials, we need more i.e. to
be taught in front of the computer "

Bearing these comments in mind, the following changes

were implemented the next year. First and foremost, the three
hours per week of contact time were divided equally between
formal lectures and structured tutorials in which students
worked on assigned problems while the professor and the
teaching assistant circulated through the class to provide them
with guidance as needed. The completed assignments would
be handed in at the end of each tutorial and, during certain
weeks, were graded. The total number of homework assign-
ments was reduced somewhat, as was the scope of the major
project. (In the latter case, the oral presentation was elimi-

Overall, I Find This Course

The Organization Of This Course Is



* 1999
Z 03
8 2000
L 0.2

Very Poor Poor Acceptable Good Excellent

By Comparison, The Workload Is


* 1999 C
* 2000 uL

Very For Poor Acceptable Good Excellent

Very Light Lighter Than Average Heavier Than Very Heavy
Average Average

Figure 3. Student responses to the course evaluation questions for years 1999 and 2000: evaluations of (a) the course
overall, (b) its organization, (c) feedback provided to students during the term, and (d) the workload.

Chemical Engineering Education





L 02



* 2000

Very Poor Poor Acceptable Good Excellent

Feedback Provided To Me On The Progress Of
My Learning Is






* 1999
* 2000

nated and students were required to complete only a single
report.) Finally, the professor began creating extensive course
notes to compensate for the lack of a suitable text.
The student evaluations following these changes are also
found in Figure 3. The overall rating of the course improved
significantly, without sacrificing the course's organization or
feedback from the students' perspective. Furthermore, it is
clear that, while students still found the workload heavier
than in other courses, the situation had improved significantly
from the previous year. The student comments likewise sup-
ported these changes.
> "Considering this was thefirst time almost any of us
have used Java (or even object-oriented type of code),
the class was quite good. The tutorials were a great
idea... I also liked the fact that we learned through
"Excellent course. Java is a LOT more interesting than
other languages we have learned."
"I liked that you used lots of chemical engineering-
relevant examples... it made it a lot easier to see just
how OOP works and what it can do."

Object-oriented methodologies offer several significant
advantages when designing and building process simula-
tors. After two years of teaching these approaches to
fourth-year chemical engineering students at the Univer-
sity of Ottawa, it appears that most of these students are
able both to grasp the basic concepts of object-oriented
programming and to apply these to design problems within
a single semester course. Further, it is clear that Java pro-
vides an effective tool with which students can explore
these concepts. Success, however, also depends to a great
extent on the opportunities for practical application of
object-oriented programming through structured tutorials,
assignments, and projects.
A course of this nature continues to face challenges. First
and foremost, chemical engineering students must be ex-
posed to object-oriented design within the context of pro-
cesses and systems familiar to them-there is little ben-
efit in illustrating object-oriented programming in Java
with a bubble-sort routine. Unfortunately, while there is a
plethora of Java-related texts available on the market, none
of these is well suited to a course in computer-aided de-
sign. This gap must then be filled by well-conceived ex-
amples and detailed notes from the instructor.
While some engineering schools have introduced ob-
ject-oriented languages such as Java and C++ during the
sophomore year of their undergraduate program, the en-
gineering program at the University of Ottawa continues
to teach a procedural language (in this case, C) to most of
its undergraduates. Therefore, when undertaking the com-
puter-aided design course described here, our students
must first learn the basic syntax of Java. More importantly,
Summer 2001

letter to the editor

Dear Editor:
Query: A colleague told me that the following equa-

f=1.375-10-3 1+ 20,000- +1

reproduces the turbulent part of the friction factor chart
very well. I have tested it; it does. It is much easier to
use than any other such function that I know of, because
it is explicit in each of the three variables.
Alas, my colleague does not know where it came from.
I would like very much to know the benefactor of our
profession who invented this function, and I would like
to give credit when I cite it.
If any of our readers can tell me who the author is and
cite a proper reference for this function, I will be most

Noel de Nevers
University of Utah

they must also rethink the programming process, often
"unlearning" approaches adopted during their study of
procedural languages. Both of these activities limit the
time available within this single semester course to present
more complex examples of object-oriented process simu-
A more effective approach, in the author's opinion,
would be to offer a programming course in Java to all
engineering students during their first year of study, leav-
ing more time to focus on practical applications of the lan-
guage in later stages of the undergraduate curriculum.

The author wishes to acknowledge the chemical engineer-
ing students at the University of Ottawa who have proven
themselves, time and again, to be equal to the challenge.

1. Booch, G.. Object-Oriented Analysis and Design with Applications,
Addison-Wesley (1994)
2. Horstmann, C.S., and C. Cornell, Core Java 1.2, Volume I-Fundamen-
tals, Prentice Hall (1999)
3. Kafura, D., Object-Oriented Software Design and Construction with
C++, Prentice Hall (1998)
4. Holman, J.P., Heat Transfer, McGraw Hill (1990).
5. Taylor. D.G., "Computer-Mediated Collaborative Learning in ChE at
the University of Ottawa," Chem. Eng. Ed., 33, 250 (1999) D

r. M classroom



The University of Texas at Austin Austin, TX 78712-1062

he Bode plot is an important tool for stability analysis
of closed-loop systems. It is based on calculating the
amplitude and phase angle for the transfer function

GOL(s)= Gc(s)Gp(s) (1)
s = jo(
where Gc(s) is the controller and G,(s) is the process. The
Bode stability criterion presented in most process control text-
books is a sufficient, but not necessary, condition for insta-
bility of a closed-loop process.-'41 Therefore, it is not pos-
sible to use this criterion to make definitive statements about
the stability of a given process.
Other textbooks"5'61 state that this sufficient condition is a
necessary condition as well. That statement is not correct, as
will be demonstrated in the following examples. In another
text17 the criterion is formulated as a necessary condition for
stability, but no definite statements can be made based on a
necessary condition alone.
Often, some statements are added for clarification,11 6'81 e.g.,
"...the Bode stability criterion only applies to systems that
cross 4 = 180 once, where 4 is the phase shift of the trans-
fer function Gc(s)Gp(s). For multiple crossings one must use
the Nyquist criterion."'31
It can be shown that the above statements about the Bode
stability criterion are not complete. For example, a system
can cross the -180 phase angle line only once, have an am-
plitude ratio of less than one at the corresponding frequency,
and still be unstable. This is because G((s)G,(s) can have an

amplitude ratio greater than unity for frequencies where 0 =
-180-n*360, where n is an integer. These conditions can
occur when the process includes time delays, as shown in the
following example.

Juergen Hahn was born in Grevenbroich, Ger-
many, in 1971. He received his diploma degree
in engineering from RWTH Aachen, Germany,
in 1997, and his MS degree in chemical engi-
1998. He is currently a PhD candidate working
as a research assistant in chemical engineer-
ing at the University of Texas, Austin. His re-
search interests include process modeling, non-
linear model reduction, and nonlinearity quanti-
Thomas Edison is a lecturer at the University
of Texas, Austin. He received his BE degree in
chemical engineering from Annamalai Univer-
sity, his MTech degree from IIT Kanpur, and his
PhD degree from the University of Maryland,
College Park. His research interests include
phase transition in liquid-liquid systems, equa-
tion of state close to critical point, and robust
process control.

Thomas F. Edgar is Associate Vice President
of Academic Computing and Instructional Tech-
nology Services at the University of Texas, Aus-
tin. He received his BS in chemical engineering
from the University of Kansas and his PhD from
Princeton University. He has concentrated his
academic work in process modeling, control, and
optimization, with over 200 articles and book
chapters in addition to coauthoring two textbooks.

Copyright ChE Division of ASEE 2001

Chemical Engineering Education


The Bode stability criterion presented in most process control textbooks is a
sufficient, but not necessary, condition for instability of a closed-loop
process. Therefore, it is not possible to use this criterion to make
definitive statements about the stability of a given process.


A reboilerprocess with a
proportional-derivative (PD) Controller

A process transfer function with inverse re-
sponse and integrating action as seen in some
reboilers is to be controlled by a PD controller
with a pre-filter.

Gp(s)= -s+ e-os
s(Ts + 1)

Gc(s)= K DS + 1

The reason that a PD controller is chosen over
a PID is that the process transfer function al-
ready contains a pole at the origin. Therefore,
the controller for this process should not include
integral action for impulse or step inputs in the
setpoint or disturbances.
The parameters are chosen to be

T = 0.1 min
e = 0.4 min
Kc = 0.2
TD = 1 min
a = 0.05

The resulting Bode plot for the open-loop sys-
tem is given in Figure 1. The phase crossover
frequency (cio80o) is defined to be the fre-
quency at which the open-loop phase angle is
-180'. Furthermore, the gain crossover frequen-
cies (co ) are defined to be the frequencies at
which the open-loop amplitude ratio is equal to
unity and Oc540o corresponds to the frequency
where the phase angle crosses -540.
The amplitude ratio corresponding to a phase
lag of -180 is 0.6. One could reach the follow-
ing false conclusions from the Bode stability

9 2 93
10 -'[ 'I -- .-- . i ----'^ ... . i -- :; -. . ..-^ -
101 100 10 : 102

-90- : c180o C540o

*e- :

-450 -

-630 0 '
10 100 0 101 102

Figure 1. Bode plot of the reboiler process.

Real part of GOL(s)

Figure 2. Nyquist plot of the reboiler process.

Summer 2001

> The system is stable.
O The gain margin is 1.67. The controller
gain can be increased by 67% without
making the system unstable.
But the amplitude ratio corresponding to a
phase angle of -540 is 2.0 and thus greater than
unity. Therefore, we conclude that the closed-
loop system is unstable. Increasing the control- 10o
ler gain makes the system even more unstable. _
Instead, a reduction of the gain by 50% will re-
sult in a stable closed-loop system.
These conclusions can be validated by ana- 10-' .
101 0 100 101 102
lyzing the Nyquist plot in Figure 2. It is appar- n
ent that this system is unstable due to the fact
that the curve shown encircles the point (-1,0) 90
twice in a clockwise direction. When this sys- -90
tem is implemented in MATLAB and simulated,
the finding from the Nyquist plot is confirmed. e-270

Another commonly found statement about the
Bode stability criterion is that it cannot be used -63o .
if the frequency response of the open-loop sys- 10' 1 10' lo
tem exhibits "nonmonotonic phase angles or
amplitude ratios at frequencies higher than the Figure 3. Bode plot of the time-delay process.
first phase crossing of -180'."18' Although this
statement applies to many cases and would ex-
clude the above example, it can lead to false
conclusions. For example, the amplitude ratio
of a system can be monotonically increasing for 2.5
a PID controller after the notch frequency,'5 2
while the phase angle is constantly decreasing
due to a time delay in the process. It is possible 1.5
to construct a case where the notch frequency
of the system has a phase lag of less than 180' 1
and the corresponding amplitude ratio is less d. -
than unity. Although this system behaves mono-
tonically in both phase and amplitude ratio af-
ter its phase crossover frequency, further analy-)
sis is required to determine the stability of the -o.s \
system. Example 2 illustrates this point.


