Chemical engineering education

Material Information

Chemical engineering education
Alternate Title:
Abbreviated Title:
Chem. eng. educ.
American Society for Engineering Education -- Chemical Engineering Division
Chemical Engineering Division, American Society for Engineering Education
Creation Date:
Annual[ FORMER 1960-1961]
Physical Description:
v. : ill. ; 22-28 cm.


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


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

Record Information

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

UFDC Membership

Chemical Engineering Documents


This item has the following downloads:

Full Text

Chemical Engineering Education
Department of Chemical Engineering
University of Florida Gainesville, FL 32611
PHONE and FAX : 352-392-0861
e-mail: cee@che.ufl edu


Phillip C. Wankat

Carole Yocum

James 0. Wilkes, U. Michigan

William J. Koros, Georgia Institute of Technology


E. Dendy Sloan, Jr.
Colorado School ofMines

Pablo Debenedetti
Princeton University
Dianne Dorland
Rowan University
Thomas F Edgar
University of Texas at Austin
RichardM. Felder
North Carolina State University
BruceA. Finlayson
University of Washington
H. Scott Fogler
University of Michigan
Carol K. Hall
North Carolina State University
William J. Koros
Georgia Institute of Technology
John P O'Connell
University of Virginia
David E Ollis
North Carolina State University
Ronald W. Rousseau
Georgia Institute of Technology
Stanley L Sandier
University of Delaware
Richard C. Seagrave
Iowa State University
C. Stewart Slater
Rowan University
Donald R. Woods
McMaster University

Chemical Engineering Education

Volume 37

Number 3

Summer 2003

162 University of California Berkeley,
Arup K. Chakraborty, P.B. Plouffe, Stacey Shulman

168 Introduction: Strategies for Effective Teaching in Chemical Engineering,
Gilda A. Barabino
170 Water Day: An Experiential Lecture for Fluid Mechanics,
Laura P. Ford
174 Introduction to Biochemical Engineering: Synthesis, Resourcefulness,
and Effective Communication in Group Learning, Tonya L. Peeples
180 A Course in Bioprocess Engineering: Engaging the Imagination of
Students Using Experiences Outside the Classroom,
Agnes E. Ostafin, Darcy LaClair, Hartley T Schmidt
184 Incorporating High School Outreach into ChE Courses,
Julia M. Ross, Taryn M. Bayles
188 Lab-Based Unit Operations in Microelectronics Processing,
Chih-Hung (Alex) ( 1..... Milo D. Koretsky, Sho Kimura,
Skip Rochefort, Cyndie Shaner
196 Incorporating Experimental Design into the Unit Operations Laboratory,
Eric J. Doskocil
202 Passing It On: A Laboratory Structure Encouraging Realistic Communi-
cation and Creative Experiment Planning,
S. Scott Moor, James K. Ferri
208 High-Performance Engines: Fast Cars Accelerate Learning,
Sang M. Han, Joseph L. Cecchi, John J. Russell
214 Increasing Time Spent on Course Objectives by Using Computer
Programming to Teach Numerical Methods,
David L. Silverstein

220 The Incontrovertible Logic of the Academy, Richard M. Felder

222 Sensitivity Analysis in ChE Education: Part 1. Introduction and Applica-
tion to Explicit Models, William R. Smith, Ronald W. Missen
228 Stochastic Modeling of Thermal Death Kinetics of a Cell Population
Revisited, L.T. Fan, A. Argoti Caicedo, S.T. Chou, W.Y. Chen

236 Choosing and Evaluating Equations of State for Thermophysical
Properties, Coray M. Colina, Keith E. Gubbins

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 2003 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 ChEDivision, ASEE, which body assumes no responsibility for them Defective copies replaced if notifiedwithin
120 days of publication. Write for information on subscription costs and for back copy costs and availability. POSTMASTER:
Send address changes to Chemical Engineering Education, Chemical Engineering Department., University of Florida,
Gainesville, FL 32611-6005. Periodicals Postage Paid at Gainesville, Florida and additionalpost offices.

Summer 2003

M n department

University of California


San Francisco Bay.

University of California Berkeley, CA 94720-1462

he academic year 2003-2004 marks the 56th anniver-
sary of the Department of Chemical Engineering at
the University of California, Berkeley. Developed as
an academic discipline in the United States near the turn of
the century, chemical engineering moved toward engineer-
ing science in the late 1940s and 1950s. During those years,
Berkeley's young department rose to prominence. It took the
lead in developing new areas of research such as electrochemi-
cal engineering, molecular thermodynamics, semiconductor
processing, and biochemical engineering. The department has
consistently ranked among the top three research and teach-
ing programs in the country. It continues to be at the fore-
front in developing programs in emerging areas of chemical,
materials, and biological technology.
The success of the department has been facilitated by be-
ing part of a world-famous university with leading depart-
ments in all fields of engineering, the natural sciences, social
sciences, and the humanities. Strong synergies with research
programs at the Lawrence Berkeley National Laboratory
have also added to the richness of the intellectual culture
of the department.
The Chemical Engineering Department at Berkeley is part

of the College of Chemistry, which also houses the nation's
top-ranked chemistry department. Nineteen members of the
faculty work together congenially with 115 graduate students,
333 undergraduates, and 31 postdoctoral fellows in the pur-
suit of excellence in teaching and scholarship.
The current faculty has received many awards and honors
recognizing their accomplishments in teaching and research.
These awards include four members of the National Acad-
emy of Engineering, one member of the National Academy
of Science, two Allan P. Colburn Awards (AIChE), two Pro-
fessional Progress Awards (AIChE), three Food, Pharmaceu-
tical and Bioengineering Awards (AIChE), the Award in Het-
erogeneous and Homogeneous Catalysis (ACS), the Marvin
Johnson Award in Biochemical Technology (ACS), The Col-
loid and Surfaces Chemistry Award (ACS), the EV Murphree
Award (ACS), two Emmett Awards in Heterogeneous Ca-
talysis (North American Catalysis Society), two Amgen
Awards in Biochemical Engineering (Engineering Founda-
tion), one Dillon Medal (APS), one AVS Award in Plasma
Physics, three Camille Dreyfus Teacher Scholar Awards, and
numerous Presidential Young Investigator, National Young
Investigator, and CAREER Awards (NSF).
Research groups in the department are housed primarily in

Chemical Engineerrng Education

Copyright ChE Division of ASEE 2003

Gilman Hall, a National Historic Chemical Landmark where
plutonium was discovered, and Tan Kah Kee Hall (dedicated
in 1997). These buildings provide state-of-the-art facilities
and infrastructure for research.

The appointment of the first Professor of Chemical Engi-

neering in July of 1946 marked the administrative decision
that ultimately led to the present chemical engineering pro-
gram at Berkeley. As the university began to more fully rec-
ognize the importance of chemical engineering-especially
through its contributions to the war effort in the development
of the atomic bomb and in the petroleum and chemical in-
dustries-the need for a full-fledged program became appar-

Chronology of Events in Chemical Engineering at Berkeley

1912 1- G.N. Lewis institutes a chemical technology major.

1942 1- An interdepartmental graduate group offers the MS degree in
chemical engineering.

1945 1- Provost Monroe Deutsch authorizes establishment of a
Chemical Engineering program in the College of Chemistry.

1946 1- Initial faculty members are appointed to teach chemical
engineering: Philip Schutz (deceased, 1947); LeRoy
Bromley (retired, 1972); Charles Wilke (chair 1953-63;
retired, 1987).
1- Undergraduate instruction begins.

1947 1> Theodore Vermeulen (chair 1947-53; deceased, 1984),
Donald Hanson (chair 1963-66; retired 1989), Charles
Tobias (chair 1966-71; retired, 1991; deceased, 1996), join
the faculty.
1- PhD program is approved.

1948 1- BS degree program in Chemical Engineering is approved.
1- F. Campbell Williams joins the faculty (left for Petrobas and
University of Brazil, 1952).

1949 1- The Chemistry Department is renamed the Department of
Chemistry and Chemical Engineering.
1- BS program is accredited by the AIChE.

1952 1- A Division of Chemical Engineering is created within the
Department of Chemistry and Chemical Engineering.
1- Kenneth Gordon joins the faculty (left for University of
Michigan, 1954).

1953 1- Eugene Petersen (retired, 1991) joins the faculty.

1954 1> Andreas Acrivos joins the faculty (left for Stanford, 1963).

1955 1- John Prausnitz joins the faculty.

1956 1> Donald Olander joins the faculty (left for UCB Nuclear
Engineering Department, 1961).

1957 1- Chemical Engineering is established as a separate depart-

1958 1> David Lyon (retired, 1982) joins the faculty.

1961 1> Alan Foss (retired, 1994) and Michel Boudart (left for
Stanford, 1965; appointed Adjunct Professor, 1994-97) join
the faculty. Simon Goren (chair, 1994-97; retired, 2001) and
Richard Wallace (resigned, 1965) join the faculty.

1963 1> Edward Grens (retired, 1987), C. Judson King (chair,
1972-81), John Newman, Richard Ayen (left for Stauffer
Chemical Co., 1968), join the faculty

1964 1> Robert Merrill joins the faculty (left for Cornell, 1977).

1965 > Michael Williams (retired, 1989) joins the faculty.

1966 1- Robert Pigford joins the faculty (returned to University of
Delaware, 1975).
1- Room 307 Gilman Hall is designated a registered National
Historic Landmark by the U.S. Department of the Interior
as the location for the discovery of plutonium.
1967 > Scott Lynn (retired, 1994) and Alexis Bell (chair, 1981-91)
join the faculty.

1969 1> Mitchell Shen joins the faculty (deceased, 1979).

1970 1> Lee Donaghey (left for Chevron, 1977) and Thomas
Sherwood (deceased, 1976) join the faculty.

1975 1> Clayton Radke joins the faculty

1977 1> Dennis Hess (left for Lehigh University, 1991) joins the

1978 1> Elton Cairns and Harvey Blanch (Chair, 1997-2001) join
the faculty.

1979 1- David Soane (Soong)( left for Soane Technologies 1994;
appointed Adjunct Professor, 1994) and Edward Reiff (left
for DuPont, 1982) join the faculty.

1981 1> Morton Denn (Chair, 1991-94; left for CCNY, 1999) joins
the faculty.

1982 1> Jeffrey Reimer and James Michaels (left for Mobil, 1989)
join the faculty.

1986 10 Douglas Clark, David Graves, and Doros Theodorou (left
for University of Patras, 1994) join the faculty.

1988 1> Arup Chakraborty (Chair, 2001-) joins the faculty.

1991 1- Susan Muller joins the faculty.
1- Room 307 Gilman Hall is designated a Nuclear Historic
Landmark by the American Nuclear Society as the site of
the first chemical identification of plutonium on February
23-24, 1941).

1992 1> Jay Keasling joins the faculty.

1993 10 Enrique Iglesia and Roya Maboudian join the faculty.
1- Construction of Tan Hall begins.

1997 1> Gilman Hall designated as an ACS National Historic
Chemical Landmark.
1> Dedication of Tan Kah Kee Hall (April 12, 1997).
1999 1> David Schaffer joins the faculty
2000 1- Alexander Katz joins the faculty.
2004 1- Rachel Segalman will join the faculty.

Summer 2003

S. the [Berkeley] chemical engineering faculty has contributed to the
emergence of a number of different areas that are now
considered quintessential chemical engineering.

ent. Initially, considerable con-
troversy developed as to
whether the program should
be in the College of Engineer-
ing or the College of Chemis-
try. For a while, this led to the
amusing situation where Ber-
keley had two chemical engi-
neering departments-one
housed in the College of En-
gineering and the other in the
College of Chemistry. The
stronger program in the Col-
lege of Chemistry ultimately
Philip Schutz, a professor of
chemical engineering at Co-
lumbia University, was se-
lected to head the fledging Professor John Prausnit
chemical engineering pro-
gram. To assist him, Dean Wendell Latimer appointed Charles Wilke
(PhD, University of Wisconsin) and LeRoy Bromley (MS, Illinois
Institute of Technology). Sadly, shortly after the first class en-
rolled in September 1946, Philip Schutz passed away. Theodore
Vermeulen (PhD, UCLA) joined the program in February 1947
and became its head. Don Hanson (PhD, University of Wiscon-
sin) and Charles Tobias (a Hungarian 6migrd, PhD, Budapest)
joined the faculty in the fall of 1947, and they were followed
by F. Campbell Williams in 1948.
This initial faculty group remained in place without further addi-
tions until 1952 when Ken Jordan from MIT joined the faculty.
During this period, a PhD program was formally approved (1948)
and the BS program was fully accredited (1949). Charles Wilke
succeeded Theodore Vermeulen as Chairman in 1953, and between
1953 and 1955, he recruited three more remarkable intellects: Eu-
gene Petersen from Penn State (1953), Andreas Acrivos from Min-
nesota (1954), and John Prausnitz from Princeton (1955).
With these pioneers establishing standards of excellence that have
consistently marked the department, new areas of research were
established in the period between 1953 and 1985. The sub-field of
electrochemical engineering (under the leadership of Charles Tobias)
is a notable example. During those same years, John Prausnitz de-
veloped a systematic approach for obtaining activity coefficients
and equations of state for substances central to the petrochemical
industry. The major award in this field, now called molecular ther-
modynamics, is named after Dr. Prausnitz, who continues to be
active in teaching, research, and administration.
In the 1970s, Dennis Hess headed up the first program in micro-

z and students in the unit operations laboratory.

Blanch, Clark, Keasling,
Schaffer, ( I, ii i.. -,
Muller, Prausnitz, Radke


Bell, ( I.i l....n Iglesia,
Katz, Reimer

Cairns, Newman,
Reimer, Balsara

Bell, Graves, Iglesia,
Keasling, King, Clark

Balsara, ( I.i i...-. MEMS
Muller, Prausnitz, Graves, Maboudian,
Radke, Reimer, Segalman Reimer, Segalman

Figure 1. Research areas.

Chemical Engineerrng Education


Electronics faculty

PetrochemG Govt.


1990s PhD Degrees

a Other m Facut

m Bectronics

Biotech o Cherri
m Petrochem

electronics processing within a chemical engineering depart-
ment. In that same decade, Charles Wilke and Harvey Blanch
initiated a pioneering program in biochemical engineering.
Under the leadership of Gene Petersen, Michel Boudart, and
Alex Bell, an innovative program in catalysis and reaction en-
gineering was established. When Mort Dennjoined the faculty
in 1981, he setup a world-class program in polymer process-
ing on the Berkeley campus and at the Lawrence Berkeley
National Laboratory. (The research programs that have
evolved since 1985 are described in the "Graduate Program"
section of this article.)

The BS degree in chemical engineering at UC Berkeley, ac-
credited by ABET, is acknowledged as one of the most demand-
ing majors on campus. The curriculum provides a strong back-
ground in mathematics and the basic sciences of physics, chem-
istry, and biology. The core principles of thermodynamics, re-
action engineering, and transport phenomena are often illus-
trated by examples drawn from the petroleum, petrochemical,
pharmaceutical, biomedical, and microelectronics industries.
The program includes extensive laboratory experience, with
mandatory laboratory courses in unit operations, transport phe-
nomena, and control systems. In addition, elective courses in
biochemical engineering, polymers, and microelectronics pro-
cessing include large laboratory components. Under the super-
vision of a faculty member, the students work in teams to de-
sign a process in the capstone course.
In the final two years, students choose an option that allows
them to more deeply study a sub-field of chemical engineering.
In particular, option programs are available in biochemical en-
gineering, materials science and engineering, chemical process-
ing, and microelectronics. The options program, in place for 16
years now, constantly evolves to meet student and industry needs.

Summer 2003

1970s PhD Degrees

Professor Alex Katz and students.

Figure 2.

Textbooks Published by Berkeley Faculty

Blanch & Clark Biochemical Engineering

Denn Process Fluid Mechamncs
Process Modeling
Introduction to Chemical Engineering Analysis
(with T.W.F. Russell)
.* ofReaction and Transport Processes
Optimization by Variational Methods
King Separation Processes
Freeze Drying of Foods
Hanson Computation ofMultistage Separation Processes
Newman .I ..
Petersen Chemical Reaction Analysis
Prausnitz Molecular Thermodynamics ofFluid Phase
Equilibria (with Lichtenthaler and de Azevedo)
Properties of Gases and Liquids (with B. Poling
and R. Reid)
Regular and Related Solutions (with J.H.
Hildebrand and R.L. Scott)

Reimer Chemical Engineering Design & Analysis (with
T.M. Duncan)

Sherwood, Pigford Mass
& Wilke

A special feature of the undergraduate program at Berke-
ley is the high premium placed on teaching students how to
communicate technical information effectively, both orally
and in writing. Each laboratory course and the design course
include oral presentations and written reports. In addition, a
required course in technical communication and ethics is
taught by a lecturer in the department. Alumni and em-
ployers report back to us the great advantage and value of
these experiences.
The undergraduate program in chemical engineering is con-
stantly reviewed by the faculty. This has allowed it to con-
tinuously evolve, and perhaps it is the reason why we do not
perceive an urgent need to make revolutionary changes to
the curriculum. Constant fine tuning has allowed us to make
necessary modifications to meet the needs of students who

find employment in a wide range of corporate sectors.
The department values excellence in teaching and over the
years several landmark textbooks have been written by the
faculty (see Table 2). They include Molecular Thermodynam-
ics and Properties of Liquids and Gases by Prausnitz, Sepa-
ration Processes by King, Process Fluid Mechanics by Denn,
Biochemical Engineering by Blanch and Clark, and Intro-
duction to Chemical Engineering by Reimer. The faculty's
excellence in teaching has been recognized by numerous cam-
pus awards, including four Distinguished Teaching Awards.
Faculty from all departments on the Berkeley campus are eli-
gible for this award, and only three distinguished teachers
are chosen each year.
Our commitment to teaching has continued to inspire the
next generation of teachers. Fifteen percent of the current fac-

The Emergence of High-Technology Industries

Berkeley has developed research areas on the frontiers
of chemical engineering. A Berkeley PhD, Andy Grove
(1963), together with Gordon Moore (BS Chemistry, 1950)
and Robert Noyce, founded the Intel Corporation. The grow-
ing importance of the semiconductor manufacturing indus-
try led to Dennis Hess joining the department in 1977, and
he initiated collaborations that were critical in the expand-
ing role that chemical engineers have continued to play in
many areas of electronic materials processing.
The proximity of Silicon Valley was a distinct advantage
for the development of the program at Berkeley, and op-
portunities for undergraduate and graduate student employ-
ment increased. Today, nearly one-third of the students at
all degree levels find employment in electronics materials
processing. This area has been strengthened at Berkeley by
the addition of David Graves (in 1986), with his interests
in plasma processing and Roya Maboudian (in 1993), who
studies surface processes occurring during electronic-ma-
terials processing (especially MEMS technologies). Jeff
Reimer's research has complemented the research program
in electronic materials. At the undergraduate level, Lee
Donaghey first developed an elective course in electronic
materials processing in 1970. Subsequently, Hess and
then Graves expanded the course, which has become a
key component of the options program in chemical en-
The second area to benefit from Berkeley's location was
biochemical engineering. Charlie Wilke turned his research
effort from diffusion and mass transfer to this new field in
the 1960s. His early studies on the kinetics of microbial
growth and gas-liquid mass transfer provided the engineer-
ing underpinnings for the revolution in molecular biology
that was to come. The advent of recombinant DNA and cell
fusion (hybridoma) technology in the mid- 1970s led to tre-

mendous opportunities for the production of new, high-
value therapeutic proteins and specialty chemicals. The Bay
Area soon became home to start-up companies such as
Cetus, Chiron, and Genentech. The department's research
in this area expanded with the addition of Harvey Blanch
in 1978, Douglas Clark in 1986, Jay Keasling in 1992, and
David Schaffer in 1999.
The biochemical engineering program is now one of the
largest in the United States. Jay Keasling introduced envi-
ronmental biotechnology to the undergraduate curriculum,
highlighting the increasing role of biological routes to en-
vironmental remediation.
Environmental engineering has also become a leading
career option for BS graduates. Jay Keasling has devel-
oped a major program in synthetic biology. David Schaffer
has brought cutting-edge biomedical research to the de-
partment, developing vectors for gene delivery. Arup
Chakraborty's work on how cells in the immune system
communicate and Clay Radke's interest in contact lenses
are two other examples of biomedical research in Berkeley's
Chemical Engineering Department.
In recent years, many new technologies have emerged
that require advance materials with specific functionality.
Precise control of these properties requires a detailed
knowledge of the molecular constitution and mesoscopic
structure of the material. Berkeley's Chemical Engineer-
ing Department has been able to capitalize on the strengths
of the Chemistry Department and develop collaborations
that elucidate the relationship between materials structure
and their function. Synthesis, quantum and statistical me-
chanical theory, and characterization are being employed
by several groups to develop novel catalytic and polymeric
materials (e.g., Balsara, Bell, Chakraborty, Katz, Iglesia,
Muller, Reimer, Segalman).

Chemical Engineerrng Education

ulty at MIT and 18% of the faculty at Cal Tech received their
PhDs from Berkeley's Department of Chemical Engineering.

Both the MS and PhD degrees in chemical engineering are
granted, but a preponderance of graduate students pursue the
PhD. The central focus of graduate education is research con-
ducted under the supervision of a faculty advisor. The fac-
ulty-advisee relationship is truly collaborative and is one of
the most enriching aspects of graduate (and, indeed, faculty)
life at Berkeley. As noted
earlier, the chemical engi-
neering faculty has contrib-
uted to the emergence of a
number of different areas
that are now considered
quintessential chemical
engineering. Research ar-
eas represented in the de-
partment today can be
broadly classified as
shown in Figure 1. \
A special feature of a
chemical engineering edu-
cation at Berkeley is that a
deep knowledge of phe-
nomenological concepts Professor David Schaffer a
pertinent to transport phe- working on an I-cycler,
nomena, reaction kinetics,
and thermodynamics is but-
tressed by a strong foundation in the underlying molecular
phenomena. Thus, experimental methods such as scattering
and spectroscopy, and the conceptual frameworks of statisti-
cal and quantum mechanics, are as much a part of the basic
tools in the arsenal of a PhD chemical engineer as are con-
tinuum models, macroscopic reaction studies, and measure-
ments of macroscopic mechanical properties of materials. As
shown in Figure 2, graduate research at Berkeley can be
viewed as bringing to bear the chemical engineer's rare
ability to think across molecular, mesoscopic, and mac-
roscopic scales to develop knowledge that can aid the de-
velopment of new technologies and discover new phenom-
ena in the related sciences.
Research in areas pertinent to biotechnology and micro-
electronics processing are facilitated by our geographical lo-
cation. The San Francisco Bay area is a hotbed for biomedi-
cal research and companies focused on developing new bio-
chemical and microelectronic products, and there are many
collaborations between research groups in the department and
these companies.
Another facilitator of research in biological engineering has
been the establishment of an institute called QB3, a
multidisciplinary institute that involves the Berkeley cam-

Summer 2003

a qu

pus and the U.C. San Francisco Medical School. QB3 aims
to bring leading scientists and engineers on the Berkeley cam-
pus together with medical researchers at UCSF Medical
School to help solve important problems in biomedicine.
Three faculty in the Chemical Engineering Department have
appointments in QB3.
Several faculty in the department are working to develop
materials that can function with precision. Such materials,
particularly those that can carry out functions in response to
changes in external stimuli, are important components of
many emerging technolo-
gies. The way to make such
materials is to learn how to
manipulate molecular com-
ponents such that the de-
sired function is obtained.
Strong and collaborative re-
search programs exist in the
department wherein synthe-
sis, characterization, and
theory are brought together
to design new catalytic and
polymeric materials as well
as those pertinent to micro-
electronics and MEMS
technologies. In this regard,
graduate students in the de-
raduate student Karen Lai apartment benefit from
antative PCR machine, strong synergies with the
Department of Chemistry as
well as the Molecular Foundry at LBNL. The Molecular
Foundry is a new national facility with the aim of developing
"designer" materials by manipulating molecular-scale phe-
nomena. Two centers of excellence-the Center for Cataly-
sis and the Center for Biochemical Engineering-further fos-
ter collaborative research.
Chemical engineering draws from various areas and is per-
haps the broadest of all engineering disciplines. Preparing
students to meet today's challenges for energy-efficient pro-
cesses, pollution abatement, and rapid translation of chemi-
cal and biological discoveries into products and processes
requires faculty dedication to research and teaching. This kind
of dedication assures that Berkeley graduates will continue
to contribute to the solution of important societal problems.
Future chemical engineering students will need skills not
even contemplated when the department was founded. The
Berkeley Chemical Engineering Department has been tradi-
tionally at the forefront in new fields and will remain so
through research that challenges the boundaries of contem-
porary science and engineering.
Acknowledgments: We thank all faculty colleagues in the
department and Yvette Subramanian for helping us write this
article. O1

Summer School


Strategies for Effective Teaching

in Chemical Engineering

Northeastern University Boston, MA 02115

My first formal introduction to effective teaching
strategies in chemical engineering was a National
Effective Teaching Institute that I attended in 1991.
It was codirected by Richard Felder, James Stice, and Rebecca
Brent, and has been held every year since 1991 just prior to
the annual meeting of the American Society for Engineering
Education (ASEE). Workshop participants are given the op-
portunity to gain firsthand knowledge of the elements of ef-
fective teaching, including lecturing, active and cooperative
learning, course planning, and evaluation.
Of all the things I learned in the workshop, the most
impactful was that the most widely used teaching method-
lecturing, as traditionally practiced-is actually the least ef-
fective in promoting student learning and that students learn
best by doing. My own experience confirms the numerous
published demonstrations that the standard lecture format is
less effective than number of alternative teaching methods.E,12]
Faculty interest in effective teaching has
increased significantly over the past decade
as a consequence of a variety of factors. Vari-
ous constituencies, including employers, leg-
islators, accrediting bodies, university admin-
istrators, and parents, have sought greater ac-
countability for the educational training of
skilled engineers equipped to solve 21 st cen-
tury engineering problems. The new Accredi-
tation Board for Engineering and Technol-
ogy (ABET) engineering criteria for accredi-
tation that became standard in 2001 require Gilda Barabino
of Chemical En
engineering departments to demonstrate that em university, s
their graduates have a command of engineer- gree from xaviE
ana in chemis
ing fundamentals as well as communication, Rice University
critical thinking, and other lifelong learning ing. she previous
skl, P]Provost for Unc
skills.3P] at Northeastern
Faculty have responded to this increased ing and research
chemical and bi
focus on learning outcomes by re-evaluating reactor design,
how engineering is taught and by seeking engineering

ways to evaluate and improve teaching. It is increasingly rec-
ognized that traditional approaches to teaching often fall short
of achieving the desired outcomes and that there are benefits
in moving from a faculty-centered teaching model to a stu-
dent-centered learning model.[41
Venues for faculty to learn about effective teaching con-
tinue to increase as more workshops dedicated to teaching
are offered by universities, professional organizations, gov-
ernment agencies, and private foundations, as universities
establish or expand teaching and learning centers, and as the
literature on teaching and learning continues to grow. Every
five years, the Chemical Engineering Division of ASEE of-
fers a Summer School for chemical engineering faculty. The
purpose of the Summer School is to disseminate innovative
and effective teaching methods to a wide spectrum of chemi-
cal engineering undergraduate programs.
During the 2002 Summer School, I organized a poster ses-
sion devoted to effective teaching strategies.
It was designed to provide a forum for the
exchange of creative ideas and approaches
to chemical engineering education and for
sharing practices related to effective teach-
ing. The poster presentations that were judged
the best by an expert faculty panel in each of
four categories (lecture-, laboratory-, and
computer-based strategies, and strategies re-
lated to general approaches, learning styles,
or outreach) are presented in this issue of
ociate Professor Chemical Engineering Education.
ng at Northeast-
eivedherBS de- Some strategies for effective teaching that
versity of Louisi-
d her PhD from are represented in the following set of pa-
mical engineer- pers include active and cooperative learning,
ruate Education project-based learning, the use of high school
rsity Her teach- outreach to enhance engineering student
rests are in bio- learning, and the use of a focus topic such as
cal engineering,
llular and tissue racing cars or biochemical/bioprocess engi-
neering to illustrate core chemical engineer-

Chemical Engineerrng Education

is Ass
*he rec
r Uni
try am
in che
usly se
'h inte.
and ce

ing fundamentals in an elective course.
My hope is that the papers presented in this special issue will serve to further
promote the sharing and exchange of ideas that began during the Summer School,
and to disseminate a set of interesting and innovative approaches and best practices
in chemical engineering education. Readers are also encouraged to investigate the
burgeoning body of engineering education literature and resources. Several resources
are recommended below as a starting point for this investigation.

Felder, R.M., and L.K. Silverman, "Learning and Teaching Styles in Engineering Educa-
tion," Eng. Ed., 78(7), 674 (1988) 1988.pdf>
Felder, R.M., D.R. Woods, J.E. Stice, andA. Rugarcia, "The Future of Engineering Educa-
tion. II Teaching Methods that Work," Chem. Eng. Ed., 34(1), 26 (2000)>
Felder, R.M., and R. Brent, "Designing and Teaching Courses to Satisfy the ABET Engi-
neering Criteria," J. Eng. Ed., 92(1), 7 (2003) pers/ABET_Paper_(JEE).pdf>
McKeachie, W.J., McKeachie s Teaching Tips: Strategies, Research, and Theory for Col-
lege and University Teachers, 11th ed., Houghton Mifflin, Boston, MA (2002)
Wankat, P. C., and F. Oreovicz, Teaching Engineering, McGraw-Hill, New York, NY (1993)
Out of print, but available free at and Publications/teaching_engineering>
Wankat, PC., The I 1r, i ,. El/ic ient Professor, Allyn & Bacon, Boston, MA (2002)
Chemical Engineering Education
Journal of Engineering Education
Magazines and Proceedings
ASEE Prism
Proceedings of ASEE Annual Conferences
Web Sites
Richard Felder's home page at < 11i lp !' ii c..I. lii llive_teaching>

Effective teachers are creative and use a range of pedagogies and learning strate-
gies designed to focus on student learning and to consider student learning styles.
They also view students as active participants in the learning process, and view
assessment as a means of driving the learning process. The papers included in this
series represent some examples of effective teaching strategies that have been used
successfully in chemical engineering.

1. Felder, R.M., G.N. Felder, and E.J. Dietz, "A Longitudinal Study of Engineering Student Perfor-
mance and Retention. IV. Instructional Methods and Students Responses to Them," J. Eng. Ed.,
84(4), 361 (1995)
2. Haile, J.M., "Toward Technical Understanding," Part 1. "Brain Structure and Function," Chem.
Eng. Ed., 31(3), 152 (1997); Part 2, "Elementary Levels," Chem. Eng. Ed., 31(4), 214 (1997); Part
3, "Advanced Levels," Chem. Eng. Ed., 32(1), 30 (1998)
3. Felder, R.M., "ABET Criteria 2000: An Exercise in Engineering Problem Solving," Chem. Eng.
Ed., 32(2), 126 (1998)
4. Barr, R.B., and J. Tagg, "From Teaching to Learning: ANew Paradigm for Undergraduate Educa-
tion," in The Social Worlds ofHigher Education: I r Teaching in aNew Century, B.A.
Pescosolido and R. Aminzade, eds, Pine Forge Press, Thousand Oaks, CA(1999) 5

Poster Session Award Paper


"Water Day: An Experiential
Lecture for Fluid Mechanics,"
"Synthesis, Resourcefulness, and
Effective Communication in
Group Learning: Introduction to
Biochemical Engineering,"

General, Learning Styles, Outreach
"Engaging the Imagination of
Students Using Experience
Outside the Classroom: A Course
in Bioprocess Engineering"
Ostafin, et al.
"Incorporating High School
Outreach into Chemical Engineer-
ing Courses,"
Ross and Bayles

"Lab-Based Unit Operations in
Microelectronics Processing,"
( t,,,i.., etal.
"Passing It On: A Laboratory
Structure Encouraging Realistic
Communication and Creative
Experimental Planning,"
Moor and Ferri
"Incorporation of Experimental
Design in the Unit Operations

"Increasing Time Spent on Course
Objectives When Using Computer
Programming to Teach Numerical
"High Performance Engines: Fast
Cars Accelerate Learning,"
Han, et al.

Summer 2003

Summer School


An Experiential Lecture for Fluid Mechanics

University of Tulsa Tulsa, OK 74104

Water Day is essentially playtime with water. Dif-
ferent stations are set up outdoors, and the stu-
dents go from station to station experiencing and
observing different basic fluid-flow phenomena. They see how
water flows from holes in a tank, how a stream narrows as it
falls and how a water stream can knock over a block. These
demonstrations are not like the usual laboratory experiments
with statistical data analysis; the students make only quali-
tative observations or rough estimations to get a feel for
the phenomena.
Students receive the most benefit from the demonstrations
when Water Day is held at the beginning of the semester.
Knowing that the students observed the phenomena during
Water Day, the instructor can refer to the different stations
when the appropriate topic comes up in class. The experi-
ments can be explored quantitatively in class examples or
homework problems throughout the semester. For an extra-
credit assignment at the end of the semester, the students can
be given the original observation sheet with instructions to
name the equation or concept that explains the observations
seen on Water Day. This forces them to review the course
material, which has now been touched upon three different
times over about three months.
If the fluid mechanics class is a spring course, Water Day
can still be held. Instead of being at the beginning of the se-
mester, it usually occurs at the end of the semester when the
weather is warmer. The benefit of being able to refer to the
stations throughout the course is lost, but the students' obser-
vations can be used to review the entire semester.
I hold Water Day during a regular class lecture for my en-
gineering science "Introductory Fluid Mechanics" course,
which is taken by juniors in chemical, mechanical, and pe-
troleum engineering. I encourage student participation with
the promise of a perfect homework score for filling out
an observation sheet, given to them at the beginning of
the demonstration. Although I usually allot an entire lec-
ture to Water Day, it is short enough to be held during
half of a 75-minute lecture.

Water Day is held to provide a minimum experience level
with fluid flow phenomena. It is especially beneficial for the
international students, who make up 10-50% of our fluids
enrollment, since many of them are from arid countries and
did not have the chance to play with water hoses as children.
The demonstrations ensure a minimum experience level for
all students, and the course then builds on this experience.
Water Day also incorporates some pedagogical tools. It
emphasizes sensing, visual, and active learning styles, which
are often neglected in engineering lectures.[1] Holding a Wa-
ter Day and referring to it during the semester appeals to so-
matic understanding, which relies on the sense of touch.'[23]
Somatic understanding can be gained by taking an object
apart, by feeling the force required to deflect a water stream,
and by making the water shoot far with a garden hose. It is
the lowest level of understanding, but it gives the students a
baseline from which they can develop the higher levels of
understanding that we want. Water Day provides an opportu-
nity for somatic understanding, and the rest of the course
works takes this understanding to higher levels.

Some assistants and special materials are necessary for
holding a Water Day. The campus physical plant can help
with hoses and a water supply, and teaching assistants for the
class, or other graduate or senior students, can help man the
stations. My current Water Day stations cover the continuity,
Bernoulli, conservation-of-linear-momentum, and moment-

Laura P. Ford is an Assistant Professor at
the University of Tulsa. She received herBS
from Oklahoma State University and her MS
and PhD from the University of Illinois at Ur-
bana-Champaign, all in chemical engineer-
ing. She does research in chemical vapor
etching of metals and intrinsic bioremediation
of oil and brine spills.

SCopyright ChE Division of ASEE 2003

Chemical Engineerrng Education

oPoster Session Award Paper

of-momentum equations, the vena contract effect, and rela-
tive and absolute velocities. The following supplies are needed
for these stations:
Observation sheets
A tank with holes and plugs
A jug with a hole in the lid
A trash can lid
A book or block
A three-arm spinning sprinkler
The materials are intentionally common and inexpensive. The
experiments are described in detail below, and suggestions
for other fluids topics are given later in this paper.
Tank-With-Holes Station To prepare for this experi-
ment, something must be found to use as a tank. For class-
room demonstrations, an 11-oz coffee can is suitable, and for
Water Day, a 50-L carboy or a 2-L soda bottle will do well.
Punch or drill at least four holes in the tank, keeping the di-
ameter of the hole in proportion to the volume of the tank
and making sure there are corks available to plug the holes.
Some holes should be at the same horizontal level, and some
of them should be at different vertical levels-a five-hole
cross pattern works fine. During Water Day, plug the holes
and fill the tank. For observation purposes, it is best to have
the tank sitting on a cart or table rather than on the ground.
Remove the corks and let the students observe the water flow.
The observation sheet should direct the students to specifi-
cally observe
1. How the water flow from a particular hole changes as the
water level in the tank drops (They should see that the flow
from a particular hole shoots a shorter distance from the
tank as the water level in the tank drops.)

Figure 1. Flow from the vertically-aligned holes in a tank.

Summer 2003

2. How the flows from the horizontal holes are the same or
different (They should see that the water streams from the
horizontal holes shoot the same distance from the tank.)
3. How the flows from the vertical holes are the same or
different (see Figure 1) (They should see that the water
shoots further from lower holes than from the upper holes.
This is true only if the tank is above a critical height from
the ground, which is why the cart or table is useful.)
4. How the diameter of a jet compares to the diameter of the
hole (They should see that the diameter of the jet is smaller
than the diameter of the hole.)
Observations 1 through 3 can be explained by applying the
Bernoulli equation to flow from the surface of the liquid in
the tank to the free jet at the hole. Observation 4 is an ex-
ample of a vena contract. In our course, this material is cov-
ered in the fifth week of class when we reach Section 3.6 of
the textbookA BriefIntroduction to Fluid Mechanics. 4
Variations can be added to the main experiment. Instead of
drilling the horizontal holes all the same size, make each one
different. The students will be able to see that the jets all
shoot the same distance (neglecting viscous effects) but that
the flowrates are different. For another variation, try round-
ing the inside of one of the holes. The students should be
able to see the difference in contraction coefficients between
the square and the well-rounded holes.
Garden-Hose Station All that is needed for this station is
a garden hose with no nozzle. Ask the students to
1. Point the opening of the garden hose upward and note how
far the water travels above the opening
2. Cover part of the opening with a thumb and note how far
the water travels above the opening (see Figure 2, next
3. Experiment with covering the opening and noting the
dilteret distance the water travels.
The students should be able to see that the water travels
further when more of the opening is covered. This can be
explained by assuming constant flowrate and applying both
the continuity and Bernoulli equations. This problem is also
addressed in the fifth week of our course and is covered
explicitly in Problem 8.24 of the text.[41 For a variation of
this station, include a flowmeter and consider the assump-
tion of constant flowrate.
Jug Station For this station, a jug with a hole in the lid is
needed-a one-gallon milk jug works well. Make a quarter-
inch or smaller diameter hole in the lid (inserting a short sec-
tion of a straw in the hole helps make the jet smoother), fill
the jug with water, and turn it upside down, as shown in Fig-
ure 3. As the water drains from the jug, the students can ob-
serve the diameter of the jet as a function of the distance
from the jug. The diameter of the jet should decrease as the
stream moves further from the jug. This can be explained by


Summer School

applying the continuity and Bernoulli equations to two points in the jet. This
material is covered at the same time as in the previous two stations and in
Problem 3.24 of the textbook.14 A variation of this station suggested by a re-
viewer is to change the viscosity of the water by adding sugar. At high viscosi-
ties, the flow should be slower and smoother.

