• TABLE OF CONTENTS
HIDE
 Front Cover
 Table of Contents
 Michael L. Shuler of Cornell...
 Rice University
 Membrane science and technology...
 Agitation and aeration: An automated...
 Integrating biology and ChE at...
 Call for papers
 We hold these truths to be...
 Improving coherence in technical...
 Teaching tips
 A simple open-ended vapor diffusion...
 Computer evaluation of exchange...
 Experiments and other learning...
 Incorporating nonideal reactors...
 Use of dynamic simulation to converge...
 Using small blocks of time for...
 Teaching nonideal reactors with...
 Back Cover


































Chemical engineering education
http://cee.che.ufl.edu/ ( Journal Site )
ALL VOLUMES CITATION THUMBNAILS DOWNLOADS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/AA00000383/00158
 Material Information
Title: Chemical engineering education
Alternate Title: CEE
Abbreviated Title: Chem. eng. educ.
Physical Description: v. : ill. ; 22-28 cm.
Language: English
Creator: American Society for Engineering Education -- Chemical Engineering Division
Publisher: Chemical Engineering Division, American Society for Engineering Education
Place of Publication: Storrs, Conn
Publication Date: Spring 2004
Frequency: quarterly[1962-]
annual[ former 1960-1961]
 Subjects
Subjects / Keywords: Chemical engineering -- Study and teaching -- Periodicals   ( lcsh )
 Notes
Citation/Reference: Chemical abstracts
Additional Physical Form: Also issued online.
Dates or Sequential Designation: 1960-June 1964 ; v. 1, no. 1 (Oct. 1965)-
Numbering Peculiarities: Publication suspended briefly: issue designated v. 1, no. 4 (June 1966) published Nov. 1967.
General Note: Title from cover.
General Note: Place of publication varies: Rochester, N.Y., 1965-1967; Gainesville, Fla., 1968-
 Record Information
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 01151209
lccn - 70013732
issn - 0009-2479
Classification: lcc - TP165 .C18
ddc - 660/.2/071
System ID: AA00000383:00158

Downloads
Table of Contents
    Front Cover
        Front Cover 1
        Front Cover 2
    Table of Contents
        Page 81
    Michael L. Shuler of Cornell University
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
    Rice University
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
    Membrane science and technology in the 21st century
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
    Agitation and aeration: An automated didactic experiment
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
    Integrating biology and ChE at the lower levels
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
    Call for papers
        Page 113
    We hold these truths to be self-evident
        Page 114
        Page 115
    Improving coherence in technical writing
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
    Teaching tips
        Page 121
    A simple open-ended vapor diffusion experiment
        Page 122
        Page 123
        Page 124
        Page 125
    Computer evaluation of exchange factors in thermal radiation
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
    Experiments and other learning activities using natural dye materials
        Page 132
        Page 133
        Page 134
        Page 135
    Incorporating nonideal reactors in a junior-level course using computational fluid dynamics (CFD)
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
    Use of dynamic simulation to converge complex process flowsheets
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
    Using small blocks of time for active learning and critical thinking
        Page 150
        Page 151
        Page 152
        Page 153
    Teaching nonideal reactors with CFD tools
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
    Back Cover
        Back Cover 1
        Back Cover 2
Full Text



'I
CEE

VOLUME38 NUMBER2 SPRING 2004


tchael L. Shuler

DOW AWARD LECTURE



Feature Articles...









and ChE at...





Visit

us

on the
Web
at


http://cee.che.ufl.edu/index.html











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

EDITOR
Tim Anderson
ASSOCIATE EDITOR
Phillip C. Wankat
MANAGING EDITOR
Carole Yocum
EDITORIAL ASSISTANT
Aimde Baum

PROBLEM EDITOR
James O. Wilkes, U. Michigan
LEARNING IN INDUSTRY EDITOR
William J. Koros, Georgia Institute of Technology


-PUBLICATIONS BOARD

CHAIRMAN *
E. Dendy Sloan, Jr.
Colorado School of Mines
MEMBERS
Pablo Debenedetti
Princeton University
Dianne Dorland
Rowan University
Thomas F. Edgar
University of Texas at Austin
Richard M. Felder
North Carolina State University
Bruce A. Finlayson
University of Washington
H. Scott Fogler
University of Michigan
Carol K. Hall
North Carolina State University
William J. Koros
Georgia Institute of Technology
John P. O'Connell
University of Virginia
David F. Ollis
North Carolina State University
Ronald W. Rousseau
Georgia Institute of Technology
Stanley I. Sander
University of Delaware
Richard C. Seagrave
Iowa State University
C. Stewart Slater
Rowan University
Donald R. Woods
McMaster University