Control of a time-delay process with an elec- .5 o_____
tronic proportional-integral-derivative (PID) -2. 5 -2 1 .5 of G .5 1 1.5 2 2.5
controller Real part of GoL(s)
Figure 4. Nyquist plot of the time-delay process.

Assume a process that consists of a pure time
delay and is controlled by a PID controller.

Chemical Engineering Education

Gp(S) = e-8s
T's +1 IIDS+l
Gc(s>= K C [ XTDS+

The parameters are given by
0 = 0.6 min
Kc = 0.1
TZ = 4 min
TD = 1 min
a = 0.05
Figures 3 and 4 show the Bode and Nyquist plots of the
open-loop transfer function of this process.

The notch frequency (co) for this PID controller is 0.5
min-' and the amplitude ratio of the open-loop process is
monotonically increasing at higher frequencies. The phase
crossover frequency is located at )c180o = 0.7 min-i and at
higher frequencies, both the amplitude ratio and the phase
angle are monotonic. It can be concluded from the Nyquist
diagram that this system is unstable, since (-1,0) is encircled
an infinite number of times in a clockwise direction. This
result has been confirmed in simulations. If the Bode stabil-
ity criterion in any of the above mentioned forms is used to
determine stability of the system, however, it would lead to
the wrong conclusion.

The foregoing examples indicate that it is not possible to
formulate a Bode stability criterion that is simple to use and
applicable to all possible cases at the same time. Therefore
we conclude:

> A system should only be analyzed for
stability using the Bode plot, if it has at most
one phase crossover frequency. Additionally,
if it has only one gain crossover frequency
and the amplitude ratio as well as the phase
angle are decreasing at the gain crossover
and afterward, then the gain and phase
margins can be calculated in a way found in
control textbooks.
A system that has only one phase crossover
frequency but multiple gain crossover
frequencies is stable if the amplitude ratios,
corresponding to frequencies where t =
-180-n*360', are all less than unity and the
open-loop system is stable. The gain margin is
calculated from the crossover frequency or a
frequency corresponding to a larger n,
whichever exhibits the largest amplitude ratio.

Summer 2001

> If the Bode plot information is inconclusive,
the Nyquist stability criterion should be
applied for stability analysis of closed-loop

From these conclusions we propose a

Revised Bode Stability Criterion

A closed-loop system is stable if the open-loop
system is stable and the frequency response of
the open-loop transfer function has an amplitude
ratio of less than unity at all frequencies corre-
sponding to 0=-180-n*360, where

The proof of the revised Bode stability criterion follows
directly from the Nyquist criterion. When this definition of a
stability criterion is recast in a form for use in a Nyquist dia-
gram, the resulting set of closed-loop stable systems is given
by the curves that do not cross the real axis to the left of
(-1,0) and are open-loop stable. Therefore, all these curves
do not encircle (-1,0) in either direction, and this set is a sub-
set of all stable closed-loop systems described by the Nyquist
stability criterion.
This revised stability criterion is a sufficiency condition
for stability. It is not a necessary condition, since a system
can have multiple phase crossover frequencies (some of
them with amplitude ratios larger than unity) and still be
stable. If a case arises that is not covered by the revised
criterion, then the Nyquist stability criterion should be
used for stability analysis.

1. Coughanowr, D.R.. Process Systems Analysis and Control, McGraw-
Hill, New York, NY (1991)
2. Coughanowr, D.R., and L.B. Koppel, Process Systems Analysis and
Control, McGraw-Hill. New York, NY (1965)
3. Ogunnaike, B.A., and W.H. Ray, Process Dynamics, Modeling, and
Control, Oxford University Press, New York, NY (1994)
4. Stephanopolous, G., Chemical Process Control, Prentice-Hall,
Englewood Cliffs, NJ (1984)
5. Seborg, D.E., T.F. Edgar, and D.A. Mellichamp, Process Dynamics
and Control, John Wiley & Sons, New York, NY (1989)
6. Skogestad, S., and I. Postlethwaite, Multivariable Feedback Control,
John Wiley & Sons, New York, NY (1997)
7. Smith, C.A., and A.B. Corripio, Principles and Practice ofAutomatic
Process Control, John Wiley & Sons, New York, NY (1985)
8. Marlin, T.E., Process Control: Designing Processes and Control Sys-
tems for Dynamic Performance, McGraw-Hill, New York, NY (1995)



Ba learning



Imperial College of Science, Technology, and Medicine

What is it that makes the student experience in en-
gineering so challenging? Although it is accepted
that the applied nature of any science is likely to
be a critical test of understanding concepts, it is particularly
true in regard to training engineers, where elements of the
key sciences are interwoven with mathematics and manage-
ment studies. This integration of disciplines is needed to pro-
vide a framework for professional practice so that, for ex-
ample, an engineer's insight into process design embraces
not only an overall understanding of scientific feasibility but
also the economic, safety, and controllability issues.
So, what constitutes effective training in undergraduate en-
gineering? It is easy to provide disparate courses covering an
identified curriculum, but effective training can be judged
only through students' overall mapping of the course and ul-
timately the holistic knowledge schema drawn by the stu-
dents. It is not then surprising that psychological and bio-
logical principles in memory and cognition should provide
some basis in the design, delivery, and evaluation of a course;
see, for example, the discussions of Haile1m on the educa-
tional implications of brain structure and function. While the
issue is complicated by individual student motivation for
learning, by using some of the basic psychological principles
that describe learning and motivation, some common prac-
tices for good teaching can be demonstrated.
To the experienced lecturer, such practice may be either
second nature or common sense, but the underlying psycho-
logical principles may not be apparent. Of course, the litera-
ture in chemical engineering has been rather thorough in de-
scribing good teaching methods, such as student-centered,
cooperative, and problem-based learning (see the following
section). But when a psychological background to memory
and learning is understood, these teaching methodologies can

* London SW7 2BY England

be better evaluated and applied by the new teacher, or indeed
even by a teacher who adheres to the traditionalist teaching
approach. In the following sections, a brief overview of these
principles is given and some exemplary teaching practices in
engineering education highlighted.

Through the centuries, human perception of the nature of
the mind has been constantly redefined to reflect the beliefs,
social norms, and even the fashions of the time. But recently,
the dominating perception has changed from one of a spiri-
tual entity to one with an analogy to computer operation. This
new perception may be most satisfying to engineers in that it
implies an underlying scientific basis for memory and learn-
ing as well as a logical and structured organization of the
human mind. For example, we can readily relate our own
thinking to recent developments in computer systems, such
as parallel processing, neural networks, and the interactive
separation of working and long-term memory facilities.
One such popular model of human memory is the multi-
store (or modal) model of Atkinson and Shiffrin,121 a descrip-
tion of which will be given later in this paper. But the impor-

Esat Alpay is Lecturer in Chemical Engineer-
ing at Imperial College, University of London.
He received his BSc from the University of Sur-
rey and his PhD from the University of Cam-
bridge. His research interests include gas sepa-
ration through adsorption processes, combined
(in-situ) separation and reaction processes,
and structured reactor engineering. As an
Education Development Coordinator, he has
ongoing and wide interests in undergraduate
and postgraduate training.

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

tance of memory, and indeed why psycholo-
gists consider its study so critical in under- The impi
standing human development and behavior, memory
will be addressed first. realized I
Why Memory? generic 4
The importance of memory can be real- that u
ized through its relation to three generic Wet
concepts'31 that underpin Western psychol- psyche
ogy: associationism, constructivism, and associa
rationalism. Associationism involves relat- construct
ing responses (R) to stimuli (S), which may ratio
be unconditioned, such as a reflex jerk to Associ
pain, or conditioned, such as emotional involve,
anxiety in anticipation of pain or, indeed, a
new learning challenge. The latter leads to sti
learning through the reinforcement of cer-
tain responses. The conditioned behavior e unco
may have originated from a previously en- such as
coded negative S-R set, but adapted to re- jerk to
late to different, and perhaps abstract (non- condition
obvious), stimuli. Such learning is also re- as em
ferred to as operant learning, i.e., behavior anxi
that is a consequence of reinforcement by anticil
pleasant or unpleasant experiences rather pain or,
than distinct stimuli, new I
Encoded memories of conditioned or un- chal
conditioned S-R sets are obviously impor-
tant in governing many everyday actions
and form the basis of the behavioral ap-
proaches to learning. But, while some simple forms of ani-
mal life rely solely on behavioral associations, humans are
also led through rational conduct. That is, we have the capac-
ity to logically argue the pros and cons of a situation and to
make connections with, or extrapolations from, previous
knowledge. Such rationalization could, for example, over-
come any inherent S-R sets through conscious recognition
and re-evaluation of S-R related behavior, so that in effect
there is continual development of existing knowledge.
This then reflects a constructivist view of learning; it is an
important basis of cognitive learning, i.e., use of reasoning,
planning, and problem solving to advance understanding. In
constructivist theories, learning is viewed as more than
adapted S-R behavior; it is also a synergetic assemblage of
cognitive operations, with perception and perceptual experi-
ences postulated to be derived from data systems within the
memory banks of the brain (see the discussion by Hailei'l on
the biological nature of such learning). An implication of the
constructivist view is that how well we learn will be depen-
dent on our existing knowledge, i.e., effective integration of
new material will rely on connections or mappings to exist-
ing knowledge. Thus, memory, whether in implicit (automatic
or associated) or episodic form, is a requisite for learning
Summer 2001

ortance of
y can be
through its
Sto three
ivism, and
s relating
nses to
vhich may
Sa reflex
pain, or
ned, such
ety in
nation of
*S-na fl n