Trash-Can-Lid Station For this station, an assistant, a trash-can lid, and a
garden hose with a nozzle are needed. The student should hold the trash can lid
perpendicular to the ground at all times and should be about 10 feet from the
assistant. Using a nozzle, the assistant should form a narrow stream of water
and aim it at the trash can lid (see Figure 4). The student holding the lid should
be directed to observe the forces required to hold the lid in place as he moves
the lid from perpendicular to the water stream to nearly parallel to the water
stream. It seems to be easier to feel the difference in forces if the lid is kept
perpendicular to the ground at all times (see Figure 5). The student should
notice that more force is required to hold the lid in place when it is perpendicu-
lar to the water stream than when it is parallel. This can be explained by apply-
ing the conservation of linear momentum equation to the water stream. This
material is covered in the eighth week of the course when we reach Section 5.2
of the text.['1 Problem 3.40 in FluidMe-
ci,/J, ..1 addresses the perpendicular
part of this station. A second part on the
nearly parallel case could be added.

Book-or-Block Station A book or
block, a garden hose, and a nozzle are
needed for this station. If a book is used
(telling the students it is their fluid me-
chanics book is particularly effective at
the end of the semester), wrap it well in
plastic bags so it doesn't get wet. Place
the book or block vertically on its short
side and adjust the nozzle to a narrow jet
of water. Direct the students to observe
what happens when they aim the jet at
the book, starting at the bottom of the
book and moving to the top. Nothing
should happen when the stream is at the
bottom, but the book will tip over as the
stream is moved upward. This can be ex-
plained by applying the conservation of
linear momentum equation to the water
stream to find the force of the book on
the water. This force is turned around to
find the force of water on the book, and
a moment balance is done on the book.
This station is considered explicitly in
Problem 5.30 of the text. 14 An interest-
ing variation would be to knock over
books of the same thickness but differ- Figure 2. Water shooting from a
ent weights or blocks of the same size garden hose with the opening
but different densities, partially covered.

Three-Arm-Spinning-Sprinkler Station
For this station, a two- or three-arm spin-
ning sprinkler is needed. Connect it to the gar-
den hose and instruct the students to determine
which direction the nozzles point (clockwise
or counterclockwise), which direction the
sprinkler rotates (clockwise or counterclock-
wise) when the water is turned on, and the ab-
solute and relative velocities of the water jets.
The students should be able to determine that
the sprinkler rotates opposite to the direction

Figure 3. Water flowing from an
upside-down jug.

Chemical Engineerrng Education

Poster Session Award Paper )

the nozzles point and that the relative velocity (rela-
tive to the sprinkler) is normal to the nozzle exit area.
They probably will not be able to determine that the
absolute velocity (relative to a fixed frame of refer-
ence) is radial if the resisting torque is negligible. This
can be explained by the moment-of-momentum equa-
tion, covered in the seventh and eighth weeks of the
course, in Sections 5.1.3 and 5.2.4 of the text.[41 Avaria-
tion would be to add a flowmeter and observe the ro-
tation rate as a function of water flowrate.

Other Stations
Adding pressure gauges at various locations in the
hoses could demonstrate pressure drop due to fittings

Figure 4. Trash-can-lid station. Note that the student
is tipping the lid towards the ground instead of keep-
ing it perpendicular to the ground.

I a. View from the side

Perpendicular Parallel

b. View from above



Figure 5. Sketches of the trash can lid in positions
perpendicular and parallel to the stream. Note that
the plane of the lid remains perpendicular to the
ground at all times.

Summer 2003

and friction. Other stations suggested by reviewers are observation
of delta formation in a muddy or sandy parking lot, the location of
eddies in a creek relative to flow disturbances, and vortex formation
in a draining tub. With latex paint and a paint gun, they could see the
effects of viscosity and particulates on fluid flow. Anything new the
students can see or feel will help their understanding of the material
later. The challenge in developing new stations is keeping them simple,
safe, and inexpensive.

Water Day can be used with groups other than fluid mechanics
classes. The idea for Water Day came from Dr. Christi Luks, who had
done similar experiments with middle-school students at a Young
Scholars Program and Engineering Summer Academy. Water Day
could be used in an Open House or done when recruiting prospective
freshmen. The stations could also be used with a freshman introduc-
tion-to-engineering class to whet their appetites for things they will
understand later. Some of the stations are easily modified for use
indoors, which expands their range of usefulness.

One of the pedagogical goals of Water Day is to include sensing,
visual, and active learning in a lecture course that might not include
them otherwise. Sensing is used in the trash-can-lid station when the
students feel the difference in force, and visual learning is in all the
stations. This watching is not "mindless" since the students must evalu-
ate what they see and record their evaluations on an observation sheet
(e.g., they note that the stream narrows as it falls from the water jug).
Active learning is included in the trash-can-lid and water-hose sta-
tions because the students must do something. Taken together, the
stations add sensing, visual, and active learning to the course.
Another pedagogical goal is to provide an opportunity for somatic
understanding as a baseline to build upon. Water day is experiencing
fluid flow phenomena, so students leave with a minimum level of
somatic understanding.
No formal assessment of Water Day has been done. Nearly 100%
of the students participate since they are encouraged by the promised
homework score. Informal feedback indicates that the students enjoy
it. One comment from a group doing this at the end of a semester
was, "Cool! You can see the vena contracta" One student said that
he liked having the experience to draw on later, i.e., he liked having
the somatic understanding to build on.

1. Felder, R.M., and L.K. Silverman, Journal of Engineering Education, 78, 674
2. Haile, J.M., Chem. Eng. Ed., 34, 48 (2000)
3. Haile, J.M., Chem. Eng. Ed., 34, 138 (2000)
4. Young, D.F., B.R. Munson, and T.H. Okiishi, A BriefIntroduction toFluidMe-
chanics, 2nd ed., John Wiley and Sons, New York, NY (2001)
5. White, F.M., Fluid Mechanics, 5th ed., McGraw Hill, Boston, MA (2003) 5

Summer School



Synthesis, Resourcefulness,

and Effective Communication in Group Learning

The University of Iowa Iowa City, IA 52242

he current nature of engineering requires its practi-
tioners to be comfortable using information from sev-
eral fields to solve problems. Successful engineering
design requires the use of several possible tools to view, de-
cipher, and address problems of interest. Interdisciplinary ef-
forts result in innovative solutions to many complex prob-
lems in the health and environmental sciences. These ap-
proaches also expand the current chemical process industries
to include the most recent advances in fields such as biotech-
nology, nanotechnology, and "green" chemistry. Today's
graduates must be equipped with tools that will enable them
to evaluate complex, open-ended problems using their engi-
neering perspective.
We must encourage students to develop an understanding
of engineering problem solving and expose them to diverse
areas where their skills may be applied. Tools that enable
students to be effective in the global community include 1)
an ability to deconstruct and solve complex problems, 2) an
ability to research and find answers to original questions, and
3) an ability to effectively communicate new discoveries.
These tools are formed by expanding laboratory opportuni-
ties for students, adding open-ended design projects to courses
(including discussions of topical issues in coursework), and
fostering individual investigations.

As students reach their junior and senior undergraduate
years, they are already well-versed in chemical engineering
fundamentals and are beginning to apply them specifically in
unit operations and process design. Critical aspects of learn-
ing can be addressed by immersion into unfamiliar territory
where students are required to apply chemical engineering
principles to solve new and unusual problems. The course
"Introduction to Biochemical Engineering" (IBChE) repre-
sents an opportunity for this kind of engineering synthesis

and design. Synthesis, the assembly of components to form a
new whole, is demonstrated in biochemical engineering as
students design products and processes based on combining
chemical engineering with microbiology, molecular biology,
and biochemistry.
Teaching a biochemical engineering course is not a novel
concept. Several useful textbooks have been written that en-
able students to apply chemical engineering approaches to
analyze, model, and apply biological systems.[1-3] Requiring
biochemical engineering as part of the synthesis courses of
the undergraduate ChE program has proven to be quite use-
ful in meeting several education objectives. The explosion of
information and expansion of molecular techniques in the
biological engineering disciplines have also presented a chal-
lenge for instructors to keep pace and expand course resources
beyond the classical texts to address bioprocess and
bioproduct design on the molecular, cellular, and process lev-
els. The challenge is in defining the key aspects of biological
and biochemical sciences to which students should be exposed.
There is great interest in applying advances in the biologi-
cal sciences to a wide range of interdisciplinary applications.
Current chemical process industries are expanding to include
the most recent advances in biotechnology for chiral mol-
ecule synthesis and environmentally beneficial catalysis.
Biobased feedstocks and biological catalysts are being inte-
grated into industrial processes for the manufacture of foods,
pharmaceuticals, and commodity chemicals. Further, the ad-
vent of genomic information and advanced molecular analy-
sis tools has stimulated great interest in bioinformatics,
biomicroelectronics, and molecular medicine. In this context,
it is difficult to strictly define the breadth of content encom-
Tonya L. Peoples is Associate Professor in Chemical and Biochemical En-
gineering at the University of Iowa. She received her BS from North Caro-
lina State University and her PhD from the Johns Hopkins University, all in
chemical engineering. Her research includes biological and interfacial ca-
talysis with enzymes and microorganisms in extreme environments.
Copyright ChE Division of ASEE 2003

Chemical Engineerrng Education

Poster Session Award Paper

passed in biological engineering. There is great interest in
elevating aspects of biology as an enabling science founda-
tional to chemical engineering. Evidence of this interest is
the number of chemical engineering programs that are pur-
suing research efforts in the biological sciences and are chang-
ing the name of their departments to encompass biomolecularr
engineering." But the mechanism for including biology in
the earlier stages of the chemical engineering curriculum has
yet to be developed.[4]
For the purposes of exposing ChE undergraduates to the
explosion of opportunities in the biological sciences, bio-
chemical engineering is introduced as the application of ChE
principles (including material and energy balances, thenno-
dynamics, transport phenomena, and reaction engineering)
to processes based on living cells or components of cells.
The goal is to meet students where they are in terms of knowl-
edge base and usher them into new territory where their unique
contributions as chemical engineers can be established.
"Introduction to Biochemical Engineering" has been de-
veloped and improved to address critical learning goals for
maturing ChE undergraduates. It also serves graduate stu-
dents seeking exposure to the bioengineering field. In the
course, students define biochemical engineering by doing
research and self-learning about the impact of biotechnology
in the technological, economic, societal, and global contexts.
Open-ended design problems, group problem solving, writ-
ten and oral presentations, and laboratory experiments based
on current research significantly enhance the learning expe-
rience. A typical class includes ChE juniors and seniors, a few
biochemistry undergraduates, and a few first-year graduate stu-
dents from chemical and biochemical engineering and the chem-
istry departments. The diversity of the class participants often
enables an interdisciplinary approach to group learning.
Published prerequisites for the course include organic chem-
istry and math through differential equations. Chemical en-
gineering students are encouraged to take the course after
having chemical reaction engineering and other fundamental
ChE courses. ChE graduate students have taken the prereq-
uisite materials and usually enroll at the same time as the
core graduate ChE courses. Students from chemistry and bio-
chemistry take the course with permission and awareness that
additional review of ChE concepts may be required. Their
experience with core biological and chemical concepts gives
them an advantage in other aspects of the course, however.
The result is a blended course where students with comple-
mentary skills work together to solve open-ended problems.

In order to guide students in the learning process and also
to develop a basis for course assessment, several course learn-

Summer 2003

ing goals were established. They related strongly to the edu-
cational outcomes of the undergraduate chemical engineer-
ing program[5' (see Table 1, next page).
Elements of the course were developed to accomplish
course learning goals. They included
Introducing the biotechnologyfield with guided ..........
web resources to enhance familiarity with field language
Practicing problem-solving skills through in-class group
Providing hands-on learning experience through laboratory
Challenging students to perform a design project on a
topical issue eor h.l W choosing, analyze current technologi-
cal problems, and pose a solution
Assigning written and oral report assignments to communi-
cate learning experiences

Problem-solving skills for maturing engineers evolve from
merely looking up answers to studying situations, reviewing
necessary background information, posing hypotheses, and
testing them to develop appropriate solutions. Students must
learn to find answers to open-ended questions by applying
an iterative design process.[6] This should be done throughout
the curriculum and not be relegated to a capstone course. To
facilitate the process within IBChE, students were assigned
several tasks that required excursions into the world-wide
web. Web links accessible from the IBChE course page]71
guided students to a variety of sites, including National Sci-
ence Foundation Bioengineering and Environmental Systems,
the Swiss Protein Database, and educational links for micro-
biology. Questions about the web information necessitated
that the students review these sites. They took web-based quiz-
zes developed on a password-accessible WEB Course Tools
(WebCT)I1I account page.
For independent study, additional resources were provided
on the IBChE course page in the form of supplementary web
links. These resources included links to writing guidelines,
to special biochemical engineering topics, to chemical reac-
tion engineering topics, and to employment opportunities in
bioprocessing and biotechnology fields.7]1

Within the IBChE course, students have several opportu-
nities to work in groups through in-class problem solving,
laboratory experiments, and design projects. Group sizes are
adjusted depending on the size of the class and the distribu-
tion of graduate versus undergraduate or chemical engineer
versus non-chemical engineer. Class sizes have ranged from
15 to 30 students, and group sizes have ranged from 2 to 5

Summer School

students, depending on the assignment. Groups are assigned IN-CLASS GROUP PROBLEM SOLVING
in a variety of ways, including random selection during class, Solving problems as groups served to encourage students
sign-up sheets for laboratory time slots, and student self-se- to communicate with each other and to work together as teams
election. Typically, all the different methods for group assign- to solve problems. This problem-based active learning mode
ments are used during a given semester.

Course Outcomes Worksheet

Course goals Outcomes4'" Course Activity Basis for Goal Assessment

By the end of the course, the student will be able to identify
enzymes and microorganisms used in bioprocessing technologies
and to describe the basics of biological structure and function.
By the end of the course, the students will be able to choose and
apply the simple models of enzyme kinetics and cell growth that
yield accurate results for the problem under consideration,
including: the Michaelis-Menten equation with and without
competitive inhibition and the Monod growth equation.

By the end of the course, the student will be able to apply
chemical engineering principles of kinetics, transport phenom-
ena, and thermodynamics in the design and analysis of batch
and continuous biochemical reactors

By the end of the course, the student will have participated in
group laboratory and design experiences and will have described
this work in several written reports and one poster presentation.
By the end of the course, the students will have been exposed
to topical issues in biotechnology covering societal and
and economic impacts of new technologies (e.g., genetically
modified foods, environmentally benign manufacturing,
gene therapy).

By the end of the course, the student will have had opportunities
to further his or her professional development through team
experiences, studying ethical issues in biochemical engineering
practice, practicing oral and written communication skills, and
using modem computer tools.


Lectures; in-class examples; web
resources; web quizzes; design project;
lab experiment; exams; homework

a(O);b(O);c(0); Lectures; in-class examples; design
e(*) project; lab experiments; exams;

a(O);c(O);e(O); Lectures; in-class examples; design
k(o) project; lab experiment; exams;

a(O);b(O);c(O); Lab experiments and design project


In-class group issues discussions;
written project reports (including
iteration of introduction, economic
societal impact); ventures in course
web resources.

g(O);h(0);i(O); Ventures in course web resources,
j(o) including employment links; poster
presentations; homework; design

Homework; lab and design reports; web
quiz data; exams

Homework; lab and design reports;

Homework; lab and design reports;

Lab and design reports; report grading
handouts; design description homework;
poster presentation evaluation forms
Design project part A; design description
homework; web-learning assignments

Ethics links in assignments page; home-
work with data analysis; design and lab
reports; web resource page

* denotes moderate contribution to the outcome
* denotes substantial contribution to the outcome

*Letters "a-k" represent desirable program outcomes for student attributes as described by the Accreditation Board for Engineering and and
modified by the University of Iowa,
University of Iowa Chemical Engineering Program Outcomes
a. Each graduate will have the ability to apply knowledge of mathematics, science, and engineering fundamentals.
b. Each graduate will have the ability to design and conduct experiments and to analyze and interpret experimental results.
c. Each graduate will have the ability to design systems, components, or processes to meet specified objectives in chemical engineering.
d. Each graduate will have the ability to work as a member of a multidisciplinary team and have an understanding of team leadership.
e. Each graduate will have the ability to identify, formulate, and solve chemical engineering problems.
f Each graduate will have an understanding of professional and ethical responsibility.
g. Each graduate will have the ability to communicate effectively in written (w), oral (o), and graphical (g) forms.
h. Each graduate will have an education that is supportive of a broad awareness of the diversity of the world and its cultures, and that provides an
understanding of the impact of engineering practice in the global/societal context.
i. Each graduate will recognize the need for and have the ability to engage in lifelong learning.
j. Each graduate will have knowledge of contemporary issues.
k. Each graduate will have the ability to use the techniques, skills, and modem engineering tools necessary for engineering practice.
1. Each graduate will have a thorough grounding in chemistry and a working knowledge of advanced chemistry, including organic and physical and
inorganic analytical, polymer or biochemistry, depending upon their individual educational goals.
m. Each graduate will have a working knowledge of chemical process safety.

Chemical Engineernng Education

also increased student engagement in class activities.[9,10] Class
diversity contributed to unique approaches to defining prob-
lems and making assumptions. Students were given textbook
classical bioprocessing examples to discuss and solve during
class time. They presented their work to the class, answering
key questions about the specific problem. Key questions for
discussion included, "What are the assumptions?" and "What
is the strategy for solving the problem?" Group work was
collected at the end of class and bonus points were awarded
based on student participation.

During the course, students were required to participate in
an enzyme kinetics lab based on the characterization of en-
zyme activity for a thermophilic a-amylase. This enabled
incorporating research efforts from Peeples' lab into the teach-
ing environment. After self-learning about enzymes through
visiting sites such as the Swiss Protein Database,[11 the website
of Novozymes[121 (a major enzyme company), and the websites
of chemical vendors,[131 students learned about activities in
the cloning, isolation, and characterization of thermophilic
enzymes.["41 They were given a sample of amylase enzyme of
unknown characteristics-source was either a commercial
thermostable amylase (Sigma Chemical Co.) or a recombi-
nant hyperthermophilic amylase from Peeples' research lab.
In the most recent course, students were asked to discuss
how they would market such a product. After brainstorming
ideas for enzyme characterization, they were instructed to
collect data about the enzyme and to produce a product in-
formation sheet reporting its characteristics. This required
writing a different type of document than is required in other
laboratory courses. Students opted to study salinity, tempera-
ture, and pH effects as well as kinetic parameters. Working
in groups of four to five, students collected data and reported
their information to the class. Class data were pooled so that
all of the students had all of the analyses. Each individual
wrote a product description, which was then graded.

The aspect of the course that allowed the students to inves-
tigate the field of biochemical engineering and put bioengi-
neering in an approachable context was the design project. It
represents an important element in establishing the relevance
of bioengineering to ChE students and comprises the major
portion of the course grade. Students worked in groups of
two to three on this semester-long project, and working within
different teams from the lab assignment afforded them the
opportunity to improve interpersonal communication.

Poster Session Award Paper )

Students were asked to investigate the economic, social,
and technological issues surrounding a chosen topic in bio-
chemical engineering. This form of case study exposed the
class to more aspects of the discipline than could be covered
in a traditional lecture format. Students completed several
group reports and a poster presentation of their project to fur-
ther widen the exposure of the entire class to a variety of case
studies in biochemical engineering. As design groups pro-
gressed in their research of bioengineering during the course
of the semester, each report served as a mechanism to iterate
through the writing as well as to give feedback on the design
process and focus to the area of bioengineering under study.
The iterative nature of the design reports enabled faculty and
teaching assistants to help students focus on and address im-
portant design considerations by the final report. Students
were encouraged to take advantage of patent information, pro-
cess reviews, and experts in the field through the research of
their chosen topic. Assignments were focused on identifying
the molecular, unit operations-level or process-level improve-
ments. Publications from the College of Engineering Center
for Technical Communications were distributed to help stu-
dents properly reference source material.[15"
The first design assignment consisted of a brief abstract
introducing the topic area. Students were required to choose
an appropriate topic that was manageable within the context
of an introductory course. Over several years, the projects
have included tissue engineering, drug delivery, genetically
modified foods, and gene therapy. A list of the topics that
design students chose in the 2001 course offering is shown in
Table 2. "Design Report A" focused on the economic and
social issues surrounding the development or application of
bioengineering technology, and "Design Report B" required
synthesis and chemical engineering analysis. Students were
asked to perform process or product design and analysis us-
ing biochemical engineering concepts. They were to decide
on the level (molecular, unit operation, or process) where
improvements to the technology could be made.
Given the ability to research and evaluate complex prob-
lems and to use this information in the synthesis of solutions,
graduating engineers must be able to communicate ideas to
diverse constituencies. In addition to the three written design
project assignments, students also presented their work in a
class poster session. Held at the end of the semester, it was a
source of interaction between undergraduates, graduate stu-
dents, and faculty. All were invited to visit the poster session
and learn about the activities of the IBChE class. Students
were required to evaluate other student posters and guests
were also invited to rate poster presentations and to provide
comments. Instructor and guest review, as well as student
reviews, were used to compute the final poster grade.

Summer 2003

Summer School

Course learning goals, student performance, and the con-
tribution of the course to the CBE program outcomes were
assessed through analysis of course elements and class sur-
veys. Based on the annual assessment, improvements were
made to the content and delivery, and reassessment was car-
ried out in subsequent offerings. Assessment tools included
evaluation of student work, WebCT data (hits and statistics),
and student surveys. Student work was assessed by develop-
ing metrics on the grading of all course elements and topic
areas. Achievement levels included "Mastery" (>85%), "In-
termediate Understanding" (>750%), "General Understanding"
(>65%), No\ Ice" (>55%), or "Failure." The entire class
obtained intermediate-to-advanced levels of understanding
across the course.
In 2002, most of the class (93%) completed surveys ad-
dressing the learning goals and course elements. Through the
surveys, students evaluated the priority warranted by, as well
as their achievement of each learning goal. They were asked
if the stated learning goals should be emphasized in the course
and if they achieved those goals. The standard deviation in
student perceptions reveal that there is not much statistical
significance in the differences among some of the course el-
ements. However, lower standard deviations reveal areas
where there was strong student agreement regarding the
course element. Students perceived their performance to be
highest in goals 6 and 2, respectively (see Table 1), and their
lowest areas of achievement to be on goal 3.
These scores appear to be improvements from the previ-
ous course offerings where students were less confident about
their achievement of course learning goals. One reason for
the increase in perceived achievement was the redefinition
of course learning goals. For example, based on previous as-
sessment (2001), "Goal 1" was written without a clear mean-
ing and as a result students were not sure if the goal was
achieved (4.3/6). The 2002 goal was rewritten to use more
clearly active words describing the expectation of student
behavior (see Table 3), and as a result, students expressed
more confidence about the priority of the goal and their
achievement of it (4.8/6)
In addition to surveying students regarding the learning
goals, course elements were assessed by asking them, "What
contributed most to the achievement of the course goals?"
and "What contributed most to student learning?" The re-
sults are given in Table 4.
It should be noted that students perceived Goal 4, the in-
clusion of lab experiences and design projects, as a lower
priority. At the same time, they recognized the group projects
as a very important learning tool, contributing the most (5.00/

6.00) to student learning. They regarded the lab experiment
itself as one of the lowest contributors to student learning
(4.67/6.00). Still, the narrower degree of deviation of re-
sponses reflects a more uniform agreement on the overall
utility of group work. Absolute numbers for student achieve-
ment of course goals appear to improve over the three-year
offering of the course. Within the ChE curriculum at the time
of IBChE, students have typically taken numerous lab-inten-
sive courses, including unit operations laboratory, chemical
process safety with lab, and materials science lab. They have
also had many lab experiences in the various chemistry re-
quirements. Thus, the enzyme lab purely for the sake of of-
fering experience is a lesser contributor. Students also found
it challenging, and in some cases frustrating, to analyze data
collected by the entire class, but the opportunity to push stu-
dents beyond their comfort zone into a different type of writ-
ing and data analysis is an important aspect of the course
delivery. Lessons can be learned about answering questions

Topics for Biochemical Engineering Design Projects
Topic Desripgtion
Anthrax Evaluation of B acillus antrhacis, treatment and
prevention options
Nature Works TMPLA Evaluating the production of Cargill-Dow's
polylactic acid-based polymer
PCR The principles, application, and details of the
polymerase chain reaction process as a new
technology in waste-water treatment
pHBA Biosynthesis Evaluating the production ofp-HBA using
genetically engineered Pseudomonas putida
Poultry Rendering Studying the fermentation of poultry waste to
produce protein meal
Protein Engineering How mutation of the glucoamylase enzyme affects
its binding characteristics
Starch Hydrolysis Studying the effect of high-temperature enzymes on
starches and understanding how the enzyme is used
in the pulp and paper industry
Xanthan Gum Studying the production of the polysaccharide by
the bacterium, Xanthomonas camplestris

Improvement of Course Learning Goal Clarity
2001 Goal 1: By the end of the course, the student will understand basic
concepts regarding the behavior of enzymes and microorgan-
isms used in bioprocessing technologies (e.g., food
processing, pharmaceutical production, environmental
remediation, and chemical synthesis).
2002 Goal 1: By the end of the course, the student will be able to identify
enzymes and microorganisms used in bioprocessing
technologies and to describe the basics of biological structure
and function.

Chemical Engineerrng Education

Poster Session Award Paper

and making decisions with less-than-perfect information.
Perhaps one way to improve student perception of the lab
experience may be to better articulate the specific learning
objectives of the lab project.

Because the course goals were designed for contribution
to program educational outcomes, as reflected in Table 1, the
learning goals could be used to reflect the larger program
goals. IBChE showed strong contributions in specific areas
of communication, problem solving, and lifelong learning skill
development. Based on the design of course elements and
assessment of student performance, the goals of enhancing
student communication and problem-solving skills, and im-
proving laboratory skills and student familiarity with con-
temporary issues were met by improved offerings of IBChE.


While there is much activity in the area of including bio-

Assessment of Learning Goals and Course Elements'

Student Perceptions of Course Learning Goals
(Out of points, 1 low priority, high priority)
2000 2000 2001 2001
Priority Achievement Priority Achievement



2002 2002
Priority Achievement



What contributed most to achievement of course learning goals?
(Out of points, I=low priority, high priority)
2000 2001 20
Group projects 5.1 4.70.80 5.4
In-class problem solving 4.2 4.2+1.4 5.2+
Homework 4.8 4.8+0.56 4.8d
Lab experiment 4.3 4.3+1.2 4.7d
Recall of facts 4.5 4.5+0.87 4.4

What contributed most to student learning ?
(Out of points, l=low priority 6=high priority)
2000 2

Group projects
In-class problem solving
Lab experiment
Recall of facts






Standard deviation from class surveys available for 2001 and 2002, 15 students
2 1. .. I in 2002 to include student evaluation of lectures

logical sciences earlier in the ChE undergraduate curricula,
opportunities exist to use bioengineering as a point of im-
mersion for engineering students. Enabling students to in-
vestigate this unfamiliar discipline and to apply ChE skills in
synthesis can enhance the educational experience. Assess-
ments showed the highest contributor to learning was the
group-learning activities where students investigated and
defined the aspects of biotechnology relevant to their inter-
est. Course activities mapped well to the overall program ob-
jectives by enabling students to analyze problems by using ChE
principles to synthesize solutions to problems in a new disci-
pline. Further, using a variety of additional course elements,
including in-class problem solving, lab experiments, and tradi-
tional lectures, homework, and tests could be integrated with
the independent learning process to facilitate development of
critical skills for a lifetime of engineering exploration.


1. Blanch, H.W., and D.S. Clark, Biochemical Engineering,
Marcel Dekker, New York, NY (1996)
2. Shuler, M., and F. Kargi, Bioprocess Engineering, Prentice-
Hall, Upper Saddle River, NJ (2002)
3. Bailey, J.E., and D.F. Ollis, Biochemical Engineering Funda-
mentals, 2nd ed., McGraw-Hill, New York, NY (1986)
4. Armstrong, Robert, "Frontiers in Chemical Engineering Edu-
cation: New Directions and Opportunities Creating the Fu-
ture," presented at the NSF/Chemical Engineering Curricu-
lum Revitalization Workshop, Orlando, FL (2002)
5. Accreditation Board for Engineering and Technology, Inc.,
"Criteria for Evaluating Engineering-Related Programs,"
last modified 02/
03/2003, accessed 02/15/2003
6. Edie, A.R., R.D. Jenison, L.H. Mashaw, and L.L. Northrup,
Engineering Fundamentals and Problem Solving, 4th ed.,
McGraw Hill, New York, NY (2002)
7. Peeples, T.L., Introduction to Biochemical Engineering,
, last updated 12.2002 (2002)
8. WebCT, Faculty Resources, , ac-
cessed 2.15.03 (2001)
9. Woods, D.R., Problem Based Learning: Helping Your Students
Gain theMostfrom PBL, 3rd ed., Waterdown, On, Canada (1996)
10. Felder, R.M., "Active and Cooperative Learning,"
html> accessed 02/15/03 (2003)
11. Bairoch. A., "The ENZYME Database in 2000," Nucleic Ac-
ids Res., 28, 304 (2000)
12. Novozymes, "Enzyme Solutions" and "Discover Enzymes,"
, accessed 9/27/01 (2001)
13. Sigma-Aldrich, "Biochemicals and Reagents," Interest/Biochemi-
calsReagents.html> accessed 2/15/2003 (2003)
14. Kim, J-W, L. Flowers, M. Whiteley, and T.L. Peeples, "Bio-
chemical Confirmation of the Family 57-Like a-amylase of
Methanococcus jannaschn," Folia Microbiologica, 46, 467
15. Coffel, Scott, "Engineering Communications Guides," Center
for Technical Communications,>, last updated 3/2003, accessed
10/2002 7

Summer 2003

Summer School



Engaging the Imagination of Students

Using Experiences Outside the Classroom


University ofNotre Dame Notre Dame, IN 46556

Bioprocess Engineering has the potential to become
one of the leading chemical engineering fields due
to the explosive discoveries that are occurring and
continue to occur in the field of life sciences. Since the rate
of progress is so accelerated in this field, we must train future
engineers to be willing to go beyond the conventional class-
room education with initiative, motivation, and creativity.
We have developed a course that engages students' imagi-
nations and participation outside the classroom, that offers
real-life experiences through field trips, and that provides a
chance for students to express their enthusiasm by publish-
ing their own newsletter. Through these activities, students
find that although life sciences is not an exact science, it is
filled with tremendous need for challenging and novel con-
ceptual development, which can be exciting to pursue through
quantitative numerical analyses and expression.
In the chemical engineering department at the Notre Dame
University, which has no formal bioengineering degree pro-
gram, the second-semester senior/graduate-level bioprocess
engineering course attracts students with various backgrounds.
By the time they have completed the course, some will have

established a strong foundation for embarking on graduate
research, while others will have acquired sufficient practical
knowledge to become entry-level bioprocessing engineers.
To fulfill the educational goals outlined in the ABET as-
sessment criteria, this course introduces essential biology,
biotechnology, and computational concepts while providing
an environment in which the student can improve their skills
in design, analysis, problem solving, teamwork, and com-
munication. In order to develop a fast-paced, attractive, and
self-motivated course, the leading roles for learning are shared
between the students and the instructor. The students attend
traditional lectures on the fundamentals of bioprocessing and
are then asked to take the initiative in using the laboratory
learning modules, in organizing a field trip to a local
bioprocessing factory, and in publishing a newsletter on
bioprocess engineering. They are evaluated based on their
performance in all of these aspects, and the course is improved
by assessing the results of a standard student Course Evalua-
tion and additional questionnaires.


Recent advancements in molecular biology demand a new
Agnes Ostafin is Assistant Professor of Chemical and Biomolecular En- way of thinking in bioprocess engineering. During the last
gineering at the University of Notre Dame. She obtained her PhD at the decade, explosive amounts of information have become avail-
University of Minnesota (1994) and her BS degrees in Chemistry and Bio- able in the life sciences and it has become almost impossible
logical Sciences from Wayne State University. Her research interests in-
volve the synthesis and characterization of nanoscale materials for bio- for engineering students to know all there is to know about
logical and biomedical applications and application of photosynthetic bac- the field. In order to introduce such a wealth of information
teria for use in industrial waste treatment and environmental remediation.
Darcy LaClair is a chemical engineering PhD candidate at the University in one semester to students with little biological background,
of Notre Dame, where she studies the bioregenerative capabilities of purple while maintaining an emphasis on traditional bioprocessing,
photosynthetic bacteria. She received herBS in chemical engineering at there must be a carefully crafted set of course objectives. We
Northeastern University where she also had the opportunityfor two years there must be a carefully crafted set of course objectives. We
of on-the-job internship experience with a variety of companies. have identified the following goals:
Hartley T. Schmidt is a chemical engineering PhD candidate at the Uni-
versity of Notre Dame. He received his BS in chemical engineering at Students should know the structures, natural habitat,
Florida Institute of Technology in 2000. He is currently working on nano- and industrial use of various microbes. Since most of
scale drug delivery systems using calcium phosphate ceramic capsules
as carriers for biomolecules. the bioprocessing topics discussed in class will be
Copyright ChE Division of ASEE 2003
'80 Chemical Engineerring Education

Poster Session Award Paper

Course Syllabus

Goals By the end of this course, you should be able to
Identify basic structure and types of microbes found in nature
and used in industry
Culture microbes-sterile method
Design and operate bioreactors
Harvest and characterize microbial products
Describe intercellular function and structure models of cell
Describe the strategy of genetic engineering and molecular
cloning to improve product yields
Describe metabolic engineering with solid understanding of
enzyme kinetics and its historic evolution
Communicate with the others and work as a team
Text Required: none
Supplemental Reading: Hand-Out Materials
Class Tuesdays and Thursdays 9:30 10:45
Instructor Agnes E. Ostafin, Assistant Professor
182B Fitzpatrick Hall
Phone 631-3798, e-mail:
Office hours: By appointment
Grades Grades will be based on your performance in
Take-home exam (30%)
Homework problems, including interactive exercises with
reports (55%)
Field trip with a follow-up lab (5%)
Participation in publishing Bioeng. Newsletter (10%)
Course Outline I Introduction toBioprocessing: Concept through Application
II Traditional Bioprocessing
a. Biological cells, bioenergetics, and cell metabolism
b. Microbial growth and modeling
c. Bioreactor design
d. Recovery and Purification of products
III Enzymes andApplications
IV Contemporary Bioprocessing
a. Genetic engineering
b. Metabolic engineering
c. Nonconventional systems

Source Textbooks Used in Course
Bioprocess Engineering, W. Veith; Wiley Interscience, New York, NY
Bioreaction Engineering Principles, J. Nielsen and J. Villadsen; Plenum
Press, New York, NY (1994)
Cell andMolecular Biology, 2nd ed., G. Karp; Wiley, New York, NY
Biochemical Engineering Fundamentals, J. Bailey and D. Ollis,
McGraw-Hill, New York, NY (1986)
Bioprocess Engineering Principles, P. Doran; Academic Press, San
Diego, CA (1995) (Out of print)
Fundamental Laboratory Approaches for Biochemistry and Biotechnol-
ogy, A. Ninfa and D. Ballou; Fitzgerald Science Press, Bethesda, MD
Bioprocess Engineering, 2nd ed., M. Shuler and F. Kargi, Prentice Hall
International Series in Physical and Chemical Sciences, Englewood
Cliffs, NJ (2002)

related to microbes, students should have a good
understanding of the subject area.
Students should have a good grasp of the
analysis of microbial cultures and be able to
describe quantitatively their growth and scale-up.
Students should be able to design bioreactors for
different microbes when provided with sufficient
knowledge about the growth culture supply, gas
supply, release of waste products, and heat
Students should be familiar with microbial
product recovery and purification. They need to
know the common methods used in product
Students should be made aware of genetic
engineering methods to design microbes to
achieve improved product yield.
Students should have some laboratory experience
in microbial culturing and enzyme kinetics with
.ullttc ielnt computational, statistical, large-scale
bioprocessing, and error analysis skills.
* Students should learn to work as a team, improve
their communication skills, and be aware of the
outside activities related to their fields.
The course syllabus that reflects such goals is shown in
Table 1, and a list of useful texts is given in Table 2.
In order to accomplish these goals, two general postu-
lates were introduced to set the direction of the lecture
section of the course. First, biological phenomena and
concepts were deductively introduced in the class.[1,21 By
approaching the subject from organisms to molecules, it
was possible to focus continually on the theme of
bioprocess engineering, while introducing the subjects
of molecular and cellular events as needed. Future engi-
neers must be trained to direct process development and
to pioneer new discoveries in biology and genetics by
bringing traditional engineering functions to bear in this
novel field. Bioprocessing in a cell can be described in
terms of various linked cellular "unit operations" (or
ganelles) in which internal functions are driven by mo-
lecular-scale bioprocessing units (proteins and nucle-
otides) within the cell. This approach differs from the
order of presentation typical in many bioprocess engi-
neering textbooks, which start with an in-depth descrip-
tion of the chemical basis of biological systems and en-
zyme kinetics before introducing the cell, growth mod-
els, and unit operations. The choice between the two ap-
proaches may well be decided by the number of credit
hours students are allowed to take in biology.