Chemical Engineering Education


Volume 38


Number 2


Spring 2004


> EDUCATOR
82 Michael L. Shuler of Cornell University, Claude Cohen

> UNIVERSITY
88 Rice University, Kyriacos Zygourakis

> DOWAWARD LECTURE
94 Membrane Science and Technology in the 21st Century,
William B. Krantz

> LABORATORY
100 Agitation and Aeration: An Automated Didactic Experiment,
Alberto C. Badino, Jr., Paulo I.E De Almeida, Antonio J.G. Cruz
132 Experiments and Other Learning Activities Using Natural Dye Materials,
Veronica A. Burrows

> CURRICULUM
108 Integrating Biology and ChE at the Lower Levels,
Kathryn A. Hollar Stephanie H. Farrell, Gregory B. Hecht,
Patricia Mosto
136 Incorporating Nonideal Reactors in a Junior-Level Course Using
Computational Fluid Dynamics (CFD),
Benjamin J. Lawrence, Jason D. Beene, Sundararajan V Madihally,
Randy S. Lewis

> RANDOM THOUGHTS
114 We Hold These Truths To Be Self-Evident, Richard M. Felder

> CLASSROOM
116 Improving Coherence in Technical Writing, G.K. Suraishkumar
126 Computer Evaluation of Exchange Factors in Thermal Radiation,
Redhouane Henda
142 Use of Dynamic Simulation to Converge Complex Process Flowsheets,
William L. Luyben
150 Using Small Blocks of Time for Active Learning and Critical Thinking,
Stephen J. Lombardo
154 Teaching Nonideal Reactors with CFD Tools,
Luis M. Madeira, Manuel A. Alves, Alirio E. Rodrigues

P CLASS AND HOME PROBLEMS
122 A Simple Open-Ended Vapor Diffusion Experiment,
David Whitmire, Wayne Blaylock

113 Call for Papers
121 Teaching Tips

CHEMICAL ENGINEERING EDUCATION (ISSN 0009-2479) is published quarterly by the Chemical Engineering
Division, American Society for Engineering Education, and is edited at the University of Florida. Correspondence
regarding editorial matter, circulation, and changes of address should be sent to CEE, Chemical Engineering Department,
University of Florida, Gainesville, FL 32611-6005. Copyright 0 2004 by the Chemical Engineering Division, American
Society for Engineering Education. The statements and opinions expressed in this periodical are those of the writers and not
necessarily those of the ChE Division, ASEE, which body assumes no responsibility for them. Defective copies replaced if
notified within 120 days of publication. Write for information on subscription costs and for back copy costs and availability.
POSTMASTER: Send address changes to Chemical Engineering Education, Chemical Engineering Department., University
of Florida, Gainesville, FL 32611-6005. Periodicals Postage Paid at Gainesville, Florida and additional post offices.


Spring 2004









educator


Michael L. Shuler

of Cornell University


CLAUDE COHEN
Cornell University Ithaca, NY 14853
he School of Chemical (now Chemical and
Biomolecular) Engineering at Cornell University is
proud to have Michael L. Shuler, a true educator and
a visionary, among its faculty members. Mike is a great ex-
ample of what it means to be an educator. He is imaginative,
creative, and prolific; most importantly, he is sincerely inter-
ested in the advancement of the field. He has a genuine curi-
osity and enthusiasm in guiding his students to become inde-
pendent researchers, and he gives them the freedom to ex-
plore new research directions.
82