u, u These issues must therefore be given due
n.g importance in an educational context. En-
ge' hancement of student self-esteem will be
considered in a future paper, but the human-
istic approach of Carl Rogers'5' may be par-
ticularly relevant here.16' For example, some specific quali-
ties of the teacher that are likely to promote high student self-
esteem include
Non-judgmental acceptance of the student
Genuineness; i.e., being a real person rather than
wearing a professional mask
Empathy; i.e., understanding and appreciating the
student's feelings; understanding the feelings behind
a student's words
High teacher self-esteem in the teaching itself
These qualities in a teacher are likely to promote a trusting
and communicative learning environment and a positive
teaching ethos. Some of these attributes have also been iden-
tified in chemical engineering education literature, e.g., the
influence of positive expectations of the lecturer on student
learning'" and "showing concern for the students as per-
sons.'"'8 It is interesting to note that in the psychotherapeutic
treatment of depression or anxiety, one approach has involved
the conscious recognition and reinterpretation of past memo-
ries, which are deemed to be inhibiting psychological devel-
opment (i.e., psychoanalytical methods), whereas other meth-
ods have concentrated on changes in behavior or thought to


and development and will govern our re-
sponses, or even feelings, toward daily oc-
Although it is clear that cognitive devel-
opment can only proceed from a knowledge
base, at any particular stage of development
self-concept will also arise from this base.
In other words, human perception of self is
related to knowledge of the specific physi-
cal and social environment and the students'
perception of their standing within this con-
text. In this sense, issues of low self-esteem
(where self-perception falls short of sought-
perception) arise from the memories and
thoughts of the individual. For example, poor
self-concept may be closely related to an
overly negative recollection of past experi-
ences, which may in turn be indicative of a
depressive state. Such depression, in turn, is
likely to be closely associated with anxiety
disorders.'41 Anxiety, depression, or low
self-esteem will, of course, influence stu-
dent motivation for learning, and may also
dictate the approach to learning, as will
be discussed below.

alleviate negative associations (i.e., cognitive-behavioral
methods). In both cases, reevaluation of memories plays an
important role in therapeutic treatment.

The Modeling of Memory
As mentioned above, the multi-store model of Atkinson-
Shiffrint12 has analogies to computer processing. Memory is
modeled as three storage blocks. The first is a high-capacity
sensory buffer in which the parallel input of visual, auditory,
and haptic (pertaining to the skin senses) information is re-
tained. The lifetime of such information is very short, but
enables the individual, for example, to perceive continuous
vision from discrete but frequent images. Attended informa-
tion is fed to a short-term store (STS), whereas unattended
information (such as background noise) is lost. In other words,
attention to sensory input results in information transfer from
the sensory buffer to STS memory.
Information in STS memory has been demonstrated to have
a relatively short lifetime of approximately 30 seconds, un-
less a deliberate attempt is made to retain it. The capacity of
STS is also small, in which typically 5 to 9 items of informa-
tion, such as numbers, names, or grouped items, can be re-
tained.[9] STS provides time for evaluation and processing
information before the information is appropriately encoded
into the long-term store (LTS).
In contrast with STS, LTS has a large storage capacity, but
with substantial limitations in the rate of input and retrieval.
The rehearsal of information within STS, or effective con-
nection or relation to existing LTS memory, will result in ef-
fective information transfer. While the structural features of
the memory store are considered as fixed, the control pro-
cesses governing, say, attention, rehearsal, coding, and re-
trieval can be learned and are flexible and variable across


Evidence for short-term and long-term store modes arises
from experimental observations in learning as well as from
neuropsychological experiences.[19 For example, free recall
of memories appears to consist of separate short- and long-
term components, and information coding within STS and
LTS memory appears to be of a different nature. Neuropsy-
chological evidence from brain-damaged patients also shows
that STS and LTS memory can be separately impaired, physi-
cally indicating there are two separate banks of memory.
Other clinical observations (in which, for example, prob-
lems in short-term memory do not necessarily result in long-
term learning problems) have led to memory theories involv-
ing a multicomponent STS and even unstructured short-term
and long-term memory processing. Likewise, previously held
views of distinct coding methods within the STS and LTS is
now recognized as an oversimplification of the storage pro-
cesses. Nevertheless, the simple multi-store model does yield
some practical insight into learning, as will be discussed
below. One notable extension of the simple multi-store
model involves the postulation of working memory, in
which the STS plays an important role in cognitive ac-
tivities by relating external information with recalled in-
formation from the LTS.

Approaches to Learning
Given the above description of memory, the relationship
between memory and learning can be considered and ap-
proaches to effective student learning proposed. In general
terms, effective learning can be described as

The encoding of information from STS memory into
existing knowledge schema within LTS memory
The effective use or recall of information within LTS


Bases associationism
conditioning to a stimulus
operant learning through reinforcement

* reasoning, thinking, problem solving, planning
* use of a body of knowledge (schema) for understanding
* assimilation of new information into existing schema,
i.e., construct understanding

Implications small (simple) operant learning tasks to leads to meaningful rather than rote learning
build up learning (i.e., gradual conditioning) need to take into account existing student ideas and
students need to master an operant task prior concepts
to moving to a more complicated task activating prior knowledge will expedite learning
behavioral interactionism possible, student may make inferences (correct or otherwise) to
e.g., through manipulation of the learning fit new knowledge into an existing schema
environment helpful to use knowledge anchors or bridges using
helpful to use topic maps for lessons, courses, and

* learning through interaction
* learning by guidance of skilled teacher or
peer to help internalization of thinking "tools"

* facilitator or peers to emphasize connections,
incite motivation, or structure task solution
* teachers act as models and learning "scaffolds"

Chemical Engineering Education

Principles of Behavioral, Cognitive, and Social Learning101'


The abstraction or processing of this knowledge in
postulating new understanding

The above definitions encompass the common student con-
ceptions of learning,"0"' namely, a quantitative increase in
knowledge, memorizing, acquiring facts and meth-
ods for subsequent use, the abstraction of mean-
ing, an interpretative process for understanding
reality, and developing as a person. In considering
how learning is achieved, the two general processes
mentioned above are applicable, i.e., behavioral
and cognitive learning.
A third process is attributed to social learning,
which pertains to the influences of the peer-group
or the teacher-student relationship, and encom-
passes the ideas of Vygotskyl251 on the socio-cul-
tural influences on cognitive development. In Table
1, key concepts and implications of these three processes are
summarized. Although of continued debate, cognitive and
social learning are generally viewed as most relevant to edu-
cational psychology. For some recent perspectives on the
behaviorist and cognitivist (constructivist) approaches to
learning, the articles of Wheldall and Glynnil2' and
Glasersfeld"3' are recommended.
While the dominance of a learning process may be situa-
tion dependent (and perhaps individual and maturation de-
pendent), it is not too surprising that several cross-process
mechanisms can be active during a learning situation. For
example, the first time a student is asked to derive a material
balance for a chemical reactor, the approach may rely on cog-
nitive processes that relate past knowledge to the current prob-
lem, planning a solution methodology, and interacting with
the lecturer for guidance from the base knowledge to the new
desired knowledge. Effective retention of this learned mate-
rial may rely on the STS rehearsal of an underlying prin-
ciple, such as mass in mass out = mass reacted + mass
accumulated, as well as connections to existing LTS compo-
nents, such as definitions of mass flux or the rate of reaction.
Of course, problems in learning can arise at any process point,
and the efficacy of the learning will strongly depend on stu-
dent motivation. Motivation does not necessarily refer to the
eagerness of a student in tackling a new learning challenge,
but defines the actual drivers for learning. Biggs and Moore 4'
for example, suggest four categories of motivation:

Instrumental such as reward and punishment
extrinsic to the task; e.g., fear of failure due to
financial or parental implications
Social the desire to please peers and lecturers
Achievement a concern for the self to enhance
position relative to others
Intrinsic an interest in the activity itself; e.g.,
curiosity-driven learning

In other theories of student motivation, specific orienta-
tions to learning are suggested. For example, these may in-
clude task orientation, where students concentrate on learn-
ing and gain pleasure from the progress in their learning (in-
trinsic and achievement motivation), ego-orientation, where

This paper [presents] an
overview of psychological bases in
student learning and [shows] them to be
complex interactions of social, cognitive,
and motivational issues, with
possible underlying
behavioral traits.

students are more concerned about their performance rela-
tive to others (social motivation), and work-avoidance orien-
tation, where students gain comfort from doing as little work
as possible. The third orientation could possibly arise from
some inherent (perhaps unconscious) and extrinsic factor.
Whatever the student motivation may be, it is likely to in-
fluence the learning approach. Indeed, most lecturers are fa-
miliar with the concepts of deep and surface approaches to
learning, and there is evidence to suggest that a student's in-
tention to learn can govern the particular approach. The
student's conception of learning will also be of importance
to the approach, however. For example, students who believe
that learning involves a quantitative increase in knowledge
will pay more attention to memorizing text than to the in-
tended content of the learning material. On the other hand,
students who adopt a deep approach to learning are likely to
view learning as an interpretative process for understanding.'"
Closely related to motivational issues, the concept of locus
of control (LC) has also been proposed in explaining deep
and surface approaches."l4 The LC defines the students' per-
ception as to how their learning is controlled. At one extreme,
students perceive themselves as being controlled by external
events (instrumental and social motivation), whereas at the
other extreme, the perception is one of internal control of
events (achievement and intrinsic motivation). Deep learn-
ing is likely to be favored by an internal locus and surface
learning by an external locus.
Even with a favorable learning locus, however, external
factors can lead to a surface approach in learning. For ex-
ample, surface learning could be promoted by assessment
criteria that are heavily biased toward factual reiteration, or
by poor enthusiasm by a lecturer; for students who are so-
cially motivated, either would signal a lack of importance for
deep understanding."1 An overview of the interrelations of
the above mentioned concepts of memory, motivation, and
learning is illustrated in Figure 1. Although the figure is based