Summer 2003

Summer School

The second postulate used in course development was that
while traditional bioprocess engineering addresses the issues
of development and optimization of organism growth and
product recovery, contemporary bioprocess engineering must
also address the issues of engineering new organisms to better
suit the purpose. The impact of recent advances in genomics
and proteomics is quickly coming to bioprocess engineering
and must be usedby bioprocess engineers. For this, discussions
of proteins, DNA, enzyme kinetics, and other biochemical pro-
cesses both inside and outside the cell are critical.

While classroom instruction continued, the students initi-
ated activities outside the classroom to enhance their under-
standing of concepts and their knowledge of real-world ap-
plications. Active participation of the students made this
course a more exciting learning experience.
Take-Home Examination Atake-home examination was
given about midway into the semester to test the students on
their ability to model and design fermentation processes. They
should show good grasp of microorganisms and their regu-
lated growth. Students appreciated the opportunity to express
their knowledge in a relaxed environment.
Regular Homework Traditional problem-solving home-
work exercises were assigned to develop and test student
comprehension of the various analytical approaches used in
this field.
Interactive Homework In addition to traditional prob-
lem-solving homework, interactive modules were developed
in order to engage student participation in the learning pro-
cess. Rather than providing students with a set of numbers to
plug into equations and submitting the computational results
as homework, we arranged opportunities for them to find these
numbers experimentally. We hoped the students would have
a better feel for the significance and the range of errors that
come with these numbers and how they would affect the final
outcome if they were provided with interactive homework.
Relying on the laboratory instruction provided, students
performed two laboratory experiments outside the regular
classroom hours to generate data. These data were subse-
quently used to answer the homework problems. Students
were divided into groups of 3-4 team members and used the
learning laboratory modules. Each module was a mobile unit
equipped with analytical instruments to conduct a specific
laboratory experiment. Consulting with the graduate teach-
ing assistants, students planned and set up the experiments
by themselves, arranging their work schedule as demanded
by the experimental procedures. With proper management of
student schedules and the learning modules, this flexible ar-

rangement could accommodate 22 students to complete the
two experiments described below before the end of the se-
mester. A high degree of student initiative was the key to the
success of this arrangement.
Growth Analysis of E. coli: Students had already been
exposed to the growth analysis of bacteria in the lecture. The
objectives of this experiment was to 1) learn the use of a pH
meter, a spectrophotometer, and a centrifuge, 2) learn sterile
techniques, 3) numerically analyze growth of bacteria, and
4) determine the mass transfer of glucose to the organism.
Using a well-established methodN31 in microbiology, the teach-
ing assistant provided the original colony of E. coli to the
students and they cultured bacteria in Luria-Bertani liquid
culture mediumI41 while intermittently determining the growth
spectrophotometrically. A similar experiment was conducted
with glucose, observing the rate of its uptake with Benedict
solution.15] Students reported quantitative kinetic analyses with
degrees of errors on the lag phase, exponential growth phase,
declining phase, stationary phase, and death phase during bac-
terial culture. The rate of transfer of glucose to the organisms
was also reported. Most of the teams were able to obtain reli-
able data, but some failed and had to borrow data from col-
leagues. This highlighted the need to perform careful, well-
recorded experiments and encouraged students to work to-
gether to achieve the long-term goal. The entire process took
a little over two weeks, during which the students expended
a similar number of hours as they would have spent doing
regular homework.
Collision and Diffusion Limited Enzyme Catalysis:
Through this experiment the students also become familiar
with the use of a pH meter and spectrophotometer. In addi-
tion, they learn how to run a LabView module, to understand
the importance of pH in enzyme kinetics, to construct the
three basic plots used in enzyme analysis (Lineweaver-Burk
Plot, Eadie-Hofstee Plot, and Hanes Plot), and to understand
the effect of collision frequency and kinetic energy of colli-
sion on the rate of enzyme catalysis. After becoming familiar
with the conventional enzyme kinetic study in a buffer, the
students performed a similar experiment in glycerol to mimic
real-world bioprocessing where diffusion-limited reactions
would take place in non-Newtonian fluids. Since all the re-
agents were prepared in advance, well-coordinated teams
completed the experiments in two sessions of 2-3 hours.
First, students carried out the enzyme kinetics analysis of
urease in water, using urea as the substrate, to understand the
importance of buffered solution for enzyme kinetics. The rest
of the experiments were conducted in HEPES buffered solu-
tion at pH 9.0. The rates of reactions were determined using
Berthelot color assayN61 and the concentrations of the enzyme
and substrate were adjusted to obtain the constants for the

Chemical Engineerrng Education

Michaelis-Menton model at room temperature. This was a
relatively simple system and the students were able to obtain
data with a relatively small range of standard deviations.
In the second part, similar experiments were performed in
a different concentration of glycerol to demonstrate the ef-
fect of increasing viscosity of the solvent on the enzyme re-
action. The students were asked to explain the significance
of high viscosity solvent on enzyme kinetics and its effect on
the various constants.
The students considered this type of homework as some-
what taxing, but in general appreciated the firsthand experi-
ence in wet biochemical experiments.
Real-Life Bioprocess Manufacturing and Laboratory
Practice About halfway into the semester, a mandatory 3-
hour tour to a local microbrewery was scheduled.[7] The stu-
dents experienced scaling up of laboratory procedures, inte-
gration of chemical and mechanical engineering in flows of
liquids and a number of regulatory systems, and communi-
cation between the administrative personnel, engineers, and
technicians. The students embarked on a small-team project
to brew three types of bottled beer, starting with the kits
purchased from the brewery. Brewing took up to four
weeks, and the grade was judged by student participation
as recorded by the TA. In general, the excitement level
for this module was highest, and an increase in registra-
tion was noted in the following year, with specific requests
for a repeat of this module.
Publication ofBioengineering Newsletter A major se-
mester-long team project, the Bioengineering Newsletter, was
assigned as the final project for the course. It provided an
opportunity for the students to practice long-term team-based
project execution and negotiation skills, to use written and
verbal communication skills, to develop better literature re-
view techniques, and to become aware of current events in
the area of bioprocessing. The entire class of 22 individuals
comprised the publication staff of the newsletter. Class mem-
bers were assigned specific roles and responsibility levels,
such as editor, writing team, design and layout team, finan-
cial team, and printing and distribution teams. Each respon-
sibility level was given a list of duties, and team leaders were
told to report regularly on their progress to their immediate
senior in the publishing-team hierarchy. General guidelines
and a timeline were provided, and the team leaders were asked
to set and enforce deadlines for their own teams. The goal
was to produce the final publication on time.
In order to keep some control over the progress of the project
throughout the semester, each team leader was expected to
provide copies of their regular reports to the instructor and to
the editor, indicating the accomplishments of their team for

Poster Session Award Paper

the preceding reporting period and reporting any difficulties
that may have arisen. The editor was asked to provide a final
comprehensive report on the project on the activities of each
team and suggestions for the future. The concept of multi-
layered reporting within a large team, although realistic, is
sometimes difficult to achieve in the classroom since some
students may be reluctant to assess the performance of their
peers or to pay strict attention to deadlines. When a very de-
tailed calendar of events was provided, there was more com-
pliance with the structure.
Each student was asked to write a one-page single-spaced
article with references and properly sourced graphics relat-
ing to a subject relevant to bioprocess engineering. The writ-
ing team was to provide a specific theme for that year's news-
letter along with a list of suggested topics for the students to
write about. Students were free to choose any topic, but if it
was not on the list, they had to obtain approval from the writ-
ing team. Eventually, students reached a consensus and in-
structor intervention was minimal.
The design and layout team was responsible for creating
the electronic publication using Microsoft Publisher. This
group defined the manuscript size and graphical elements,
and provided a mock-up to the editor, who proofread the copy
before printing. The financial team raised money to cover
the cost of printing by soliciting donations from student en-
gineering organizations on campus, such as AIChE, SWE,
MEP, and various departments. Excess funds were used to
provide pizza and drinks for the inevitable last-minute "push
to publish" just before finals week. Each member of the class
and each financial donor received a paper copy, and five
went to the instructor. Distribution of printed newsletters
takes place during finals week. Wider distribution was
achieved through website dissemination at>.


Students were evaluated based on
Their ability to comprehend and express bioprocess
engineering subjects quantitatively: take-home
examination (30%)
Problem-solving homework (30%)
Taking initiative in conducting two experiments and
the ability to analyze the results quantitatively:
Interactive homework (25%)
Ability to apply knowledge to the real world: site visit
and follow-up experiments (5%)
Communication skills and ability to work as a team:
Continued on page 207.

Summer 2003

Summer School



University of Maryland, Baltimore County Baltimore, MD 21250

Three years ago, an undergraduate-level Introduction
to Biomedical Engineering elective was created at the
University of Maryland, Baltimore County (UMBC).
One goal of the course was to include a group project that
allowed an opportunity for the students to delve deeply into
an area of interest that was not covered in the lecture mate-
rial. A second objective was to provide a forum for the stu-
dents to hone their presentation and group interaction skills.
We decided to try something new and different that would
challenge the students in a way they had not yet experienced.
The idea was to integrate a high school outreach presentation
into the course that required the students to participate in prob-
lem-based learning. In addition, using the high school setting
for the presentation challenged the undergraduates to present
highly technical material to an audience with little technical
knowledge. We believe this latter skill is critical for success
in the "real world." Based on the success of the outreach
projects in the elective course, the idea was also tailored to a
required course in heat and mass transfer.

While the content of the projects varies between the two
classes, many aspects of the exercise are the same. The pre-
sentations are planned for a high school class period that lasts
for approximately 45 minutes. In most cases, presentations
are given to an upper-level science or math class. Typically, a
segment of the presentation is in a lecture format to intro-
duce students to the topic. Interactive components are criti-
cal to the success of the presentations, however, so lecture
time is kept to a minimum. Exercises to get the high school
students involved are an important part of the project and
have encompassed such ideas as games ("Who Wants to be a
Biomedical Engineer?", "Engineering Jeopardy"), skits ("The
Doctor, the Scientist, and the Engineer"), demonstrations
(geyser eruption, air freshener diffusion, conduction and con-
vection experiments, electroplating), and puzzles. Student
participation is encouraged by using small treats (approved

by the teacher) for those who volunteer answers or partici-
pate in activities in front of the class.
At the end of the presentation, the group members give the
high school students information on their backgrounds, how
they chose engineering as a career, and any summer indus-
trial or research experiences they have had. The undergradu-
ates also describe their future plans, if known. The high school
students are then given time to ask career-related questions.
Before concluding, a survey is performed to help the group
determine the effectiveness of its presentation. Finally, infor-
mation related to the presentation topic is left in the class-
room for any student who wants to learn more. Examples of
"leave behinds" are bookmarks with definitions and web ad-
dresses, a bulletin board that stays up in the classroom, infor-
mational CDs, and web sites with links related to the projects.

In each class, the projects carry significant weight in the
final course grade (15% in Heat and Mass Transfer and 40%
in Introduction to Biomedical Engineering). The projects last

Julia M. Ross is an Associate Professor of
Chemical and Biochemical Engineering at
UMBC. She has a BS degree from Purdue
University and a PhD from Rice University,
both in chemical engineering. Her research
interests are in the area of cellular engineer-
ing. In particular, her work focuses on bacte-
rial adhesion to physiological surfaces.

Taryn Bayles is a Visiting Lecturer of Chemi-
cal and Biochemical Engineering at UMBC.
She has spent half of her career working in
industry and the otherhalf teaching engineer-
ing. She received her BSChE from New
Mexico State University, and her MSChE,
MSPetE, and PhD in chemical engineering
from the University of Pittsburgh. Her re-
search interests focus on engineering edu-
cation and outreach.
Copyright ChE Division of ASEE 2003

Chemical Engineerrng Education

for a full semester and points are distributed throughout the
semester for reaching milestones leading up to the actual
outreach presentation. Groups are chosen by the students
within the first week of class, and project topics are assigned.
Groups contain a minimum of three students and a maximum
of four, depending on the class size (we have integrated these
projects into classes with up to 25 students). We found it very
important for each group to have at least one member who
attended high school in the U.S, and it was also helpful if at
least one group member had a car. All group members are
required to spend equal time leading the presentation.
In the Introduction to Biomedical Engineering course, top-
ics that are current or emerging research areas in which chemi-
cal engineers participate are chosen. Examples include tissue
engineering, bioinformatics, biosensors, metabolic engineer-
ing, and controlled release. In this case, the projects are an
exercise in problem-based learning because the outreach top-
ics are not covered in the course content.
In the Heat and Mass Transfer course, the projects must
include demonstrations and activities that demonstrate the
fundamentals of transport phenomena. Part of one class pe-
riod at UMB C is taken to demonstrate what the projects could
include. The following are example demonstrations.

[ The session is begun by asking the students to discuss how
they can tell when air pressure changes. (Possible answers may
include flying in an airplane, going up a mountain and feeling
your ears pop, etc.)
[J The students are divided into pairs and given a straw, two pieces
of string, and two balloons. Ask them to blow up the balloons
to equal sizes and tie them to the ends of the strings. The trail-
ing ends of the strings should then be tied to the straw, so that
the two balloons are near each other on the straw, but not touch-
ing. One student in the pair will be told to hold the straw and
the other student will blow so that his/her breath goes directly
between the balloons. Ask the students to predict what will
happen. (The students usually expect the balloons to separate
further, but the opposite is true...the balloons move together!
Blowing between the balloons creates a stream of air that is
moving faster than the surrounding air The pressure between
the balloons is lower than the pressure of the air surrounding
them, so they come together)
[J The students are given a clear plastic cup filled with water,
along with two straws (one 3-inch and one 4-inch straw), and
are challenged to blow across the top of a straw in the water
and cause water to come out of it. (In order for the water to
spray (- .i -Fil ed straw that is in the water, the students must put
the 3-inch straw into the water in such a way that it does not
touch the bottom or hold it near the opening and perpendicular to the top of the
straw that is being held in the water Blowing through the longer
straw across the top of the immersed straw causes a spray of
water to come out o< ,il straw in the water The faster the blast

Poster Session Award Paper )

of moving air from the student's mouth, the lower the pressure
exerted around the blast. When the blast flows across the top of
the lower straw, the pressure on the liquid in that straw is re-
duced, and since elsewhere on the water s surface the air pres-
sure is atmospheric, the water is pushed upward through the
upright straw toward the low-pressure area. When the water
gets into the air stream, it is pushed and scattered by the quickly
moving air)
EN Give the students a Ping-Pong ball, two small paper cups (2 to
3 inches tall), and some masking tape. Ask them to tape the
cups onto the top of a table, one behind the other, four inches
apart. Put the Ping-Pong ball in the first cup and challenge the
students to get it out of the first cup and into the second cup-
neither the cups nor the ball can be touched by any solid or
liquid. (Blowing across the top of the cup will produce lift on
the ball and it will pop up out of the cup when the air speed is
high enough. Controlling the ball's motion to get it to land in
the second cup is not easy!)
EN Give the students a round piece of cereal and a bendable drink-
ing straw and challenge them to get the cereal to hover above
the straw. (Bend the straw to a right angle and insert it into the
mouth like a pipe. Take a deep breath and, with the cereal held
above the straw opening, gently and with control, blow air at
the cereal. Let go of the cereal once you start blowing. This
takes practice! The cereal will stay roughly at the same height
because forces upward [the air stream] and downward [grav-
ity] on the cereal are balanced. This is more easily demon-
strated using toy blow pipes or with a hair dryer and a Ping-
Pong ball the angle of the air dryer can be varied to further
demonstrate the effect of velocity, air pressure, and gravity.)
These are just a few examples of hands-on experiments
that can be used to demonstrate Bernoulli's principle (which
the heat and mass transfer students have studied the previous
semester in fluid mechanics). Then, the students are encour-
aged to explain what math and science skills they learned in
high school to study Bernoulli's principle and how it relates
to industrial applications (design of pumps, airplanes, etc.)
The first milestone of a project is to find a high school
classroom that is within 30 miles of campus for the presenta-
tion (5% of the project grade). In general, the UMBC stu-
dents find schools and classes on their own, with some choos-
ing to return to a group member's own high school. Other
common choices include schools that are near UMBC or that
target a specific demographic (such as an all-girls school or a
predominantly minority inner-city school). To date, presen-
tations have been made in six different counties in the greater
Baltimore area. A compiled list is kept of classes we have
previously visited, along with contact information to be given
to groups that encounter difficulties finding a venue for the
To meet the first milestone, groups must turn in the name
of the high school, the course name and time, the teacher
name and contact information, the number of students en-

Summer 2003

Summer School

rolled, and the grade level of the students. We have found high school
teachers to be quite accommodating and enthusiastic about our par-
ticipation in their classrooms.
Within the first month of the semester, the groups are required to
meet with the professor for a preliminary meeting (5% of the project
grade). At this meeting, the group must present preliminary ideas for
the presentation and have a general idea of the information they will
present. Each group must turn in one typed page outlining their ideas.
We have found this initial part of the project to be very challenging
for the students because it requires significant independent reading
and synthesis of information. The topics assigned are purposefully
broad so that this step requires a significant effort.
At this point, students must make choices as to what material would
be both interesting to high school students and at an appropriate level.
Approximately two weeks later, the students are required to have a
second meeting with the professor (5% of the project grade) to present
a full outline of the project presentation, including specific details
about what material will be presented and how it will be presented.
The plan at this point must incorporate feedback given after the pre-
liminary meeting.
Approximately two weeks before the outreach presentation takes
place, at least one member of the group is required to visit the school
classroom to meet the teacher, verify the date and time for the project,
and to assess if the room is adequate for the planned exercise (5% of
project grade). We have found this step to be critical for a successful
outreach project. It was added after the first year of the projects
when we experienced several classrooms that were not adequate or
appropriate for our presentation. Examples of classrooms that were
inadequate or inappropriate included classrooms without screens for
the projection of the presentation, open classrooms that made hear-
ing difficult, lack of seating for all the observers, and lack of lab
space for the demonstration.
At least two weeks before the presentation date, each group is re-
quired to present the outreach project to the professor (5% of the
project grade). This presentation is expected to mimic exactly what
is planned for the high school classroom, to be carefully timed, and
to include any props, games, and handout material. The groups must
also turn in the "leave behind" information that has been developed,
along with a copy of the student and teacher surveys they plan to use
for assessment. This meeting typically results in significant feed-
back on content level, clarity of explanations, and quality of presen-
tation materials (overheads, slides, etc.). In addition, the meeting com-
monly identifies any problems that arise with the organization of
planned activities or games. At the end of the practice presentation,
groups are either given permission to go ahead with their presenta-
tion or are required to practice again in front of the professor.
The outreach presentations are typically held within the last sev-
eral weeks of the semester and account for 60% of the project grade.
The UMBC students are asked to dress in typical "everyday" clothes
for the presentation to help the high school students relate to them. In

the Introduction to Biomedical Engineering course, one
lecture is cancelled for each outreach presentation.
In addition to performing an outreach project, all stu-
dents are required to attend one other outreach presen-
tation as observers. We have found this to be especially
helpful for students who are nervous about presenting
in front of a group or who did not attend high school in
the United States and are unsure of what to expect. The
student groups are solely responsible for running the
project presentation. The professor does not participate,
instead observes and evaluates the presentation. See
Table 1 for a typical evaluation sheet.
Assessment is a vital aspect of the project (15% of
grade). Each group must collect survey information from

Sample Outreach Project Evaluation

Group Members



# High School Students in Attendance

Poor 1 2 3 4 5 Excellent
Clarity of speech overall 1 2 3 4 5
1 2 3 4 5
1 2 3 4 5

Volume of speech overall
1 2 3 4 5

1 2 3 4 5
Organization 1 2 3 4 5
Flow 1 2 3 4 5
Group participation 1 2 3 4 5
Student participation 1 2 3 4 5
Slides/Overheads 1 2 3 4 5 N/A
Game 1 2 3 4 5 N/A
"Leave behind" 1 2 3 4 5
Professionalism 1 2 3 4 5
Ability to relate to audience 1 2 3 4 5
Enthusiasm 1 2 3 4 5
Technical content
Technically correct 1 2 3 4 5
Clearly explained 1 2 3 4 5
Appropriate to topic 1 2 3 4 5
Appropriate level 1 2 3 4 5
Terminology 1 2 3 4 5
Held student interest 1 2 3 4 5

Chemical Engineerrng Education

both the high school teacher and the students. All teacher
surveys must include the question "Would you be willing to
have another group perform an outreach project in your class
in future years?" The response to that question is turned in to
the professor for use in building a list of contacts for groups
in future years that may have difficulty finding a classroom.
The remainder of the survey is not turned in, but rather
serves as feedback for the group. Information from the sur-
vey is expected to be incorporated into the group's self-as-
sessment statement to provide support for conclusions that
are drawn. The self-assessment statement is a critical assess-
ment of the efficacy of the outreach project (three typed pages
maximum, one statement per group, with signatures of each
group member). The statement must address several issues:
What was done well? What could have been done better?
Was the presentation effective? How do you know? What
would you do differently if you were to do it again? How
well did your group function together? How could you im-
prove the way your group functioned? What overall grade
would you give yourself?
Points are distributed for "accuracy" of the assessment as
judged by the professor and for writing clarity (not for strength
of the project itself). For example, a project with major diffi-
culties that were accurately assessed would receive a low
project score, but a high score for self-assessment. We be-
lieve this element of the project is critical because it forces
the students to begin self-evaluation. The professor's evalua-
tion of the project is given back to the students after the self-
assessment is turned in so the students can gauge how well
they identified their own strengths and weaknesses.
We have estimated that we spend approximately 2 to 3 hours
per group in order to implement this program. This time in-
vestment has certainly been worthwhile considering the posi-
tive feedback we have received from the high school teach-
ers and students and the UMBC participants.

We have witnessed several desirable outcomes from our
outreach presentations. In nearly every case, the level of en-
thusiasm of UMBC students has been very high. They have
been highly motivated to put on an interesting presentation
and participation within the groups has been surprisingly
uniform. We have witnessed only a few cases where one or
two group members performed all the work. To the contrary,
since each member is required to spend an equal amount of
time leading the presentation, in most cases all group mem-
bers have contributed equally to the project. The comments
relating to the projects on the end-of-semester course evalu-
ations have been overwhelmingly positive.
The initial reaction to the project at the beginning of the

Poster Session Award Paper )

semester has typically been, "This will be really easy," fol-
lowed shortly by, "This is going to be really hard!" In par-
ticular, the students have been challenged by learning a new
topic on their own, narrowing the topic, and choosing what
material to present. In addition, it has been challenging for
them to determine the level of their audience and to tailor a
presentation to that level. While these skills can be critical in
the "real world," they are often overlooked in the undergradu-
ate experience. We believe the outreach format can success-
fully target and enhance these skills.
Since each group is required to develop its own survey, no
standardized results can be reported here. Each group did have
to survey the teachers and students on the effectiveness of
their presentation, however, and in all cases the surveys indi-
cated the students' understanding of the topic had significantly
improved. They also indicated that they had a better under-
standing of what chemical or biochemical engineers do.
An unexpected outcome for the UMBC students has been
an increased pride in what they are doing and a sense of ac-
complishment. As a result of returning to the high school set-
ting, many students have an increased awareness of how much
they have learned and how far they have come during their
undergraduate experience. They often speak with pride about
their impending graduation and future plans for work or gradu-
ate school. In one case, at the end of a presentation three
group members described their plans to pursue PhD degrees
(two at M.I.T. and one at Georgia Tech). Quite unexpectedly
and without prompting, the class of thirty high school stu-
dents stood up and applauded. After the presentation, the
group members expressed pride at what they had accom-
plished and expressed an understanding for the first time that
they were role models for the high school students.
Providing role models to high school students has been an
ongoing goal for the outreach projects. In our experience, the
high school students have been very interested to learn more
about the experiences of the undergraduates. In particular,
UMBC attracts a significant pool of highly talented minority
students (graduating classes in chemical engineering have
averaged 47% minority students and 47% women in the last
three years). We believe that these students in particular
project powerful images to young high school students. It is
not unusual to hear the high school students comment, "You
don't look like engineers." We believe that this type of direct
contact can powerfully combat stereotypical images of the
engineering profession.

We would like to thank Dr. Michael Sierks, Arizona State
University, for his input into the initial concept for the out-
reach projects. EO

Summer 2003

Summer School



Oregon State University Corvallis, OR 97331-2702
he semiconductor industry has grown rapidly in the
last three decades, and chemical technologies have
played a central role in its continuing evolution. His-
torically, chemical engineering has focused on petrochemi-
cal and bulk chemical production, but more and more chemi-
cal engineers are working in the microelectronics and related
industries. The most recent AIChE placement survey shows
that the number of BS graduates placed in the electronics
industry has increased since 1998 from 7.0% to 15.9% in
2001 (it decreased dramatically last year to 4.2% due to the
economic slowdown).
The percentage of ChE graduates with advanced degrees
hired into the semiconductor industry is even larger. For ex-
ample, more than 25% of PhD graduates have been hired by
the electronics industry since 2000.[11 Chemical engineers have
the advantage of a solid background in chemical kinetics,
reactor design, transport phenomena, thermodynamics, and
process control that allows them to meet the challenges in

Chih-hung (Alex) Chang is Assistant Professor of Chemical Engineering
at OSU. He received his BS degree from National Taiwan University and
his PhD degree from the University of Florida, both in chemical engineer-
ing. His research interests include phase equilibria, photovoltaics, X-ray
absorption, fine structure, electronic materials, and nano- and
Milo D. Koretsky is Associate Professorof Chemical Engineering at OSU.
He received his BS and MS degrees from UCSD and his PhD from UC
Berkeley, all in chemical engineering. His research interests are in thin
film materials processing, including plasma etching, chemical vapor depo-
sition, electrochemical processes, and chemical process statistics.
Sho Kimura is Professor of Chemical Engineering at OSU. His research
interests cover high-temperature materials synthesis, nano-sized materi-
als synthesis, surface modifications, applications of high-temperature flu-
idization technology, reaction kinetics, catalytic effects on gas-solid reac-
tions, and reactor design and simulations.
Skip Rochefort is Associate Professor of Chemical Engineering at OSU.
He received his BS degree from UMass, his MS degree from Northwest-
ern, and his PhD from UCSD, all in chemical engineering. His research
interests are in all areas of polymer engineering and science and in engi-
neering education.
Cyndie Shaner served as the Linus Pauling Distinguished Engineer at
OSU from 2000 to 2002. She received her BS degree in chemical engi-
neering from Northwestern University and has had 22 years industrial ex-
perience with Chevron.
Copyright ChE Division of ASEE 2003


The most recent AIChE placement survey
shows that the number of BS graduates
placed in the electronics industry
has increased since 1998 from
7.0% to 15.9% in 2001

microelectronics processing.
Many chemical engineering pioneers in this field have rec-
ognized this symbiotic relationship[124' and a number of
schools have started incorporating microelectronics process-
ing into their curriculum. For the most part, however, the ma-
terial tends to be presented in specialized, elective courses
without a laboratory component. One common approach to
provide hands-on experience is to have students make a tran-
sistor in a lab typically offered by the electrical engineering
department. Students go through a series of microfabrication
processes (either hands-on or observed) to build a complete
integrated circuit, or a set of test transistors.
While this lab can be an exciting and successful approach
to introduce the students to the field of microelectronic pro-
cessing, there are some limitations. First of all, it is often
taught with a cookbook approach, and second, there is lim-
ited time devoted to each process, so process analysis is usu-
ally ignored. The resulting device functionality supersedes
the process engineering. Finally, not all schools have the nec-
essary resources to offer this lab to their chemical engineer-
ing undergraduates.
A complementary approach is to incorporate the unit op-
erations in microelectronics processing into the existing
chemical engineering curriculum. An example of such an
approach is the incorporation of thermal oxidation of silicon
into the unit operations lab at Georgia Tech.[15 This approach
can provide students with depth as well as breadth by pre-
senting these unit operations in the context of core chemical
engineering science. ,

Chemical Engineerrng Education

Poster Session Award Paper

Development of programs in this area has led to innova-
tive and improved education practices.8-'101 A successful ex-
ample is the curriculum developed by the Chemical and
Materials Engineering Department at San Jose State Uni-
versity. The essence of their project is to abandon the tradi-
tional laboratory cookbook instruction method and create a
team-oriented and open-ended laboratory where students de-
velop the types of skills they will later use in industry. The
content of their laboratory includes having students make a
field effect transistor and perform open-ended experiments
to improve the process.[10]

While the approach at San Jose relies on coordination be-
tween students in three different engineering disciplines
(electrical, material, and chemical), at Oregon State Univer-
sity (OSU) we are implementing the same type of learning
environment solely within chemical engineering. In this way,
we can leverage off the fundamental research in microelec-
tronics processing to develop unit operations accessible to
undergraduate students based on their core engineering sci-
ence background.

Integration of unit operations into microelectronics has oc-
curred in conjunction with a transformation in the senior unit
operations laboratory that began during the 2000-2001 aca-
demic year. A newly created endowed chair, the Linus Pauling

Engineer, was hired from
industry to identify and in-
corporate the highest prior-
ity professional practices in
the senior lab. She serves as
"project director" for this
class that helps graduates
become prepared for indus-
trial practice. Professional
practices are incorporated
into the lab through lectures,
classwork assignments, and
homework assignments.
Eight lectures cover project
management, meeting
skills, technical writing, oral
presentations, safety, ratio-
nal management processes
(situational, problem, deci-
sion, and potential problem
analysis), personality self-
assessment, and conflict
resolution.["1 All students
complete writing assign-
ments and oral presentations
to practice the professional

skill as well as to demonstrate technical understanding of
the unit operation. The instructor, the student, and the
student's peers assess each student's work process skills,
safety performance, and team behavior.

Hundreds of individual process steps are used in the manu-
facture of even simple microelectronics devices, but the fab-
rication sequence uses many of the same unit processes a
number of times. A list of unit operations that are common
for the fabrication of microelectronic devices, along with
the curricular material related to each topic, are given in
Table 1. These unit operations rely on core chemical en-
gineering science.
We are developing unit operations modules for integration
into the chemical engineering curriculum and unit operations
laboratories at OSU, including plasma etching, chemical va-
por deposition (CVD), spin coating, electrochemical deposi-
tion, and chemical mechanical planarization (CMP). These
unit operations contain complex systems that involve inter-
action of many physical and chemical processes. Fortunately,

Unit Operations in Microelectronic Device Fabrication

Unit Operations
Bulk Crystal Growth from Melt

Surface Reactions
Cleaning and Oxidation
Plasma Etching, Wet Etching
Thin Film Deposition
* Physical Vapor Deposition
* Chemical Vapor Deposition
* Electrochemical Deposition
* Photoresist spin coating
* Photoresist baking
* Photoresist exposure and development

Doping and Dopant
* Ion Implantation
* Thermal Diffusion
* Chemical Mechanical Polishing

ChE Core Courses
Fluid Dynamics; Heat Transfer; Mass
Transfer; Thermodynamics; Reaction
Engineering; Process Control
Kinetics; Fluid Dynamics; MassTransfer

Mass Transfer; Kinetics; Reaction
Engineering; Process Control
Kinetics; Fluid Dynamics; Mass Transfer;
Heat Transfer; Thermodynamics;
Electrochemical Engineering; Reaction
Engineering; Process Control
Fluid Dynamics; Mass Transfer; Polymer
Rheology; Kinetics; Process Control

Mass Transfer; Heat Transfer; Process

Fluid Dynamics; Mass Transfer; Electro-
chemical Engineering; Process Control

there have been extensive re-
search efforts in these areas,
and many of the fundamen-
tal mechanisms have been
elucidated. For example,
plasma etching processes
have been modeled based on
the fundamental transport
and reaction processes oc-
curring within the glow dis-
charge to understand issues
of etch rate, selectivity, uni-
formity, and profile.12- 15]
Similarly, chemical vapor
deposition reactors have
been modeled in analogy to
porous catalysts,[161 incorpo-
rating transport and reaction
processes.[17-20] Control
schemes have been based on
the fundamental CVD reac-
tor models.[211 The fluid dy-
namics of photoresist spin
coating has been modeled
and studied experimentally
to predict coating thickness
and uniformity as a function

Summer 2003

Summer School

of spin-speed, fluid properties, and spin duration.[22-
2" Similarly, fluid-dynamics-based models of chemi-
cal mechanical polishing are being developed."26-281
We are synthesizing the research results in the
literature and applying them to the five unit op-
erations discussed above to make them accessible
to undergraduate chemical engineers. At the same
time, we are reinforcing the fundamental engineer-
ing science taught in the curriculum. To accom-
plish this objective, we are developing both lab
based and classroom based instruction.
Integration into the lab occurs through the two
required unit operations laboratories (ChE 414 and
415) as well as an elective, Thin Film Materials
Processing. The first quarter of the two-quarter
lab sequence is highly structured and focuses on
the students completing three unit operation ex-
periments. We intend to have each student com-
plete at least one microelectronics unit operation
during this rotation. The second quarter of the se-
nior lab course builds on the work done in the first
quarter. The focus is on working independently,
developing a project proposal, completing experi-
mental work, and writing a final technical memo-
randum that includes recommendations for future
work. The microelectronics unit operations are de-
signed to be flexible enough so that each year the
group of students has a new, unique, and creative
experience. The first four unit operations were in-
tegrated into the second-quarter lab in the spring
of 2002 and are described below. The unit opera-
tions experiment in chemical mechanical
planarization was added in the spring of 2003.
Students who are interested in pursuing high-
tech careers can obtain a transcript visible micro-
electronics or materials science and engineering
option in the chemical engineering department.
In either of these options, Thin Film Materials
Processing is required. Starting in the winter of
2003, the thin films course was expanded from 3
to 4 credits to enhance the laboratory component
and in the lab the students rotated through six ex-
periments. Three experiments-plasma etching
and spin coating, silicon nitride deposition, and
copper electrodeposition-are based on the unit
operations mentioned earlier. Students also study
wet and dry silicon oxidation, have a vacuum com-
ponents lab, and undergo a "virtual" lab based on
the semiconductor device applets developed by
Wie and coworkers at SUNY, Buffalo.J291

In the processing labs (plasma, CVD, electrodeposition, and oxidation),
students are introduced to the unit operations. For efficacy, these labs are
conducted in a well-prescribed manner where the process parameters and
lab procedure are given to the students. Each group runs the process at
different parameter settings, and once the entire class has rotated through a
given unit operation, the students are presented with the data collected
from the entire class for analysis.
For example, the silicon nitride deposition is operated at three tempera-
tures and three flow rates (NH3 rich, stoichiometric, and NH3 lean). Stu-
dents measure growth rate, film thickness (from which they calculate den-

Reactor Design
ChE 432 Detailed Design Project LPCVD Silicon Nitride
Spring, 2002
This project is designed for senior students in ChE 432 who are interested in
experiencing a detailed design/simulation project in microelectronics processing.
The topic selected is LPCVD (Low Pressure Chemical Vapor Deposition) that has
been used for the deposition of silicon nitride on silicon wafers in the process for
producing ICs.
Students who work on this project are required to
1. Propose a design idea for a piece of equipment that handles 200 silicon wafers
of 300mm in diameter
2. Develop a model to simulate the performance of the equipment
3. Determine proper operating conditions, i.e., temperature distributions, operat-
ing pressure, feed rates, and reaction time for controlling the deposit thickness
at 1000A with variations across wafers and from wafer to wafer within +3%.

Figure 1. The CVD senior design-project assignment,

Implementation Grid of
Microelectronics Unit Operations and ChE Classes

Chemical Electro- Chemical
Plasma Vapor Spin chemical Mechanical
Course Etching Deposition Coating Deposition Polishing
Materials Balances X X
Energy Balances
Chemical Process Statistics X
Thermodynamic Properties & Relationships X X
Phase and Chemical Reaction Equilibrium X X X
Applied Momentum and Energy Transfer X
Chemical Engineering Laboratory I X X X
Chemical Engineering Laboratory II X X X X
Chemical Plant Design I X
Chemical Plant Design II X
Chemical Reaction Engineering X X X
Thin Film Materials Processing X X X X X
Polymer Engineering and Science X

Chemical Engineerrng Education

sity), and index of refraction for their particular run. Then they look at the entire
data set of the (nine) groups in the class to explore the effect of temperature and
concentration on nitride deposition.
The vacuum lab gives students experience with constructing a vacuum system
and shows them the basic components in any vacuum system, including pumps,
pressure measurement, mass flow control, and different types of flanges. The
virtual lab provides a workshop on carrier physics, crystal structure, and inte-
grated circuit processing with subtractive pattern transfer. For more detail, go to
Classroom examples are developed based on the labs as well as the research

'A ali Boat

Reactor Plasma Glow as a function of Pressure

Figure 2. Plasma barrel etcher and schematic.

Tmaqrf-nmen Aibsti4WGA) Ssptm

Figure 3. Silicon nitride CVD reactor and schematic.