BACKGROUND AND EDUCATION
Michael L. Shuler grew up in Joliet, Illinois, where as an
8t' grader he had already decided to pursue biochemical en-
gineering. All students were required to write a description
of the career they wanted to pursue. Mike chose chemical
engineering and wrote specifically about building processes
to make antibiotics. He felt that making a life-saving drug in
large enough quantities to really help people was a noble and
ennobling enterprise. It combined his interests in science, so-
ciety, and service.
Mike's choice of career path was a natural one given his
family situation. His father Louis put himself through col-
lege (Bradley University) during the depression, majoring in
chemistry, and began his worklife at the Army Ammunition
Plant in Joliet before joining the Army Air Force in World
War II. He spent his Army career as a civilian employee work-
ing as an industrial engineer, ultimately coordinating a pro-
duction schedule of 700 different chemical plants. At the end
of the war, Louis married Mary Boylan, who taught English
to seventh graders in West Des Moines, Iowa. Not surpris-
ingly, both Mike and his younger brother, Patrick, ended up
being chemical engineers. Pat worked for Chevron prior to
becoming manager of Oil Field Chemistry and Production
Chemicals in the Petroleum Energy and Environmental Cen-
ter administered by Caltech.
Growing up as a Roman Catholic, Mike's religious beliefs
have always been an important part of his life and have greatly
influenced his thoughts on service. He always had an interest
in teaching, and while at the University of Notre Dame he
became committed to the idea of teaching at the university
level. Kramer Luks and Nick Sylvester were new, young fac-
ulty members there and they both strongly encouraged him
to pursue graduate studies at the University of Minnesota.
When Mike graduated from Notre Dame in 1969, the Viet-
nam War was at its peak; Mike had participated in ROTC
both in high school and at Notre Dame and graduated as a
second lieutenant in the U.S. Army. He was able to obtain an
educational delay to go to graduate school. At that time the
only other male U.S. students who would enter graduate
Copyright ChE Divison of ASEE 2004


Chemical Engineering Education






















MiKe wlrn uornei jaculty members ar me zUUz annual IiunrI meeting in
Indianapolis, pictured at the NCAA Hall of Champions where Cornell's re-
ception was held. Front row: Fernando Escobedo, Paulette Clancy. Back row:
Matt DeLisa, Kelvin Lee, David Putnam, Bill Olbricht, Mike Shuler, Yong Joo,
and Abe Stroock.
school had either failed the Army's physical exam, were veterans, or had joined the
U.S. National Guard or Reserves and had completed six months of training. While
Mike had no interest in the Army as a career, he felt that the experience was useful in
developing leadership and teaching skills (the only formal instruction he ever re-
ceived in pedagogy).
At the University of Minnesota he decided to work with Professors Rutherford
"Gus" Aris and Henry Tsuchiya. Gus was a theoretician with exceptional mathemati-
cal skills and Henry was a bacteriologist. Mike soon realized that he knew more math
than Henry and more biology than Gus, and that part of his role was to help bridge
the interdisciplinary divide. This experience prepared Mike well for conducting and
leading interdisciplinary research teams as he advanced along his career path. Mike
also completed a minor (supporting program of courses) in mathematics, biochemis-
try, and microbiology and was only one course short of fulfilling the requirements
for a Masters of Science in microbiology. A strong, formal background in biology
was unusual for a chemical engineering student of that day and time and it provided
Mike with an advantage that proved to be important to his early academic success.
When the time came to interview for an academic job, Neal Amundson, head of
Minnesota's Chemical Engineering Department, took a great deal of interest in help-
ing Mike (along with Gus and Henry). His perspective and insight were crucial;
Amundson thought that chemical engineering at Cornell had great potential and that
the faculty there would welcome someone with Mike's talents. The department at
Cornell at that time was in transition from a department with a focus on undergradu-
ate education to one that was developing a top-rated research department.
Mike's selection of Cornell was based not only on the advice of the faculty at
Minnesota, but also on the supportive environment he found when he visited there.
In particular, he found that Bob Finn (a founding father of biochemical engineering)
had a well-established laboratory with the entire infrastructure necessary for bio-
chemical engineering. During those days, academic offers did not include start-up
packages and none of Mike's three academics offers had included any money for
equipment or supplies.
Spring 2004


Mike

is a great

example

of what it

means

to be an

educator.

He is

imaginative,

creative,

and

prolific;

most

importantly,

he is

sincerely

interested

in the

advancement

of the

field.









Even more impor- -
tant, Bob was by nature
a very generous person
and was a natural men-
tor for Mike. When I in-
terviewed at Cornell
three years later, I re-
member visiting with
Bob and Mike in Bob's
office. Mike remained
rather quiet during my
visit, and his boyish
looks (he is a few years
younger than I) led me
to the mistaken conclu-
sion that he was one of
Bob's students-Julian -
Smith corrected my er- Research Group The Shuler re
roneous impression in the Shuler home. Each y
later in the day. for his re
When Mike arrived
at Cornell, Jim Stevenson and John Anderson were young
assistant professors, and their presence and active support
helped make Cornell an attractive and stimulating place. Mike
was also happy to have the able administrative assistance of
Bonnie Sisco right from the beginning.