Summer 2001

on the multi-store model mentioned above, the features of mo-
tivation and student approach are incorporated to emphasize
the student, teacher, and organizational influences on learning.
Although some insight into student learning and motiva-
tion can be gained through the various theoretical and re-
search contributions in the literature, what is it that makes
individuals amenable to certain types of education and not
others? Earlier in this article, the difficulty of engineering
education was attributed to its multi-subject and applied na-
ture. So the question could be rephrased as: What are the
underlying aspects of applied and cross-subject work that
favor certain types of learning styles? Here, issues of cogni-
tive styles arise, i.e., a student's habitual mode of perceiving,
thinking, and problem-solving, or the characteristic style in
which the cognitive tasks are approached or handled.
Several distinct cognitive styles have been identified by
researchers and are typically defined through polar dimen-
sions such as serialist-holistic, convergent-divergent, field
dependent-field independent, reflective-impulsive, and lev-
eling-sharpening."r61 Specific consideration will be given here
to the convergent-divergent and holistic-serialist styles since
they have the most obvious implications in engineering edu-
cation and practice.
The convergent-divergent style considers students to have
a preference for problems that either have a well-defined so-
lution (convergent problems) or that are open-ended with
several possible solutions (divergent problems). Convergent
thinkers are likely to be attracted to courses in mathematics
and the physical sciences, even early in their education,
whereas divergent thinkers are likely to pursue studies in the

external motivators:
- teacher lack of interest
- lack of responsibility
- work overload
- poor or no feedback
- assessment of recall /
trivial procedures

external motivators:
- stimulating teaching
- student responsibility
- clear objectives
- reasonable work load
- assessment for active /
long-term engagement

poor recall

: surface leamin
focus on signs
poor connection
no underlying


arts or in the social or biological sciences.
Engineering undergraduates are thus expected to be con-
vergent thinkers, and this is perhaps reinforced by elements
of the coursework and the exam-based assessment. Often,
however, the actual integration of various convergent-type
procedures for, say, engineering design, is carried out with
limited physical or chemical data, as well as loose or very
general design specifications. Many engineering undergradu-
ates do find it difficult to tackle such problems, and will typi-
cally narrow the decision-space, often using superficial cri-
teria to converge on a solution. Thus, successful engineering
education requires some training in divergent thinking so as to
enable a flexible (creative) approach to knowledge application.
The second relevant cognitive style considers a student to
have a preference for either overall (holistic) or step-by-step
(serial) learning. Holistic learning arises when connections
are sought to previous knowledge and other subject matter in
order to attain comprehensive understanding (deep learning).
In serialist learning, the student learns the details of specific
operations and procedures of the subject field and makes use
of serial strategies to logically connect these operations in
addressing a problem.
Of course, many versatile learners will adopt both holistic
and serialist styles as and when needed. In the context of
engineering, a versatile style is necessary, but there is often a
student tendency to adopt a serialist approach whereby a con-
vergent style can be exercised at each task stage. Although
such a tendency may be adequate for some problems, the
outcome can be rather deterministic and therefore possibly
lacking in an awareness of global issues. As for the case of

......................................... ...........F ig u r e 1 .
Overview of
memory, motivation,
ng: and learning,
incorporating the
n multi-store

and Shiffrin.'2]

- internal motivation or
learning orientation


deep learning:
underlying concepts

Chemical Engineering Education

divergent thinking, a student awareness of holistic approaches
to learning is needed. Finally, it is interesting to note that a
holistic and divergent approach is likely to lead to a student
commitment for the critical evaluation of material as well
as the appreciation of multiple viewpoints. Thus, such an
approach is consistent with the higher levels of attain-
ment of Perry's Model of Intellectual Development as
described by Felder.[J7
Finally, learning styles that reflect the particular abilities
of the student should also be appreciated. For example,
Gardner["81 postulates the theory of multiple intelligence that
include linguistic, logical-mathematical, spatial, bodily-ki-
nesthetic, musical, and inter- and intra-personal intelligence.
An educational implication here is that different students may
learn in different ways, with some, for example, preferring
visual material such as charts, graphs, and illustrations, and
others preferring auditory material such as listening to an ar-
ticulate lecturer, participating in class discussions, or using
mnemonics to aid memory.


The above discussions on the theories and issues of stu-
dent learning demonstrate the complexity of the process.
Aspects of student motivation, self-concept, perceived locus
of control, and cognitive styles, as well as differences in the
knowledge schema of individuals, suggests that the learning
experience should be student-centered. While it is easy to
relate to the multitude of ideas on learning, practical imple-
mentations for effective learning may not be so obvious.
The approach in the following sections will be to summa-
rize some key principles from the above discussions and high-
light teaching practices that can promote favorable learning
situations. Many of the recommendations arise from the teach-
ing experiences gained within the Department of Chemi-
cal Engineering and Chemical Technology at Imperial
College. Some reference to the literature in chemical en-
gineering is also given to further illustrate the consistency
of good teaching practice with the psychological argu-
ments presented above.
While several of the recommendations below imply a con-
ventional lecture-based teaching approach, it is important to
recognize that formal lectures themselves have been identi-
fied as having a negative influence on deep learning."91 Nev-
ertheless, Giralt, et al., [20 describe several factors that exist
in the current higher-education climate that favor the lecture-
based approach, such as the significant effort and time needed
by the teaching staff to implement, for example, student-cen-
tered teaching strategies, appropriate staff motivation when
ultimate promotional benefits may be perceived to be attained
through research rather than teaching efforts, and the logis-
tics of alternative (e.g., small-group) teaching arrangements
Summer 2001

when student-to-staff ratios are relatively high. Even so, there
is scope for improvement on traditional lecture styles, and
indeed to complement existing lecture-based learning through
pragmatic programs.

Cognitive Aspects

Students should be encouraged to develop cogni-
tive skills and effective cognitive learning should
be facilitated.

E[ The knowledge base and concepts held by individual
students require adequate assessment, particularly in the
foundation years of study. Prior-knowledge question-
naires should be supported by review classes before the
introduction of new material. This is especially true for
mathematics courses because of the wide variation in
school-level mathematics curricula among national and
international students. Assumed prior knowledge should
also be clearly indicated to students for each lecture
course, and if necessary, students should be provided
with adequate time and references to prepare for the
The use of graded problem sheets soon after the start of
a course could be used to test the students' basic under-
standing of assumed concepts, such as the meaning of
equilibria and kinetics in reaction engineering or mass
conservation in process analysis. The teaching of courses
by lecturers who have a similar background and focus
as the students is also beneficial, such as the teaching of
numerical methods and physical chemistry by relevant
engineering staff. Similarly, managerial courses should
be placed in the context of engineering by means of in-
dustrial case studies.
For all courses, an element of feedback is required to
enable students to evaluate acquired concepts. Where
possible, detailed feedback on examinations could be
given and compulsory post-exam tutorials arranged for
students who do not achieve a minimum standard. Like-
wise, supplied with an outline solution set, all students
could be asked to re-answer unsatisfactory examination
questions, irrespective of their overall exam mark. Such
a procedure may help to alleviate the carryover of mis-
conceptions and ensure adequate background knowledge
for follow-on courses.
As a direct means of prior knowledge assessment, Felder
and Brent211] also suggest that lecturers new to a subject
could ask students to anonymously compile lists on what
is known about the course content as well as any spe-
cific questions they may have about the course.
L Students should be made aware of connections within
and between lecture courses. Where possible, material
should be related to practical knowledge through use of

analogies or physical (visual) models. Furthermore, stu-
dents should be given opportunities to infer connections
through project work and cross-subject examinations.
For example, within a particular course, a coursework
element that integrates intended learning outcomes is
beneficial. On a wider scale, yearly mastery exams that
assess general (integrative) and essential engineering
knowledge are useful, as are final-year projects in, for
example, overall plant, process, or equipment design.
Students could also be asked on occasion to draw con-
cept-maps to link knowledge and skills attained among
different courses.E8] Particular care is needed in modular
degree programs to ensure course coherency and an ef-
fective student perception of course connectivity.
El Projects or learning tasks requiring a holistic and/or a
divergent approach could be supplemented by work-
shops or tutorials that highlight subject inter-relation-
ships and the need for an open-structure solution ap-
proach. For example, a design exercise could be pre-
ceded by a session that enables students to explore the
complexities involved in defining a successful design
and perhaps emphasize the iterative rather than the lin-
ear (serial) approach that is needed.
Solution methodologies can sometimes be devised by
working backward from the design specifications so as
to generate an expanding, branched outline of pertinent
issues, highlighting any interrelationships between them.
Particularly relevant here are the creative problem-solv-
ing modules in engineering design described by
Mackenzie, et al.,"22' e.g., problem-statement definition
techniques, brainstorming, and potential-problem analy-
sis. In addition to core engineering courses, students
should be encouraged to attend courses or to carry out
activities that have a natural element of creativity and di-
vergence, such as creative writing, art, and philosophy.
El For courses of particular conceptual difficulty, greater
effort is needed on course design and delivery for dem-
onstrating physical relevance. In such cases, problem-
based learning, where theory is developed during the
course of the problem solution, may be particularly use-
ful. Problem-based learning here is congruent with the
definition of Woods,18] where the student learns because
of a need to solve a real problem, as in the cases of
design projects and interpretating laboratory results.
Problems with real-world connections are likely to in-
cite student interest.121 Also of benefit is dividing some
general courses, such as thermodynamics or engineer-
ing mathematics, and integrating the relevant compo-
nents into practical or applied course material.
Where possible, visual imagery may be particularly ben-
eficial in illustrating underlying theory or concepts. For
example, the McCabe-Thiele diagram for teaching dis-
tillation-column design (where the memory of the stair-

case effectively activates memories of material balances
and VLE relationships) has proven effective for gen-
erations of students.[7] Such graphical and analytical
methods should be sought where possible, particularly
in foundation or introductory courses.

Social Aspects
Social learning should be encouraged through op-
portunities ofpeer learning and guided participa-
EN Students should have opportunities to demonstrate their
own understanding to peers and to discuss underlying
issues of course material. This could be through stu-
dent-led seminars and small-group tutorials.
EL Students should have opportunity to share and apply
understanding. The use of design projects throughout
the degree program, as well as group-based laborato-
ries, pilot plants, or field work, are of particular value.
Another interesting approach, described by Newell,'23
involves peer-review of undergraduate laboratory re-
ports for improving the oral and written communica-
tion skills of both the reviewer and the reviewee. But
in any approach, care is needed to avoid undue student
competition, which could otherwise cause student anxi-
ety. Furthermore, effective student work groups are
generally characterized by strong interpersonal inter-
actions within the group and with the teaching staff."'7
Team development workshops, or away-days, as well
as dedicated team facilitation by teaching staff, are
likely to help promote an effective team approach to
group work.
E Teaching staff (lecturers and postgraduate demonstra-
tors) should have appropriate training in group super-
vision so as to act as effective learning scaffolds and
bridges. This can be achieved through guided experi-
ential programs and dedicated support groups on teach-
ing.[241 The teachers themselves need to be of high self-
esteem in their teaching and to demonstrate a genuine
care (empathy) for the students.