Poster Session Award Paper )

literature. Each of the four unit operations
listed above will include at least two ex-
ample exercises or homework problems to
be integrated into a core chemical engineer-
ing science or design course. By integrat-
ing the technical content in this manner,
the future process engineers in this indus-
try will be able to draw upon core funda-
mentals as they go about problem solving.
A grid of target courses for classroom in-
tegration is given in Table 2. Those marked
with an "x" represent targeted courses. For
example, the design problem offered in
Process Design II in the spring of 2002 is
shown in Figure 1.
* Plasma Etching
Glow discharge plasmas are used for a
variety of surface manufacturing applica-
tions, especially in integrated circuit manu-
facturing where up to 30% of all process
steps involve plasmas in one way or an-
other. A plasma barrel etcher has been in-
corporated into the projects in the second-
quarter lab and the thin film course. This
plasma barrel etcher unit and supporting
systems were donated by Intel (see Figure
2). In the barrel etcher, ion bombardment
is suppressed since the substrate holder is
contained within a Faraday cage. Thus, the
etch rate depends on the concentrations of
free radicals that react at the substrate sur-
face. Uniform etching only occurs when
mass transport to the surface is much
greater than the inherent reaction rate. By
measuring the etching rate as a function of
radial position, the relative importance of
mass transfer to surface reaction can be
backed out. The variation of etch rate as a
function of the sample radius allows stu-
dents to interpret etch data in terms of fun-
damental chemical engineering principles
(e.g., transport and reaction). Industrially,
obtaining a uniform etching rate is also
a central problem in plasma etching re-
actor design.
Other examples of student lab experi-
ments include: finding optimal process set-
tings for etching polyphenylene oxide ma-
terials using SF6 and 02 feed gases by us-
ing Design of Experiments (DOE) and ana-

Summer 2003

Summer School

lyzing the problem using a well-mixed reactor model; the
effects of wafer spacing on etch rate; the effect of the num-
ber of substrates, i.e., loading, on etch rate; and transient
analysis of temperature effects on the etching rate.
In the spring of 2002, two groups of three students each
were given an assignment to develop a process that mini-
mized interwafer, as well as wafer-to-wafer, variation in
etching rate. This lab included several processes to pattern
and etch a wafer, including cleaning, spin coating, photoli-
thography, and plasma etching. Additionally, students de-
veloped their own artwork to serve as a mask in photoli-
thography. The experimental design focused on etching pa-
rameters, however, while the other processes were un-
changed. One group did a 2x2 design in which they varied
pressure and wafer spacing while the other group varied
power and spacing. Thickness measurements before and
after etching were made using a profilometer. The plasma
reactor does not have temperature control. Thus the tem-
perature rises as power is input during the etch process.
This facet provided a good opportunity to use chemical
engineering analysis to understand the system. The group
was advised to record temperature during the process, but
needed faculty help to develop an averaging method based
on the Arrhenius expression for the activated process. The
plasma lab analysis can be found at>.

* Chemical Vapor Deposition
In this module, the gas-solid reaction kinetics are eluci-
dated through real-time rate measurements using a modi-
fied thermogravimetric analyzer (TGA)(see Figure 3). Stu-
dents can measure an increase in mass of the silicon wafer
sample specimen (about 10 mm by 20 mm) with time, re-
sulting from the deposition of silicon nitride at different
reactant concentrations and reaction temperatures at atmo-
spheric pressure. Inert argon is mixed with the two gas-
eous reactants (ammonia and dichlorosilane). Since the
reaction is at atmospheric pressure, as opposed to vacuum,
students must account for the effect of resistance to diffu-
sion through the gas film on the silicon surface and find
ways to eliminate the mass-transfer effects. In this con-
text, they were asked to discuss the difference between the
low-pressure CVD and atmospheric-pressure CVD. The
kinetics obtained using the modified TGA were integrated
into the senior capstone design course via designing a CVD
reactor and simulating its performance for achieving uni-
form film thickness. Students were challenged to develop
a simple mathematical model that incorporates the fluid
flow, diffusion, and reaction that take place simulta-
neously. Students used their model to predict the growth
of silicon nitride films on two hundred 300-mm wa-

Spin Coating of Photoresist-Fluid, Mass, and Heat Transfer Problems
Spin coating of photoresists is found in most photolithography steps in microelec-
tronics processing. One general formulation of photoresist involves a low molecular
weight polymer dissolved in a low boiling point (high vapor pressure) solvent at
various concentrations (typically 60-80wt% polymer). The photoresist is spun onto a
silicon wafer (currently 6-, 8-, or 12-inch nominal diameter) at various spin speeds
for various spin times. The material properties (such as the viscosity) change with
time as the solvent evaporates. Typical coating thickness values after spinning are 1-
5gim, with residual solvent in the film. Residual solvent removal is achieved by a
high-temperature "baking" step. There are several interesting and challenging trans-
port problems that can be derived from this process that are suitable for undergraduate
courses. These problems are not "plug and chug" but rather require that the students
reason through the physical situation and apply simplifying assumptions to arrive at
reasonable models (see two recent textbooks by Middlemani3,34,351 for examples).
Fluid Mechanics Problem In the industrial setting, the coating process typically
begins by dispensing a "puddle" of liquid at the center of a rotating wafer. The puddle
very quickly spreads to the edges of the wafer. The spreading of a "viscous drop" has
been presented by Middleman.[34,p 2721 The analysis most relevant to the spin coating
problem presented here was first studied by Emslie, et al.,[221 where the author as-
sumed that the film was initially at a uniform thickness across the entire wafer and
that the fluid was Newtonian and nonvolatile. These assumptions led to a rather simple
model for film thickness (H) in terms of spin speed (m), spin time (t), initial film
thickness (Ho), density (p), and viscosity (It): H Ho (1 + 4pW2Ho2t/3t))1/2. The
students can be given typical values of spin speed (1000-5000 rpm), spin time (30-
120 sec), initial film thickness (Ho ~ 100gm), density (800 1000 kg/m3), and
viscosity (10 1000 mPa-s) to generate curves of film thickness (H) vs. time (t).
These calculations can be used to point out the relative importance of the various
parameters in a spin coating application.
Mass and Heat Transfer Problem The transient mass transfer problem for the
rate of removal of solvent from the film during the spin coating process is a difficult
problem to solve. Once again, the students are challenged to make simplifying as-
sumptions, as was done with the fluid mechanics problem, to arrive at models that
can be tested. In addition to the information given above, the students are told that
there must be 90% removal of the solvent in the spinning step. One approach to
solving this problem is to start with the assumption as was made above of a uniform
initial coating thickness (Ho ~ 100lim). The problem then reduces to an unsteady-
state mass transfer problem, similar to solvent evaporation from a polymer film de-
scribed in Middleman.[35'p 128] The students must identify the simplifying assump-
tions of no internal convection, one-dimensional diffusion (thin film), and rapid ex-
terior convection due to the high speed spinning, and comprehend the limitations of
these assumptions. For instance, if the film is spinning at high speed and thinning,
how can there be no internal convection? The residual solvent removal in the "bak-
ing" step can be modeled in a similar manner.

Figure 4. (a) Spin coater, and (b) spin coating
classroom example.

Chemical Engineerrnng Education

P W q

fers, varying with temperature-profile settings, reactant feed
rates, and operating pressures.
* Spin Coating
Spin coating has come into widespread use in the micro-
electronics industry for coating the photoresists used to de-
fine patterns in the films on a silicon wafer. It will also be
used in future technologies as polymers become incorporated
as dielectric materials. The underlying principles of spin coat-
ing (fluid flow, fluid properties, surface phenomena) and the
process itself make it a natural for inclusion in the chemical
engineering curriculum. The precursors to coating, surface
wuenin_ and adhesion are also classical problems. The spin
coating of solid substrates with various viscous liquids and
surface wetting phenomena (surface tension, contact angle,
viscosity) is done using a "state-of-the-art" programmable
laboratory spin coater from Specialty Coating Systems (SCS
Model P6700) (see Figure 4a) and highly polished oxide-
coated 6-inch wafers. Examples of engineering projects in-
clude: experiments on viscous, Newtonian liquids to test the
Emslie model and compare data to published spin coating
results for Newtonian liquids;1221 coating photoresist on sili-
con wafers as the first step in the photolithography process.

OD. ---------

ChE 312 Phase and Chemical Reaction Equilibrium
Electroplating is used to deposit the copper metal lines in integrated circuit processing. In this case,
dissolution of a copper anode is used to provide cupric ions for reaction, i.e., the solid copper-cupric ion
reaction is used both at the anode to generate cupric and at the cathode to grow the resulting film. Please
calculate the minimum electrode potential to achieve copper growth for the following cell

using (a) ideal solution model, and (b) modified Debye-Hiickel theory
A number of concepts in chemical reaction equilibrium can be taught through this classroom example,
* How to write electrochemical reactions
* How to incorporate the concept of electrical work into the equilibrium formulation. How to calculate the
electrochemical potential from the Gibbs energy
How to estimate activity coefficients for ionic species in an electrolyte solution (Debye-Hfickel theory)

Figure 5. (a) Copper electrodeposition reactor and schematic, and
(b) electrodeposition classroom example.

Poster Session Award Paper )

This unit process has been used as a project in several out-
reach activities. Classroom examples for momentum, heat,
and mass transfer are given in Figure 4b.

U Electrochemical deposition

The electrochemical deposition system includes a computer-
controlled bipotentiostat, PineChem software, rotator, elec-
trodes, and a standard voltammetry cell (see Figure 5a). A
variety of experiments could be designed using this system.
Examples of such experiments are
Diffusion coefficient determination by .i ,i,, i. elec-
trode cyclic voltammetry
Measurement of the kinetics and the flux of copper ions
to an electrode surface by means ofj. ,ianis ring-disk
Study of mass transfer using j.ii i. electrodes
The effects of additives on deposition rates
Leveling effects of additives
Superfilling phenomena
Resistive seed effect

In the spring of 2002, the student
team was taught how to use this sys-
tem by going through a "cookbook"
experiment using cyclic voltammetry
and the rotated disk electrode to char-
acterize the redox reaction of potas-
sium ferricyanide solution. After the
training, they were asked to propose
an experimental plan using this setup.
They decided to study the copper mass
transport using different copper elec-
trolytes (CuCl2 and CuSO4) and the
influence of a sulfur-containing addi-
tive (thiourea). The experiments were
performed using acid copper solutions
prepared from CuC12 and CuSO4 with
and without thiourea. The electro-
chemical reactions were characterized
by sweeping the voltage and measur-
ing the current. The boundary layer
thickness was controlled by the rotat-
ing speed of the working electrode and
the Levich equation was used to de-
termine the diffusivity.
A classroom example based on cop-
per electrodeposition is given in Fig-
ure 5b. In addition, many interesting
laboratory problems on copper plat-

Summer 2003

Summer School

ing can be found in a manuscript by Talbot.1301
* Chemical Mechanical Polishing
The experimental setup of a bench-scale CMP module is
shown in Figure 6a. The setup is adopted from the research
literature,[31-33] which has been useful in understanding the
reaction mechanisms during CMP. In this setup, the copper
CMP will be studied in a three-electrode electrochemical cell
by using a copper-plated rotating disk electrode. The polish-
ing downward force will be measured by a balance, which
supports the entire electrochemical cell. The DC electrochemi-
cal measurement will be carried out by using a potentiostat.
A variety of experiments can be designed to study the copper
CMP process. For example, the effect of HNO3 or NH4OH
on the chemical etching mechanism, the effects of additives
(e.g., inhibitor, oxidizer), the relationbetween downforce, and
removal rates through the Preston equation. The Preston equa-
tion[311 relates removal rate to driving force in the same way
that mass transfer coefficients relate mass transfer to driving
force. Studies on the bench-scale system will be scaled up to
the industrial-scale system shown in Figure 6b.

Three modules, including plasma etching, spin coating, and
copper electrodeposition, were implemented
in the outreach programs that are currently in
place in the chemical engineering department:
1) Summer Experience in Science and Engi-
neering for Youth (web site at>), and 2) the Sat-
urday Academy and Apprenticeships in Sci-
ence and Engineering program (web site at
index.html>). Each of these programs has a
somewhat different focus, but share several
common underlying themes: exposure of high
school students to careers in science and en-
gineering through research experiences and
other opportunities that are typically not avail-
able to them in the high schools; recruitment
and retention of underrepresented groups
(girls and ethnic minorities) into science and
engineering; and a goal of increasing the tech-
nological literacy of high school students so
they can be empowered to make educated
career choices.

The measurable student outcomes for each
unit operation will include Figure 6.

The students will demonstrate communication skills. For
example, they will be required to master written and oral
The students will demonstrate technical synthesis in each
of the unit operations. For example, in CVD they will use
kinetic data in reactor design problems.
The students will demonstrate professional practices. For
example, they will be required to demonstrate project
planning before performing experiments.
Each of these outcomes will be assessed by three methods:
Student self-assessment and peer-assessment, e.g., sur-
vey of effectiveness of educational materials.
Evaluation of student performance by instructors.
Feedback from industrial constituency, e.g., survey of stu-
dent performance from industrial employers.

The integration of microelectronics-based unit operations
into the chemical engineering curriculum at Oregon State
University has been presented in this paper. We are develop-
ing both lab-based and classroom-based instruction. Five new
unit operations are being implemented in the senior lab, in-
cluding plasma etching, chemical vapor deposition, spin coat-

(a) Bench-scale CMP and schematic (b) industrial-scale CMP

Chemical Engineerrng Education

ing, electrochemical deposition, and chemical mechanical
planarization. These labs are also included in an elective
course, Thin Film Materials Processing. Classroom examples
are being integrated into chemical engineering core courses.
In addition, the students are learning professional practices
that include effective oral and written communication, project
planning, time management, interpersonal interaction, team-
work, and proactive behavior. These modules are also being
used effectively in outreach programs.


The authors are grateful for support provided by the Intel
Faculty Fellowship Program and the National Science
Foundation's Course, Curriculum, and Laboratory Improve-
ment Program under grant DUE-0127175. We also acknowl-
edge the Dreyfus Special Grants Program (SG-97-075), OSU
Precollege Programs, Kelley Foundation, and Bridges Foun-
dation for their support of the outreach program Summer
Experience in Science and Engineering for Youth, and Pete
and Rosalie Johnson for the endowment of the Linus Pauling
Engineer. SEH America graciously donated the highly pol-
ished and oxide-coated 6-inch silicon wafers. Helpful dis-
cussions with Emily Allen of San Jose State University
and Chuck Croy of Intel Corporation are greatly appreci-
ated. The authors would also like to acknowledge the re-
viewers for the critical reading and suggestions they made
regarding this paper.

1. "Initial Placement of Chemical Engineering Graduates,">
2. Hess, D.W., and K.F. Jensen, eds., Microelectronics Processing: Chemi-
cal Engineering Aspects, Advances in Chemistry series, Vol. 221,
American Chemical Society, Washington, DC (1989)
3. Middleman, S., andA. Hochberg, Process EngineeringAnalysis inSemi-
conductor Device Fabrication, McGraw-Hill, New York, NY (1993)
4. Lee, H.H., Fundamentals of Microelectronics Processing, McGraw-
Hill, New York, NY (1990)
5. Hess, D.W., S. Bidstrup-Allen, P. Kohl, M. Allen, and G. May, "Ther-
mal Oxidation of Silicon: A Unit Operation for ChEs," Session 19,
ASEE Summer School for Chemical Engineering Faculty, Snowbird,
UT (1997)
6. Fogler, H. ...n ..ring, 3rd ed.,
Prentice-Hall PTR, Saddle River, NJ (1999)
7. Welty, James, Charles Wicks, Robert Wilson, and Gregory Rorrer,
Fundamentals of Momentum, Heat, and Mass 4th ed., John
Wiley & Sons, New York, NY (2001)
8. Dang, S.S., R.A. Matthes, and C.G. Takoudis, "A Web-Based Course
in the Fundamentals of Microelectronics Processing, Chem. Eng. Ed.,
34(4), 350 (2000)
9. Chang, J.P., "A Hands-On Laboratory in the Fundamentals of Semi-
conductor Manufacturing, Chem. Eng. Ed., 36(1), 14 and references
therein (2002)
10. Muscat, A.J., E.L. Allen, E.D.H. Green, and L.S. Vanasupa, "Interdis-
ciplinary Teaching and Learning in a Semiconductor Processing
Course," J. Eng. Ed., 87, 413 (1998)

Summer 2003

Poster Session Award Paper )

11. Koretsky, M.D.,C.-H. (... k i..... '.chefort,andC. Shaner,
"Integration of Microelectronics-Based Unit Operations into the ChE
Curriculum," Session 1313, Proceedings of the 2003 ASEE Annual
12. Bell, A.T., "Fundamentals of Plasma Chemistry," in Techniques and
Applications ofPlasma ( A.T. Bell and J.R. Hollahan, eds.,
John Wiley & Sons, New York, NY (1974)
13. Graves, D., and K.F. Jensen, "A Continuum Model of DC and rf Dis-
charges," IEEE Trans. Plasma Sci., P5-14, 78 (1986)
14. Alkire, R.C., and D.J. Economou, "Transient Behavior During Film
Removal in Diffusion-Controlled Plasma Etching," J. Electrochem.
Soc., 132, 648 (1985)
15. Plumb, I.C., and K.R. Ryan, "A Model of the Chemical Processes
Occuring in CF4/02 Discharges Used in Plasma Etching," Plasma
Chem. Plasma Proc., 6, 231 (1986)
16. Levenspiel, 0., The Chemical Reactor Omnibook, Oregon State Uni-
versity, Corvallis, OR (1993)
17. Hess, D.W., K.F. Jensen, and T.J. Anderson, "Chemical Vapor Depo-
sition: A Chemical Engineering Perspective,"Rev. ( ,, 3, 97
18. Jensen, K.R., and D.B. Graves, "Modeling and Analysis of Low Pres-
sure CVD Reactors, J. Electrochem. Soc., 130, 1950 (1983)
19. Roenigk, K.F., and K.F. Jensen, "Low Pressure CVD of Silicon Ni-
tride, J. Electrochem. Soc., 134(7), 1777 (1987)
20. Fogler, H. Scott, Elements .. .ring, 3rded.,
Prentice-Hall PTR, 789 (1999)
21. Edgar, T.E, S. Butler, W.J. Campbell, C. Pfeiffer, C. Bode, S.G. Hwang,
K.S. Balakrishnan, and J. Hahn, "Automatic Control in Microelec-
tronics Manufacturing: Practices, Challenges, and Possibilities,"
Automatica, 36, 1567 (2000)
22. Emslie, A.G., F.T. Bonner, and L.G. Peck, "Flow of a Viscous Fluid
on a Rotating Disc," J. Appl. Phys., 29, 858 (1958)
23. Flack, W.W., D.S. Soong, A.T. Bell, and D.W. Hess, "A Mathematical
Model for Spin Coating of Polymer Photoresists,"J. Apply. Phys., 56
1199 (1984)
24. Bornside, D.L., C.W. Macosko, and L.E. Scriven, "Spin Coating: One-
Dimensional Model," J. Apply. Phys., 66, 5185 (1989)
25. Strong, L., and S. Middleman, "Lubricant Retention on a Spinning
Disk,"AIChEJ., 35(10), 1753 (1989)
26. Subramanian, R.S., and L. Zhang, "Some Transport Phenomena Is-
sues in Chemical Mechanical Polishing," 3rd Annual Workshop on
Chemical Mechanical Polishing (1998)
27. Runnels, S.R., and L.M. Eyman, "Tribology Analysis of Chemical
Mechanical Polishing," J. Electrochem. Soc., 141, 1698 (1994)
28. Stein, D.J., D.L. Hetherington, and J.L. Cecchi., "Investigation ofthe
Kinetics of Tungsten Chemical Mechanical Polishing in Iodate-Based
Slurries," J. Electrochem. Soc., 146, 376 (1999)
30. Talbot, J.F., "Electrochemical Engineering in the Process Laboratory
Course," Chem. Eng. Ed., 35(1), 74 (2001)
31. Steigerwald, J.M., S.P. Murarke, and R.J. Gutman, ChemicalMecham-
cal Planarization of Microelectronic Materials, John Wiley & Sons,
New York, NY (1997)
32. Tsai, Tzu-Hsuan, and Shi-Chern Yen, "Electrochemical Effects of Vari-
ous Slurries on the Chemical Mechanical Polishing of Copper-Plated
Films," submitted to J. Electrochem. Soc.
33. Kneer, E.A., C. Raghunath, V. Mathew, and S. Raghavan, "Electro-
chemical Measurements During the Chemical-Mechanical Polishing
of Tungsten Thin Films," J. Electrochem. Soc., 144, 3041 (1997)
34. Middleman, S., An Introduction to Fluid Mechanics: Principles of
Analysis andDesign, John Wiley and Sons, New York, NY (1998)
35. Middleman, S.,An Introduction toMass andHeat Principles
ofAnalysis andDesign, John Wiley & Sons, New York, NY (1998) 5

Summer School




University ofMissouri, Columbia Columbia, MO 65211

Typically, the unit operations laboratory is one of the
least-liked classes in the chemical engineering cur-
riculum at the University of Missouri-Columbia. Un-
fortunately, many students view laboratory experiments as
unfocused and haphazard investigations of various types of
equipment that provide mere busy-work during their demand-
ing course load. Although the concepts investigated are use-
ful for all students, most of them think that the circumstances
behind the investigations are not representative of those typi-
cally encountered in industry. In an attempt to provide the
students with a more industrially meaningful experience when
performing an experiment in this class, design-of-experiment
(DOE) techniques have been combined with one classic unit
operations experiment to present a more "real-world situa-
tion" to the student.
DOE is the application of screening designs and statistical
sampling to obtain a desired result, such as minimizing the
number of experiments necessary to obtain the answer to a
problem or minimizing the variance of estimated coefficients
obtained through regression to a model that predicts the out-
put from some system. The first result above affects quantity
of effort, which is directly related to cost, while the second
affects accuracy of the result. Therefore, DOE is useful in
industrial settings as a powerful tool that effectively reaches
a compromise between cost and accuracy so an informed
decision can be made.
A designed experiment tests how changes made to chosen
input variables for a process or system affect some output
response. Experimental design methods play an important
role in process development and improvement. In-depth
knowledge of the process is not required to make improve-
ments in the process when using this approach-a statement
that was confirmed in class by multiple students who had
industrial co-op assignments where such experiments were

Three class lecture periods are scheduled for teaching the
basics behind an effective DOE approach, including statisti-
cal and error analysis, blocking, replication, randomization,
screening designs, full- and half-factorial experimental de-
sign, and central composite design. The material is elemen-
tary enough that each student can easily understand how DOE
works and complements the coursework covered by students
who have taken the elective course "Experimental Design
and Statistical Quality Control for Chemical Engineers." Ad-
ditional material detailing the experimental design approach
and factorial analysis can be found in the text by Box, Hunter,
and Hunter 11 used for this course as well as in other widely
used sources.[2,3] For the sake of those who are unfamiliar
with DOE principles, however, and to adequately intro-
duce the material used for the sample analysis for the ex-
periment presented later, the class material covered for a
three-factor, two-level, half-factorial experimental design
is discussed below.
In many cases, a large number of variables exist that can
have an effect on a system. The main effect is a change in the
response that results from a change in level from a low value
to a high value of the variable and is calculated as the aver-
age of the difference between all the responses at the high
value of a variable and those at the low level of that variable.

Eric J. Doskocil is Assistant Professor of
Chemical Engineering at the University of
Missouri-Columbia. He received his BS in
1994 from Florida State University and his
PhD in 1999 from the University of Virginia.
His research interests are in the area of solid
heterogeneous catalysts, with current focuses
on epoxidation and acid-base bifunctional

Copyright ChE Division ofASEE 2003

Chemical Engineerrng Education

As seen in Figure 1, the eight trials indicated by all the gray
and black points are required to perform a two-level, three-
factor, experimental design that effectively models the main
effects and interaction effects associated with changes to the
three-input variables. Equations can be written to calculate
the main effects and interaction effects between variables
when three or more variables are being investigated. The
greater the magnitude for a calculated main or interaction
effect, the more significant the effect on the response that
factor exhibits.
When a large number of potential input variables exists,
the system has high variability, or additional factor levels need
to be studied due to nonlinear relationships, the number of
experiments that are required to completely understand the
system can become quite large, causing a large time commit-
ment for optimization of the system being investigated. Ex-
perimental design techniques can be used to maximize the
amount of information obtained from a reduced number of
experimental trials. When high accuracy is not necessary, a
half-factorial experimental design can be performed, which
requires half the amount of the 2" trials (2n-1) needed for the
full-factorial experimental design. Therefore, for the half-fac-
torial three-factor, two-level design, four trials are performed
at operating conditions indicated by the black points in Fig-
ure 1 or at the complementary half-fraction conditions indi-
cated by the gray points. Conducting trials at either of these
sets of four operating conditions gives a balanced design nec-
essary to determine the optimum. The equations used to cal-
culate the main effects for the three input variables become

Figure 1. Geometrical representation of three-factor, two-
level experimental design strategies. A full-factorial experi-
ment is achieved using trials performed at all the black and
gray points, while a half-factorial experiment is achieved
by performing trials at either the black or the gray points.
Letters in parentheses indicate results obtained for an out-
put response from a trial performed at a high value for each
variable indicated and a low value for any variable not in-
dicated. For example, trial (ac) indicates the output re-
sponse measured when factors A and C were set at their
high (+) value and factor B was set at its low (-) value.

Poster Session Award Paper )

MainEffectofVariableA=2p, = (ab) +(ac) (1) (bc)
2[ 2b ] 2
or {[(a)+(abc) (b)+(c)j
o 2 J L 2 J

MainEffectofVariableB= 2p2 {(ab(bc)j [( l)(ac)

'[(b)+(abc)] (a)+(c)]
r 2 J L 2 J

MainEffectofVariableC=2p3 {2(ac)+(bc) (1)- (ab)
2 2
or { (c)+(abc) (a)+ (b)j 1
o 2 J L 2 J]
The coefficients for all interaction effects are set to zero in
the process model. When performing a half-factorial experi-
ment, interaction effects must be assumed to be negligible
since Eqs. (1) through (3), which are used to determine the
main effect for each of the three input variables, are identical
to the equations used to determine the interaction effect be-
tween the other two variables. Proving that these equations
are indeed the same makes an excellent exercise for those
students who have not had any previous experience with DOE.
For this design's mathematical model, the origin for the
model is relocated to the center of the investigated three-di-
mensional volume pictured in Fig. 1 and measurements for fac-
tors A, B, and C are normalized so that value of +1 for A, B, or
C corresponds to that factor's value at the level where the fac-
tor was the maximum value investigated, and -1 corresponds to
that factor's value at the level where the factor was the mini-
mum value investigated. Therefore, the model becomes
Output Response = 30 + PiA + 32B + j33C + error (4)
where B0 is the average of the four completed trials and the
error is determined from replicates of trials completed at the
conditions of the new origin or at the conditions of at least
one of the four completed trials (see Box, Hunter, and Hunter
for more in-depth analysis['). This relationship can also be
determined via multilinear regression analysis conducted us-
ing various statistical and mathematical software.

Not all laboratory experiments are adequate candidates for

Summer 2003

Summer School

successful incorporation of the DOE approach. When deter-
mining if a current experiment has the potential for inclusion
of the DOE approach, the following factors should be con-
Experimental trials should not take a significantly
large amount of time to complete. One or two trials
should not take the entire laboratory period unless
multiple days will be given to complete the necessary
Effort required to convert the system to perform
additional trials with the new input variables should be
minimized to keep the students interested in the
A variety of potential input variables should exist such
that each student group can potentially investigate
different variable combinations and avoid repetition.
Including variables that have little-to-no effect on the
process keeps the exercise realistic.
A maximum, minimum, or discontinuity in the output
response should not exist when each input variable is
changed over the range of values given to the students
when only two-level design strategies are used.
Students may become confused in their analysis when
these types of behavior exist within the range being
studied, especially if the result contradicts theory.
When time does not permit a full-factorial experiment
to be performed during the assigned laboratory time,
interaction effects between variables should be small
or negligible. Since this is an introductory exercise for
the students, strong interaction effects should not
complicate the analysis.
A highly variable system requires a sufficient number
of trials to adequately determine the error. If only a
limited number of trials can be performed, however,
the trials should be highly reproducible so that experi-
mental errors are minimized. In the case discussed here,
performing trials for determining the error component to
the model is not possible during the allotted laboratory
time; therefore, students are required to speculate on the
error component in their report.
It is of the utmost importance that, when incorporating the
DOE approach into a current experiment in the laboratory,
the original technical objective of the experiment must still
be achieved while meeting each of the conditions listed above.
Otherwise, the equipment may be viewed as a "black box"
for generating data and the original focus of the experiment
is replaced by a focus on DOE analysis.

A unit operations experiment titled "Residence Time Dis-

tribution for Stirred Tanks in Series" was chosen for incorpo-
rating the DOE approach into the curriculum. A diagram of
the apparatus used for this experiment is shown in Figure 2.
It consists of three Plexiglas stirred tanks in series, where the
tank outlets are overflows to the next tank or to drain. The
third tank can be physically changed between two different
tank sizes (3.70 or 7.15 L). Stirring speed is adjusted using
speed controllers, which control the percentage of power sup-
plied to each motor used for stirring the solution in each tank.
At the start of the experiment, tank #1 contains an initial
salt concentration (Morton, kosher). Each tank is fitted with
a 10 k2 thermistor conductivity cell (Cole-Parmer Model
#U-19500-00) connected to a digital readout (Cole-Parmer
Model #U-19300-00), which is used to measure and display
the conductivity near the outlet of each tank. Conductivity is
measured on a scale of 0-10 mS and is correlated to the salt
concentration using a student-generated calibration curve ob-
tained prior to performing the first trial. The concentrations
determined from the time-dependent conductivity measure-
ments are used to determine the mean residence time for each
tank in the series.
Distilled water is fed from a large feed tank to the bottom
of tank #1 using a pump. A rotameter located between the
pump and tank #1 is used to regulate the flow rate of distilled
water to the system. A "T" configuration of valves located
prior to the entrance of tank #1 allows for the option to by-
pass the series of tanks in order to establish a consistent flow
rate prior to introduction to the tank. A LabView-based data
acquisition "virtual instrument" (VI) is used to display to
screen and record to file the conductivity measurements as a
function of time. Data analysis used to determine mean resi-
dence time is performed using the material in Chapter 13 of

Feed Location / nk3
Tank3 (3) Sim
Feed Location

Figure 2. Diagram of the apparatus for the "Residence Time
Distribution for Stirred Tanks in Series" laboratory. The
eight potential input variables that could be investigated
are indicated in the figure. The symbol "" indicates the
location of conductivity probes in the stirred tanks.

Chemical Engineerrng Education

Feed Flow
Rate -

Poster Session Award Paper

Elements of Chemical and Reaction Engineering by Fogler4]
and will not be presented here.

The original objective for this lab experiment was to deter-
mine the residence time distribution (RTD) for each tank in
the series from one set of experimental data, which on aver-
age takes between 30 and 45 minutes. The objective for the
lab experiment was modified to incorporate experimental
design into the procedure so students perform additional
analysis on how changes to various input variables affect their
specific output response. The students were also able to see
how these changes affected the mean residence time and the
residence time distribution for the four trials investigated,
which remains the main objective for the experiment. In this
way, students play an active role in determining the direction
in which the experiment proceeds, while the original techni-
cal objective of the experiment is still achieved.
Since the results from this experiment are typically pre-
sented in an oral presentation, the objective for each group
performing the experiment is different, to avoid duplicate
presentations and keep the students involved in the pre-
sentations. Examples of two objective statements given
to groups are:
Objective 1:As the head engineer at Max Product
Chemical Company, you are required to determine
which ,'j 1i,,i- conditions would allow for the
shortest amount of time for the maximum product

Operating Conditions Available for Potential Input Variables
for the "Residence Time Distribution for
Stirred Tanks in Series" Experiment

Original Operating

Feed flow rate
Initial salt concentration
Tank #1 stirring speed setting
Tank #2 stirring speed setting
Tank #3 stirring speed setting
Tank #2 feed location
Tank #3 feed location
Size of tank #3b

a Setting based on power-regulated percentage applied to stirrer motor.
b Tank is ', .,- ., changed between two -. sizes.

Input Variables

0.3 gal/min
6 g/L
Bottom of tank
Top of tank
3.70 L

LowLevel High

0.5 ga
Top o
Top o

0.3 gal/min
4 g/L
Bottom of tank
Bottom of tank
3.70 L

concentration to be released to the downstream
separator (from ;,,,ni 3). To accomplish this, an initial
product concentration is flushed from a CSTR (tank
#1), using distilled water into a second, and ulti-
mately, a third tank before heading to the separator
located in a remote part of the plant. For this experi-
ment, you will use salt to simulate the product as it
travels ;l,. i, il the series of stirred tanks. Note: The
initial salt concentration in ;,,i- 1 can be varied.
Objective 2: As the lead engineer in the chemical
waste group at Enviromate Industries, you are
required to determine which 7' ...., s conditions
would allow for the lowest maximum concentration of
a toxic substance to be released from the plant (from
;,i1/, 3). To accomplish this, the waste stream under-
goes dilution in a series of three tanks. For this
experiment, you will use salt to simulate the toxic
substance as it travels ;l,. ,i, li the series of stirred
tanks. Note: The initial salt concentration in ;,,, i 1 is
a constant.
The premise of the experiment is that, as engineers work-
ing for a particular industry, each lab group is given a 3-hour
time frame (i.e., one laboratory period) in which the process
equipment is available for their investigations. This period
of time is meant to simulate plant downtime that is scheduled
for them to obtain the necessary data from the trials to reach
their assigned objective. Each group must choose three fac-
tors from the list of eight potential variables given in Table 1
in which adjustment within the given high and low limits
might bring about some desired output response. Students
are encouraged to review the theory and follow
their own curiosity when making their choices. For
example, many groups have performed experi-
ments with the tank #2 stirrer speed set to 0 in
order to investigate nonideal mixing, a behavior
that is not typically covered in lectures. Since this
condition is not covered in the lab procedure ei-
Level their, this leads them to seek additional sources to
help explain their results. The students are actively
l/ min involved in the design of the experiment, so they
/L learn more about the process they are investigating.

Since the experiment is performed in a fixed
0 three-hour time period, students are unable to per-
0 form a complete two-level full-factorial experi-
ftank ment in the allowed timeframe and must perform
ftank a half fraction of the two-level full-factorial ex-
5 L periment (a total of four trials), making relevant
assumptions concerning interaction effects in or-
der to rank the effectiveness of the three chosen
parameters toward achieving the desired response.

Summer 2003

Summer School

The high and low levels for the potential input variables are
indicated in Table 1. The groups are required to keep the in-
put variables that were not chosen at the settings indicated by
the original operating conditions.
Before developing a design, the students inspect the ex-
perimental setup to familiarize themselves with the equip-
ment. Each lab group is required to complete an experimen-
tal design to present to the instructor prior to being cleared to
perform the experiment. The presentation is required to pro-
vide an exact list of the factors that will be investigated and
why, plus provide an order in which the experimental trials
will be performed so a well-developed plan is specified.
Each group submits a one-page summary of their design,
which typically includes a figure, and responds orally to ques-
tions concerning their design approach. Incomplete or unor-
ganized presentation materials are to be updated and resub-
mitted for approval. Any failure to successfully present a de-
sign strategy before the allotted laboratory time results in
forfeiture of the group to perform the experiment at that time
and a penalty in their grade for having to reschedule time for
performing the experiment.

Sample "Objective #1" (given previously) was investigated.
The feed flow rate (variable A), tank #1 stirring speed (vari-
able B), and tank #3 volume (variable C) were chosen as the
three input variables to be investigated to minimize the time
needed for the maximum concentration to leave tank #3. The
output response that was measured was time measured in
seconds. Table 2 gives the experimental conditions and
output responses for a half-factorial experiment that was
performed for trials represented by the gray points given

in Figure 1. Using Eqs (1)
main effect for each
variable gives Bl = -
78.5, 2 = +11.5, and B3
= +41.0. Taking the av-
erage of the time
needed to reach the
maximum concentra-
tion in tank #3 for the
four trials gives a value
for 10 of +362. There-
fore, by plugging in the
determined coefficients
into Eq. (4), the model
for determining the
time in which the con-
centration in tank #3 is a
a maximum becomes

through (3) to solve for the

Experimental Conditions and Outpu
Experiment Investigating Fe
Stirring Speed, and T

Feedflow Tank #1 Tank
rate stirrer volu
Trial (gal/min) setting (L)

(1) 0.3 20 3.71
(ab) 0.5 50 3.7'
(ac) 0.5 20 7.1.
(bc) 0.3 50 7.1:

Time needed to reach a maximum concentra
Theoretical residence times are indicated in

Time (in seconds)= 362-(78.5 -A) +(11.5-B) +(41.0 -C) (5)
For this particular investigation, the allotted time did not al-
low for replicates to be performed, so the error term could
not be determined. Students are required to address this issue
when they are only able to complete four trials during the
laboratory period.
Ranking the chosen input variables in terms of importance
shows that, over the given range available for each variable,
changes in the feed flow rate had the greatest effect on the
output response, while changes in the stirring speed for tank
#1 had the smallest effect. Increasing the size of tank #3 and
the stirring speed for tank #1 had a positive effect on the time,
while increasing the feed flow rate had a negative effect on
the time. Therefore, the model equation given in Eq. (5) clearly
shows that, in order to reduce the time required to elute the
salt from tank #3, the system should be operated at a high
level for the feed flow rate, while operating at a low level
both for the size of tank #3 and for the stirring speed in tank
#1. It is important to note that these optimum operating con-
ditions were determined without obtaining data from that
specific set of operating conditions.
By looking at the experimental residence times determined
for each of the experimental trials investigated, given in Table
2, one sees that the ordering of the residence times follows that
of the elution times for these trials. Therefore, the original ob-
jective of determining the residence time distribution for the
system of three tanks in series remains intact and the role of
residence time can be used in the analysis of the objective.

Student comments on year-end course evaluations have
been positive for using DOE in the unit operations labora-
tory. Overall, they found the "Residence Time Distribution
for Stirred Tanks in Se-
ries" lab performed from
E 2 a DOE approach to be of
it Responses for Half-Factorial industrial importance,
ed Flow Rate, Tank #1 especially those who had
'ank #3 Volume used DOE techniques in
earlier co-ops and intern-
#3 Residence time ships without exactly
me Elution time for systent knowing that they were
(-S- (S- using these techniques.
0 388 544 (586) Course evaluations in-
0 254 357 (352) dicated that the DOE
5 313 454 (461) approach used in the
5 493 714 (769) laboratory was "inter-
esting" and "useful." In
tion in tank #3. fact, one group imme-
parentheses. diately applied their

Chemical Engineerrng Education

I Poster Session Award Paper

new knowledge of experimental design in a subsequent
lab experiment.
In informal discussions, students expressed their enjoyment
of the opportunity to choose their own pathway for the ex-
periment based on the outlined conditions set forth by the
specific objective. The report grades associated with analy-
sis and discussion of the results for this experiment were con-
sistently one of the highest of the semester, presumably due
to the students being more actively involved in the up-front
preparation of the laboratory procedure followed to achieve
the objective and the subsequent analysis.
Future assessments will quantify the success of this ap-
proach by surveying the students on the following goals,
where students
Recognize applicability of DOE in other labs and, most
especially, its potential usefulness in future career
Have a better understanding of how to efficiently
optimize a system.
Appreciate how DOE can enhance the understanding
of how changes to input variables on/,.,. t a process.