LEADERSHIP IN
BIOCHEMICAL ENGINEERING
Mike has been a leader in biochemical engineering since
joining Cornell thirty years ago and he continues in that role
today. His work has had an extraordinary influence on aca-
demic research, on teaching and curriculum development, and
on the industrial development of biotechnology.
Early on, Mike undertook the ambitious and challenging
work of building comprehensive mathematical models for
the kinetics of metabolic and synthetic pathways in single
cells. He and his students integrated a large amount of infor-
mation from microbiology and cell biology into a theoretical
and computational framework to predict cell growth and how
the formation of products depends on bioreactor variables.
His paper with Mike Domach on single-cell mathematical
models for the bacterium, Escherichia coli, published in 1984,
was recognized as one of the twenty most influential papers
published in Biotechnology and Bioengineering in the 40th
anniversary issue of that journal. Over the years, Mike has
been a leader in developing quantitative, numerical models
of cellular functions and applying those models to the design
and scale-up of biochemical processes. The computer mod-
els have been used successfully in identifying conditions that
enhance production of desirable compounds while minimizing
formation of undesirable byproducts. Gene Network Sciences
(Ithaca, NY) uses Cornell's single-cell model as the base for


sear
ear
sear


bacterial models. The
company is developing
these models to find and
evaluate new antimicro-
bial drugs.
The area of plant cell
S culture for the produc-
tion of bio-chemicals
was pioneered in the U.
S. by Mike. His group
was the first to explic-
itly demonstrate that
the response of cell cul-
tures containing cell
aggregates depends on
the interplay between
mass transfer and
ch group at a 2003 holiday party chemical reactions in
Karen and Mike host a party the culture. By combin-
*ch group. ing immobilization,
medium optimization,
and in situ product extraction, the group was able to obtain
significant extracellular production of ajmalicine, a compound
previously thought to be constrained to the intracellular com-
partment. This work opened the door for others to produce a
variety of compounds.
The best example of how Mike strives to convert new dis-
coveries into large-scale production of compounds that will
benefit people and society is the story of Phyton, Inc., a com-
pany founded by two of Mike's former PhD students, Bobby
Bringi and Chris Prince. Mike headed an industrial and aca-
demic collaboration aimed at understanding the factors that
would allow the effective production of the important anti-
cancer compound Taxol using plant cell culture techniques.
Phyton, Inc., was one of two industrial collaborators. At Phy-
ton, Bobby and Chris applied the principles developed in their
theses and described in a joint monograph with Mike, Plant
Cell Tissue Culture in Liquid Systems,"l to the successful,
large-scale commercial production of Taxol from Taxus sp.
The company has grown in ten years from three to approxi-
mately a hundred employees and boasts the world's largest
dedicated plant-cell-culture facility with bioreactors of up to
75,000 liters in capacity. Mike's ten-year service on Phyton's
Board of Directors has been an important factor in the
company's growth. The company has recently been acquired
by DFP Pharmaceuticals, Inc.
A major issue in bioprocess engineering has been the pro-
duction of proteins, particularly therapeutic proteins, using
recombinant DNA technology. Mike's group has worked ex-
tensively on a variety of host cells to make various proteins:
bacterial, yeast, insect, plant, and mammalian. Mike's breadth
of experience with this wide-range of host cells is unique.
This work not only has increased our fundamental understand-


Chemical Engineering Education









ing of protein expression, but it has also led to new approaches
that enhance productivity and product quality. Working with
Al Wood and Bob Granados of Boyce Thompson Institute,
Mike developed an insect cell line that is widely used and
sold commercially by Invitrogen as the "High Five" cell line.
Working with Dave Wilson, Mike demonstrated the possibil-
ity of producing large amounts of a desired protein that could
unexpectedly be released by the commonly used bacterium,
Escherichia coli, into the extracellular medium where recov-


ery or purification is relatively inexpensive.
In another area of innovative research, Mike
has proposed cell culture analog (or CCA)
models that combine detailed cellular/molecu-
lar models with traditional physiology-based
pharmacokinetic models (PBPK) to guide the
construction of a physical analog of the PBPK.
The physical analogs can act as human or ani-
mal surrogates for estimating the response of
an organism to challenges by pharmaceuticals
or potentially toxic chemicals. Mike's group
is currently constructing microscale devices
with four compartments: "liver-lung-fat-other
tissue" and other microscale models of the
blood-brain-barrier and of the gastrointestinal
tract. The expectation is that these devices will
facilitate evaluation of new drugs prior to ani-
mal studies and/or human clinical trials.
HmRELTM, a company cofounded by Greg
Baxter (who collaborated with Mike's group),
has licensed patents developed in Mike's group
to commercialize these concepts.