Motivational and Behavioral Aspects
Appropriate mechanisms should be in place to ad-
dress student personal developmental and motiva-
tional issues. Inherent behavioral characteristics
that are inhibiting favorable motivation for learn-
ing should be recognized and addressed.
LE Personal (non-academic) tutorials should be frequent
and should promote adequate recognition of, and guid-
ance on, counseling matters. This can be achieved
through close collaboration with an appropriately
trained senior tutor or a college-based counselor.
Chemical Engineering Education

[ To promote self-development and positive motivational
drives, students should be made aware of individual
and social psychology principles. This could be
achieved indirectly through courses on industrial and
organizational psychology, or through field courses in
team development. Alternatively, to prevent increas-
ing the students' workload, programs in these areas
could be incorporated into existing personal tutorial
systems. Such efforts may be helpful in promoting meta-
learning strategies among the students."'21
In order to discourage a surface approach to learning,
as well as instrumental motivations, overwork should
be avoided and assessment criteria clearly defined; see
also the discussions of Felder and Brent12[ on active
learning vs. covering the syllabus. Likewise, to main-
tain the favorable stance of an internal-locus of con-
trol, student responsibility and choice should be made
apparent. Students should be made aware of core areas
for examination in a course and the expected level of
competency demonstrated by examples. A choice of
areas for student specialization should also be indicated
and evaluated through an option-based component in
the exam and/or a coursework element.
E[ It should be recognized that students in tutorial and
project groups of mixed or similar academic ability are
still likely to vary widely in their motivational drive.
Here, careful academic counseling to illustrate indi-
vidual components on the group dynamics, and to guide
the peer learning of poorly motivated individuals, may
be beneficial. In practice, academic supervisors and
counselors could work together on some key group-
based projects during the foundation years of study. This
could also offer an opportunity to observe students for
inherent or inhibiting behavioral traits that may be in-
dicative of, for example, social phobia or anxiety.

This paper has presented an overview of psychological
bases in student learning and has shown them to be complex
interactions of social, cognitive, and motivational issues, with
possible underlying behavioral traits. Most established and
successful courses in engineering have, in many ways, suc-
cessfully addressed cognitive and social learning issues
through design and problem-based learning. But greater rec-
ognition is needed on the constructivist view of education in
assessing prior knowledge and concepts and integrating course
Issues concerning student motivation, cognitive style and
personal development require attention that has, to date, only
been addressed superficially through inappropriately trained
personal tutors or a facile (non-proactive) college counseling
service. Improvements can also be made in controlling the
peer learning environment through, for example, detailed
Summer 2001

consideration of the technical and social qualities of the indi-
viduals. Indeed, there may be an educational benefit in hav-
ing mixed motivational groups that are carefully monitored
and facilitated by an experienced supervisor, or on occasion,
by academic and counseling supervisors.


1. Haile, J.M.. "Toward Technical Understanding: Part 1. Brain Struc-
ture and Function." Chem. Eng. Ed., 31(3), 152 (1997)
2. Atkinson. R.C., and R.M. Shiffrin, "Human Memory: A Proposed
System and Its Control Processes." in The Psychology of Learning
and Motivation: Advances in Research and Theory; K.W. Spence. ed.,
Academic Press. New York. NY. Vol. 2. p. 89, (1968)
3. Richardson, K., Understanding Psychology, Open University Press,
Milton Keynes (1989)
4. Tyrer, P.. Anxiety: A Multidisciplinary ReviewM Imperial College Press,
London, England (1999)
5. Rogers, C.R., On Becoming a Person, Houghton Mifflin, Boston, MA
6. Lawrence D., Enhancing Self-Esteem in the Classroom, 2nd ed., PCP
Ltd.. London, England (1996)
7. Wankat. PC., "What Works: A Quick Guide to Learning Principles,"
Chem. Eng. Ed., 27(2). 120 (1993)
8. Woods. D.R.. "Three Trends in Teaching and Learning," Chem. Eng.
Ed., 32(4). 296 (1998)
9. Baddeley. A., Human Memory: Theory and Practice, Psychology Press
Ltd., Hove (1999)
10. Ireson J., and D. Male, Psychology of Education I, University of Lon-
don Press (1999)
11. Marton. E, D. Hounsell. and N. Entwistle, eds, The Experience of
Learning, Scottish Academic Press (1997)
12. Wheldall. K., and T. Glynn. "Contingencies in Context: A Behavioral
Interactionist Perspective in Education," Ed. Psychology, 8, 5 (1988)
13. Glasersfeld. E.V.. "Learning as a Constructive Activity," in Develop-
ments in Learning and Assessent,. P. Murphy and B. Moon, eds.,
Hodder and Stoughton. Open University Press. London, England,
14. Biggs, J.. and P.J. Moore, The Process of Learning, Prentice Hall,
Englewood Cliffs, NJ (1993)
15. Ramsden. P., Learning to Teach in Higher Education, Routledge, Lon-
don, England (1992)
16. Sternberg. R.J.. and L.F. Zhang. eds. Perspective on Thinking, Learn-
ing, and Cognitive Styles, Lawrence Erlbaum Associates, Inc., Lon-
don, England (2001)
17. Felder, R.M., "Meet Your Students: 7. Dave, Martha. and Roberto,"
Chem. Eng. Ed., 31(2), 106 (1997)
18. Gardner. H.. Frames of Mind: The Theory of Multiple Intelligences,
Fontana, London, England (1993)
19. Entwistle. N., Styles of Learning and Teaching: An Integrative Out-
line ofEducational Psychology for Students, Teachers, and Lecturers,
David Fulton Publishers, London, England (1996)
20. Giralt, F, J. Herrero, M. Medir, F.X. Grau, and J.R. Alabart, "How to
Involve Faculty in Effective Teaching," Chem. Eng. Ed., 33(3), 244
21. Felder, R.M.. and R. Brent. "Getting Started," Chem. Eng. Ed., 29(3),
22. Mackenzie. J.G.. R.M. Allen, W.B. Earl, and I.A. Gilmour, "Teaching
Creative Problem-Solving Skills in Engineering Design." Chem. Eng.
Ed.. 33(2), 150 (1999)
23. Newell, J.A., "Using Peer Review in the Undergraduate Laboratory,"
Chem. Eng. Ed., 32(3), 194 (1998)
24. Alpay, E., and M.A. Mendes-Tatsis, "Postgraduate Training in Stu-
dent Learning and Teaching," European J. Eng. Ed., 25(1), 83 (2000)
25. Vygotsky, L.S., Thought and Language, MIT Press, Cambridge, MA
(1962) 0

MR% = laboratory


A Classroom Demonstration

Northeastern University Boston, MA

his demonstration is ideal for a presentation topic re
lated to flashback and laminar flames. Fires require
four components (the fire tetrahedron):"' fuel, an oxi-
dant, an ignition source, and a chain reaction. A good refer-
ence source on flammability is Crowl and Louvar's'21 Chap-
ter 6.
Flammability of liquids is characterized by their flash point
and their upper and lower flammability limits. The flash point
is the minimum fluid temperature at which the vapor over the
liquid, in air, can be ignited. The flash point is easily mea-
sured and easily understood; it is used by the NFPA and by
fire departments across the USA for hazard ratings of flam-
mables. Any fuel with a flash point below 37.7C (1000F) is
termed a "Class 1" flammable. Both gasoline and ethyl alco-
hol are Class 1 flammables with flash points of -430C and
13'C, respectively. Note that the vapor over gasoline is flam-
mable at temperatures far lower than alcohol. At ordinary
ambient temperatures, however, the equilibrium vapor over
alcohol can be far more hazardous than the equilibrium va-
por over gasoline. This can be demonstrated in the manner
described below.

Required equipment:
Two, one gallon "F"-style metal cans, with closures. (One
gallon plastic containers may be used; see details below.)
0 A small supply (100 ml) of reasonably fresh gasoline.
(Coleman fluid works well.)
> A similar supply of denatured alcohol. The denatured
alcohol sold in hardware stores is suitable.
To carry out the experiment, add about 10 ml of gasoline to
one of the cans and a similar amount of alcohol to the other.
Close the cans, shake well, and wait about 60 minutes at
ambient temperature to permit vapor equilibration. Remove
the cap from the gasoline-containing can and pass a small
ignition source (match flame, lighter flame, etc.) over the
opening. A small, stable laminar flame will appear and will

continue to burn quietly for many minutes.
The experimentalist should wear a face shield, a long-sleeve
shirt, and leather gloves when igniting the cans. Also, a por-
table visible shield can be placed in front of the cans for fur-
ther protection.
Next, remove the cap from the alcohol-containing can and
present a similar flame at the opening, taking care to keep
fingers, etc., well away from the opening. A flashback will
occur, accompanied by an angry noise (a "whoosh").
CAUTION: Make sure that the can opening (the spout) is
unobstructed. Ignition of alcohol vapor in an unvented can
or through a small opening may cause the can to burst with
considerable violence. Cans with openings less than 2 cm in
diameter can deform, and cans with openings less than 1.3
cm in diameter can burst."1
The demonstration can also be carried out in a plastic 1-
gallon juice or milk high density polyethylene (HDPE) con-
tainer (spout opening of about 3 cm) for an even more visual
effect. In our testing, the HDPE didn't ignite during the dem-
onstration (recall the fire tetrahedron requirements). In fact,

Ed Shanley is a retired Vice President ofArthur
D. Little, Inc., in Cambridge, Massachusetts,
who specializes in chemical process safety.

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

Ronald Willey is a Professor of Chemical
Engineering at Northeastern University. He
is a member of SACHE (Safety and Chemi-
cal Engineering Education) and a director in
the AIChE Safety and Health Division.