Incorporation of design-of-experiments techniques in the
undergraduate unit operations laboratory provides students
with an industrially useful approach to test how improve-
ments can be made to an existing system without intimate
knowledge of how the process works. Incorporation of DOE
into current laboratory experiments is relatively simple, pro-
vided the experimental trials do not require a significant
amount of time to perform or modify from trial to trial and
that the original objective of the laboratory is not diminished.
Students are receptive to performing experiments using DOE
since the industrial significance of this approach is high and
has the potential to be used in future career endeavors.

1. Box, G.E.P., W.G. Hunter, and J.S. Hunter, Statistics for Experiment-
ers, John Wiley & Sons, New York, NY (1978)
2. Montgomery, D.C., Design andAnalysis ofExperiments, 5th ed., John
Wiley & Sons, New York, NY (2001)
3. Hicks, C.R., and K.V. Turner, Fundamental Concepts in the Design of
Experiments, 5th ed., Oxford University Press, New York, NY (1999)
4. Fogler, H.S., Elements of Chemical Reaction Engineering, 3rd ed.,
Prentice Hall, Upper Saddle River, NJ (1999) O

Teaching is Easier When Students Actually Read the Text

Two-Way Attack on Lab Reports Er Theses

Books in Printed Er Electronic Forms

Technical Style
Technical \ ,a in
a Di, ;t. Age
by J. M. Haile
Your students can be better
writers. This book shows
students how to improve
and explains why the
improvements work. Cov-
ers language, punctuation,
tables, graphics, equations.
208 pp. (5.38 x 8.38) Jan 02
pbk. $29.95
ISBN 0-9715418-0-9
ebook $17.95
ISBN 0-9715418-6-8

Analysis of Data
by J. M. Haile
This book shows how to
assess quality of measure-
ments, how to search for
linearizing functions, how
to test for correlations, and
more. Discussion of uncer-
tainties follows the stan-
dards adopted by the
American National Stan-
dards Institute (ANSI).

128 pp. (5.25 x 8.25) Jun 03

pbk. $19.95
ISBN 0-9728602-0-7

Lectures in
Heat and Work
by J. M. Haile
The language, organiza-
tion, and approach used in
this book demystifies ther-
modynamics and helps
students develop confi-
dence that they can use the
material successfully.
592 pp. (7 x 10) May 02
pbk. $49.95
ISBN 0-9715418-1-7
ebook $24.95
ISBN 0-9715418-5-X

Our books are in pdf for-
mat, so they faithfully
reproduce all text, equa-
tions, and figures. These
books can be read using
the (old) Adobe eBook
Reader or the new Reader
6.0 (free from Adobe ).
About the Author
J. M. Haile has 24 years of
experience in teaching sci-
ence, engineering, and
technical communications.
He has been North Ameri-
can editor of the journal
Molecular Simulation and is
the author of Molecular
Dynamics Simulation.

To learn more about these books, visit our website at http:/ /



FAX (877) 898-0287

Summer 2003

Summer School


A Laboratory Structure Encouraging Realistic

Communication and Creative Experiment Planning

Lafayette College Easton, PA 18042

ne of the difficulties of structuring a laboratory class
is how to leave problems open-ended while at the
same time provide the students with enough infor-
mation so they know what is expected of them and how to
get started. The key goal in our effort to achieve such a bal-
ance was to create a setting where students would have
A realistic communication experience, including both
memo -, ii and oral presentation
The opportunity to plan sequences of experiments, in-
cluding the use ofstatistical design of experiments
An opportunity to design their own studies
Our basic approach was to have student groups write memos
to each other and to build on each other's work. This was
done in the context of a one-semester unit operations labora-
tory course that begins with simple unit operations and builds
to running and simulating distillation. It is the middle semes-
ter of a three-semester sequence of laboratories. The first labo-
ratory is a series of bench-top heat, mass, and momentum
transport experiments that includes instruction and practice
in some basic statistical analysis, laboratory planning, and
formal report writing. In the second laboratory, the subject of
this paper, students work in our unit operations laboratory,
which includes pilot-scale chemical process equipment. It
includes simple systems for investigating the operation of
pumps, flow in pipes, heat exchangers, and flow in packed
beds, building to eventually operating and analyzing two dis-
tillation columns. Also included are experiments that teach
students how to calibrate and use a gas chromatograph and
experiments that measure vapor/liquid equilibrium using an
Othmer still.[1]
This sequential approach has been used for the last three
years. There have been some variations in the details of the
structure in various semesters. The specifics described here
were used for the spring 2001 semester. In that semester, we

ran three sections of laboratory with three groups of three
students in each laboratory section. We have also used a simi-
lar setup with four groups per section.

The semester was organized into three distinct parts: 1) a
three-week introduction to sequential experimental design
using a fluidized-bed coating experiment, with the students
working in small groups, 2) a seven-week set of simple unit
operations experiments, where the small groups pass on their
results, and 3) a four-week distillation section where larger
groups work on distillation experiments, simulation, and pre-
sentation. Table 1 shows the experimental schedule for the labo-
ratory. The three parts of the laboratory are described below.
Introduction to Sequential Experiments
The initial block uses a fluidized-bed coating process based
on a design from Rowan University.[2] A washer is heated

S. Scott Moor is Assistant Professor of Chemi-
cal Engineering at Lafayette College. He re-
ceived his BS and MS in Chemical Engineering
from M.I.T. After a decade in industry he re-
ceived his PhD in Chemical Engineering and his
MA in Statistics from the Universityof California
at Berkeley. His research interests include spray
and particle systems and the development of
materials for science, mathematics, and engi-
neering education.

James K. Ferri is Assistant Professor of Chemi-
cal Engineering at Lafayette College. He re-
ceived his BS and PhD degrees from Johns
Hopkins University His teaching areas include
transport phenomena, unit operations, and labo-
ratory applications. His research interests lie in
interfacial phenomena and bioengineering.
SCopyright ChE Division of ASEE 2003

Chemical Engineerrng Education

Poster Session Award Paper

with a heat gun and dipped in a fluidized bed of colored poly-
ethylene powder, which melts to form a coating on the washer.
The main experiment is to examine the impact of preheat
temperature and dip time on coating thickness. This is mea-
sured by the change in weight of the washer. The first week
is devoted to detailed experimental familiarization. Students
are presented with a laboratory plan that instructs them to
determine the practical operation limits of the process (e.g.,
what is the temperature range that will produce coatings and
is achievable with this equipment?), to spot check the cali-
brations (e.g., test the thermocouple reading in ice water and/
or boiling water), to develop a procedure including a written
standard operating procedure (SOP), and to test the repro-
ducibility of the system by completing three to five coatings
at the same conditions. This week helps students get used to
the idea that we will not be telling them all the details of how
to run the equipment and shows them how they can approach
a new process.
In the second week the students are expected to design and
complete a 24-1 fractional factorial experiment examining the
impact of preheat temperature, dip time, washer size, and flu-
idization air-flow rate.[31 During this week, students are learn-
ing about fractional factorial experiments and the important
idea of screening variables to determine which ones can be
safely ignored.
In the third and final week of the block, students design
and complete a two-variable central composite design.[41 Here
they are developing a detailed second-order empirical model
of the coating process. This three-step sequence allows them
to get used to the idea that experiments can build from one to

the next. This sequence gives them some ideas for their own
sequential experiments in the next part of the laboratory and
an introduction to some important statistical experimental
designs. To make some of the logistics of the laboratory work,
the third week of the initial block is actually interleaved with
the second part of the laboratory.

Simple Unit Operations Experiments
Students carry out seven simple chemical engineering unit
operation experiments. This is the heart of our building ap-
proach. There is only one apparatus of each type, so each
group is working on a different experiment in any given labo-
ratory session. The experiments are broken into two groups:
one group of four experiments (Block II) and one group of
three experiments (Block III). Experiments from a given group
are mn every other week to allow time for memos to be passed
from one group to the next (see Table 1, particularly the ev-
ery-other-week arrangements of Blocks II and III). A brief
description of the seven setups is given in Table 2 (next page).
Students are given a simple introduction to each experi-
ment. These introductions present the general nature of the
unit operation, list the capabilities of the equipment, and ref-
erence sections in their texts where they can find additional
information. Table 3 (next page) is an example experiment
There are three groups in a section, so each experiment is
mn three times. The first week an apparatus is used and stu-
dents are required to complete familiarization runs examin-
ing calibrations, limits of operation, procedure, reproducibil-
ity, and an initial base experiment. A memo to the instructor
and to the next group that will use
the equipment is due one week after

the laboratory. The interleaving of

experiments allows two weeks be-
tween runs so that the subsequent
group has time to plan experiments
based on the first group's runs. The
second group to use an apparatus
generally uses the most obvious ca-
pabilities of the equipment. This
week ends up being the experiment
we had students run when they did
not pass information. They pass their
results to a third group and it is re-
sponsible for finding ways to explore
creative new ideas based on the pre-
vious two groups' work.
During this part of the course, each
group is required to orally present
their results twice. The presentations

Summer 2003

Overall Laboratory Schedule Showing Interleaving of Experimental Blocks

Week Block Group 1 I Group 2 I Group 3
1 I Fluidized Bed familiarization and characterization
2 Fluidized Bed fractional factorial design
3 II Flow in pipes and valves I Pump Performance I Heat Exchanger
4 I Fluidized Bed central composite design (Response Surface Methods)
5 II Pump Performance Heat Exchanger GC
6 III Mixing Packed Bed Othmer Still
7 II Heat Exchanger GC Flow in pipes and val
8 III Packed Bed Othmer Still Mixing
9 II GC Flow in pipes and valves Pump Performance
10 III Othmer Still Mixing Packed Bed
11 IV Total Reflux Distillation
12 Simulation of distillation
13 Continuous Distillation
14 Presentations, Evaluations, Clean Up



Summer School

are video taped and the students are required to review the
tape and complete a simple self-evaluation.
The process of transferring memos from one group to an-
other is facilitated by using a web-based document storage
and transfer area, which is part of the Blackboard 5 software
system[51 we use at Lafayette College. We use Blackboard's
group function to set up a group for each laboratory section.
Students can then upload their reports to their section's file
exchange area and download other students' reports.
We use several mechanisms to ensure the planned experi-
ments are safe and reasonable. Each group is required to sub-
mit a pre-laboratory report to the instructor two class days
before the laboratory. It must include an experimental plan
with a complete run list, a review of the safety issues and
precautions, and an SOP they will follow. The instructor re-
views these reports before the laboratory meets. An instructor
is also present when the experiments are run (we are an under-
graduate institution and do not have graduate TAs to assist).
Some institutions may need to run more than the three or
four groups that are usual for our laboratories. Having two
sets of groups that are running out of sync could double the
number of groups per section. One set of groups could be
running the Block II experiments while another set of groups
is running the Block III experiments-the next week they
switch. The two sets of groups essentially function as two
sections within one time block. This type of arrangement could
accommodate up to eight groups per section.
Distillation Experiments
The last four weeks of the term the students are reconfigured
into two larger groups to run experiments on our two distilla-
tion columns. In this block, each group builds on its own
work. The first week they become familiar with the equip-
ment and run it in total reflux. The second week is held in the
computer laboratory and the students use the process simula-
tor, ChemCAD,16 to model their distillation column. They
compare their results from the previous week to simulation
results and predict how their column will perform in con-
tinuous operation. The third week they run the columns in
continuous operation, and during the final week they deliver
an oral presentation of the results of their distillation studies.

A first indication of the success of this approach was the
positive performance of students in the laboratory. Providing
students with only a simple introduction to the laboratories,
such as is seen in Table 3, students succeeded in developing
and planning these laboratories themselves.
The students' efforts during the third week of a laboratory
were particularly encouraging. For time-consuming experi-

ments such as the Othmer still, this structure allowed stu-
dents to accumulate a significant body of data. For the fluid-
flow experiments (pump performance, pipes/valves/fitting,
and mixing), many students chose to explore the impact of
different viscosities. They were able to adjust viscosity using
propylene glycol antifreeze solutions. Some students com-
pared measured pump efficiency to the temperature gain in
the system for the pump experiment. For the packed bed, one
group completed experiments examining the impact of tem-
perature on the pressure drop.
In some cases, students' suggestions have led us to improve
the capabilities of the laboratory. Recently, one student wanted
to look at the impact of impeller diameter on pump perfor-
mance. This led us to purchase a new centrifugal pump with
interchangeable impellers. While this was an additional ex-
pense, it improved our facilities and made them more flex-

List of Experiments in
Simple Unit Operations Blocks II and III

Flow in pipes and valves
This is a simple fluid-flow apparatus consisting of a feed tank, pump,
rotameter, and piping network. The piping network has a number of
segments where pressure drop can be measured, including an orifice
meter with interchangeable plates, two valves, a straight run of pipe,
and a run of pipe with two elbows.
Pump Performance
This apparatus consists of a turbine pump and a centrifugal pump
piped in parallel and instrumented to measure inlet and outlet
pressure, total flow, and the power drawn by the motor.
This apparatus includes three mixing tanks of various sizes with
removable baffles. We use a fully adjustable laboratory mixer with
torque and speed readouts. Three different types of impellers are
available. The mixing of solid particles was examined.
Heat Exchanger
This apparatus consists of a small shell-and-tube heat exchanger with
temperature sensors on all inlet and outlet streams. A steam injection
heater is used to provide hot water. Simple three-way valves are used
to reverse which stream is on the tube side and which stream is on the
shell side. Both flowrates can be varied.
Packed Bed
A pilot-scale absorption bed, packed with half-inch Intalox saddles, is
used for this experiment. In this laboratory, students are simply
looking at the pressure drop at various air and liquid velocities. In
addition, they are introduced to the concept of flooding.
This experiment uses a Varian 3000 gas chromatograph. Students are
expected to learn how to operate the equipment and to develop a
calibration curve for methanol/water mixtures.
Other Still
This bench-scale equilibrium still is used to develop vapor liquid
equilibrium information for the methanol/water system.

Chemical Engineerrng Education

Poster Session Award Paper

ible for future years.
TABLE 3 At the end of the term, a random sample of junior-year
Example Laboratory Handout chemical engineering students was interviewed as part of our
ABET assessment process. The interviews were about their
Heat Exchangers entire experience as chemical engineering students and were
Heat exchangers transfer heat from one fluid to another. They are conducted by professors who were not their instructors. One
essential to the operation of most chemical plants. In this equipment
we will see how a simple countercurrent heat exchanger functions of the questions asked was, "How have your experiences this
under various conditions and compare the results to usual design year developed your ability to solve engineering problems?"
correlations. As students responded to this question, one of the primary
Equipment and Capabilities areas of discussion was the value they saw in this particular
Equipment and Capabilities
The heat exchanger setup has a cold tap-water feed and hot-water laboratory. They expressed satisfaction at being able to de-
feed. Each stream can be fed to either the shell side or the tube side. sign their own experiments and felt that writing memos to
Temperature sensors allow monitoring of both inlet and both outlet each other gave them a realistic experience.
To further assess student response to this course, the stu-
With this apparatus you can
Estimate the overall heat transfer coefficient for both flow dents completed a written survey one year after they had been
configurations (hot-water shell side and hot-water tube side) in the class. Table 4 lists the results for quantitative ques-
Look at the variation in heat transfer coefficient for various flow tions. They were asked to rate on a 1-to-5 scale how well the
rates of each fluid and both flow configurations statement reflected their experience. Students generally felt
Look at the effect of hot stream temperature on heat transfer the laboratory contributed to their ability to write memos, to
Estimate the time constant of the heat exchanger the laboratory contributed to their ability to wrte memos, to
their ability to contribute to a sequential study, to their ability
References to use statistical methods to plan experiments, and to their
1. McCabe, Smith, and Harriott, Unit Operations .. ability to present their work. Scores were highest for writing
Engineering, 6th ed., pp. 315-324 (particularly 321-323), 329- and lowest for statistical design of experiments. These low
330, 346-348, 440-441, 1071 and lowest for statistical design of experiments. These low
2. Bird, Stewart, and Lightfoot, Transport Phenomena, pp. 465- scores are a result of a small number of students who rated
467 this area low. They expressed a desire for a separate course in
statistics. The impact of these few students can also be no-
Specific Requirement
Each group should measure the overall heat transfer coefficient for ticed in the higher standard deviation.
several conditions and compare the result to coefficients estimated by In addition, several open-ended questions were asked that
standard correlations. required a written response. Two of these questions are shown
in Table 5. Most students mentioned some form of commu-
nication skills as both a key goal and as what was most help-
TABLE 4 ful to them. For both questions, many mentioned the use of
Results of Quantitative Questions in Survey sequential experiments and statistically designed experiments.
Students were asked to rate how much they agreed with the statement It is interesting that a quarter of the class cited the open-ended
on a scale of ] to 5 (5 being "strongly agree"). Results are for nature of the experiments as a help to their learning but seem
responses from 26 students (out of 2) who took the class. to have missed that this was one of our goals. Almost all stu-

Standard dents mentioned at least two of the items as helpful. The re-
Mean Deviation sults presented here are categorizations (a content analysis)
During Laboratory II, my ability to write a memo of the students' written responses. That the categories corre-
that clearly communicated my results to my peers spond for the two questions shows that the communicated
improved 4.5 0.65 goals of the class and the resulting learning were connected.

As a result of my experience in Laboratory II, I feel Many of the specific responses to the question concerning
better able to contribute to a sequential exper- what the students found helpful were very encouraging. Some
mental study 4.1 0.63
of the things they noted were
I am better able to use statistical methods to plan The openness and freedom to design experiments that
engineering experiments. 3.8 1.07
investigated your personal area of interest

My ability to prepare and present my own experi- The ability to work on a piece of equipment that I have never
mental work improved over the course of the seen before and learn its functionality by setting up an
semester 4.2 0.59 experiment and seeing it through.
Having to force nim'lf to answer engineering questions

Summer 2003

Summer School

Response to Two Qualitative Questions
This table summarizes a content analysis of students'written responses to two questions.

Values are the percent ofrespondents who
included the areas below in their
discussion of these two questions

Communication skills
Planning sequential experiments
Statistical design of experiments
Understanding of important unit operations
Ability to work in teams
Understanding distillation
Ability to handle open-ended problems

Number of students responding to question

rather than following a set recipe.
*I actually realized the importance of all the background
work I had done for the previous .. -,,,, i.,.-,-, i'years.
One particularly encouraging comment that appeared was
At no other point at L aiii'e'tt' was I able to understand and
experience truer "engineering. "
In this survey, we also asked, "If I could change one thing
about the structure of Laboratory II, it would be..." Twenty
percent of the respondents indicated they would change noth-
ing. Twenty-three percent found some problems with the
structure. They noted that they had difficulties if the previ-
ous group did a poor job or was late with its report. Two
students went so far as to request that we be stricter with our
deadlines. The major drawback and difficulty of this struc-
ture is the dependence of students on each other's work. In-
structors can work to make sure the previous group is on time
and that the quality of its work is taken into account when
grading reports, but we believe this interdependence among
the students is a realistic and helpful experience.
Since this survey was taken, we have made the report dead-
lines stricter, including a strict 10%-per-day late penalty. This
has essentially eliminated late reports. Students are also given
a grading sheet that lists the key items we will be looking for
in each report. This has led to more consistent student memos.
Finally, student memos are graded considering how they have
improved upon previous work.

The key aspects of the approach we are using are
LE It allows students to build on each other's experiments

What did you see The aspects of the laboratory
as the key educational that were most helpful to my
goals of the laboratory? educational experience were...

LE It provides only minimal information on how to run the
equipment and leaves that as the first job to be done (which
develops independent thinking by requiring students to
develop their own SOPs)

LE It provides students with an example of a sequence of
experiments, including how to get started with a new piece of

LE It requires students to communicate their results via memos
between the groups.

This approach provides a setting for realistic communica-
tion, experience in working on realistic sequences of experi-
ments, and an opportunity for students to design their own
studies. Student feedback and instructor observations sup-
port that these goals are being met. Student response is gen-
erally positive, as seen by the responses in general depart-
ment interviews and in assessment surveys. A key to success
is to manage the process so that students are not unfairly im-
pacted by deficiencies in previous groups.

1. Othmer, D., Analytical Chem., 20, p 763, August (1948)
2. Slater, C.S., and R.P. Hesketh, "Fluidized Bed Polymer Coating Ex-
periment," Proc. Conf Amer Soc. Eng. Educ., Session 3413 (2000)
3. Box, George E.P, William G. Hunter, J. Stuart Hunter, Statistics for
Experimenters: An Introduction to Design, DataAnalysis, and Model
Building, John Wiley & Sons, New York, NY, p. 374 (1978)
4. Box, George E.P, and Norman R. Draper, Empirical Model-Building
and Response .. John Wiley & Sons, New York, NY, p. 306
5. Blackboard, Inc., Welcome to Blackboard, n.d.>
6. Chemstations, Inc., n.d., 5

Chemical Engineerring Education

Summer School

A Course in Bioprocess Engineering
Continued from page 183.
1, II,. iIt,, I in publishing a BioengineeringNewslet-
ter (10%)

This course has been taught only once under its present
title, "Bioprocess Engineering"-in the spring of 2002. As-
sessment of whether or not the learning goals were achieved
in this class was determined through exit surveys and teacher
course evaluations. Some examples of questions that were
asked of the students were

Did they feel the amount of learning they experienced
was better as good as, or less than in other core
engineering courses?
Did they think this course helped them to see where
key issues of transport, thermodynamics, and reaction
kinetics play a role in biological systems?
Did they feel confident in their ability to begin
employment in afield related to bioprocess engineer-

The responses were divided: about 2/3 of the class partici-
pants responded positively to the course and felt it was a good
introductory course to enable their further advanced studies
in this area, while the other 1/3 did not find the course par-
ticularly beneficial. Some commented that the course was
"very interesting," "tied in concepts previously learned in
other courses in bioprocessing," and "the homework labora-
tories made the corresponding parts of the lecture more col-
orful." Because there was no formal text for the course, some
students "had difficulty in finding additional information to
solve some homework problems." A reserve list of library
texts covering the course material was always made avail-
able in the Engineering Library, but access to on-line resources
may be a more effective way to disseminate these materials.

To our surprise, most of the ambitious goals set for the
course were accomplished. Such excellent accomplishments
were due to the students' willingness to participate in the
outside classroom activities. We could have done even
better if more students had had some university-level bi-
ology background.
The student engineers adapted well in the conceptual and
computational analyses in the classroom dealing with the bio-
logical system, but some had difficulty in growing microbial
culture. This was quite understandable, and the Department
of Chemical and Biomolecular Engineering is moving to-
ward requiring one semester of biology for all engineer-

Summer 2003

ing students in the future.
Many students who took the course were excited about the
prospect of molecular-biology driven bioprocess engineer-
ing and appeared to have enjoyed the less stressful approach
to the subject, although it was a demanding course. We in-
tend to improve the quality of the course by analyzing formal
and informal assessments of the course. In the future, it would
be advisable to solicit an external evaluation of the course.

1. Schreuders, P., and A. Johnson, "A Systems Approach for Bioengi-
neering," Int. J. Eng. Ed., 15, 243 (1999)
2. Coger, R., and H. DeSilva, "An Integrated Approach to Teaching Bio-
technology and Bioengineering to an Interdisciplinary Audience," Int.
J. Eng. Ed., 15, 256 (1999)
3. Sambrook, J., E. Fritsch, and T. Maniatis, eds., Molecular Cloning: A
Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press
4. R. Altas, and L. PaI CRC Press,
Boca Raton, FL (1993)
5. Benedict, S., "A Reagent for the Detection of Reducing Sugars," J.
Biol. Chem., 5, 485 (1909)
6. Berthelot, J., F. Fournier, and M. Fournier, "Convenient One-Pot Syn-
thesis of 4-Bromo-1,1-Diphenyl-1-Butene-3-One," Comm., 15,
213 (1985)
7. Farrell, S., R. Hesketh, J. Newell, and C. Slater, "Introducing Fresh-
men to Reverse Process Engineering and Design Through Investigatiion
of the Brewing Process," Int. J. Eng. Ed., 17, 588 (2001) 5

e letter to the editor

To the Editor:
The recent article (Pitt, M.J., and J.E. Robinson, "Using a
Commercial Movie for an Educational Experience," Chem.
Eng. Ed., 37(2), 154, 2003) on the use of movies for educa-
tional experience is interesting, and may I suggest another
possible movie for study...The Man in the White Suit. A bril-
liant young polymer chemist, working unofficially in a cor-
porate research laboratory, develops a fabric that "never gets
dirty, never wears out." At first this is seen as a blessing, but
very soon it is seen as a threat to the entire textile industry,
and attempts are made to suppress it.
Although The Man in the White Suit was made as a com-
edy, it says much about the role of industrial research and its
economic and social impact. In some ways it is very dated,
as it was made over fifty years ago; but today's students who
see it could be asked to explain how it is dated and also to
identify the aspects that are still highly topical. When I first
saw it as a teenager, it reinforced my interest in applied chem-
istry, and that eventually led me to chemical engineering.

Malcolm Baird
McMaster University

Summer School


Fast Cars Accelerate Learning

"FEEL THE NEED FOR SPEED? Ifyou think the sound ofan F-1 engine cranking up to
18,000 rpm coming out of turn 8 at Hockenheim is up there with Vivaldi and Mozart,
you 're probably going to like this course. Ifyou know there is no such ;llg as too much
horsepower and torque under your right foot, this may just be the course for you. If you
would rather go to an Auto Show than go to a casino to gamble, hop aboard. Ifyou think
that a car is really for getting between Bernalillo and the main campus, there are prob-
ably other courses you would enjoy more."

The University of New Mexico Albuquerque, NM 87131-0001

he "High-Performance Engines" (HPE) course at The
University of New Mexico integrates problem-[1,21 and
simulation-based[3-51 learning approaches. The course
encourages students to synthesize core chemical and mechani-
cal engineering principles (e.g., thermodynamics, transport
phenomena, kinetics, catalysis, mechanics, and dynamics) to
analyze historically famous racing engines. With indepen-
dent analysis based on computer simulations, the students in
groups of 3 or 4 get to understand engineering innovations
that lead to the engines' superior performance in auto racing.
The analysis transforms abstract engineering concepts into
concrete, object-oriented outcomes.
The students also discuss the design parameters (e.g., com-

Sang M. Han is Assistant Professor of Chemical and Nuclear Engineer-
ing at the University of New Mexico. He earned his BS in Chemical Engi-
neering with honors from the University of California Berkeley (1992) and
his PhD in Chemical Engineering from the University of California Santa
Barbara (1998). His primary research interest is in micro- to atomic-scale
surface science and engineering on Si and Ge based systems.
Joseph L. Cecchi is Dean of the School of Engineering at the University
of New Mexico and Professor of Chemical and Nuclear Engineering. He
earned his BS from Knox College and his PhD from Harvard University.
His primary research is in semiconductor and MEMS fabrication technol-
ogy on the micro and nano scales.
John J. Russell is Interim Chair of the Department of Mechanical Engi-
neering at the University of New Mexico. He earned his BS in Aeronauti-
cal and Astronautical Engineering from the Ohio State University and his
PhD in Aerospace Engineering from the University of Michigan Ann Arbor.
His research interest lies in vehicle dynamics and controls.

The format of the HPE
course is a problem-based
learning environment centered
on case studies of specific engines,
many with great historical significance.

pression ratio, intake manifold geometry, and valve timing)
to improve engine performance. The groups compete through
modeling and simulation to achieve the best improvement.
The educational competition augments students' interest in
the course subjects and promotes their active participation.m']
To compare the improvement, students rely on WAVE, an
engine design software provided by Ricardo Software. It en-
ables custom engine design and performance simulation by
simultaneously solving momentum, energy, and mass bal-
ance equations. Since auto manufacturers such as Ford Mo-
tor Company use the software to design production engines,
the software training advantageously increases students' ca-
reer potential in automotive design engineering.
The course also prepares a group of University of New
Mexico students for the collegiate Formula Society of Auto-
motive Engineers (FSAE) Competition.[6] The participating

Copyright ChE Division ofASEE 2003

Chemical Engineerrng Education

Poster Session Award Paper

students optimize the competition car design by using WAVE.
Figure 1 illustrates how the course exploits the racing en-
gines as the centerpiece to integrate core chemical and me-
chanical engineering concepts while promoting active learn-
ing and preparing our students for the national FSAE Com-

A technical elective course such as High Performance En-
gines provides a convenient venue to explore problem-P1,2'
and simulation-basedE3'51 learning. These learning approaches
have been proven to promote active student participation and
enrich students' learning experience.[1 We introduced the
course in the fall of 2001 and taught it again in the fall of
2002 to the upper division students in the departments of
chemical engineering and mechanical engineering.
During the conception of the course, our strategy was to
exploit high performance engines as the centerpiece to draw
students' interest and integrate the two learning approaches
around a central topic. If the course became successful, we
anticipated that it might lead to additional courses that ex-
plore the rest of the automobile.
Our first task was to identify the potential audience and to
promote the course. We advertised that the course would be
about what makes automobiles go, and in particular, what
makes them go fast. We also placed the focus of the course
squarely on high-performance internal combustion engines.
Although much more goes into an automobile or a racecar to
make it perform well (e.g., transmission, suspension, steer-
ing, tires, brakes, many forms of electronic instrumentation,

Underlying Principles
Transport phenomena

Sound wave dynamics

Design Software WAVE
Design parameters

and controls), our perception was that the high-performance
engine would be a good place to start. After all, the engine is
the power source, or more precisely the "transformer," of
chemical energy into mechanical energy.
The HPE course appeals largely to students who view cars
as more than just transportation and who regard an engine as
the soul of an automobile. Since the course is geared toward
car enthusiasts, these students are highly motivated to learn
the subject matter. The course also reflects local history and
general public interest. Albuquerque is the hometown of the
legendary auto racing family, the Unsers. Auto racing is a
tradition in Albuquerque and has a loyal following, includ-
ing Formula 1 (F-1) and National Association for Stock Car
Auto Racing (NASCAR) circuits. Because the University of
New Mexico serves largely in-state students deeply rooted in
the local community, the HPE course certainly appeals to a
wide audience, including some of the alums who wish to re-
turn for continuing education.

With the target audience clearly identified, our intention
was to help students see how fundamental engineering prin-
ciples come to life in internal-combustion, spark-ignition en-
gines. The engine operation and design require integrative
understanding of thermodynamics, transport phenomena, ki-
netics, catalysis, mechanics, and sound wave dynamics. Stu-
dents get to capitalize on the application of these engineering
fundamentals to engine operation and design and see how
they come to life. The course relies heavily on computa-
tional tools and information technology for modeling,
simulation, visualization, and design of historically fa-
mous racing engines.

Figure 1.
A conceptual diagram
of the HPE course
that exploits students'
interest in concrete
objects to integrate
core chemical and
mechanical engineering
principles via problem/
Problem-based Learning simulation-based,
- Classical racing engines active learning while
- Group competition preparing them for
national FSAE
(Picture of Ferrari FX
is available via

Summer 2003

Summer School

We chose Lumley's Engines: An Introduction81 and
Heywood's Internal Combustion Engine FundamentalsE'1 as
the main textbooks for the course. Lumley's book surveys
the relevant topics of engine operation and carries enough
depth for the students with limited hands-on experience in
auto engines. This simplicity allows students with very basic
mathematics to appreciate the engineering fundamentals be-
hind engine operation. Overall, Lumley's book is an excel-
lent option to broaden the potential audience by lowering the
barriers for those who may not have extensive knowledge of
auto engines. It also uses the Stanford Engine Simulation Pro-
gram (ESP) available via . This
freeware lets students explore the engine's operating charac-
teristics and thermodynamic performance as a function of
geometric parameters of engines. The program is a straight-
forward and effective way to visualize methods for improv-
ing engine performance. We discuss other software tools
later in this paper.
In comparison, Heywood's textbook thoroughly discusses
operating principles with in-depth rigor. The topics in both
books range from ideal models of engine cycles, to engine
breathing, to engine heat transfer, to engine operating char-
acteristics. These topics naturally lead to a discussion of sub-
jects that are largely familiar to chemical and mechanical en-
gineers, such as pressure vs. volume (P-V) diagrams of ideal
Otto cycle, flow separation, thermal boundary layer,
combustion, and mechanical vibrations and balanc-
ing. We added Heywood's book as a response to im-
prove the course, based on a student survey con-
ducted in the fall of 2001. The students expressed a
need to see more rigorous analysis of engine opera-
tion in addition to the cursory demonstrations in
Lumley's book.

The first topic discussed in HPE is the Otto en-
gine cycle. Although the ideal Otto cycle is reviewed
in numerous engineering thermodynamics courses,
students often fail to appreciate its engineering sig-
nificance. Our goal is to relate the P-V diagram to
tangible engine performance so that students depart
from abstract understanding of the diagram. We
achieve this realistic understanding by first visual-
izing how each stroke (i.e., intake, compression, ig-
nition, expansion, blowdown, and exhaust) corre-
sponds to the piston position between the top dead
center (TDC) and the bottom dead center (BDC) of
the cylinder, as illustrated in Figure 2.

We simultaneously introduce students to a concept of mean
effective pressure (mep) as an artificial measure of thermo-
dynamic performance of an engine. The mep value is inti-
mately related to the P-V diagram. The product of mep and
total engine displacement (Vd) is work done by the engine
per power cycle, and the product is simply the shaded area
carved out by the piston trajectory in the P-V diagram.
The power generated by the engine, a number that we all
look at before purchasing a vehicle, is merely a product of
the work produced per power cycle and the engine speed (N)
in revolutions per second. This proportionality explains pre-
cisely why the brake horsepower (bhp) increases more-or-
less linearly with the engine speed, often given in revolu-
tions per minute (rpm), until the engine breathing becomes
the limiting factor at high rpm.
This example demonstrates how an abstract concept can
come to life by associating it with concrete objects and visu-
alization. We expand on this approach with subsequent course
materials by leading the students to make such associations.
For instance, students directly relate intake manifold tuning,
intake runner design, intake/exhaust valve size and number,
intake/exhaust valve lift and overlap control by cam, and su-
per/turbo charging to the engine performance figures. Mak-
ing such connections requires fundamental understanding of
sound-wave dynamics, turbulence and combustion, con-
tinuum fluid flow, mechanics, and advanced thermodynam-

D Z Mean Effective
Satm placement Pressure (mep)

t Volume t

Figure 2. P-V diagram of ideal Otto engine cycle. Mean effective pres-
sure (mep) is an artificial measure of an engine's thermodynamic
performance intimately related to the P-V diagram. The generated
power is simply a product of mep, total engine displacement (Vd),
and engine speed (N) in revolutions per sec divided by the number of
cycles per power stroke (x).

Chemical Engineerrng Education

Work=mepxVd; Power = mepxV N
I x

I Poster Session Award Paper

ics. When the abstract engineering concepts are closely linked
to tangible parts and concrete engineering outcome (e.g.,
horsepower and torque), students become capable of retain-
ing the learned information, and their desire
to understand the engineering subjects re-
mains strong throughout the course. Th

The instructors formulate the case studies synth
that will be the focus of class discussion. c
These case studies require students to inves-
tigate the engineering topics mentioned pre- e
viously. The instructors act primarily as fa- eng
cilitators to provide some degree of continu- prince
ity and structure for the course. They essen- therm,
tially provide a forum for the extensive par- tr
ticipation of students in this learning process. ph e
Each case study starts with a brainstorm- ki
ing session moderated by the instructors. Stu- catalysi
dents identify relevant engineering topics and and di
areas that they wish to focus on. From time analyze
to time, the instructors themselves may be famo
resources of knowledge in helping to explain
concepts, but the faculty do not lecture in the
conventional manner. They are, however, re-
sponsible for introducing modeling software
(e.g., ESP, WAVE, and Working Model 2D) and ensuring that
students have adequate access to running these codes. The
faculty members are expected to seek productive feedback
from the students and to act on the feedback to improve the
Finally, the faculty evaluate student learning based on mea-
sures that truly reflect what students have gained in knowl-
edge and competence. Some of the evaluation criteria are
The team 's understanding of newly acquired
The depth and diversity of learning resources
provided by the students
The students'active participation during each case
Thorough documentation of quoted references
Students are expected to have taken engineering thermo-
dynamics before enrolling in the HPE course and instructor
approval is required for those who have not taken it. Stu-
dents are expected to attend all classes and to be thoroughly
At the beginning of the course, students form teams of 3 to
4 members each, and all activities during the semester are
team-produced. Students can name their teams, for instance,

e co
s, m

Summer 2003

after Fl, IRL, and NASCAR race teams and compete during
each case study. They are expected to participate extensively
and fully in all aspects of working out the cases. This may
include initial brainstorming sessions, con-
cept identification, problem solving, and
urse resource discovery. Teams are responsible
the for developing their own case study with
es guidelines provided by the faculty.
ts to
'e core Although no one taking this course should
11 and be concerned about his or her grade, for the
ical record, the grade is based on classroom par-
ticipation, creative contribution to analyz-
ering ing case studies, production of creative ma-
?s (e.g., trials and learning resources, and the qual-
namics, ity of case studies on performance engines,
ort each of which is approximately equally
iena, weighted. For instance, students can bring
ics, real mechanical components to class, such
echanics, as superchargers or camshafts, to demon-
mics) to state their working principles relevant to
torically the corresponding case studies. Such par-
racing ticipation reflects not only initiative, but
acing also the students' creative approach to class-
es. room presentation.
Since individual performance needs to be
weighed appropriately in addition to the
overall team performance, each team is responsible for a self-
assessment of the performance of each team member. The
grade for each student is based on overall team performance,
peer evaluation, and self-assessment. The evaluation metrics
chosen for the HPE course is available from
Finalized%20Materials%20for%20CD/cats.htm>. the "Peer
Rating Factor"10 ] on this website is particularly useful in for-
mulating the evaluation metrics.