The flash point is the minimum fluid temperature at which the vapor over the liquid,
in air, can be ignited... [it] is easily measured and easily understood;

it took a direct flame from a propane torch to get the HDPE
to ignite, and then only a mild flame ensued.
Differences in the ignition behavior of equilibrium alcohol
vapor and gasoline vapor follow from differences in flamma-
bility limits. Vapor mixtures containing just the right propor-
tion of fuel to combine with the available oxygen, so-called
oxygen-balanced mixtures, are obviously most energetic.
Mixtures on either side of the oxygen balance may also be
flammable. The range of compositions on either side of the
oxygen balance that will support combustion is characterized
by a lower flammability limit (LFL) and an upper flamma-
bility limit (UFL). It is conventional to define these limits in
terms of volume percent of "fuel" in the vapor. For example,
handbook data indicate that gasoline has an LFL of about 1.3
volume % and a UFL of about 7.6 volume %. Ethyl alcohol
has a LFL of about 3.3 volume % and a UFL of about 19
volume %.12J
We are now in a position to understand the difference in
flammability of gasoline and alcohol vapor as demonstrated
in the experiment. The equilibrium vapor pressure over gaso-
line at common ambient temperatures is quite high. For ex-
ample, at 21 C the vapor pressure is about 0.46 bar. Accord-
ingly, gasoline vapor concentration above the liquid is about

Figure 1. Photograph of demonstration. Peak jet flame
appeared about 400 ms after the butane lighter
flame was brought near the spout.
(Oxford Lasers, Inc., Littleton, MA, is gratefully acknowledged
for assistance in obtaining this photograph.)
Summer 2001

46 volume % at this temperature, far above the 7.6 % UFL of
gasoline. (Those who have tried to start a "flooded" gasoline
engine have experienced the UFL phenomenon). If a gaso-
line container is open to the atmosphere at its spout, there
will be a diffusion region containing a flammable air/gaso-
line mixture. Ignition will lead to a fire front in the form of a
calm laminar flame, as observed in the demonstration. In
contrast, the equilibrium vapor over ethyl alcohol at 21C
contains about 5.9 volume % alcohol (vapor pressure of 0.059
bar), well within the flammable limits and very near the sto-
ichiometric ratio for complete combustion to CO, and H,O
(5.7% alcohol in air). In a closed container, or one with a
restricted opening, the mixture inside the container is flam-
mable, as observed.
One may ask under what conditions will gasoline-can flash-
back occur? This can happen if the can is equilibrated at tem-
peratures well below 0C, something not easily done in the class-
room. Another possibility, confirmed by our testing, is to re-
duce the quantity of gasoline to match the oxygen balance in-
side the can. This quantity is around 1 ml, well below the 10 ml
recommended. Thus, shaking the can in the classroom before-
hand and listening to the gasoline swirl around inside ensures
that enough gasoline is present to prevent flashback.
Professors can designate as homework an assignment that
consists of looking up the flammable properties of gasoline
and ethanol (denatured alcohol). Later, the students can write
a short essay as to why the results are what they observed.
Other assignments could include the determination of the
vapor composition in a confined container via the Antoine
equation and comparing it to the flammability limits of the
materials. Or alternatively, using the published LFL, stoichio-
metric (oxygen balance) amounts and UFL, students can cal-
culate (via the Antoine equation) the temperatures that match
the vapor pressure/composition given by these three quanti-
ties for each material.

The authors gratefully acknowledge Mr. Richard Slagle,
Oxford Lasers, Inc., Littleton, MA, for assistance in taking
the photograph shown in Figure 1.

1. See, for example,
2. Crowl. D.A., and J.F. Louvar. Chemical Process Safety, Prentice Hall,
Engelwood Cliffs. NJ (1990)
3. Shanley, E.S., "The Safety of Small Containers for Flammable Flu-
ids," J. Chem. Edi., 65. A6 (1988) 1

W laboratory




Michigan Technological University Houghton, MI 49931-1295

he environmental consequences of chemical produc-
tion and use are becoming more important to chemi-
cal engineers. Increasingly stringent environmental
regulations are motivating chemical manufacturers to take a
proactive approach to environmental management. Often this
takes the form of preventing pollution and reducing health
and environmental risks of chemical production at the source,
rather than relying exclusively on treatment of pollution at
the end of the pipe.
This new approach will be more effective if chemical en-
gineers increase their environmental literacy, especially in
areas that chemical manufacturing has a dominant influence.
Some areas of great environmental concern are natural re-
source depletion, the quality of air, water, and soil, hazard-
ous waste management, toxic chemical use reduction, and
cleanup of prior soil and groundwater contamination. Many
of the environmental aspects of chemical manufacturing are
being addressed by the member companies of the American
Chemistry Council (formerly the Chemical Manufacturers
Association) in one or more of the six Codes of Management
This article presents a set of experiments, both laboratory
and numerical, to aid engineering students in understanding
pollutant transport in groundwater, which is a topic of great
concern for water quality and resource conservation. The
transport of both organic chemicals and biocolloids is dis-
cussed, the procedures and laboratory equipment are fully
explained, and the theory needed to interpret the experimen-
tal results is developed. Representative results from two col-
umn transport experiments and one numerical experiment are
presented, and interpretation of their practical importance to

* Current Address: Bayer Corporation, Clayton, NC 27520

chemical engineers is given. Further information on ground-
water transport of pollutants can be found in a number of
excellent texts.12-4]


Shaker table; 250-ml side-arm flasks; visible spectro-
photometer (Model D21, Milton Roy); Marathon 22K
centrifuge (Fisher Scientific); 1-inch glass chromatogra-
phy column (Adjusta-Chrom, Ace Glass Inc.); metering
pump (model QG 150, Fluid Metering Inc.); personal
computer; Labview software

Experimental Procedures
Sand Preparation The subsurface material used for these
experiments is a coarse-grain silica sand (Agsco Corp.) hav-

David R. Shonnard is Associate Professor of
Chemical Engineering at Michigan Technologi-
cal University. He received his BS from the Uni-
versity of Nevada-Reno in 1983 and his MS
(1985) and PhD (1991) from the University of
California, Davis, all in Chemical Engineering.
His research and teaching interests include en-
vironmentally conscious design of chemicalpro-
cesses and bioremediation of environmental

Prasanna A. Deshpande received his BS from
the University of Bombay (UDCT) in 1992, his
MS from the Indian Institute of Technology-
Bombay, in 1994, and his PhD from Michigan
Technological University in 1999, all in Chemi-
cal Engineering. He is currently a Sr. Process
Engine aterat layer Corporation. His research and
professional interests include blood plasma frac-
tionation, bioprocessing, fermentation, andpro-
cess modeling.

Copyright ChE Division of ASEE 2001

Chemical Engineering Education

This article presents a set of experiments, both laboratory and numerical, to aid engineering
students in understanding pollutant transport in groundwater ... The transport of both
organic chemicals and biocolloids is discussed, the procedures and laboratory
equipment are fully explained, and the theory needed to interpret
the experimental results is developed.

ing a weighted mean diameter of 740 microns. For each col-
umn experiment, approximately 250 g of sand is cleansed of
residual clay particles, thus avoiding interference with sub-
sequent absorbance measurements. This is accomplished by
first shaking it multiple times, using distilled water in a large
bottle (1/2 gallon of water in a 1-gallon bottle) followed by
decanting until the turbidity (measured as absorbance at 500
nm) drops below 0.01. The sand is allowed to soak overnight
in a phosphate buffer salt (0.01 M PBS) solution containing
1.62 g/L NaHPO, .7 H,O and 0.53 g/L KH,PO4 (pH=7.0). It
is again shaken in the bottle, using the 0.01 M PBS, until the
turbidity level drops down to an absorbance of around 0.005.

Porous Media Properties The bulk density of the sand
can be measured using a 250-ml glass graduated cylinder and
a knowledge of the sand particle density (P,=2.6 g sand/cm3
sand solids). The dry cylinder is weighed empty. Approxi-
mately 50 to 100 g of dry sand is added and gently tapped to
settle the sand particles. The weight of the cylinder plus sand
is taken and the sand volume is noted. The ratio of sand weight
to volume is the bulk density (Pb) of the sand. The porosity
(n) of the sand is obtained using the equation

Bacterial Peristaltic Salt solution collection
suspension pump

Computer data

Figure 1. Schematic diagram of column transport
apparatus for collecting breakthrough data.
Summer 2001

n = l-b /Pp

Column Assembly The column used in the studies is a
glass chromatography column (Adjusta-Chrom, Ace Glass
Inc.) with an internal diameter of 2.5 cm. The column is closed
at both ends using O-rings and Teflon plungers. The inlet
plunger is connected at the bottom. Air is flushed out of the
inlet tubing and the plunger using the 0.01 M PBS before
packing the column with sand. Some solution is allowed to
stand in the column base. A slurry of pre-cleaned sand in
0.01 M PBS is transferred to the column using a pipette gun
(Pipette Aid) and 25-ml disposable pipette and is discharged
into the column with the tip of the pipette below the liquid
level in the column. This procedure is an effective way of
packing the sand in the column without trapping any air. The
column is packed to a settled bed height of approximately 20
cm. The outlet plunger is then connected at the top of the

Column Operation A schematic diagram of the experi-
mental column apparatus is shown in Figure 1. An adjustable
flow pump (model QG 150, Fluid Metering Inc.) delivers the
solutions to the inlet of the column. Two visible spectropho-
tometers (model D21, Milton Roy), each containing a 0.5 ml
flow-through absorbance cell (Fisher Scientific), continuously
measure the solution concentration of a non-binding tracer
(phenol red) or of the colloid (a soil gram-negative bacteria)
at the column inlet and outlet. At readings less than 0.8, the
absorbance is found to be linearly related to the solution con-
centration. A personal computer running the Labview data
acquisition software acquires absorbance data from the spec-
trophotometers at 30-second intervals through two serial con-
nections. (The data acquisition program can be obtained from
the authors by request at .) An alterna-
tive configuration for the apparatus would entail collecting
small volumes (2 to 5 ml) of column effluent manually or
using a fraction collector (Gilson Medical Electronics Model
FC80 Microfraction Collector), which would be analyzed
subsequently in the spectrophotometer.
Before the experiment is started, the 0.01 M PBS is passed
through the column for one hour at a flow rate of about 5 ml/
min. This is sufficient to stabilize the pH of the sand and to
flush out residual fine particles. The phenol red solution (20
to 50 gM in 0.01 M PBS, pH 7) is then pumped until 3 pore

volumes is introduced to the column, followed by roughly 2
pore volumes of 0.01 M PBS. Immediately thereafter, the
bacteria solution (in 0.01 M PBS) is pumped for an addi-
tional 3 pore volumes, again followed by 2 pore volumes of
the PBS. The 0.01 M PBS is sufficiently high in ionic strength
to facilitate attachment of the bacteria to the solids.'5'
Cost and ordering information for components used in these
experiments can be found at SubSurface/subsurface.html>. Total cost is about $5,000 with
instrumentation but excluding cell-culturing and data-acqui-
sition equipment. With suitable modification, this procedure
could be used in adsorption studies.