The case studies used in the course are based on histori-
cally famous racing engines with significant engineering in-
novations of the period. These engines provide very success-
ful examples of excellent engineering design. The posed prob-
lems generally require analyzing engine design parameters
and their impact on performance. A few case studies are listed
here as examples.

> 1954 Jaguar XK 3.4 L Inline Six
1. Provide additional background information and
references) for the XK engine.
2. Estimate the indicated mean effective pressure for

Summer School

this engine.
3. Estimate the peak power.
> Honda RA122E/B 3.5-Liter V12
1. Provide additional historical and technical
information on the Honda engine.
2. Estimate cylinder wall, piston crown, and exhaust
valve temperatures as a function of engine speed
3. Estimate the cycle-averaged heat loss rate in the
given engine as a function of engine speed (rpm).
4. Discuss how the thickness of the cylinder wall
and piston crown affects the heat loss.
5. Nii. ,11- semi-quantitatively how the exhaust valve
diameter would affect valve temperature.
6. Provide an alternate way of,. ,tiit,,i the cycle-
averaged gas temperature and the exhaust gas
> Timing and Camshaft Design of 1989 Mustang GT
1. Provide nomenclature for main parts of the
2. Relate cam profile and phasing to valve actuation
and ultimately engine performance.
3. Select a commercially available cam for the 1987
Mustang and provide reasons for selection.
4. Use available software to simulate the engine

The third case is a primary example that can easily lead to
group competition to produce the best performing engine.
The group competition brings fun back to class, while at the
same time promoting students' active participation.P1] "Who
can produce an engine for the best quarter-mile time?" has
proven to be an all-time motivator.
For those instructors who wish to implement a case study
strictly limited to chemical engineering principles, the key
points of the second case study are discussed in detail for
illustration. The Honda F-1 engine case study focuses on heat
transfer from the engine cylinder to the surrounding environ-
ment when chemical energy is released from the combustion
process. The students are encouraged first to gather relevant
information on gasoline, such as average molecular weight,
specific gravity, heat capacity, and heat of combustion. For
instance, one can assume complete combustion of octane to
estimate the upper limit on the released chemical energy

C8H18 + -2 A 8CO2 + 9H20 (1)
For the adiabatic complete combustion, the students can cal-
culate the exhaust gas temperature within the cylinder. The
adiabatic gas temperature translates to the first estimate on
the actual exhaust gas temperature since in the real situation,
the released heat flows from the exhaust gas to valve assem-
bly, cylinder wall, and piston head. The instructors must si-
multaneously guide the students to recognize that the amount

Figure 3. Screen Shot of WAVE simulation software. The image shows modular compo-
nents of a four-cylinder engine with complete intake and exhaust systems.

Chemical Engineerrng Education

VIN-1 I112qll_11



c 2.............:: i.... S. Et
All 2H_=- 7. ,ec.
Id 0:43 o~i= cc 2sr\e~rX~Ttt- l- d-t

Poster Session Award Paper

The course appeals largely to students who view cars as more than just transportation and
who regard an engine as the soul of an automobile. Since the course is geared toward
car enthusiasts, these students are highly motivated to learn the subject matter.

of released chemical energy per unit time increases as the
engine speed increases and that the heat transfer coefficient
(he) for the turbulent flow of gasoline and combustion prod-
ucts in the cylinder also increases with increasing engine
speed. That is, the increasing engine speed equals increasing
Reynolds number (Re) for the working fluid in the cylinder,
which in turn increases h. according to
0 75
hb= 10.4 Gb(2)
kg btg)

where b, kg and jig denote cylinder bore, thermal conductiv-
ity of working fluid, and viscosity of working fluid.?1] G is
defined as rha / Ap, where Ia and A denote mass flux per
unit time through the cylinder and piston area. Gb/|Pg essen-
tially equals Re. The heat transferred from the working fluid
to the surrounding mechanical parts by forced convection is
subsequently conducted through mostly metal engine com-
ponents. In the case of the cylinder wall, the heat is finally
transferred to the cooling jacket, whose temperature can be
fixed for the purpose of calculations at a cycle-averaged tem-
perature. By equating the heat flux from the convective heat
transfer to the heat loss by conduction to the surrounding
engine components, one can determine the surface tempera-
ture of valve assembly, cylinder wall, and piston head.
The remaining task is to account for the heat loss to the
surroundings. The initial approximation on the exhaust gas
temperature by adiabatic combustion no longer holds true.
Thus, the students need to solve reiteratively for the exhaust
gas temperature, which in turn affects both convective and
conductive heat transfer. The students can consider additional
complexity of heat conduction based on the varying cross-
sectional areas of engine components available for the con-
ductive heat transfer.

We introduce a number of software tools in the course.
Although the effectiveness of introducing more than one soft-
ware tool in a semester is still in debate, students often find
their favorite from a selection of available tools, depending
on their level of willingness to learn increasingly more com-
plex software. For instance, ESP allows students to quickly
grasp the key design parameters to improve engine perfor-
mance with much ease and simplicity. For this reason, ESP is

introduced in the early stages of the course.
The next software that we have considered introducing is
Dyno/Drag 2000 by Motion Software, Inc., which contains
an extensive list of commercially available engines and their
components to simulate engine performance. The level of
complexity comparatively increases from ESP to Dyno/Drag
2000. Dyno/Drag 2000 tests students' ability to choose the
"right" components based on sound engineering decisions.
In comparison, WAVE (manufactured by Ricardo Software)
is the most comprehensive and flexible software for simulat-
ing engine performance by simultaneously solving momen-
tum, energy, and mass balance equations with a Wiebe-based
combustion model. The software requires approximately a
month of training to bring students to the point of simulating
their own engines. Figure 3 shows a screen shot of WAVE
simulation that students initially work on as a part of the train-
ing. The image illustrates four cylinders placed in the center
with a full intake system on the left and a full exhaust system
on the right. The user can specify details such as choice of
fuel, exact geometry and operating temperature/pressure of
all components, intake/exhaust valve lift as a function of crank
shaft angle, rate and amount of heat release during combus-
tion, and even complex muffler arrangements. Because some
auto manufacturers use WAVE to design and optimize pro-
duction engines, this training improves students' potential to
pursue a career in auto design and manufacturing.
The last tool introduced is Working Model 2D by MSC
Software. This tool allows students to construct moving com-
ponents of an engine such as the cam and analyze their dy-
namic mechanical behavior. The exposure to mechanics helps
students see that constructing a working engine requires more
than understanding its thermodynamics, transport phenom-
ena, and combustion.
The software tools additionally benefit the students who
belong to the FSAE team. Using WAVE, these team mem-
bers optimize their racecar design. The design rules of FSAE
competitionN61 are comprehensive, including chassis rules,
crash protection, safety rules, and power-train restrictions.
Among these rules, the design constraint of a 20.0-mm-di-
ameter intake flow restrictor is the most limiting factor for
engine breathing and hence, for its performance. The team
members have been using WAVE to optimize the intake mani-
fold to overcome this constraint and to maximize the engine
Continued on page 219.

Summer 2003

Summer School

Increasing Time Spent on

Course Objectives by




University of Kentucky Paducah, KY 42001

he chemical engineering curriculum is a crowded
agenda, with students needing to allocate limited time
resources to focus on educational objectives. A typi-
cal curriculum includes instruction in numerical methods, and
programming assignments are one way of ensuring that stu-
dents understand the methods underlying modeling calcula-
tions. Developing applications from "scratch" requires too
much time spent on tasks not central to course objectives,
however, leaving too little time for the engineering problem.
An innovative approach to this problem, "Template-Based
Programming," minimizes time spent on the elements of pro-
gramming outside of course and assignment objectives.[2]
Template-based programming (not to be confused with the
software engineering term "template") provides the student
with a fully functional application, in all respects but one-
the subroutine containing models and numerical methods is
deliberately empty, with variables necessary to communicate
with the user exposed and well defined for the student. The
student is required to write only the codes necessary to imple-
ment the model and appropriate numerical methods. This
enables students to focus on the assignment objectives with-
out unnecessarily concentrating on the use of syntax and struc-
tures not germane to the engineering assignment.

Newly graduated chemical engineers will potentially be
called upon to perform a wide range of tasks involving com-
puters. They will need to use current technologies requiring
little programming skill, but they may also be called on to
develop models and simulations that require some ability to

Template-based programming
provides a means for incorporating
computer programming into courses while
minimizing time the students spend on
programming tasks unrelated
to the course objectives.

develop procedural computer code. At times this will involve
legacy codes, often in FORTRAN, dating back several de-
cades. Sometimes this will require working in modem envi-
ronments, such as writing complex procedural scripts in
MATLAB, or integrating spreadsheet calculations with pro-
cedures written in Visual Basic.
The limited space available in the curriculum for computer
training must be used to provide as broad a base as possible
to enable graduates to adapt to the specific computing-re-
lated requirements of their employer as rapidly as possible.
At the same time, it must be recognized that most chemical
engineers are not expected to be applications developers,

David L. Silverstein is currently Assistant Pro-
fessor of Chemical Engineering at the University
of Kentucky, Paducah. His special interests in-
clude learning-style centric educational software
development, pedagogy-driven design of dis-
tance-learning classrooms, improving retention
of mathematics for engineering courses, and co-
operative experiences in process control between
engineering and technology students. He holds
a PhD and MS in Chemical Engineering from
Vanderbilt University and a BSChE from the Uni-
versity of Alabama.

0 Copyright ChE Division ofASEE 2003

Chemical Engineerrng Education

writing complete user-friendly programs from "scratch."
To maximize the value of each course in the curriculum,
the integration of programming must be directed toward meet-
ing educational and program objectives as defined by the in-
dividual department or program. Accreditation Board for
Engineering and Technology (ABET) accredited curricula
require that students learn "appropriate modern...computing
techniques."121 In most cases, chemical engineering programs
have addressed that criterion in part by incorporating a pro-
gramming course.Y3] The most common language used is
FORTRAN, with C, C++, and Visual Basic taught in many
other programs. 4]1 Teaching procedural elements of script-
ing languages, such as those in MATLAB, Maple, and
Mathematica, is proving to be an increasingly popular option
among chemical engineering programs. [4
The decision on how to incorporate programming into the
curriculum should take into consideration why programming
is important to chemical engineers. A CACHE Corporation
suIn I 0-1 provides some interesting insight. Of the practicing
chemical engineers surveyed, 92% never use FORTRAN or
another computer language in their work. Furthermore, 86%
of employers did not expect literacy in different computer
language paradigms. Faculty members revising chemical en-
gineering curricula to meet the needs of industry will find it
challenging to determine the computing skills that should be
taught as part of a core curriculum.[6] When the requirements
of the ABET Engi-
neering Criteria
2000 (EC2000) are ,
also considered, the
need to specifically Solving a non-linear equation of state using Newtol
address program- Pressure [atm (abs)] I
ming and computer Specific Volume [L/mol]
use in the curricu- Temperature [K] I
lum is even more
pronounced. When you press the Solve button, the program
should use the SRK Equation of State to solve
Since there ap- for whichever of the above values is equal to
p- zero. If more than one value is equal to zero, an
pears to be little error message will appear.
need for practicing
engineers to pro- Solve
gram w e m ust ask ... ............. ............... ........ status
why we teach pro-
gramming in most
chemical engineer- E-it
ing curricula. The
most likely answer
seems to be, be- Figure1. The user interface resultinA
cause developing a plate. The program is fully functions
program requires thing when the button is clicked. E
that the program au- still displayed, even without the stu

dl, e

thor break down a complex problem into a logical series of
steps in a syntactically rigorous language, with a flow of logic
that should reflect the mental discipline an engineer should
be capable of to solve challenging problems. Programming
concepts are expected to strengthen two key facets of engi-
neering education-problem formulation and problem solv-
ing,[7' although some studies indicate this may not be the
case.[8' It can also be argued that programming languages are
"a novel formal medium for expressing ideas about method-
ology,"[91 and thereby strengthen the communications com-
ponent of the curriculum.
General purpose mathematical software (such as MATLAB,
Maple, Mathematica, or MathCad) has been adopted by many
chemical engineering programs with the intent of teaching
students problem-solving skills requiring computer usage.
Typically, this requires some degree of programming. The
benefit of this approach is that the "overhead" of applica-
tions development is handled by the host application. One
difficulty with this approach is teaching numerical methods
with these packages when the program already has most nu-
merical methods incorporated directly into the software. In
this case, use of a high-level programming language may be
more appropriate. In all cases, the opportunity for sufficient
instruction of students in the software or programming lan-
guage they are expected to use is critical.
Applications development requires that all details regard-
ing a program be ad-
dressed, including
_memory manage-
ment, user interfaces,
thod. buffer protection,
-System-Specific PhysicalConstants graphics, file man-
Critical Temperature [K] agement, and operat-
1 ing system integra-
Critical Pressure [tm] tion. Engineering stu-
I dents should be fo-
Accentric Factor caused on the prob-
lem, addressing nu-
merical methods,
model accuracy and
assumptions, stabil-
ity, limits of applica-
bility, communicat-
ing the model and
results, and integrat-
ing the results with
m compiling and executing the tern- reality. The method
except that it does not calculate any- described herein en-
messages regarding user input are ables focus on the
t code. topics important to

Summer 2003

n's me


Poster Session Award Paper )

Summer School

chemical engineering students.


The template-based approach begins with a fully functional
application. The student is provided with the complete
workspace for the integrated development environment (IDE)
under which they have previously worked in their program-
ming course. Immediately after loading the template, they
can compile and run the program, which will appear as in
Figure 1. One key element is, however, missing. When the
sol C'" button is pressed, nothing is returned. That button
calls a subroutine that, while fully commented and linked to
the application, is devoid of executable code. The student is
expected to write all necessary code to provide the subrou-
tine with the functionality required by the assignment. Atypi-
cal subroutine is shown in Figure 2. All variables needed as
input for the routine are exposed by global variables, as is the
variable declared to store the result calculated by the students,
which is returned to the application interface.
The result of the assignment is an attractive, uniform, fully
functional, stand-alone application that students can then use
for solving additional problems. Time spent on non-objec-
tive activities, such as preparing "FORMAT" statements or
reading and writing to files has been minimized or eliminated.
The high-level programming language has been used in such
a way that it shares many of the benefits of math packages,
including the ability to assume that I/O management, memory
management, and the user interface are taken care of without
need for consideration by the subroutine programmer.


Use of these templates develops along the same lines as
other course material. First, expected outcomes for the as-
signment were derived from the course objectives. An as-
signment was developed and application constructed sans
components required for the student to develop to achieve
the expected outcomes.
Students were required to take a FORTRAN course to
fulfill programming requirements until a policy change at
the University of Kentucky (UK) in 2001 allowed them to
take a course in Visual Basic to satisfy that requirement.
To maximize the value of the student's previous training,
early uses of this approach were coded in Compaq's Vi-
sual FORTRAN environment. Most recent uses of this ap-
proach have been assigned with a student option of using
FORTRAN or Microsoft Visual Basic.NET templates. In
all cases, students have been given the option of writing
their own program from scratch, using any programming
environment. The assignment is typically rather detailed,
since some significant instruction is required to get started

using the template because students have never had to work
within someone else's application prior to this assignment.


Students did not write routines to collect model parameters
from a console or data file, design structures to pass data
between routines, or format numerical output and send it to
the console or to a file. They gained no experience in inherit-
ance, encapsulation, polymorphism, or GUI programming.
They did design a model, specify required input and output,
write subroutines and functions involving the model and a
numerical method, debug their code, validate the model, and
verify the numerical method. They focused, thereby, on the
course objectives-modeling chemical processes or perform-
ing engineering calculations, selecting and implementing
numerical methods, and practicing fundamental logic and
problem-solving skills.
An additional benefit of the approach includes ease of grad-
ing. Since all students started from a common set of vari-
ables, the code was easier to trace. The results presented
in the form of screen captures were consistent and easier
to follow.
Template-based programming has been applied in two
courses so far: Process Principles, the sophomore material
and energy balance course, and Process Modeling, a junior-
level course in modeling principles and numerical methods.
The class size in all courses in which this approach has

.Routine to actually run the SRK EOS calculations
It obtains values of 3 of 4 variables (PVT) and Tc,Pc,and the accentric factor
From the dial og box
It must provide a value of the missing variable while not changing the variables
! minimal error checking is performed
1 The following two statements incorporate this file into the overall project
use dfwin
use SRKGlobals
!The following global variables are available to you.
!double precision Pressure Pressure given in the dialog box
double precision Volume Specific Volume given in the dialog box
!double precision Temperature Temperature given in the dialog box
double precision TCrit Critical Temperature given in the dialog box
'double precision PCrit Critical Pressure given in the dialog box
double precision Omega Accentric Factor given in the dialog box (may be -0)
!You must assign a value to whichever variable from amongst Pressure, Volume,
Temperature was =0
!You will need to create some variables for use within your subroutine
!You may choose to create another subroutine INSIDE THIS FILE to handle equation
You will need to first figure out what to solve for,and then branch to actually
!solve for that unknown
!That gives you three cases,one of which is trivial. For the other two, you will
need to use Newton's method
!Remember to calculate the SRK parameters which you will use in all three cases.
!You will likely need to declare other variables to complete your program


Figure 2. Typical student subroutine template.

Chemical Engineerrng Education

been applied has been ten or less. The UK engineering pro-
grams in Paducah are an extension of the main campus in
Lexington. Students pursue a BS in chemical engineering
under what is currently a curriculum identical to that of the
Lexington program. Chemical engineering courses at UK
Paducah are taught by UK faculty stationed in Paducah, with
the remaining courses taught by Murray State University and
Paducah Community College.1101 ABET accreditation is ex-
pected in July 2003.
The first use of template-based programming was in Pro-
cess Modeling. One of the objectives for this course is that
students "use computer software and programming languages
to solve complex mathematical systems." Two assignments
were given. The first was a simple determination of the ma-
chine epsilon, the smallest difference between two numbers
that the computer can distinguish. The purposes of the as-
signment were to reinforce the idea of inherent limitations of
computer calculations, to reintroduce computer programming
to the students (who had not programmed during the 1 2
years since their programming course), and to familiarize the
students with the template-based approach. The second project
was an implementation of Gauss-Siedel elimination to solve
for a solution to a system of linear equations.
The first time this course was taught at the Paducah cam-
pus, students were unable to complete the epsilon calcula-
tion due to difficulties with passing variables to subroutines,
with using DO loops, with reading and writing input and out-

Poster Session Award Paper )

put, and other serious programming deficiencies. The options
left to the instructor were to reteach these topics every time
the modeling course was offered, to eliminate the program-
ming task, or to develop a method that would allow students
to focus on the skills they needed to retain.
The second time the course was taught, templates were used.
Students were surveyed regarding their perception of their
programming skills and the utility of the template-based ap-
proach.m1 Table 1 summarizes the results. Students demon-
strated improvement in perception of their individual skills
in various programming concepts after the template-based
assignment in all cases except for the question about basic
mathematical operations and on the question regarding DO
loops. The second project required a sequence of nested DO
loops that tested their understanding of this construct, which
was not as strong as they had previously perceived. Other
results of the survey indicated more time was spent on
engineering objectives than would have been spent with-
out the template.
The students were enthusiastic about this approach, stating
that they were more confident about starting to program since
the "busy work" had been completed for them. The visual
nature of the output was more satisfying than the console-
based output of programs they had written previously. Stu-
dents from the previous section of the course expressed envy
that students using templates did not have to struggle through
the issues they did with fundamentals of applications devel-

I am comfortable writing programs using FORTRAN
I understand how to use variable arrays in FORTRAN
I understand how to use if-then-else structures in programs
I understand the use of DO loops and how they are terminated
I understand the use of comparison operators in FORTRAN
I can perform mathematical operations on variables in a
FORTRAN program
Using a template allows me to focus more on the engineering
problem compared to writing a program from a blank file
The time spent programming was reduced by using the template





N/A 4.5 NA 3.8


After having worked through the issues involved in starting to use N/A
the template, the template was easier to use than writing a
program from a blank file

4.3 NA 4.3
4.1 N/A 4.5

The biggest improvements ob-
served by the instructor were the
reduction in time spent during of-
fice hours instructing students in
programming fundamentals and
the fact that all students turned in
programs that met instructional
objectives. Students spent far less
time working on the programs us-
ing the templates than before, but
the time they did spend was qual-
ity time, focused on the assign-
ment objectives.
The second course in which this
technique was applied was Pro-
cess Principles in the fall of 2001
and 2002. During coverage of
equations of state, students were
required to solve the Soave-
Redlich-Kwong equation of state
for an unknown PVT variable us-
ing Newton's method. Again, stu-

Summer 2003

Summary of Survey Results
("1" student strongly disagrees: "5" strongly agrees)
(Courses: CAE200-Process Principles; CAE420-Process Modeling)

Summer School

dents had issues with the numerical method, but not with the
programming technique. In part, this was due to the fresh-
ness of the programming course in these sophomore's ex-
perience, but students still expressed a preference for
working from a template instead of writing a program from
a blank file. Submitted results were of superior quality
compared to those of an assignment in the fall of 1999
that required complete development of an application to
solve a similar problem.

Surveys were again conducted involving students in both
sections of Process Principles in which templates were imple-
mented. Table 1 summarizes those survey results. The key
finding of the survey is confirmation that students again per-
ceived that they spent more time focused on meeting course
expected outcomes while practicing and improving their pro-
gramming skills.
The template-based approach is flexible, readily applied
not only to the Compaq Visual FORTRAN IDE and the
Microsoft Visual Basic.NET environment, but applicable
within MATLAB and other packages. In the fall of 2002, the
Process Principles class had a group of students, half of whom
had instruction in FORTRAN, and the other half a course in
Visual Basic. Templates forboth languages were made avail-
able and were used by students. The complete templates, in-
cluding the code for the user interface, also serve as an ex-
ample for students interested in developing their own appli-
cations for research or personal use.

There are some difficulties associated with the use of tem-
plates. The primary issue in implementing them is actually
developing a reliable, debugged interface. With Compaq Vi-
sual FORTRAN, the interface development requires under-
standing of Microsoft Windows GUI development concepts,
which is not part of the typical chemical engineering
professor's background. More recent FORTRAN environ-
ments (Leahy Fortran.NET, for example) make this process
much simpler. When students work under different languages
or environments, multiple templates may be required, as they
were during the fall 2002 offering.
The implementation of templates will benefit from some
technical improvements. The subroutine should be encapsu-
lated as an object, so that the student does not directly modify
any variables from the host routines. Another alternative
would be having the students develop the core of a dynamic
link library that compiles independently from the application
and may be called from a variety of other programs. Pro-
cess simulators and data acquisition systems can also be
integrated with the templates to further develop the
student's computing skills.
To simplify adoption of this approach, the author has made

templates freely available for download.J111 Other faculty who
wish to share their templates with the engineering commu-
nity at large are invited to submit their templates to the au-
thor for inclusion in a web-accessible library.

Template-based programming provides a means for incor-
porating computer programming into courses while minimiz-
ing time the students spend on programming tasks unrelated
to the course objectives. It allows the student to model pro-
cesses and apply numerical methods in a logical manner,
breaking down complex procedures into syntactically rig-
orous steps, resulting in a usable application. It also capi-
talizes on the programming courses students take early in
the curriculum.
Templates have been successfully incorporated into two
courses at the University of Kentucky's extended campus in
Paducah, with very positive results, particularly with respect
to time spent on engineering tasks as described in the course
objectives. It is a flexible approach, as demonstrated by use
with both FORTRAN and Visual Basic, and can be applied
to many other languages or computing environments. Stable
and robust template development does require substantial
time, but results in a more efficient educational experience
for the student.

1. Silverstein, D.L., "Template-Based Programming in Chemical Engi-
neering Courses," Proceedings of the 2001 ASEE Annual Conference
& Exposition, American Society for Engineering Education (2001)
2. Criteria for Accrediting Engineering Programs, Accreditation Board
for Engineering and Technology, Inc., Baltimore, MD: www.abet. org/> (2002)
3. CACHE Survey Re-
4. Dahm, K.D., R.P. Hesketh, and M.J. Savelski, "Is Process Simulation
Used Effectively in ChE Courses?" Chem. Eng. Ed., 36(3), 192 (2002)
5. Davis, J., G. Blau, and G.X. Reklaitis, "Computers in Undergraduate
Chemical Engineering Education: A Perspective on Training and Ap-
plications," Technical report, CACHE Corporation, Draft 3.1 (1993)
6. Kantor, T.J., and T.F. Edgar, "Computing Skills in the Chemical Engi-
neering Curriculum," in Computers in Chemical Engineering Educa-
tion, B. Carnahan, ed., CACHE Corp, Austin, TX (1996)
7. Stephanopoulos, G., and C. Han, "Languages and Programming Para-
digms," in Computers in Chemical Engineering Education, B.
Carnahan, ed., CACHE Corp, Austin, TX (1996)
8. Urban-Lurain, M., and D.J. Weinshank, "Do Non-Computer Science
Students Need to Program?", J. Eng. Ed., 88, 535 (2001)
9. Abelson, H., and G. Sussman, Structure and Interpretation of Com-
puterPrograms, MIT Press, Cambridge, MA(1985)
10. Smart, J.L., W. Murphy, G.T. Lineberry, and B. Lykins, "Develop-
ment of an Extended Campus Chemical Engineering Program, Pro-
ceedings of the 2000 ASEE Annual Conference & Exposition, Ameri-
can Society for Engineering Education (2000)
11. 5

Chemical Engineerring Education

Poster Session Award Paper

High Performance Engines
Continued from page 213.

performance of Yamaha R6.
In addition to the software tools, we use WebCT, a web-
based teaching tool. It enables prompt distribution of course
materials, fast real-time communication with students via chat
rooms, immediate feedback on student evaluation, and com-
pilation of research topics and information relevant to the
discussion topics. Although creating and improving a WebCT
site requires more hours of preparation than conventional lec-
tures, it proves an effective medium to keep the course more
accessible and malleable for improvement.

We expect the students to be able to perform the following
tasks during the course of the semester:
Construct ideal Otto as well as nonideal P-V diagram
describing each stroke of the internal combustion
Identify key elements that contribute to the nonideality
and how the nonideality impacts engine efficiency
Recognize the importance of engine hi and
propose the methods of improving engine ...ii,,hi
based on full understanding of fluid mechanics
Compare supercharging and turbocharging to natural
aspiration and draw the piston trajectory on the P-V
Perform convective and conductive heat transfer
calculations to approximate the temperature of
mechanical components surrounding engine cylinders
Design and virtually build a mechanically stable
camshaft valve assembly, using Working Model 2D
Design a working 4- to 8-cylinder engine, using
WAVE, and baseline the performance compared to
commercially available production engines
The majority of students are capable of performing the
above tasks, evidence of which is based on their classroom
presentation and full discussion of case studies, including the
mathematical calculations. The narrow final grade distribu-
tion, ranging from 80% to 96% in cumulative scores based
on the evaluation criteria described previously, support that
the student groups collectively meet the expected course out-
comes better than the majority of engineering classes.
The latest student course evaluation (Fall 2002) rates the
course at 6 on a scale of 1 through 6, where 6 is the most
favorable rating. The main categories of evaluation are course
content, instructors, and effectiveness in learning. Overall,

students enjoy the introduction to industry-developed soft-
ware such as WAVE, the seminar format of the course that
gets everyone involved, and the assignments that lead to good
discussions. Students request lessening the gear head talk,
however, and increasing the availability of instructors to an-
swer the questions in regards to the case studies.

The format of the HPE course is a problem-based learning
environment centered on case studies of specific engines,
many with great historical significance. Students and faculty
jointly explore the fundamental chemical and mechanical en-
gineering aspects that govern the operation of the internal
combustion engines, leading to an understanding of how en-
gines work, how to model them, and how to design them for
maximum performance. While engineering rigor is not sacri-
ficed, the course is about applying good engineering and its
tools to high-performance engines, and in the process, mak-
ing engineering fundamentals come to life.
This course is not primarily about plugging numbers into
formulas, although students do some extensive calculations
at times, but rather about understanding, through implemen-
tation of computing and information technology for model-
ing, simulation, visualization, and design. The set of resources
compiled over the semester, with contributions from all faculty
and students, represents a rich legacy that will be useful to those
who have taken the course as well as students who follow.

1. Sansalone, M., "Teaching Structural Concepts through Case Studies
and Competitions," Eng. Ed., 80, 472 (1990)
2. Dolmans, D.H.F.M., I.H.A.P. Wolfhagen, C.P.M.v.d.Vleuten, and
W.H.F.W. Wijnen, "Solving Problems with Group Work in Problem-
Based Learning: Hold on to the Philosophy,"Med. Ed., 35, 884 (2001)
3. Yurcik, W., index.html> (2001)
4. Fishwick, PA., "Modeling the World: How It's Blocked Out on the
Web," IEEE Potentials, 19, 6 (2000)
5. Budhu, M., "Enhancing Instructions Using Interactive Multimedia
Simulations," Simulation, 76, 222 (2001)
6. SAE International Soci-
ety of Automotive Engineers, Inc. (2003)
7. Wankat, P.C., The. ... ... ** .-Professor: Teaching, Scholar-
ship, and Service, Allyn & Bacon, Boston, MA (2002)
8. Lumely, J.L., Engines: An Introduction, Cambridge University Press,
Cambridge (1999)
9. Heywood, J.B., Internal Combustion EngineFundamentals, McGraw
Hill, Inc., New York, NY (1988)
10. Felder, R.M., and R. Brent, "Effective Teaching: A Workshop," Peer
Rating Factor from teamwork/peerRating.pdf>, Bucknell University (2001) O

Summer 2003

Random Thoughts...




North Carolina State University Raleigh, NC 27695

I've spent decades at one academic institution and regu-
larly visit others, but I still sometimes have trouble un-
derstanding academic institutional logic. Whenever my
bewilderment reaches a critical point, I like to check in with my
old grad school buddy Kreplach, who always seems to be on
top of everything. I ran into him at a conference the other day.
"So, Kreplach-I hear your Engineering School has raised
the first 10 million dollars in its latest fund drive. "
i,htli i, r 'Forward to the Future' campaign is off and run-
ning. We 're going for 150 million big ones. "
"But last month you told me that you justfinished raising 50
million in your 'Tackling Tomorrow's Technology Today'
drive and you haven't spent all dut m. i,,i.'yyet."
i',tlit what s your problem?"

"Why do you need another 150 million? "
"So we can build new state-of-the-art bionanoinfotechnology
facilities. "
"I;h. ',.i, t you already had them-wasn 't your biotech cen-
ter built just last year? "
"Well sure, but what we've got now does 't have nearly
enough room to put in the labs and offices that the new
graduate students are going to need. "
"What new graduate students? "
"The ones we're going to go after with those 60K fellowship
programs that will come out of the 150 million. "
"Sixty thousand? But that s more than we pay our new assis-

tant professors."
"Yep, same with ours, and with that kind of money we should
be able to pull a bunch of hotshot undergrads away from
Stanford and MIT "
"That would be nice, but don't you and your colleagues al-
ready have as many graduate students as you can handle
can you just keep on adding new ones indefinitely?"
"Probably not, but that s where the new faculty members come
in. "
"What new faculty members?"
"The ones in the new d, ,iiii,, h,..1 chairs, of course."
"You mean the ones you endowed from the 50 million?"
', the eight new ones that will come from the 150 million.
We're going to be able to pay those puppies a lot more
than the chancellor gets and almost as much as the foot-
ball coach-the Stanford and MIT faculties will befalling
all over themselves to get some of that action."
"That would make sense, since you'll be -,. m,,i. their

Copyright ChE Division of ASEE 2003

Chemical Engineerrng Education

Richard M. Felder is Hoechst Celanese Pro-
fessor Emeritus of Chemical Engineering at
North Carolina State University. He received
his B ChE 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
National Effective Teaching Institute

students...but look, Kreplach why do you need so many
new professors? Most of you are already teaching only
two courses a year "

i,'oli. and with those new professors on board, we can get
it down to one. Pretty sweet, eh? Besides, who'll advise
the new graduate students if we don't bring in more fac-

"I forgot about that. "

"See, that your problem -you keep missing the big picture.
Now, what do you think will happen when we've got all
those new faculty members and graduate students on
board? "

"It will get even harder for you to find a parking place on
campus? "

1 vyell, yes, but never mind that. Think about all the grants
we'll pull in and the papers we'll churn out and the re-
search dollars we'll spend we'll be right up there with
Stanford and MIT!"

"You should be ahead of them since you'll have most of their
faculty and graduate students...anyway, I'm still not sure
I see the point. "

"It simple. Since U.S. News & World Report bases its rat-
ings on grants and papers and research expenditures, we'llI
move right up into the top ten..."

"And then... "

"And then our graduate school applications from other top
ten schools will go up and we can boost our productivity
and cut down our teaching loads even more, and we'llI
become attractive to faculty at..."

"Cal Tech and Minnesota because there 's no one left at
Stanford and MIT? "

i,'.I I/Is that a plan, or i /ha "

"Let me see ifI have it straight. You're trying to squeeze 150
million dollars out of companies and alumni so you can
build new facilities to house new faculty members and
graduate students who will increase your research fund-
ing and output, enabling you to recruit even more new
faculty and graduate students?"

"I could 't have put it better myself"

"But where will you put those people? "

"Aha glad you asked! The Dean has put me in charge of
the 'Engineering the Dream of a Brighter Tomorrow at
the Dawn of the New Global Millennium 'campaign we 're
i,. ui, ready to announce."

"Catchy title."

"Yeah, the Dean ;hi-.oiht of it himself We're ;hiiing that
400 million dollars should cover most of our needs for at
least two years."

"I like how you ;h1, i1 -. .. Kreplach, but where are you going
to find any more companies and alumni willing to kick in
after all those other campaigns? "

"We ;hil, ,;it about that, and we came up with two untapped
sources with a strong vested interest in making us numero
uno. Here, check this out. "

"Let see whoa, a pledge form for the faculty with check-
off boxes for salary tithes and a free legal service to help
them redraw their wills. Outstanding!"

"Yeah, and don't overlook the space for the number of boxes
of cookies they commit to sell every year. "

"Brilliant ;hliili-, Kreplach I know everyone on your fac-
ulty will jump at the chance to help out...but what your
other group of eager donors? I can't believe there s a tur-
nip you haven't already tried to squeeze blood from. "

"Lots of them, my boy. Think who is the most important
person on campus, the one the faculty and administration
are primarily there to serve, the one who gets more ben-
efits from the university than anyone else?"

"That would be the football coach. "

"Don't be absurd it the undergraduate student. The value
of a college education to every undergraduate over the
course of a lifetime is over a million dollars, i 'i

"Yes, but..."

"So if we just ask each of them to kick in a mere $100,000,
which is only a tenth of what we're giving them, we're
home free."

"Fiendishly clever Kreplach-our fundraising people would
be green with envy. "

"And who could blame them? Oh, before you go, our choco-
late macadamia shortbread wafers are particularly fine
this year. How many boxes can I put you down for? "

Summer 2003

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




Part 1. Introduction and Application to Explicit Models*

University of Guelph Guelph, Ontario, Canada NIG 2W1

Engineering analysis and design typically proceed by
construction of system models, followed by incorpo-
ration of required input quantities and the solution of
the model equations to obtain values of output quantities of
interest. The construction of models is an important aspect of
engineering pedagogy, as is familiarity with numerical meth-
ods of solution for their more complex forms.[1] In addition
to teaching students to construct and solve models for their
outputs, we believe it is also important to emphasize consid-
eration of the effects on outputs when input quantities are
subject to variability and/or uncertainty. This is accomplished
by sensitivity analysis (SA), which canbe defined as the study
of the uses of, and methodologies for, quantitatively calcu-
lating the effects of changes in input quantities on model out-
puts, especially for a large number of inputs. (Of course, one
can directly calculate values of output quantities as functions
of input quantities, but this only reveals sensitivities qualita-
tively and is only satisfactory when the number of inputs is
small, say, < 3). In addition, abroad pedagogical use of SAis to
provide a unifying theme for the study and understanding of
topics arising in a number of areas in chemical engineering.
The fundamental concepts of SA are accessible to students
who have a basic background in multivariable calculus, but
full development requires some knowledge of linear algebra,
which may not be a requirement in all undergraduate pro-
grams. Thus, although some concepts can be introduced in
the undergraduate curriculum, the full treatment is more ap-
propriate at the graduate level.
In chemical engineering, SA has a tradition of importance
in the field of optimization.[2,3] Our own experience with SA
has been principally in equilibrium models for chemically
reacting systems,[4-6] which can be considered as optimiza-
tion problems. In recent chemical engineering literature, SA
Part 2 will be published in the Fall issue ofCEE.
** Present address: University of Ontario Institute oJ .. Oshawa,
Ontario, Canada L1H 7K4
*** University of Toronto, Toronto, Ontario, Canada M5S 3E5

has been applied to the design of reactive distillation units[7'
and to process simulator models.[8] Varma, et al., have con-
sidered SA in the context of chemical reactor models and
related situations,[9] and a recent general treatment of SA (not
specifically oriented to chemical engineering) was given by
Saltelli, et al.110o
Questions addressed by SA in a broad sense include
1. What are appropriate quantitative measures of
sensitivity of model outputs to changes in inputs?
Particular measures addressing this question involve
the marginal rates of change or marginal sensitivities
of each output with respect to each input. These
particular measures include the following items, which
form the basis of this paper:
a. A measure of the relative importance of each input
for a given output is the relative value of the
output 's marginal sensitivity with respect to each
b. A measure of the uncertainties in model outputs

William R. Smith is Professor of Engineering
and of Mathematics and Statistics at the Uni-
versity of Guelph. He received his BASc and
MASc in chemical engineering from the Univer-
sity of Toronto and his MSc and PhD degrees in
applied mathematics from the University of
Waterloo. His research is in classical and statis-
tical thermodynamics. He is coauthor of Chemi-
cal Reaction Equilibrium Analysis (1982, 1991).

Ronald W. Missen is Professor Emeritus
(chemical engineering) at the University of
Toronto. He received his BSc and MSc degrees
in chemical engineering from Queen's Univer-
sity and his PhD in physical chemistry from the
University of Cambridge. He is coauthor of
Chemical Reaction Equilibrium Analysis (1982,
1991) and Introduction to Chemical Reaction
Engineering and Kinetics (1999).