Growth and Harvesting of Bacteria
The microorganism used in these experiments is Pseudomo-
nasfluorescens UPER-1, a gram-negative soil bacterium. The
cell line is maintained by transferring cell samples (1 ml) at
two-day intervals into autoclaved 250-ml side-arm flasks
containing 100 ml of autoclaved (121'C for 15 min) basal
salt media (BSM) supplemented by glycerol (1.5 vol%, the
carbon source). The flask is then placed on a shaker table and
held at a constant temperature of 320C and at 200 rpm. BSM
is comprised of
4.0 g/L K2HPO4
4.0 g/L Na,HPO4
2.0 g/L (NH )2SO4
0.001 g/L CaC1,
0.001 g/L FeSO4 7 HO
0.2 g/L MgSO4 7 H20
and its pH is adjusted at 7.0 using 12 N HC1. Immediately
prior to conducting the cell transport experiment, the side-
arm flask cultures are allowed to grow until that absorbance
of the culture (A) is between 1 and 2, as measured by insert-
ing the side-arm into a visible spectrophotometer (Model D21,
Milton Roy) at a wavelength of 500 nm. This corresponds to
a cell concentration of approximately 2 to 3 x 10' cells/ml,
obtained by using the equation

X(cells/ ml) = (l x09)(A00) for 0 A500 < 0.8 (2)

At the harvesting stage, the bacteria are centrifuged using
a Marathon 22-K centrifuge (Fisher Scientific) at 11,000 rpm
(13,900 x gravity) for five minutes. The cells are resuspended
using vigorous pipetting in roughly 40 ml of deionized water
and spun down one more time to wash the media off the cells.
The cell pellet is suspended a final time to a total volume of
200 or 250 ml of 0.01 M PBS in a glass beaker (Aso approxi-
mately 0.35). This suspension is constantly stirred to keep
the cells well mixed prior to being used in sand column trans-
port experiments.
If conducting bacterial transport experiments appears too
much of a challenge, alternative abiotic systems can be used,

including polystyrene latex particles (Interfacial Dynamics
Corp., Portland, Oregon), kaolinite powder (VWR Scientific),
anatase colloids (Degussa Corp.), boehmite colloids (Vista
Chemical Co.), and iron oxide (hematite colloids) synthe-
sized from ferric hydroxide gel.

Analysis ofData from Column Breakthrough Experiments
The column absorbance data must be manipulated in order
to calculate the relative concentration, CR = C/Co, where C is
the concentration at the column inlet and C is the outlet con-
centration. CR is plotted against the number of pore volumes
pumped into the column after initiating the experiment, cal-
culated by taking the ratio of the volume pumped (Qt) di-
vided by the pore volume (ntAn). Q is the volumetric flow
rate (cm3/sec), t is the time after initiating the experiment, A
is the cross-sectional area of the column (cm2), and n is the
porosity. Because the detectors are not located exactly at the
column inlet and outlet, an amount of time equal to the non-
column system volume divided by the flow rate must be sub-
tracted from each time value. The non-column system vol-
ume can be calculated by knowing the inlet and outlet tubing
lengths and diameters.


Phenol Red (A Non-Sorbing Tracer)

The isothermal one-dimensional continuity equation that
describes a balance between accumulation of dissolved com-
pound and the difference between transport in and out by
convection and by mechanical dispersion is

ac ac +D (3)
at x- x +L ax2
where C is a volume-averaged concentration (mg/cm3 of so-
lution) based on the liquid phase,[61 vx is the interstitial ve-
locity (v,=Q/(An) cm/sec), t is time, x is the coordinate in
the direction of flow, and DL is the longitudinal dispersion
coefficient (cm2/sec). This equation can be integrated subject
to the following conditions:

Initial Condition
Boundary Conditions

C(x,0) = 0 for all x
C(0,t) = C for all t20
C(o,t) = 0 for all t>0

to obtain the relative concentration (CR = C/Co) at the column

C I LL-v t v L L+v t
C = = L erfc -v + exp -- erfc / (4)
R Co 2 22 Xt Dj 2Dt1
where erf = complimentary error function= -erf, and where
where erfc = complimentary error function = 1-erf, and where

Chemical Engineering Education

the error function acting on an argument "a" is defined as

erf =- e-n2 dq (5)

The equation for CR can be simplified when transport by bulk
flow is large in comparison to dispersion transport, i.e., when
the column Peclet number is large (Pe = vxL / DL > 100)

1 L- vxt
C = erfc (6)
R 2 V2DL t

With the definition of U = vxt / L as the number of pore
volumes pumped into the column, we can recast Eq. (6) as

S0.5 erf(Pe)'_/2(l U) (7)
C = 0.5 erfc 2 U2 (7)

Since the erfc(a) is normally distributed with respect to an
argument, a, a plot of C/Co versus (U- 1)/U"' on linear prob-
ability paper will be a straight line whose slope is related to
the dispersion coefficient, DL, or to a close approximation 21
=( vL 2
DL 0 .84J0.16) (8)

where J0 84=(U- 1)/U when C/C =0.84, and J,,,=(U-1)/Um"
when C/C =0.16. Since the dispersion coefficient is related
to the pore velocity through the dispersivity
(DL = cLvx + D*), we can solve for the dispersivity (XL)
knowing DL and the effective diffusion coefficient in the po-
rous media, D* (for phenol red = 5.2x10' cm'/s)
DL -D*
(XL V (9)

Sorbing Compounds
Sorption is a process by which solutes partition between
the ground water and the surface of soil and aquifer solids.
Non-polar organic compounds tend to partition preferentially
onto solid surfaces and into the natural organic matter present
on soil and aquifer mineral surfaces. This partitioning is of-
ten described using a linear sorption isotherm
CS = KdC (10)

C sorbed concentration of solute (mg solute sorbed/g dry solid)
Kd linear sorption coefficient (L water/g dry solid)
C solute concentration in the pore water (mg solute/L water)
The linear sorption coefficient, Kd, is often related to the frac-
tion of natural organic matter (on a carbon basis [g organic
carbon/g dry solid], fc) in the soil by Kd = focKo., where Ko is
an organic matter partition-water coefficient on a carbon ba-
sis (L water per g organic carbon). Ko. is often correlated

with another important environmental property of organic
compounds, the octanol-water partition coefficient, Ko,,, which
is a measure of a chemical's hydrophobicity.'7'
The transport of a solute that sorbs to aquifer solids is de-
layed or retarded compared to a non-sorbing solute. This
retardation occurs because the solute molecules spend a
fraction of their time immobilized on the solid surface
and are thus not fully available for transport by the bulk
pore water flow.
The equations above must be modified for the accumula-
tion of solute on the solid surfaces by including another ac-
cumulation term on the left-hand side, as shown by

ac + Pb aCs
at n at

Da2C ac
D v
L ax2 x ax

By introducing the equilibrium sorption isotherm, this equa-
tion can be recast as

ac DL a2C V (12)
at Rf, x2 Rf ax
where R, is the retardation factor; defined by

R-= 1+- Kd (13)

The solution to this equation, assuming the same initial and
boundary conditions as stated above, is (Pe>100)

C = 0.5

Suspended Colloids
Colloids are micron-sized (10-6 m) particles suspended in
groundwater, of either bio-
logical (microorganisms) or
geological (clay particles, Colloid
organic macromolecules) 1 .Transport step
origin. The transport of col- [- 1 2. Attachment step
loids in groundwater is a
concern because of, for ex- flow line
ample, the migration of
pathogenic microorganisms
away from septic-tank Figure 2. Multi-step
leach fields and the ability attachment process for
of mineral colloids to facili- colloids to soil grains.
tate the migration of radio-
nuclides or other tightly bound pollutants.
Colloids can be removed from the pore water, however,

Summer 2001

and may become immobilized on the surface of the solid ma-
trix in the subsurface. This process occurs in two steps, as
shown in Figure 2: transport from the pore water to the solid
surface, followed by attachment at the solid surface. The
mechanisms occurring in the transport step are affected by
the size of the colloid and may be dominated by diffusion,
inertial, or surface-interaction forces."' The attachment step
is dependent on the surface characteristics of the colloid
and of the solid matrix, such as electric charge and hy-
drophobicity, in addition to solution properties, such as
pH and ionic strength.
Taken together, these properties affect the rate of colloid
removal from the pore water and attachment to the solid sur-
faces, which can be modeled as a first-order process similar
to chemical reaction. The convection-dispersion equation can
be modified to include this first-order removal process by
aC a C 2c
= -v + D -kC (15)
at x L ax2
where k is a first-order attachment rate coefficient (sec-l). The
solution to this equation for one-dimensional transport oc-
curring in a laboratory column experiment subject to the ini-
tial and boundary conditions listed earlier is"'1


exp{(2a ) [




The results to be presented were obtained by groups (of 4
to 8 students) in either the senior chemical engineering labo-
ratory course or in an elective senior-graduate-level course
covering the fundamentals of subsurface remediation. The
remediation course was co-listed in both the Chemical Engi-
neering Department and the Geological Engineering and Sci-
ences Department at Michigan Technological University.'8'
The experimental procedures, along with handouts on trans-
port in porous media, were given to the groups a week prior
to the experiment, and the transport experiments were con-
ducted within a 3-hour laboratory period. (The procedures
and handouts can be obtained by contacting

Phenol Red and Bacteria Transport
An analysis of phenol red breakthrough provides the lon-
gitudinal dispersion coefficient (DL) of the dissolved com-




C/C.06 --. *---



00 05 10 15 20 25 30
Pore Volumes
Phenol Red Bactena ---Theory Bactena Theory Phenol Red

Figure 3. Comparison between theory and experiment for
Experiment 1, Table 1.

Figure 4. Comparison between theory and experiment
for Experiment 2, Table 1.

Summary of Results from Column Transport Experiments

Experiment I Experiment2
Experimental Parameters Phenol Bacteria Phenol Bacteria
Red Red
Column Length, L (cm) 20 20 21 21
Sand Porosity, n 0.38 0.38 0.38 0.38
Flow Rate, Q(cm1/min) 2.6 1.57 6.8 6.8
Velocity, vx=Q/(stD'/4.n)(cm/min) 1.35 0.82 3.59 3.59
Dispersion Coefficient, D, (cm/s) 3.3x10-' 1.2x10-2
Dispersivity, CL (cm) 1.48x10' 2.0x10 -
Attachment coefficient, k(sec ) 3.2x 10' 1.2x10-

Chemical Engineering Education


10 _------------

C/C.0 6


00 05 10 15 20 25 30
Pore Volumes Pumped (U)
Phenol Red Bacteria --- -Theory Bactena Theory Phenol Red


The experimental results in this paper
were obtained using a laboratory
column, but the knowledge
gained can be applied to
real-world situations.