Copyright ChE Division of ASEE 2003

Chemical Engineerring Education

that result from uncertainties in the inputs. In
general, this matter is called uncertainty analysis,
and international bodies have published documents
describing standards for the expression of such
uncertainties. 1'"21 Uncertainty analysis is also
discussed in some undergraduate texts.[13-161
c. A measure of the overall effects on the outputs of
changes in combinations of inputs. This measure
shows the relative importance ofparticular
combinations of inputs.
2. What is the direction of change (sign) of outputs for a
specified direction of input changes, given only the
model structure (i.e., without obtaining a solution)?
3. Is there a large qualitative change in the outputs for
small input changes?
4. Do the outputs change ,,- ,it,...,tli li,,.i, the underly-
ing assumptions of the model change?
Question 1 is usually considered to be the primary scope of
SA, but it may include Question 4,1101 and we add Questions
2 and 3 as being in the same spirit and further broadening its
scope. Question 2 is important in requiring only a minimal
amount of model information (signs of quantities) as, for ex-
ample, in a form of Le Chatelier's Principle.[171 Question 3
arises, for example, in the context of models described by
differential equations and is addressed by stability and bifur-
cation theory.E181 A chemical engineering example concerns
finding circumstances under which a chemical reactor changes
its behavior from steady state to periodic.
Question 4 arises in statistics; the term robustness refers to
the effects on a statistical analysis of changing the assump-
tions concerning underlying probability distributions.[191 An-
other chemical engineering example concerns situations when
parameters in models are estimated from experimental data,
requiring assumptions regarding measurement errors. It is
desirable that the resulting parameter values do not depend
strongly on departures from such assumptions.
In these two papers (Parts 1 and 2), we focus on only the
first question. In this paper (Part 1), we introduce SA, de-
scribe two different classes of engineering models, and ap-
ply SA to address items la, lb, and Ic for one of these classes.
In Part 2, we will apply SA to the same items for the other
class of model. Conclusions, including what is appropriate
for the undergraduate curriculum and what can be deferred
to the graduate curriculum, will be given at the end of Part 2.
In the following sections, we first describe two types of
inputs and two types of models. For one type of model, we
then consider, in turn, item la for both types of inputs, lb
for one type of input, and Ic for the other type of input (al-
though, in principle, all three items could be considered for
both types of inputs).
The reasons for applying items lb and Ic to different input
quantities are: we are usually interested in determining the
effects on the outputs of uncertainties in the type considered

for specified values of the other inputs (item lb); conversely,
we are usually interested in determining the effects on the
outputs of changes in inputs over which the designer has con-
trol (item Ic). We will conclude with a pressure drop example
for numerical illustration.
The treatment we describe and the examples we use inboth
Parts 1 and 2 relate to models that are amenable to analytical
treatment to obtain values of the sensitivity quantities in-
volved. The methodology, however, is not limited to such
cases. In very complicated situations involving many vari-
ables, it may be possible to use computer algebra software to
obtain values, and if necessary, values can be obtained by
numerical differentiation, e.g., using a process simulator.

It is useful to consider two different types of input quanti-
ties in models: system variables (x, j = 1,2,...,J) and constitu-
tive parameters (Pk, k = 1,2,..,K). System variables (such as
temperature, T, and pressure, P) are input quantities that can
be manipulated by the designer or are imposed by the exter-
nal environment; their values are usually considered to be
specified precisely, but they may be subject to change. Con-
stitutive parameters (such as viscosity and thermal conduc-
tivity) are those that must be obtained externally to the model,
either by direct measurement or from correlations; their val-
ues are usually subject to uncertainties.

It is also useful to consider two different types of models
that arise in engineering-explicit and implicit models, terms
that denote mathematically how the outputs are available from
the model in terms of the inputs. In principle, we can also
consider further model classifications involving, for example,
continuous versus discrete variables, and deterministic ver-
sus stochastic models. Here, however, we treat only the sim-
plest situation of a deterministic model involving continuous
variables, and we further assume that any functions involved
in the models are differentiable. Although SA canbe used for
more complex situations, the methodology is correspondingly
more complex. Furthermore, continuous-variable and differ-
entiable deterministic models constitute a large class of engi-
neering models.
An explicit model with N outputs, yl, is one that can be
y,= f,(x;p); i= 1,2,...,N (1)
where x is the vector of J system variables and p is the vector
of K constitutive parameters. A simple example is a pres-
sure-explicit three-parameter equation of state (EOS), with
output quantity P, system variables molar volume (v) and T,

Summer 2003

and constitutive parameters critical temperature (To), critical
pressure (Po), and acentric factor (io)
P = P(v,T;, T,P, co) (2)
A more complex class of models is that of implicit models,
which arise when the output quantities are implicit functions
of the inputs, expressed formally as

f,(y,x; p)= 0;

each input (i.e., each system variable and each constitutive
parameter) on each output is the absolute value of the rel-
evant sensitivity coefficient, as shown below.
The effects of small changes in the inputs on the outputs
can be approximated by a Taylor expansion using the first-
order sensitivity coefficients of the outputs with respect to
the inputs

i= 1,2,...,N

where y is the vector of outputs. Implicit models can take
many forms. Their distinguishing property is that Eq. (3) can-
not be solh cd analytically" for y, (although it is usually as-
sumed that the solution of the equations is unique). A simple
example of an implicit model arises in the context of a pres-
sure-explicit EOS when v is the output quantity and (P,T) are
system variables; then v(T,P) is defined implicitly by

P -P[v(T,P),T,Tc,P,col ]= 0 (4)
More complex examples of implicit models include sets of
nonlinear algebraic or transcendental equations, systems of
differential equations (ordinary or partial), and optimization
The details of SA methodology are different for explicit
and implicit models, but the general mathematical tools are
similar. In Part 1, we will consider applications of SA to ex-
plicit models, while deferring implicit models to Part 2.

Fundamental quantities used in SA are the marginal rates
of change of output quantities in the model in terms of input
quantities, i.e., their (partial) derivatives. This addresses item
la. The first derivatives dy/dx and dyl/dPk are called first-
order sensitivity coefficients. The normalized first-order sen-
sitivity coefficients are a In yl/d In x and a In yl/ In Pk' which
have the advantage that their values are independent of the
units used. Furthermore a normalized coefficient can be in-
terpreted as the % change in the output yl for a 1% change in
the input x1 or Pk. The two types of coefficient are related by

ny X ( i=1,2...,N j= ,2,...J (5)
D1nxj y1 Dxj

Dn Pk _Y i=1,2,...,N k=l,2,...,K (6)
a1n pk Y1 a(Pk

In some situations, it is appropriate to consider second de-
rivatives as second-order sensitivity c etti ienrt'. d2yl/x2,
2y/ 1 j2, and corresponding normalized forms.

Relative Importance of Changes in Input Quantities
To address item la, a measure of the relative importance of

y, = x(yx, XJ + i Pk
j= J) k=1 k

i= 1,2,...,N (7)

In order to consider all outputs and inputs on the same (rela-
tive) basis, the logarithmic form of Eq. (7) can be used:

Y 6Iny l = t ln x1 )
y, ^l=J nx J

+_ nyl n81nPk

i = 1,2,...,N (8)

For a given value of each input change, the relative impor-
tance of each input on the magnitude of each output is the
magnitude of the relevant sensitivity coefficient.

Effects of Constitutive Parameter Changes on Outputs:
Uncertainty Analysis
Here we use the sensitivity coefficients with respect to the
constitutive parameters to address item lb, since uncertainty
analysis concerns the effects on outputs of uncertainties in
these parameters.
The (standard) uncertainty of measurement of a quantity,
wl, is an associated quantity, u(wl), that characterizes the dis-
persion of the measured values that could reasonably be as-
cribed to the quantity.J111 u(wl) is taken as the square root of
the variance, oa2, if this is available, and approximate 95%
uncertainty limits (assuming the uncertainties arise from a
normal distribution with zero mean) for w are then (w -
2u(w), w, + 2u(w)). (If Co2 is unavailable, an estimate of u(wl)
must be used. 11])
We identify the average value of 8y12 or (8 In y)2 obtained
from Eq. (7) or (8) with o2(yl) or o-2(ln y), respectively, with
analogous identifications for the covariances. In terms of
uncertainties, this yields

K ( i2 K K
u2(y j)=_ U2(p)+ 1 i'Y1 U(pjpk)
j= / 1 1 kj = P k
i = 1,2,...,N (9)

where u(ip p,) is the joint uncertainty of pJ and Pk. When the
constitutive parameters are uncorrelated (the typical case),
the last part of Eq. (9) is absent. In this case, the relative

Chemical Engineerrng Education

importance of each parameter is given by the square of the
relevant sensitivity coefficient. An expression analogous to
Eq. (9) can be written in terms of relative uncertainties. We
note that u(ln y) u(y)/yl, which represents the relative un-
certainty in y, (and correspondingly for pj). Reporting nu-
merical values of uncertainties in such a standard way, in
addition to values of the quantities themselves, is an emerg-
ing requirement.[11,12]

* ITEM 1c
Overall Effects of System Variable Changes on Outputs
To address item ic, we consider only the system variables,
which (unlike the constitutive parameters) are normally un-
der the control of the investigator. The sum of squares of the
changes in the model outputs for given changes in the system
variables is an appropriate overall measure. In what follows,
we give expressions in terms of the sensitivity coefficients
themselves; corresponding expressions can be written in terms
of the normalized sensitivity coefficients (Eqs. 5 and 6). The
change in the overall sum of squares due to small system
variable changes is given from Eq. (7) by
aSS= Sy2 =Y Pjk6XjXk (10)
1=1 j=l k=l
where Pjk are entries in a matrix P:

Pjk x J(lk j,k =1,2 ...J (11)

For a given 8x, 8S can be calculated from Eq. (10), but
further insight can be obtained by expressing 8S in a simpler
form in terms of a new set of system change variables, 80, as
follows. The right side of Eq. (10) is a quadratic sum of the
system variable changes, and it is a standard exercise in lin-
ear algebra to express this as a weighted sum of squares.[201
This is essentially the approach used by Seferlis and Grievink
in design and sensitivity analysis for reactive distillation.]71
An arbitrary change vector dx can be expressed in terms of
a set of normalized (i.e., in the mathematical sense of unit
length) linearly independent eigenvectors of P, {z, j=1,2,..,J},

8x = zjOj (12)
where 860 is the coordinate of 8x with respect to z. We then
express 8S in Eq. (10) in the simplified form

aS= j z Pz 60k6
1=1 j=1 k=1 f=1

Ii J

where X (which is non-negative[201) is an eigenvalue ofP cor-

Summer 2003

responding to z,. Equation (10) represents a J-dimensional
ellipsoid involving the variables 8x for a given value of 8S.
Equation (13) expresses this ellipsoid in "standard form" in
terms of the new variable 80. Equation (12) resolves 8x into
its coordinates in terms of the new coordinate directions z.
Equations (12) and (13) show that if 8x is proportional to
an individual eigenvector z, 8 S is given by the product of the
square of the proportionality constant 86 and the correspond-
ing eigenvalue X,. We can thus assess the relative importance
of changes of 8S from a nominal value in the direction of
each eigenvector by ordering the eigenvalues of P.

A fixed-bed catalytic chemical reactor consists of a bed of
catalyst particles through which the reacting system flows.
As part (for a simple illustration) of the overall design/analy-
sis of such a reactor, we consider an explicit model that ap-
proximates the pressure drop (-AP) of a fluid flowing through
a cylindrical bed of spherical particles:[211

(-AP) = 4 1 pD6 [1.75+150( 4 p
3 p.7fdD 4md14)

= f (1.75+ 12) (14)


64 (1--EB M2V
f1 = 4 3 pfdpD6

f2 =150(l- B)fD2
4 imdp

V bed volume
D bed volume diameter
u superficial linear fluid velocity
Pf fluid density
dp particle diameter
EB bed voidage
pif fluid viscosity
m mass flow rate through the bed.

The values at the reactor inlet are pf and Tji The bed depth, L,
is given by
L (17)
Equations (14) and (17) constitute an explicit model for
the output quantities {(-AP),L} in terms of the system vari-

ables {D,V, mi } and the constitutive parameters {EB, dp, pf,
fj}. For this example, N = 2, J = 3, and K = 4.
Expressions for the first-order sensitivity coefficients, to-
gether with the corresponding normalized coefficients are
given in Table 1; the numerical values shown are obtained
from the following data, relating to the first stage of a par-
ticular sulfur dioxide converter.E211
D =4.31 m; V= 12m3; rn = 38.0 kg s-1;
pf = 0.548 kg m-3; f = 3.8 x 10-5 kg m-1 s-1;
d = 0.015 m; EB = 0.45
From these data, f1 = 4097 kg m-1 s-2 (-Pa) and f1 = 0.08024.
With regard to item la, Column 4 of Table 1 indicates that

the vertical line at 4.31 m.
For item lb, from the values in Column 5 of Table 1, the
uncertainty in ln(-AP) in terms of the uncertainties in the con-
stitutive parameters is given by the normalized version of
Eq. (9) as

u2 [in(-AP)] = u2 (In pf) + (0.0438)2 u2 (In pf) +

(1.044)2 u2 (ln dp) + (3.854)2 u2 (n EB) (18)

where the covariances are assumed to be zero (a typical case).
This further indicates the relative importance of the variable
Eg. Equation (18) requires values of the uncertainties in the
constitutive parameters, u(pl), as described under item lb.

the most important system
variable affecting changes
(in an absolute sense) in
both (-AP) and inL is D; the
most important constitutive
parameter in this sense af-
fecting (-AP) appears to be
li; L is independent of all the
constitutive parameters.
For relative changes, Col-
umn 5 indicates that the
most important system vari-
able for both (-AP) and L is
also D; the most important
constitutive parameter for
(-AP) is E Of these two
types of comparison, that in-
volving normalized quanti-
ties is more informative,
since the coefficients are in-
dependent of the units.
To gain further insight
concerning the dependence
of the outputs on the inputs,
we calculate the sensitivity
coefficients as functions of
the input quantities. For ex-
ample, since the most im-
portant system variable is D,
we show in Figure 1 the
relative sensitivity coeffi-
cients of (-AP) as functions
of D. They are all weak
functions of D, and their
relative importance does not
change appreciably from
that at the conditions of the
problem statement, indi-
cated by intersections with

Nonzero* First-Order Sensitivity Coefficients for the Pressure Drop Example**



Coefficient (A)
(from Eqs 14-17)

f(10.5 + 4f2)

Expression for
Coefficient (B)
(from Eqs. 5-6)

2 5.25 + 2f2
S1.75 + f2

Value of(A) Value of(B)

10.29 kPa m-1 -5.912

AP (175 + f,) 1 0.625 kPam- 1
av v

a(-AP) f 3.50 + f,
( ) 1 (3.5 + f2) .50+ f2 0.386 kPa s kg-' 1.956
am mI 1.75 + f2

(-AP) (175+f) -1 13.7 kPa m kg -1
Dpf Pf
D(-AP) filf f2
-fff2 8650 kPa ms kg- 0.0438
ap, P, 1.75 + f2
a(-AP) fl (1.7 2f2) 1.75+2f2 52
d d(1.75+2f5) [ f -522kPam-1 -1.044
add d 1.75 + f

a(-AP) f x B)1.75+2f2
aB EB(1 EB) L -EB 1.75 +f2
-64.2 kPa -3.854
[1.75(3- 2B)+ (3 EB)f2]

aL 8V
D 3 -2 -0.3817 -2
aD eD3
DL 4
-- tD1 0.06854 m-2 1
V TtD-

rit Dpf Datf d p DEB
" f,=4.097 Pa; f,-0.08024; f, and f, are defined by Eqs. 14 and 15.

Chemical Engineerrng Education

For example, a 1% relative uncertainty in each parameter results in a
relative uncertainty in (-AP) of 4.1%.

For item Ic, to assess the overall effects of relative changes in the
system variables on relative changes of the outputs, the equivalent of
the matrix P defined in Eq. (11) in terms of normalized sensitivity
coefficients is calculated from the values in Column 5 of Table 1:

38.955 -7.912 -11.565'

P= -7.912 2 1.956 (19)

1-11.565 1.956 3.827 )

The eigenvalues of P, X, and their normalized eigenvectors, z1, are[22]

k1 = 43.999; z1 = (0.941, 0.190, 0.280)T

k2 = 0.783; Z2 = (0.060, 0.721, 0.691)T

k3 =0; Z3 = (0.333, 0.667, 0.667)T

The normalized equivalent of Eq. (13) gives

8 In S [8 In(-AP + (A InL)]2 =
43.999 802 +0.783 802 +0802 (21)

This indicates that, when a change vector for the system variables
(8 In D, 8 In V, 8 In mn)T is proportional to an individual eigenvector
z1 (with proportionality constant 860), the relative change in S (8 In S)
is given by \86012. Since X, = 43.999 is the largest eigenvalue, Eq.
(21) indicates that the largest relative change in S occurs in the direc-
tion of z1. For example, a unit change of the system variables (8 In D,
8 In V, 8 In mi)T in the direction of z1 leads to a 44-fold relative change

2 --------------------



... . . .

1 2 3 4 5
D = 4.31 m

6 7 8 9 10
D / m

in S. Note also that there is one zero eigenvalue; in gen-
eral, when N < J, the number of zero eigenvalues is at
least J N. For a system variable change vector
proportional to z3 (corresponding to X3 = 0), the relative
change in S is zero.

Financial assistance has been received from the Natu-
ral Sciences and Engineering Research Council of

1. Cutlip, M., and M. Shacham, Problem Solving in Chemical En-
gineering with Numerical Methods, Prentice-Hall PTR, Upper
Saddle River, NJ (1998)
2. Fiacco, A.V, Introduction to Sensitivity and .1 Analysis
in Nonlinear Programming, Academic Press, New York, NY
3. Edgar, T.F., D.M. Himmelblau, and L.S. Lasdon, Optimization
Processes, 2nd ed., McGraw-Hill, New York, NY
4. Smith, W.R., Can. J. Chem. Eng., 47, 95 (1969)
5. Smith, W.R., and R.W. Missen, Chemical Reaction Equilibrium
Analysis, Chapter 8, Wiley-Interscience, New York, NY(1982);
Krieger, Malabar, FL (1991)
6. Norval, G.W., M.J. Phillips, R.W. Missen, and W.R. Smith, Can
J.Chem.Eng.,67,652(1."". ; Catal., 54, 37(' ,".
Eng. Chem. Res., 28 1884 (1989)
7. Seferlis, P., and J. Grievink, Ind. Eng. Chem. Res., 40, 1673
8. Xin, Y., and W.B. Whiting, Ind. Eng. Chem. Res., 39, 2998
9. Varma, A., M. Morbidelli, and H. Wu, Parametric Sensitivity in
( .. Cambridge University Press (1999)
10. Saltelli, A., K. Chan, and E.M. Scott, Sensitivity Analysis, John
Wiley & Sons, New York, NY (2000)
11. European Cooperation for Accreditation ofLaboratories, Re-
port EAL-R2, 27pp. (1997)
13. Figlioa, R.R., and D.E. Beasley, Theory and Design for Me-
chanical Measurements, Chapter 5, John Wiley & Sons, New
York, NY (1991)
14. Holman, J.P., ExperimentalMethods for Engineers, 6th ed., pp.
49-56, McGraw-Hill, New York, NY (1994)
15. Taylor, J.R., An Introduction to Error Analysis, 2nd ed., Uni-
versity Science Books, Sausalito, CA (1997)
16. Coleman, H.W., and W.G. Steele, Experimentation and Uncer-
taintyAnalysisforEngineers, 2nd ed., John Wiley & Sons, New
York, NY (1999)
17. Le Chatelier, H., Compt. rend., 99, 786 ( i 4 100, 50 (1885);
196, 1557 (1933);Ann. desMines, Sen 8, 13, 157, 200, 362 (1988)
18. See, for example, Boyce, W.E., and R.C. DiPrima, Elementary
1 '--. Equations and Boundary Value Problems, 7th ed.,
Chapter 9, John Wiley & Sons, New York, NY (2001)
19. Huber, P.J., Robust Statistics, John Wiley & Sons, New York,
NY (1981)
20. Anton, H., and C. Rorres, Elementary Linear Algebra, Appli-
cations Version, 7th ed., pp. 483-485, John Wiley & Sons, New
York, NY (1994)
21. Missen, R.W., C.A. Mims, and B.A. Saville, Introduction to
ChemicalReaction Engineering andKinetics, pp. 516-519, John
Wiley & Sons, New York, NY (1999)
22. MAPLE is a registered trademark of Waterloo Maple, Inc. O

Summer 2003

Figure 1. Pressure drop example: Normalized sensitivity coeffi-
cients for (-AP) as functions of bed diameter D; values at
D = 4.31 m correspond to those in Table 1.






Kansas State University Manhattan, KS 66506

Numerous biochemical processes occurring in nature
or man-made systems (e.g., biochemical reactors)
involve randomly behaving viruses, fungi, bacteria,
and plant and animal cells that are discrete and mesoscopic
in size. This gives rise to incessant fluctuations in their char-
acteristic properties, including their number concentrations
(density), movements, and metabolic activities. Such fluc-
tuations are profoundly magnified when the number concen-
trations of pores, spores, or cells are very low as found, for
instance, in the tail end of thermal disinfection of foodstuffs.
Medical needs and public health concerns often demand
that the disinfection process be complete or nearly complete.
Hence, it is indeed appropriate that the notion and methodol-
ogy of stochastic processes have been introduced in most of
the major textbooks on biochemical engineering through
analysis and modeling of the disinfection process;1-41] the dis-
course in these textbooks is based on the original contribu-
tions of FredricksonN15 and Aiba and Toda.161 According to
Ramkrishna,171 "Fredrickson was the first to point out the
importance of stochastic analysis in dealing with steriliza-
tion processes." The crux or essence of stochastic analysis
and modeling is in their capability to estimate or predict in-
herent fluctuations of the characteristic property of a random
phenomenon or process and the distribution of this fluctuat-
ing property. In these textbooks, however, only the mean (the
first moment), or at most, the variance (the second moment
about the mean) of the fluctuating number concentration of
cells during disinfection is given. The evaluation of additional
quantities defined in terms of the moments of the number
concentration of cells higher than the second moment is use-
A Address: Kun Shan University of Technology, Yung-Kang ( .1. Tainan
Hsien, 71003 Taiwan
2 Address: University ofMississippi, Univer i 38677

ful and often necessary to gain insight into the stochastic, or
random, nature of the phenomenon or process of interest.
In fact, mesoscopic entities in the form of bubbles, drop-
lets, and particles are ubiquitous in many phenomena, pro-
cesses, and operations taught in various courses in chemical
engineering besides biochemical engineering. Some of these
courses are chemical reaction engineering, transport phenom-
ena, separation, particle technology, material science and
engineering, and surface science. The phenomena, processes,
and operations involving mesoscopic entities includes het-
erogeneous reactions,'81 gas absorption,91] distillation, liquid-
liquid extraction, adsorption,[10,111 fluidization,[12,131 filtration,[141
crystallization,[151 solids mixing,[16] and grinding and attri-
tion.[17'181 With slight adaptation, the current contribution can

L. T. Fan is University Distinguished Professor, holds the Mark H. and Mar-
garet H. Hulings Chair in Engineering, and is Director of Institute of Sys-
tems Design and Optimization at Kansas State University. He served as
Department Head of Chemical Engineering between 1968 and 1998. He
received a BS from National Taiwan University, his MS from Kansas State
University, and his PhD from West Virginia University, all in chemical engi-
neering, in addition to an MS in mathematics from West Virginia Univer-
Andres Argoti Caicedo is a graduate research assistant in the Depart-
ment of Chemical Engineering at Kansas State University. He received his
BS in Chemical Engineering from the Universidad Nacional de Colombia,
Bogota. His research interest is in the application of stochastic processes
in chemical engineering.
Song-tien Chou obtained his BS in Chemical Engineering from National
Taiwan University, MS's in Chemical Engineering and Statistics and PhD
in Statistics from Kansas State University. His research interests include
the application of stochastic processes, risk analysis, and environmental
Wei-Yin Chen received his PhD in Chemical Engineering from the City
University of New York, his MS in Chemical Engineering from the Poly-
technic Institute of New York, an MS in Applied Mathematics and Statistics
from the State State University of New York at Stony Brook, and a BS in Chemi-
cal Engineering from Tunghai University. He is presently a Professor of
Chemical Engineering at the University of Mississippi.

Copyright ChE Division ofASEE 2003

Chemical Engineerrng Education

surely be made useful for several chemical engineering
The current effort aims at presenting the fluctuations in the
microorganism's number concentration in the course of dis-
infection not only in terms of the variance but also the skew-
ness (the third moment about the mean over the third power
of the standard deviation) and the kurtosis (the fourth mo-
ment about the mean over the fourth power of the standard
deviation). The results would be useful as supplementary
materials to what is already available. Following the text-
books mentioned above, the example adopted is the thermal
disinfection of a cell population, specifically that of Staphy-
lococcus aureus,[19] where the number concentration is con-
sidered to decrease according to the first-order rate law.
Interest in stochastically analyzing and modeling bacteria,
or cell populations in general, has been steadily growing in
recent years in view of their importance in different areas of
biochemical engineering and biotechnology. Various research-
ers have resorted to the master-equation and other closely
related algorithms; the majority, if not all, of them have con-
sidered linear systems or processes. Tsuchiya, et al.,[20] re-
viewed their works on the growth and replication of cultures
of unicellular organisms. Ramkrishna[211 wrote an informa-
tive exposition containing a variety of stochastic algorithms
for modeling the dynamics of cell populations including the
master-equation algorithm. Stephanopoulos and Fred-
rickson[221 analyzed the extinction process by the prey-preda-
tor model involving both deterministic and stochastic com-
ponents. Nassar, et al., [231 stochastically modeled the dynam-
ics of a unicellular organism population; they[241 also mod-
eled the enzymatic degradation of cellulose. Lauffenburger
and Linderman[251 published a monograph based on their ear-
lier works on receptor/ligand trafficking by the master-equa-
tion algorithm.

As in any stochastic analysis and modeling, a mathemati-
cal model characterized by a random variable or variables is
required for the system under consideration. It is formulated
according to the procedure outlined bc lo '-I 'I
Description of the System The system under consider-
ation is the population of microorganisms, or cells, that are
thermally deactivated. The status, or state, of the system is

specified by the number (or size) of the population, that de-
creases due to the death of cells that have ceased to grow, one
at a time, that do not reproduce throughout the deactivation.
It is assumed that each member of the cell population alive at
t = 0 is independently subjected to the same risk of dying.
Moreover, the current size of the cell population depends
solely on the size of the immediate past population. In other
words, the system possesses the so-called Markovian prop-
erty, implying that only the current state of the process or
system is relevant in determining its future behavior;[26-281 in
fact, the system under consideration constitutes a special type
of Markov processes (time-continuous Markov chains) called
"the pure-death process."[27,281
Identification ofRandom Variable and State Space The
number of live microorganisms, simply termed cells hereaf-
ter, at time t is taken as the random variable of the process of
deactivation, N(t); a realization of N(t) is denoted by n. All
the possible numbers of live cells are the states of the process
and the collection of these numbers, {no,no-1,...,2,1,0}, is the
state space, where no is the initial number of cells susceptible
to thermal disinfection, i.e., n at t = 0.
Construction of Transition Diagram Since no cells will
be produced, the size of the population will decrease through-
out disinfection. In the transition diagram of the process pre-
sented in Figure 1, the circles indicate the possible states of
the system and the arrows describe transitions of the system
at any moment. The figure is a typical representation of the
time-homogeneous pure-death process.
Definition of Transition-Intensity Functions The rate
law adopted here for thermal deactivation (disinfection)
kn (1)
where n is the number of cells at a specific time t; this ex-
pression is known as Chick's law.[321 The intensity of transi-
tion (intensity function) is defined as the instantaneous rate
of change of the transition probability.[27-28] Hence, the inten-
sity of death, P n, for the thermal disinfection under consid-
eration can be assumed to be of the form

n = kn (2)
where k is a proportionality constant. This expression im-
plies that the cell-death rate is proportional to the number of

Figure 1. Transition diagram of
the pure-death process: pu is
the intensity of death, and
no,no-1,...,n+ ,n,n-1,...,2,1,0 are
the states of the process.

Summer 2003

0 *** n-1 n n+1 n*

1 n-1 n n1 n+2 n0

live cells. Clearly, the intensity of death, P n, is only a func-
tion of n and independent of time, and thus time homoge-
neous. [27 28]

Derivation ofthe Master Equation The governing equa-
tion of any Markov process expresses the infinitesimal change
of the transition probability, pl (s,t), between state i at arbi-
trary time s and state j at time t.[26,27] For convenience, the
governing equation is often written in terms of the state (ab-
solute)36 absolute probabilities, p (t) and pl(t), are related through the
transition probabilities, p1(s,t)'s, as

S(t)= P (S)p (s, t)

This renders it possible to transform the transition probabili-
ties in the governing equation into the absolute probabilities,
thereby yielding the gain-loss or probability-balance equa-
tion, or master equation;[28,331 for the pure-death process un-
der consideration, it is derived as follows:
With N(t) = n given, it is assumed that during time inter-
val (t, t+At): (1) the conditional probability that a death will
occur, i.e., a live cell will die, is n At + o(At), and (2) the
conditional probability that more than one death will occur is
o(At), which is defined such that

Lim = 0 (3)
At-*O At
Naturally, the conditional probability of no change in the
number of live cells during this time interval is [1 Pn At -
Let the probability that exactly n cells are alive at time t
be denoted as pn(t) = Pr[N(t) = n], n = no, no 1,...,2,1,0.
Then, for the two adjacent time intervals,
(0,t) and (t, t+At), the occurrence of exactly
n cells being alive at time (t+At) can be ;
realized in the following mutually exclu- 0.9
sive ways:
1) With a probability of pn +(t)
[tn +At+o(At)], exactly one cell will die 0.30
during the time interval (t,t+At), provided
that exactly (n+1) cells are alive at time t. 025
2) With a probability of o(At), exactly j cells v 0.20
will die during the time interval (t, t+At),
provided that exactly (n+j) cells are alive 15
at time t, where 2 < j < (no-n).
3) With a probability of p,(t)[1 gnAt -
o(At)], no cell will die during the time in- .o05
terval (t, t+At), provided that all n cells are
alive at time t.
Summing all these probabilities and con- Figure 2
solidating all quantities of o(At) yield

Pn(t + At) = Pn(t)(1- gnAt) + pn+I(t)(tn+1At) + o(At) (4)
* Rearranging this equation and taking the limit as At -> 0
yield the master equation of the pure-death process given
below (see Figure 1):

- pn(t) =n+lPn+l(t)- nPn(t) n=n, n0 -1,...,2,1,0 (5)

For n = no, we have lno+1 =0; thus

dPno (t)
or, by virtue of Eq. (2),
dt no (t)

-no Pno (t)

-knoPno (t)

Forn=no- 1, no-2,...,2,1, Eq. (5) is


d Pn (t) = k(n + 1)pn+1 (t) knpn (t)

Finally, for n = 0, we have po0 = 0; thus,

d Po(t) = gI(t)

Spot) = kp (t)

Solution of the Master Equation As can be discerned
from Eq. (2), the intensity of death, Pn, is of linear form; as

--- = 0.02

. ...- x = 4c

10 20 30 40 50 60 70 80 90 100
. Temporal evolution of the binomial distribution of the
number of live cells, N(t), with no = 100.

Chemical Engineerrng Education

a result, the master equation, Eq. (5), can be solved recur-
sively, thereby yielding the probability that n cells will be
alive at time t, pn(t), as 5'341 (see Appendix A, available at

nt) no. e-kt)n e-ktno-n
Pn(t) [(no n)! n!(e

)( 1kt e -ktn (6)

The above expression indicates that the distribution of ran-
dom number N(t) is binomial with two parameters. One of
the parameters, no, representing the initial value of N(t), sig-
nifies the total number of the events that can possibly occur,
and the other parameter, e-kt, signifies the probability of oc-
currence of one event. For the thermal disinfection under
consideration, no is the number of cells alive at t = 0, which
will eventually die, and e-kt is the probability of an individual
cell being alive at time t. The temporal evolution of the bino-
mial distribution as given by Eq. (6) is illustrated in Figure 2;
for simplicity, no is specified to be 100 and kt is lumped as T.

Moments about the Mean With the solution of the mas-
ter equation in hand, we can proceed to calculate the mean
and variance of the process that should constitute the core of
any stochastic analysis and modeling. Furthermore, higher
moments about the mean, such as skewness and kurtosis, are
determined; they provide additional information useful for
characterizing the stochastic and statistical properties of the
process.[27,34-361 For illustration, the derivation of the skew-
ness is elaborated in Appendix B (available at>)
Mean. The mean, E[N(t)] or m(t), which is the expected
value (first moment) of the distribution of random variable
N(t), is defined as

m(t) = E[N(t)] = npn(t) (7)
The mean or expected value, m(t), is the weighted sum of the
realizations of the random variable where the weights are the
probabilities corresponding to those realizations. 341
For the thermal disinfection under consideration, the mean
in terms of dimensionless time T is
m(T) = noe-' (8)

The normalized, i.e., dimensionless form of the mean, m(r),

rm(T) m( e

Variance. The variance, Var[N(t)] or 2 (t), is the second
moment of the distribution of random variable N(t) about the

mean, m(t); thus

2 (t) = E{[N(t) -E[N(t)]]2 }= {n-E[N(t)}2 Pn(t) (10)

By expanding the above equation, 2 (t) can be related to the
mean, m(t), as

2(t) = EN (t) -m (t) (11)

In the above expression, E[N2(t)] is the second moment of
N(t), i.e.
E[N2 (t)]= n t) (12)
For the thermal disinfection under consideration, the vari-
ance in terms of dimensionless time T, 02 (), is, from Eqs.
(8) and (11)
0 n ()=noe- (i-e-") (13)

Standard Deviation. The standard deviation, o(t), of the
process is the square root of the variance, 02 (t); thus

(t)= [2 (t)]12 (14)
The variance, c2 (t), or more specifically, the standard de-
viation, o(t), signified the fluctuations, i.e., scatterings, of
the values of the random variable about their mean.
For the thermal disinfection under consideration, the stan-
dard deviation in terms of dimensionless time T, (o(j), is,
from Eqs. (13) and (14)

o(T)=n0e (1- e- ) (15)
GCettic int of Variation. The coefficient of variation, CV(t),
is the quotient (or ratio) of the standard deviation, o(t), and
the corresponding mean, m(t); thus[351

CV(t)= (t) (16)
For the thermal disinfection under consideration, the coef-
ficient of variation in terms of dimensionless time T, CV(j),
is, from Eqs. (8) and (15),

CV(T)= Ie (17)

1 1 1 1
CV(T)= no m) no
noe' no = m(T ) no


Skewness. The skewness, y(t), is the quotient (or ratio) of
the third moment of the distribution of random variable N(t)
about the mean, m(t), and the third power of the standard
deviation, u-(t); thus[361

Summer 2003

y(t) E{[N(t) E[N(t)]]3 }

tE{[N(t) E[N(t)]2

1 {Yn-E[N(t)]}3pn(t) (18)

By expanding the above equation, y(t) can be related to the
mean, m(t), and the standard deviation, o(t), as

7(t) = -1{E[N3 (t)] -3m(t)G2 (t)- m3 (t)} (19)

where E[N3(t)] is the third moment of N(t) defined by

E[N3 (t)]= n3n(t) (20)
Skewness characterizes the degree of asymmetry of the dis-
tribution of random variable N(t) about the mean, m(t). Posi-
tive skewness indicates a distribution with a longer tail to the
right of the mean than to the left, and negative skewness in-
dicates a longer tail to the left of the mean than to the right. It
vanishes for any symmetric distribution.
For the thermal disinfection under consideration, the skew-
ness in terms of dimensionless time T, y7(), is, from Eqs.
(8), (15), and (19)

y( -e= (21)

noe- (i-e )

Kurtosis (Curtosis). The kurtosis, k(t), is the quotient (or
ratio) of the fourth moment of the distribution of random vari-
able N(t) about the mean, m(t), and the fourth power of the
standard deviation, o(t); thus[361

k(t): E{[N(t) E[N(t)]]4 }

[E{[N(t) E[N(t)]2 t)

1 {n- E[N(t)p (t)
' ( n

By expanding the above equation, k(t) can be related to the
mean, m(t), the standard deviation, o(t), and the skewness,
(t), as


41(t) {E[N4 (t)]-4m(t)y(t)G3 (t)- m2 (t)[6G2 (t) + m2 (t)]}


where E[N4(t)] is the fourth moment of N(t), i.e.,


l Mean
| i(t) C -(r)1no
0.800 .- -- n =10
..-.. n,=100
n= 100,000

0 ] IExperimental data
06o o T = 52C
0 T = 54TC
SA T = 56C





oooD ~ -
0 2 4 6 8 10 44 46 48 50
Dimensionless time,

Figure 3. Normalized mean, -i, and standard deviation,
( as functions of the dimensionless time, T, exhibiting
the entire range of the number concentration of live cells.


- S5


Experimental data
O T= 52"Cy

5 T = 54C
I A T = 56C

0 2 4 6 8 10 12 14 16 18 20 44 46 48 50
Dimensionless time, T

Figure 4. Normalized mean, m, and standard deviation,
( as functions of the dimensionless time, T for the low
range of the number concentration of live cells.

Chemical Engineerrng Education

E[N4(t)]= n4pn(t) (24)
Kurtosis is a measure of how outlier-prone, or peaked, the
distribution of random variable N(t) is. The kurtosis of a dis-
tribution that is less outlier-prone than the normal distribu-
tion tends to be greater than 3, which is the kurtosis of the
normal distribution, and the opposite is the case for a distri-
bution that is more outlier-prone.
For the thermal disinfection under consideration, the kur-
tosis in terms of dimensionless time T, is, from Eqs. (8), (15),
(21), and (23)

k()=) 1-6e -e +3 (25)
noe [1-e )

The mean, variance, skewness, and kurtosis of N(t) have
been computed from their corresponding analytical expres-
sions, Eqs. (8), (13), (21), and (25), respectively. They are
functions of both the size of the initial cell population, no,
and the time, t. The value of the proportionality constant, k,
in the transition intensity, Eq. (2), can be recovered through
the least-square fitting of the expression for the mean, Eq.
(8), to the available experimental data of the thermal disin-
fection by means of a nonlinear minimization method, e.g.,
the Levenberg-Marquant method.[371

The model derived is illustrated with the same set of the
available experimental data for the thermal death of S. aureus
strain S-1 in neutral phosphate buffer[191 as those adopted by
the majortextbooks in biochemical engineering.[1-41 These data
have been obtained at the temperatures of 325.15 K (520C),
327.15 (540C), and 329.15 (560C), thereby yielding the val-
ues ofk as 0.0192 s-1 (1.15 min-1), 0.0362 s-1 (2.17 min-1), and
0.0678 s-1 (4.07 min1), respectively. With these values of k,
the mean, as well as the variance or standard deviation, skew-
ness, and kurtosis, of the number concentration of live cells
have been computed. These quantities are graphically pre-
sented and their significance is discussed.
Mean The mean, m(j), and the normalized mean, m(j),
have been computed according to Eqs. (8) and (9), respec-
tively. Only the latter, which is the exponential decay func-
tion independent of any parameters, is graphically plotted in
Figures 3 and 4 as a function of dimensionless time T. The
former exhibits the entire range of m(z) and the latter, the
low range. Naturally, Eq. (8) or (9) as well as m(r) are in
accord with those given in the available textbooks[1-41 as well
as in the original contributions.[5 61 The experimental data[191
are also superimposed in both figures for comparison.
Variance Stanardr Deviation and C'teic ient of Variation *

The expression for variance, o2, Eq. (13), is given by
Blakebrough11m and Fredrickson.['] The variance is a measure
of the variability, spread, or dispersion of the values of a ran-
dom variable. Naturally, the larger the value of the variance,
the greater the dispersion of the values of the random vari-
able about their mean.
The standard deviation, o, is obviously the square root of
the variance, (2 The value of o(T) as given by Eq. (15)
varies from 0 at T =0, reaches its maximum at 1/2, or z= n2,
where m(r)=1/2, and eventually vanishes as uT->-, as ex-
pected. This trend of the o(T) 's variation in terms of
m(r) + o(T) /n. for three values of n, is also illustrated in Fig-
ures 3 and 4. Note that in these figures, especially in the
latter, the deviations of the majority, if not all, of the avail-
able experimental data are substantially more pronounced than
those of the deviations predicted by the model in view of the
reported n, between 7.5 x 106 and 19 x 106 in obtaining the
experimental data.[19] This is almost always the case: the over-
all deviations of the experimental data include not only those
attributable to the internal or characteristic noises of the pro-
cess as predicted by the stochastic model, but also to the exter-
nal noises due to instrumental deficiencies and errors that can
never be totally eliminated.
The coefficient of variation, CV(u), is defined to provide a
meaningful relative measure of the variability, spread, or dis-
persion of the values of a random variable about their mean.
Note that CV(r) expresses the random variable's dispersion
as a fraction of the mean, or frequently, as a percentage.
For the thermal disinfection under consideration, CV(T),
specifically, the relative variation of size of the cell popula-
tion about its mean, at any dimensionless time T is inversely
proportional to the square root of the initial cell population
size, n., as indicated by Eq. (17) or (17a). By evaluating CV(t)
for various values of n., m(j), and their combinations, it can
be readily shown that
At any T, the larger the no, the smaller the CV(T), or
the relative extent of the fluctuations
For any no, CV(T) is initially zero and increases
monotonically with T, and the smaller the m(j), the
larger the CV(Q)
For any m(j), CV(z) increases monotonically with the
increase in no and asymptotically approaches a constant
value, and the smaller the no, the smaller the CV(j).

Skewness Skewness y(t), as defined by Eq. (18), mea-
sures the extent of asymmetry of the distribution of random
variable N(t) relative to its extent of deviation or dispersion.
Thus, it is indeed a meaningful measure of asymmetry; nev-
ertheless, y(t) depends on parameter no.
For the thermal disinfection under consideration, y(T) has
been evaluated according to Eq. (21) for no of 10, 100, and
100,000; the values of y(T) obtained are illustrated in Figure

Summer 2003

5. The configurations of the resultant curves can
be discerned at least qualitatively from that of the
probability distribution of N(t) presented in Figure
2; it is binomial in nature. Note that when t is nearly
0, N( ) can assume only those values immediately
to the left of n.; thus, its distribution is highly skewed
to the left and consequently, gives rise to y7() with
an appreciable negative value. The opposite is the
case as uT->o-. Naturally, in between these two ex-
tremes, y7() approaches from either direction to 0,
which is the value of y7() for the normal distribu-
tion, at T=T t1/2 = n2.
Kurtosis Kurtosis k(t), as defined by Eq. (22),
measures the degree of peakedness of the distribu-
tion of random variable N(t) relative to its extent of
deviation or dispersion. Thus, it is indeed a mean-
ingful measure of peakedness; nevertheless, as for
y(t), k(t) depends on parameter no.
For the thermal disinfection under consideration,
the values of k(T) have been computed by Eq. (25)
for no of 10, 100, and 100,000 and plotted in Figure
6. Similar to y7() presented in Figure 5, the con-
figuration of the resultant curves can be readily in-
terpreted in the light of Figure 2. When T is nearly
0, N( T) can assume only those values in the imme-
diate vicinity of no, and thus its distribution is highly
peaked, giving rise to k(z) with a large positive
value; the same is the case at ->- when the values
of N( T) are nearly 0. Inbetween these two extremes,
k(z) approaches from either direction to 3, which is
the value of k(T) for the normal distribution, at
TT1/2 = n2.

A stochastic model for the thermal death kinetics
of a cell population as a pure-death process has been
derived based on the first-order rate law. The solu-
tion of the governing differential equation of the
model, termed the master equation, yields the prob-
ability distribution of the number concentration (den-
sity) of live cells during disinfection, which is re-
garded as the random variable of the process; the
resultant distribution is the binomial distribution
whose two parameters are the number of cells ini-
tially alive, which will eventually die, and the prob-
ability of an individual cell being alive, expressed
as the exponential decay function. In addition to the
mean of live cells, various higher moments about
the mean have been derived to characterize this dis-
tribution. These higher moments include variance
(second moment about the mean), skewness (third
moment about the mean over the third power of the
standard deviation), and kurtosis (fourth moment

about the mean over the fourth power of the standard deviation). Only the
expressions for the mean and variance are available in one or more of the
existing major textbooks inbiochemical engineering. Naturally, augment-
ing them with the skewness and kurtosis would better characterize the
distribution, thereby deepening the understanding of stochastic, i.e., tem-
porally varying probabilistic, nature of thermal death of cells during their
Thermal disinfection has long been regarded as a suitable or useful
instructional example for illustrating stochastic analysis and modeling of
biochemical phenomena or processes, the majority of which deal with
discrete mesoscopic entities that are neither microscopic nor macroscopic.
To enhance its usefulness, the example has been substantially elaborated
in the current exposition. As indicated at the outset of this article, various
chemical engineering courses are richly populated with subjects involv-
ing mesoscopic entities, such as bubbles, droplets, and particles includ-
ing nanoparticles. Thus, these subjects would be suitable examples for

{M 2 -.. /- -' .............. ......-- -....... 5

S ----- ..............---- .........- 5
---- ----------- -- J -------------

0 2 4 6 8 10 12 14 16
Dimensionless time, t

-- n,=10
n= 100
n= 100,000

4- 46-48 50

5 .......4 ...............

44 46 48 50

Figure 5. Skewness as a function of the dimensionless time, t, and
the initial size of cell population, no as the parameter.

-I I I i -
I*........... *-'---- ------ -- ----------- --7........-----------

------------------ ------I-- --------------- ------------i
/- ------------

0 2 4 6 8 10 12
Dimensionless time,

5 : .
....... n.= 10
S- n= 100
n= 100,000
5:--- -

S... ....... .. .........

46 48 50
46 48 50

Figure 6. Kurtosis as a function of the dimensionless time, T, and
the initial size of cell population, no, as the parameter.

Chemical Engineerring Education

introducing the application of stochastic processes in these
courses, similar to the thermal destruction of microorganisms
for biochemical engineering.


The authors are grateful for the constructive criticisms re-
ceived from the three reviewers, including Professor
Ramkrishna of Purdue University.

E[N(t)] mean, expected value, or the first moment of the
random variable, N(t)
E[N2(t)] second moment of the random variable, N(t)
E[N3(t)] third moment of the random variable, N(t)
E[N4(t)] fourth moment of the random variable, N(t)
k proportionality constant in the intensity of death, Eq.
(2), min-1
k(t) kurtosis of the random variable, N(t)
N(t) random variable
m(t) E[N(t)]
n realization of the random variable, N(t)
no number of live cells at t = 0
pn(t) probability that the process will be in state n at time t
t time
y(t) skewness of the random variable, N(t)
Pn intensity of death for the pure-death process in state n

2 (t)

standard deviation of the random variable, N(t), as
defined in Eq. (14)

1. Blakebrough, N., "Preservation of Biological Materials Especially by
Heat Treatment," in Biochemical and Biological Engineering Science,
Vol. 2, N. Blakebrough, ed., Academic Press, London, England, p. 29
2. Aiba, S., A.E. Humphrey, and N.F. Mills, Biochemical Engineering,
2nd ed., Academic Press, New York, NY p. 247 (1973)
3. Bailey, J.E., and D.F. Ollis, Biochemical Engineering Fundamentals,
McGraw-Hill, New York, NY, p. 441 (1986)
4. Schuler, M.L., and F. Kargi, Bioprocess Engineering Basic Concepts,
Prentice Hall, Englewood Cliffs, NJ, p. 297 (1992)
5. Fredrickson, A.G., "Stochastic Models for Sterilization," Biotechnol.
Bioeng., 8, 167 (1966)
6. Aiba, S., and K. Toda, "Thermal Death Rate of Bacterial Spores," Pro-
cess Biochemistry, 35, (February 1967)
7. Ramkrishna, D., Population Balances: Theory and Applications to
Par .. .. ring, Academic Press, San Diego, CA,
p. 322 (2000)
8. Chen. W.Y., Z.P. Zhang, B.C. Shen, and L.T. Fan, "Stochastic Model-
ing of Tar Molecular Weight Distribution During Coal Pyrolysis,"
Chem. Eng. Sci., 49, 3687 (1994)
9. Fan, L.T., B.C. Shen, and S.T. Chou, "The Surface-Renewal Theory
of Interphase Transport: A Stochastic Treatment," Chem. Eng. Sci.,
48, 3971 (1993)
10. Fan, L.T., Y.Y. Chiu, J.R. Schlup, and S.T. Chou, "The Master Equa-
tion for Linear Adsorption and Desorption of Gases on Solid Surfaces,"
Chem. Eng. Commun., 108, 127 (1991)
11. Shen, B.C., L.T. Fan, and W. Y. Chen, "Stochastic Modeling of Ad-
sorption in a Batch System," J. Hazard. Mat., 38, 353 (1994)
12. Fox, R.O., and L.T. Fan, "Stochastic Modeling of Chemical Engineer-

Summer 2003

ing Systems: Application of the Generalized Master Equation to the
Bubble Population in a Bubbling Fluidized Bed," Chem. Eng. Sci., 42,
1345 (1987)
13. Fan, L.T., Y. Kang, M. Yashima, and D. Neogi, "Stochastic Behavior
of Fluidized Particles in a Liquid-Solid Fluidized Bed," Chem. Eng.
Comm., 135, 147 (1995)
14. Fan. L.T., S.H. Hwang, S.T. Chou, and R. Nassar, "Birth-Death Mod-
eling of Deep Bed Filtration: Sectional Analysis," Chem. Eng. Comm.,
35, 101 (1985)
15. Too, J.R., R. Nassar, S.T. Chou, and L.T. Fan, "Stochastic Analysis of
Crystallization in an Open Flow System," J. of Chin. I., Ch. E., 17,
303 (1986)
16. Chen. S.J., L.T. Fan, and C.A. Watson, "The Mixing of Solid Particles
in a Motionless Mixer: A Stochastic Approach,"AIChE., 18,984 (1972)
17. Nassar, R., S.T. Chou, and L.T. Fan, "Stochastic Analysis of Particle
Degradation in a Semi-Continuous Flow System Containing Solid
Particles," Hungarian J. Ind. Chem. Veszprem, 15, 73 (1987)
18. Duggirala, S.K., and L.T. Fan, "Stochastic Analysis of Attrition: A
General Cell Model," Powder Tech., 57, 1 (1989)
19. Walker, G.C., and L.G. Harmon, "Thermal Resistance of .
coccus aureus in Milk, Whey, and Phosphate Buffer,"Appl. Microbiol.,
14, 584 (1966)
20. Tsuchiya, H.M., A.G. Fredrickson, and R. Aris, "Dynamics of Micro-
bial Cell Populations," in Advances in Chemical Engineering, Vol. 6,
T.B. Drew, J.W. Hoopes, andT. Vermeulen, eds., Academic Press, New
York, NY, p. 126 (1966)
21. Ramkrishna, D., "Statistical Methods of Cell Populations," inAdvances
in Biochemical Engineering, T.K. Ghose, A. Filchter, and N.
Blakebrough, eds., Springer-Verlag, Berlin, p. 1 (1979)
22. Stephanopoulos, G., and A.G. Fredrickson, "Extinction Probabilities
in Microbial Prediction: A Birth and Death Approach," ..
Bio., 43, 165 (1981)
23. Nassar, R., S.T. Chou, and L.T. Fan, "A Probabilistic Model for the
Dynamics of a Structured Population of Unicellular Organism: Sto-
chastic Compartmental Model," Commun. Statist.-Stochastic Models,
6, 593 (1990)
24. Nassar, R., S.T. Chou, and L.T. Fan, "Stochastic Analysis of Stepwise
Cellulose Degradation," Chem. Eng. Sci., 46, 1651 (1991)
25. Lauffenburger, D.A., and J.J. Linderman, Receptors: Models for Bind-
ing, i .. and Signaling, Oxford University Press, New York,
NY, p. 134 (1993)
26. Cox, D.R., and H.D. Miller, The Stochastic Processes, John
Wiley & Sons, Inc., New York, NY, p. 165 (1965)
27. Chiang, C.L., An Introduction to Stochastic Processes and Their Ap-
plications, Robert E. Krieger Publishing Company, Huntington, NY,
p. 225, 271, 396 (1980)
28. van Kampen, H.G., Stochastic Processes in Physics and (
North-Holland, New York, NY, p. 1 (1981)
29. Fox, R.O., and L.T. Fan, "Stochastic Modeling of Chemical Process
Systems: Part 1, Introduction," Chem. Eng. Ed., 24, 56 (1990)
30. Fox, R.O., and L.T. Fan, "Stochastic Modeling of Chemical Process
Systems: Part 2, The Master Equation," Chem. Eng. Ed., 24, 88 (1990)
31. Fox, R.O., and L.T. Fan, "Stochastic Modeling of Chemical Process
Systems: Part 3, Application," Chem. Eng. Ed., 24, 164 (1990)
32. Chick, H., "An Investigation of the Laws of Disinfection," J. Hyg., 8,
92 (1908)
33. Oppenheim, I., K.E. Shuler, and G.H. Weiss, Stochastic Processes in
Chemical Physics: The Master Equation, The MIT Press, Cambridge,
MA, p. 53 (1977)
34. Lindgren, B.W., Statistical Theory, The Macmillan Company, New
York, NY, p. 50 (1962)
35. Ostle, B., and R.W. Mensing, Statistics in Research: Basic Concepts
and Techniques for Research Workers, The Iowa State University Press,
Ames, IA, p. 67 (1975)
36. Bock. R.K., and W. Krischer, The Dc. .- Springer-
Verlag, Berlin, pp. 27 and 165 (1998)
37. Bates, D.M., and D.G. Watts, Nonlinear Regression Analysis and Its
Applications, John Wiley & Sons, New York, NY, p. 80 (1988) 5

Fe R class and home problems

The object of this column is to enhance our readers' collections of interesting and novel prob-
lems in chemical engineering. Problems of the type that can be used to motivate the student by
presenting a particular principle in class, or in a new light, or that can be assigned as a novel
home problem, are requested, as well as those that are more traditional in nature and that eluci-
date 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 0. Wilkes (e-mail:, Chemical Engineering Department,
University of Michigan, Ann Arbor, MI 48109-2136.




North Carolina State University Raleigh, NC 27695

wo kinds of pressure-volume-temperaure P-v-T dia-
grams for real substances are usually introduced in
the first undergraduate thermodynamics course (here-
after, "thermo I"): a typical P-v-T diagram for a substance
that contracts on freezing, e.g., carbon dioxide, and a P-v-T
diagram for a substance that expands on freezing, usually
water (see, for example, References 1-4). The relevance of
the P-v-T data is usually discussed, and different P-v-T func-
tions or equations of state are introduced. From this moment
until the end of their professional life, chemical engineering
students (or chemical engineers) will have to deal with the P-
v-T relations, in one way or another.
Most chemical engineers will never develop a new EOS,
but they will often be in the position of having to select an
equation that is the most appropriate for a specific situation.
De Nevers and Seader5'] pointed out ten years ago that stu-
dents must learn that "computer answers may depend strongly
on which correlations for thermodynamic properties are used.
Students need to learn of the many sources of such correla-
tions, along with their limitations and recommended regions
of applicability. Also, they need to be aware of experimental
sources of data and how to make comparisons between ex-
perimental data and empirical correlations."
It is well known that practical engineering applications are
deeply affected by fundamentals of thermodynamics. Harvey
* Simon Bolivar University Caracas, 1080A, Venezuela

and Laesecke[6] recently stated, "Engineers should learn the
value of thermophysical properties in college." We agree
with this statement. Students must be able to choose from
among several methods and be aware of how their choice
may depend on the nature of the system and state conditions;
moreover, they should be exposed to this matter at an early
stage of their career.
In traditional chemical engineering curricula, students start
to evaluate different models using process simulators such as
ASPEN PLUS, ProSIm, or ChemStation in their chemical
process courses, typically during the third and/or fourth year

Coray M. Colina is Senior Assistant Profes-
sorin the Department of Thermodynamics and
Transport Phenomena at Simon Bolivar Uni-
versity. She received her BS (1993) and her
MS (1994) in chemical engineering from Simon
Bolivar University and after six years of teach-
ing she returned to school and is working on
her PhD at North Carolina State University

SK Keith E. Gubbins is the W.H. Clark Distin-
guished University Professor at North Carolina
State University, where he has been since
1998. He obtained his PhD at the University of
I London and has been a faculty member at the
University of Florida and Cornell University prior
to joining North Carolina State University.

Copyright ChE Division ofASEE 2003

Chemical Engineerring Education

of their studies. In this work, we give an example of how
students can be introduced to this topic from an early stage in
their career (thermo I), continuing progressively to their pro-
fessional life (or graduate studies). We present a thermo
project where the student will need to use the Internet, handle
some software, and read tables.

Carbon dioxide (CO2) has been studied since the days of
van der Waals, since it exhibits supercritical behavior at mod-
erate conditions. Nowdays, CO2 is used in a variety of fields,
e.g., in the oil industry (as a carrier gas for enhanced oil re-
covery) and in "green chemistry," where research efforts are
under way to identify sustainable processes and products us-
ing CO2-related technology. Carbon dioxide is extremely at-
tractive in industrial applications because it is the second most
abundant and the second least expensive solvent on earth.
In this project, the student is asked to evaluate the P-v-T
prediction capabilities of different models for carbon diox-
ide. They are asked to obtain Pv (or Pp, where p = 1/v =
density) diagrams for carbon dioxide showing at least three
isotherms, 278.15 K, 198.15 K, and 360 K, for a pressure
range from 0.51795 MPa up to 100 MPa, in the gas and liq-
uid phases, as well as the vapor-liquid equilibrium region
(from the triple point [Tt = 216.592 K, Pt = 0.51795 MPa] up
to the critical point). They are told to compare the volume (or
density) predictions in each phase (including the critical re-
gion) obtained from
Two cubic equations of state (e.g., Soave-Redlich-
Kwong17 1] and Peng-Robinson18,11)
A three-parameter corresponding states method (e.g.,
Lee-Kesler[9,11 tables)
A multiparametric EOS for CO2 (e.g., Span-
Tables of thermodynamic properties for CO2 (IUPAC[121
For extra credit, the student can comment on the use of the
virial EOS (truncated at the second virial term, in the density
expansion) and the Rackett[13,14,11 equation for the determi-
nation of volume for the gas and saturated liquid phases,
The tables mentioned in Section d, below, are from fits of
thermodynamic properties to multiparametric equations, i.e.,
they are smoothed data. But these tables1121 are based on older
carbon dioxide data. We believe that the Span-Wagner equa-
tion1101 is the best representation of the experimental data for
carbon dioxide and should be used as the standard against
which to test the other equations of state and the correspond-
ing states method.[151 Span and Wagner made a careful analy-
sis of all the experimental data for carbon dioxide, including
recent data, and excluded data that did not satisfy the ther-
modynamic consistency conditions when fitting their equa-
tion of state.

a. This project would not have been feasible for a thermo I course
ten to fifteen years ago, but with the availability of computers,
software, and the Internet, it is now rather straightforward. Text-
books come with computer-aided strategies for solving cubic
equations of state, such as a) the use of "meta-computing" soft-
ware involving the use of packages such as Maple, Mathematica,
Mathcad, and Matlab"'e,1, 3]; b) the use of spreadsheets such as
EXCEL,'e,31] or c) the supply of a "code" that can be used for a
specific cubic or multiparametric equation of state. 2,31Addition-
ally, several sites can be found on the Internet where on-line
cubic EOS software is available at no charge.['e,161
b. Corresponding-state correlations are usually introduced in text-
books[1-4] in graphical or tabular form. We believe that students
must become familiar with the use of tables and graphs, but we
should also keep in mind the pedagogic importance of the imple-
mentation of corresponding-state principles by meta-comput-
ing software, as was recently discussed by Smith, et al. [1"I
c. Students should also be exposed to state-of-the-art
multiparametric EOS available in up-to-date databases. This can
be pursued through the Internet, where leading institutions such
as NIST or DECHEMA have friendly web pages. For example,
the NIST chemistry webbook[11] contains thermophysical prop-
erties for 33 pure substances. Tester and Modell81 list in Chap-
ter 13 of their book some of the commonly used sources of pure
component and mixture data in both published and electronic
form. Additionally, they give examples of United States research
centers that focus on physical property measurements and data
Ideally, recent developments in molecular-based EOS should
also be included. Unfortunately, to the best of our knowledge,
molecular-based EOS (e.g., StatisticalAssociating Fluid Theory,
or SAFTI191) software is not yet available for use at the under-
graduate level. The availability of such software in the near fu-
ture is likely, however, as these kinds of EOS are already in-
cluded in graduate-level textbooks[18,201 and in at least one intro-
ductory textbook.[3]
d. Students should become familiar with the use of tables for the
determination of thermodynamic properties of fluids. Usually,
the steam tables are in an appendix in thermodynamic textbooks,
and sometimes a few extra tables for different substances are
also included. It should be made clear to the students that no
published thermodynamics table is a direct transcript of experi-
mental measurements, because such data must be smoothed, cor-
related, and made internally consistent by means of a mathemati-
cal model, i.e., by an equation of state. Thus, thermodynamic
tables can become obsolete as new data and improved funda-
mental EOSs become available. For carbon dioxide, several
sources are available.[10,11,21,22]

Homogeneous fluids are normally divided into liquids and
gases, but this distinction cannot always be sharply drawn
because the two phases become indistinguishable at the criti-
cal point. Several authors1-'4] define the supercritical region
as the region existing at temperatures and pressures above
the critical temperature and critical pressure; this region is
bounded by the dotted lines in Figure 1. For simplicity, in

Summer 2003

this project we will use this definition of the supercritical state. The criti-
cal conditions for CO2 are To = 304.128 K, Po = 7.3773 MPa, and p =
467.6 kg/m3. Thus, the isotherm at 360 K will be analyzed as supercritical
for pressures above the critical pressure.
In this work, we use the Internet to obtain the predictions of the Span-
Wagner EOS through the NIST webbook11l and to obtain the predictions
of the Peng-Robinson equation from Reference 16. We read the tables
presented in the Appendix of Reference 1 for the Lee-Kesler model and
also the tables presented in Reference 12 for pure carbon dioxide. Fi-
nally, we program the Soave-Redlich-Kwong EOS using the
VisualBasic macro of Excel and use that program to show the re-
sults. This is only one of several possible lines of attack that stu-
dents can choose to solve their assignment.
Vapor-Liquid Equilibria Region
The phase envelope (vapor-liquid equilibria or saturation conditions)
for carbon dioxide is shown in Figure 2. Results are shown for the Span-
Wagner, Peng-Robinson, and Soave-Redlich-Kwong equations and also
from the IUPAC tables. Results from the Lee-Kesler method are not shown
because the available tables in Reference 1 (as well as in the original
publication[9]) are provided only for the single phase (liquid or vapor).
(The reason is that the Lee-Kesler method fails to predict a well-behaved
van der Waals loop at near-critical temperatures, thus making it impos-
sible to find two fluid states with the same fugacity; in normal practice,
saturation pressures from the Lee-Kessler EOS are obtained from a sepa-
rate linear correlation proposed by the same authors.[9])
This figure illustrates the predictive capabilities of multiparametric
equations of state. The Span-Wagner EOS is the equation most frequently
used for carbon dioxide and can be taken as a reference (for a more de-
tailed discussion, see References 10 and 11). On the scale of the graph,
results from the IUPAC tables are equivalent. It is worth mentioning that
tables,P1] software, and Internet[11] were used in this stage of the project,
and they can also be used to show that "old" methods (tables) are not
necessarily less accurate than icn" methods (Internet), or vice versa.
Deficiencies of cubic equations of state in the critical region as well as in
the prediction of liquid densities are shown by the SRK and PR re-
sults. For carbon dioxide, in the region under study, the PR equation
is more accurate than SRK.
The phase envelope for carbon dioxide is also shown in Figure 3, but
in a P-p diagram, which can be used to show the practical convenience of

Figure 1. p
PT critical
diagram p
for a pure
substance liquid /
that solid / i
contracts vapor
on ap



10 100 1000 10000
v (cm3/mol)
Figure 2. Vapor-liquid phase envelope for carbon
dioxide shown as a P-v diagram. Continuous line
predicted by the Peng-Robinson EOS,f8' dashed line
predicted by the Soave-Redlich-Kwong EOS, (0)
predicted by the Span-Wagner EOS,fl0 and (A) pre-
dicted from the IUPAC tables.

0 0 001 m 0015
p (rol/emr)

Figure 3. Vapor-liquid phase envelope for carbon
dioxide shown as a P-p diagram. Continuous line
predicted by the Peng-Robinson EOS,8'l dashed line
predicted by the Soave-Redlich-Kwong EOS,1 (0)
predicted by the Span-Wagner EOS,f101 and (A) pre-
dicted from the IUPAC tables.

003 .... ..
0 026
001 --------- ---- --------
210 Tp 230 250 270 290 310

Figure 4. Saturated liquid density versus tempera-
ture diagram for carbon dioxide. Continuous line
predicted by the Peng-Robinson EOS81', dashed line
predicted by the Soave-Redlich-Kwong EOS,17 (0)
predicted by the Span- Wagner EOS,`101 (0) from the
Rackett equation,[s31and (A) from the modified
Racket equation due to Spencer and Adler.1141

Chemical Engineerrng Education

9 oA

002 002 0


S0025 00 0
6E o o


0 P'" 20 40 60 80 100
P (MPa)

Figure 5. Density versus pressure diagram for car-
bon dioxide at 298.15 K, liquid phase. Continuous
line predicted by the Peng-Robinson EOS,f8' dashed
line predicted by the Soave-Redlich-Kwong EOS,m7
(0) predicted by the Span-Wagner EOS,'101 and (*)
from the corresponding states method (Lee-Kesler
tables in [1]).







0 1 2 3 4 5 6 Ps 7
P (MPa)
Figure 6. Density versus pressure diagram for car-
bon dioxide at 298.15 K, vapor phase. Continuous
lines predicted by Peng-Robinson EOS,f'8 (0) pre-
dicted by Span-Wagner EOS,0o0 and (*) predicted
by virial equation (expanded in the density series
and truncated at its second term).

0 0 2 5 .. .. ... .. ... .. ... .. .. .. ... ..
O 02 5

'0 015



0 P. 20 40 60 80 100
P (MPa)

Figure 7. Density versus pressure diagram for car-
bon dioxide at 360 K, a supercritical state. The
continuous line is the Peng-Robinson EOS,18 (0)
is predicted by Span-Wagner EOS,'0'1 and (*) is
predicted by the virial equation (expanded in the
density series and truncated at its second term).

working in terms of density. (Density is convenient to use because it al-
ways stays finite as P -> 0, whereas V diverges.3])
The Rackett equation[13] is a simple empirical equation for determina-
tion of saturated liquid volumes and relates them to the reduced tempera-
ture and the critical compressibility factor. Molar densities obtained from
the original Rackett equation are shown in Figure 4 (closed gray
circles); closed triangles represent the prediction from the modified
Rackett equation due to Spencer and Adler.[14] Among the methods
studied here, the modified Rackett equation gives the best prediction
of saturated liquid densities.
Liquid Phase
The density versus pressure diagram for carbon dioxide at 298.15 K is
shown in Figure 5, from the saturation pressure (6.43 MPa) up to 100
MPa. Results are shown for the Span-Wagner, Peng-Robinson, and Soave-
Redlich-Kwong equations and also from the corresponding states method
(Lee-Kesler tables). Results from the 1UPAC tables are not shown for clarity
of the graph, but the results are equivalent to the Span-Wagner equation
up to 60 MPa (range of validity of the IUPAC tables). This figure corrobo-
rated deficiencies of cubic equations of state in the prediction of liquid
densities, as shown by the SRK and PR results. Prediction from the Lee-
Kesler method are also shown in this figure, and good agreement with the
Span-Wagner equation is obtained up to the range of validity of the tables
(73.7 MPa or Pr = P/Pc = 10).
Vapor Phase
Except for the virial equation, all EOSs represent the saturated vapor
region very well, as seen from Figures 2 and 3. In Table 1 we present the
relative errors with respect to the Span-Wagner prediction for the two iso-
therms under study (278.15 and 298.15), from 0.51795 MPa (triple point)
up to the vapor pressure at each temperature. Table 1 shows that all equa-
tions behave very well in the vapor phase, with an average error less than
2%, except the virial equation.
The virial equation (expanded in the density series and truncated at its
second term) is often recommended[1-4] for use at low pressures (under 1
MPa). The results in Figure 6 show that for carbon dioxide at 298.15 K the
virial equation can be used up to 2.5 MPa, with an error less than 2%.
The PR EOS predictions are shown for comparison. The second virial
coefficient was taken from the correlation due to Pitzer and Curl and
presented in Reference 1.
Supercritical State
The density versus pressure diagram for carbon dioxide at 360 K is shown

Summer 2003

Relative Errors* for the Vapor Phase
with Respect to the Span-Wagner EOS Predictions

EOS/Temperature (K) Virial SRK'1 PRIs LK9'1 IUPAC[1]
278.15 (P 3.9695 MPa) 5 2.7 1.3 0.82 1.76
298.15 (P 6.4342 MPa) 4.6 1.9 0.76 0.64 1.17

* Np
PSpan-Wagner PEOS 100

n \ PSpan-Wagner I

in Figure 7, from the critical pressure up to 100 MPa.
Results are shown for the same equations and methods
presented for the liquid phase (Figure 5) corroborating
the validity of the corresponding states method (Lee-
Kesler tables) for carbon dioxide in the homogeneous
phase up to the range of validity of the tables (73.7 MPa
or Pr = P/P = 10). Deficiencies of cubic equations of
state in the prediction of liquid-like densities (high pres-
sures) are shown by the SRK and PR results.
The "Complete" P-v Diagram
The P-v diagram for carbon dioxide predicted from
the Span-Wagner equation is shown in Figure 8. The
construction of this diagram is straightforward for the
student from the NIST webpage.[11]

The results in Figures 2-7 can be used to discuss the
advantages and deficiencies of the various methods. For
example, Figure 3 can be used to discuss the deficien-
cies of cubic equations of state in the critical region as
well as in the prediction of liquid densities, and Figure
5 can serve to show the importance of using a model
within its range of validity; e.g., if the IUPAC tables
are extrapolated to 100 MPa (results not shown), errors
up to 200% can be found.
Care must be taken to emphasize that although it is
possible to draw conclusions about the best EOS for
CO2 (or any other substance) for a given property and
range of state conditions, such conclusions cannot be
extended to other fluids, properties, or state conditions
without further calculations. Thus, although the PR
equation is found to be superior to the SRK equation
for CO2 properties studied here, quite different conclu-
sions may be drawn for another fluid. For this purpose,

i \ \
10 I \
a 8 \ T \
a. ....

4 ------------------------------- -

10 100 1000 10000
v (cm /mol)

Figure 8. Carbon dioxide P-v diagram predicted by
Span-Wagner EOS[1] The continuous line represents
the vapor-liquid phase envelope. The dotted-dashed
line is the 360 K isotherm, the dashed line is the 320 K
isotherm, the dotted line is the 298.15 K isotherm, and
the double-dotted dashed line is the 278.15 K isotherm.

different substances can be assigned to each student, or results for
other substances/regions can be shown in class.


We would like to thank our students at both Sim6n Bolivar and
North Carolina State Universities for their support, feedback,
and motivation on the implementation of this project. This ma-
terial is based on work supported in part by the STC Program
of the National Science Foundation under Agreement No. CHE

1. Smith, J.M., H.C. van Ness, and M.M. Abbott, Introduction to Chemical En-
gineering Thermodynamics, 6th ed., McGraw Hill, New York (2001)
2. Van Wylen, J.G., R.E. Sonntag, and C. Borgnakke, Fundamentals of Classi-
cal Thermodynamics, 4th ed., John Wiley & Sons (1994)
3. Elliott, J.R., and C.T. Lira, Introductory Chemical Engineering Thermody-
namics, Prentice Hall (1999)
4. Cengel, Y.A., and M.A. Boles, Thermodynamics, 2nd ed., McGraw Hill (1996)
5. de Nevers, N., and J.D. Seader, "Helping Students Develop a Critical Attitude
Towards Chemical Process Calculations," Chem. Eng. Ed., 26(2), 88 (1992)
6. Harvey, A.H., and A. Laesecke, "Fluid Properties and New Technologies: Con-
necting Design with Reality," Chem. Eng. Prog., 98, 34 (2002)
7. Soave, G., "Equilibrium Constants from a Modified Redlich-Kwong Equa-
tion of State," Chem. Eng. Sci., 27, 1197 (1972)
8. Peng, D.-Y., and D.B. Robinson, "A New Two-Constant Equation of State,"
Ind. Eng. Chem. Fund., 15(1), 59 (1976)
9. Lee, B.I., and M.G. Kesler, "AGeneralized Thermodynamics Correlation based
on Three-Parameter Corresponding States," AIChEJ., 21, 510 (1975)
10. Span, R., and W. Wagner, "ANew Equation of State for Carbon Dioxide Cov-
ering the Fluid Region from the Triple-Point Temperature to 1100 K at Pres-
sures up to 800 MPa," J. Phys. Chem. Ref Data, 25, 1509 (1996)
11. Lemmon, E.W., M.O. McLinden, and D.G. Friend, "Thermophysical Proper-
ties of Fluid Systems," in NIST ( WebBook, NIST Standard Refer-
ence Database Number 69, P.J. Linstrom and W.G. Mallard, eds., (2001) Na-
tional Institute of Standards and Technology, Gaithersburg MD 20899 />
12. Angus, S., B. Armstrong, and K.M. De Reuck, IUPAC Carbon Dioxide Inter-
national Thermodynamic Tables of the Fluid State-3, Imperial College, Lon-
don; Pergamon Press (1976)
13. Rackett, H.G., "Equation of State for Saturated Liquids,"J. Chem. Eng. Data,
15, 514 (1970)
14. Spencer, C.F., and S.B. Adler, "A Critical Review of Equations for Predicting
Saturated Liquid Density," J. Chem. Eng. Data., 23, 82 (1978)
15. Jacobsen, R.T., S.G. Penoncello, E.W. Lemmon, and R. Span, "Multiparam-
eter Equations of State," in Equations of State for Fluids and Fluid Mixtures,
Elsevier (2000)
17. Smith, W.R., M. Lisal, and R.W. Missen, "The Pitzer-Lee-Kesler-Teja (PLKT)
Strategy and Its Implementation by Meta-Computing Software," Chem. Eng.
Ed., 35(1), 68 (2001)
18. Tester, J.F., and M. Modell, Thermodynamics and Its Applications, 3rd ed.,
Prentice Hall (1997)
19. Chapman, W.G., K.E. Gubbins, G. Jackson, and M. Radosz, "New Equation
of State for Associating Liquids," Ind. Eng. Chem. Res., 29, 1709 (1990)
20. Prausnitz, J.M., R.N. Lichtenthaler, and E. Gomes de Azevedo, Molecular
Thermodynamics ofFluidPhase Equilibria, 3rd ed., Prentice Hall (1999)
21. Lemmon, E.W., R.T. Jacobsen, S.G. Penoncello, and S.W. Beyerlein, "Com-
puter Programs for Calculating Thermodynamic Properties of Fluids of Engi-
neering Interest," Report 97-1, Center for Applied Thermodynamic Studies,
University of Idaho, Moscow, ID (1997)
22. Vargaftik, N.B., Y.K. Vinogradov, and VS. Y ..O... hysicalProp-
erties of Liquids and Gases, Third augmented and revised edition, Begell
House, Inc., NY (1996) O

Chemical Engineerring Education

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 EZ2KN7TPZ_FS70GP INGEST_TIME 2011-11-02T14:42:42Z PACKAGE AA00000383_00003

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 EHOIY86R3_OFGL6C INGEST_TIME 2011-05-04T18:01:58Z PACKAGE AA00000383_00003