CIc. 06


0o 'L L ---
0 1 2 3 4 5 6 7 8 9
Dimensionless Time (v, t/L)
-- Retardation Factor = 1 -Retardation Factor = 5

Figure 5. Results from a numerical experiment showing the
effect of sorption on column breakthrough for a chemical
with a Rd = 1 and Rd = 5. In this prediction, Pe = 135 (Ex-
periment-1 parameters from Table 1).

pound and the porous media dispersivity (a ). A measure of
the bacteria breakthrough provides a contrast to the break-
through of the dissolved tracer (phenol red) and also the bac-
terial attachment rate coefficient (k). Figures 3 and 4 illus-
trate typical breakthrough results for this experiment. The
two figures show breakthrough data and comparison to theory
(Eqs. 7 and 16) for phenol red and the bacteria. The flow
rates and other experimental parameters are listed in Table 1.
The phenol red breakthrough (rising portion of the data) is
symmetrical, while the bacteria data is not. The phenol red
data achieves a value of C/C =1, while the bacteria achieves
a lower value due to the attachment of bacteria to the sand
grains in the column. The model results for phenol red were
plotted by using Eq. (7) and for bacteria by using Eq. (16).
The theory matches experiment very well for phenol red,
but it tends to over-predict the measured concentration of bac-
teria for these experiments, although the general shape is pre-
dicted correctly. Apparently the cell attachment is greater than
is predicted by the model, especially during the initial por-
tion of the breakthrough.
Table 1 provides a summary of the calculated results, i.e.,
the values for dispersion coefficient, dispersivity, and the at-
tachment rate coefficient. The results in Table 1 show that

the dispersion coefficient is higher at higher flow rates, as
expected from Eq. (9). The dispersivity of the sand is ex-
pected to be on the order of the grain size, 0.07 cm.[1]0
The dispersivity results in Table 1 are a factor of 2 to 3
high, possibly due to the effects of mixing in the column
inlet and outlet lines and from mixing in the flow-through
absorbance cells.
The attachment rate coefficient for bacteria is lower at the
higher flow rate, indicating less attachment. Also, both at-
tachment rate coefficients in Table 1 are lower than theory
[k=3/2vx(l-n)rla=1.33x10-3] by about a factor of 4 to 10,
assuming a collision efficiency of a =1.0 and a collector ef-
ficiency nr=4As'"/Pe -2/=7.3 x 10-3 for this system,'98' where["]

A = 2(1-p5)/(2-3p+3p'-2p6)
p = (1-n)"3
Pe = particle Peclet number = ud/D
u = Darcy velocity volumetricc flow rate/column cross-
sectional area)
d = sand grain diameter (0.07 cm)
D = Brownian diffusion coefficient of the bacteria
(4.3x10 cm/s)

It is expected, however, that the observed value of k is lower
by a factor of 5 to 10 compared to theory because the colli-
sion efficiency is on the order of 0.1 to 0.2 for bacteria at the
ionic strength of the solution used in this experiment.15'

Model Results: Transport with Retardation
Numerical experiments were conducted to demonstrate ef-
fects of sorption on breakthrough curves. Equation (14) was
used to perform simulations using a Microsoft Excel spread-
sheet. Figure 5 shows the effect of equilibrium sorption on
chemical breakthrough from the column. The curve with sorp-
tion (Retardation Factor = 5 curve) is delayed upon arriving
at the column exit by a factor of 5 compared to the curve with
no sorption of the transported chemical (Retardation Factor
= 1 curve). Also, the breakthrough curve for the sorbing
chemical is much more spread out in relation to the inert
chemical curve.
The experimental results in this paper were obtained using
a laboratory column, but the knowledge gained can be ap-
plied to real-world situations. Industrial chemicals might be
released into the groundwater due to leaks from underground
storage tanks or from spills on the ground surface. Hydro-
philic chemicals, characterized by high water solubilities and
low K o, might be expected to migrate in the groundwater
similar to an inert tracer, such as phenol red. For hydrophilic
chemicals, this migration would occur at roughly the same
velocity as the groundwater flow, which is on the order of 1
to 10 meters per year."' A hydrophobic chemical of low wa-
ter solubility and high Kow would be expected to migrate away
from the zone of original contamination much more slowly.

Summer 2001

Though hydrophobic chemicals are transported in ground-
water more slowly, their significant adsorption to the aquifer
solids results in a "reservoir" of bound material that is diffi-
cult to remove by remediation processes, such as pump and
treat and bioremediation.

Environmentally Benign Design: MTBE Case Study
Consider the case of methyl tert butyl ether (MTBE)."21
The use of MTBE as a gasoline additive began in the United
States in the late 1970s, when it replaced alkyl lead additives
as octane boosters. The use of MTBE increased dramatically
in the 1990s when the Clean Air Act Amendments of 1990
mandated efforts to reduce the combustion byproducts car-
bon monoxide and unburned aromatic hydrocarbons (lead-
ing to ozone formation in urban smog). In 1970, MTBE was
the 39th most produced chemical in the U.S.; today it ranks
4'h in production volume.'13'
The environmental properties of MTBE are such that when
leaks from fuel storage tanks occur, the resulting groundwa-
ter plume can extend over a kilometer from the spill site. An
MTBE spill will potentially contaminate much greater vol-
umes of community water supplies than other fuel com-
ponents, such as benzene, toluene, ethylbenzene, and xy-
lenes (BTEX), whose plumes extend only 100 m or so
from a spill site.
The primary reason for this difference in behavior is envi-
ronmental reaction (degradation) rates and sorption to aqui-
fer solids."2' BTEX and related compounds found in fuel are
relatively biodegradable in the subsurface, with reaction half-
lives of approximately one month or so, and they readily ad-
sorb to aquifer materials (e.g., toluene Kow = 450). MTBE
is resistant to degradation biodegradationn half-life of
about two years) and is very water soluble and non-ab-
sorbing (MTBE Kw = 9).
Using the aquifer solids properties of this column study as
a basis, we can use Eq. (13) to compare the retardation fac-
tors of toluene and MTBE. For this calculation, we take the
bulk density to be pb = 1.5 g dry sand/cm3 sand and a porosity
of n = 0.38 cm3 voids/cm3 sand; estimating Kd = focKo =
f(0.48)K where f0 is the fraction of organic carbon in the
soil (0.001 g organic carbon/g sand for this comparison), the
calculated retardation factors for toluene and MTBE are 2.87
and 1.04, respectively. Thus, toluene is expected to take al-
most three times as long to reach a monitoring location down
gradient from a source of contamination compared to MTBE.
Conversely, MTBE is expected to contaminate almost three
times the aquifer volume compared to toluene.
MTBE is currently being phased out of production as a
gasoline additive because of the problems discussed above.
Had the environmental properties of MTBE been considered
more carefully at the beginning, a more environmentally be-
nign fuel additive may well have been discovered, perhaps

one with a lower water solubility, a higher Kow, and shorter
degradation half-life in the subsurface. Of course, a suitable
substitution would also need to have a low degree of toxicity
to humans and also not cause other environmental impacts
(global warming, stratospheric ozone depletion, ecosystem
damage, etc.).

Chemical engineers have the responsibility to acquire a
fundamental understanding of environmental fate character-
istics for the chemicals they produce and to work diligently
toward implementing in-process modifications to minimize
the impacts of these chemicals. The experiments described
in this article, along with appropriate handout materials in
contaminant hydrogeology, can aid undergraduate chemical
engineering students in receiving this type of information.

Support for this research was provided by the National
Science Foundation's combined Research and Curriculum
Development Program award EEC-9420526. A PhD Fel-
lowship for P.A.D. was provided by Michigan Techno-
logical University.

I. CMA, the Chemical Manufacturers Association, Codes of Manage-
ment Practices, available on the internet at
2. Fetter, C.W., Contaminant Hydrogeology: Prentice Hall, Upper Saddle
River. NJ (1993)
3. Domenico, P.A., and F.W. Schwartz, Physical and Chemical
Hydrogeology, John Wiley & Sons, Inc., New York, NY (1997)
4. Freeze, R.A., and J.A. Cherry, Grouvndwate; Prentice Hall, Englewood
Cliffs, NJ (1979)
5. Deshpande, P.A., and D.R. Shonnard, "Modeling the Effects of Sys-
tematic Variation in Ionic Strength on the Attachment Kinetics of
Pseudomnonasfluorescens UPER-1 in Saturated Sand Columns," Wa-
ter Resources Res., 35(5), 1619 (1999)
6. Whitaker, S.W., The Theory of Volume Averaging, Kluvar Publishing,
Amsterdam (1999)
7. Mackay, D.. W. Shiu, and K. Ma, Illustrated Handbook of Physical-
Chemical Properties and Environmental Fate for Organic Chemicals,
1st ed., Vol. 1 to 4, Lewis Publishers (1992)
8. Elimelech, M., J. Gregory, X. Jai, and R.A. Williams, Particle Depo-
sition and Aggregation: Measurement, Modeling and Simulation,
Butterworth Heinmann, Oxford, UK (1995)
9. Gierke, J.S., A.S. Mayer, and D.R. Shonnard, "Multidisciplinary Sub-
surface Remediation Courses: Fundamentals, Experiments, and De-
sign Projects," J. of'Eng. Ed., 87(5), 555 (1998)
10. Thibodeaux, L.J., Environmental Chemodynamics: Movement of
Chemicals in Air; Water; and Soil, John Wiley and Sons (1996)
11. Shonnard, D.R.. R.T. Taylor, M.L. Hanna, C.O. Boro, and A.G. Duba,
"Injection-Attachment of Methylosinus trichlosporium OB3b in a Two-
Dimensional Miniature Sand-Filled Aquifer Simulator," Water Re-
sources Res., 30(1), 25 (1994)
12. Johnson, R., J. Pankow, D. Bender, C. Price, and J. Zogorski, "MTBE:
To What Extent will Past Releases Contaminate Community Water
Supply Wells?" Environ. Sci. Tech., 32(5), A2 (2000)
13. CMA, Chemical Industry Statistical Handbook 1998, the Chemical
Manufacturers Association, Arlington, VA (1999) 1
Chemical Engineering Education


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

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

ACKNOWLEDGMENT Include in acknowledgment only such credits as are essential.

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

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

Full Text

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EGDIHVA1S_KX57B7 INGEST_TIME 2012-02-17T16:30:56Z PACKAGE AA00000383_00151