student days. He was chair of the Food Pharmaceutical and
Bioengineering Division in 1994. He has been a member of
the publications committee for about 15 years and was found-
ing editor for Biotechnology Progress. In addition to serving
on numerous AIChE committees and task forces, Mike was a
member of the Board of Directors for the last three years
(2000-2003). As many readers are aware, the last three years
have been a tumultuous period for AIChE, making board
membership a very intense experience. The evolution of
AIChE in the last three yeas has been remarkable, and Mike


His work
[in biochemical
engineering]
has had an
extraordinary
influence on
academic
research, on
teaching and
curriculum
development,
and on the
industrial
development of
biotechnology.


Mike's research work has been well recognized by nu-
merous awards: the Marvin J. Johnson Award of the Micro-
bial and Biochemical Technology Division of the ACS in
1986; the AIChE Food, Pharmaceutical, & Bioengineering
Division Award in 1989; the AIChE Professional Progress
Award in 1991; the Amgen Award in Biochemical Engineer-
ing in 1997; and the Warren K. Lewis Award in 2003. Be-
sides being elected fellow of the AIChE in 1997, Mike has
been honored by election to the National Academy of Engi-
neering in 1989 and to the American Academy of Arts and
Sciences in 1996.
Mike has had a lasting educational influence in the field
of biochemical engineering. His textbook, Bioprocess Engi-
neering: Basic Concepts (with F. Kargi, a former student),
has been extremely successful, by any measure, and is now
the textbook-of-choice for a course in biochemical or bio-
process engineering. The second edition of the book was pub-
lished in 2002.[2] Mike is also very keen on integrating re-
search into undergraduate courses or laboratories'341 and
strongly encourages his graduate students to work with un-
dergraduate students on their research.
Mike has been actively involved with the AIChE since his


Spring 2004


hopes that his work as part of a large team
effort will serve the profession well.
Mike is currently serving on a Formation
Committee for a new "Society for Biological
Engineering" (announced at the November,
2003, AIChE meeting). It is being formed
through AIChE but will seek partner organi-
zations. The expectation is that AIChE's lead-
ership in establishing SBE will result in a
clearer understanding of chemical engineer-
ing contributions to biotechnology and bioengi-
neering and better link the profession to new
opportunities at the interface of biology and en-
gineering.

ROLE IN THE DEPARTMENT
AND THE COLLEGE
To be a successful teacher, one should have
an excellent command of the subject matter,
an ability to inspire students to high achieve-


ment, and generosity with one's time-Mike possesses all
these qualities. His success as a teacher has been recognized
by several teaching and advising awards at Cornell. He was
awarded the prestigious College of Engineering's Tau Beta
Pi Award early in his career, an award given to the best teacher
in the college (of 200 faculty members). More recently (1996),
he received the Richard F. Tucker award for outstanding teach-
ing from the college, and this past year he received the James
M. and Marsha D. McCormick award from the college in
recognition of his sustained contributions to freshmen aca-
demic advising.
Mike has also been an extraordinarily effective mentor of
students at the graduate education level. According to many
of his former students (see Table 1), most of whom have gone
on to successful careers in academia and in industry, he has
always been approachable, kind, and encouraging. Under his
guidance, students are given the freedom to take their research
in new directions and to act as research mentors to under-
graduate students-both of these opportunities have prepared
many for academic positions.
Mike served as director of the School of Chemical and
Biomolecular Engineering from 1998 to 2002. He demon-
strated excellent leadership throughout his tenure, as reflected
85









by curriculum development and resolution of dif-
ficult personnel issues during that time. He was
successful in encouraging faculty to develop fo-
cal areas of research and instruction, and he con-
stantly strived to keep the morale of the faculty,
staff, and students high.

Mike was also successful in the hiring, devel-
opment, and retention of faculty during his ten-
ure as director. He played a valuable role in sup-
porting young faculty (Kelvin Lee and Fernando
Escobedo) and in hiring new faculty (Lynden Ar-
cher, Yong Joo, David Putnam, Abraham
Stroock, and Matthew DeLisa). Besides serving
as acting director for one year in 1986 and for
one semester in 1993, Mike repeatedly resisted
serving as director of the school on a more con-
tinuous basis before 1998. Having served as di-
rector of the school from 1990 to 1993, I under-
stood his earlier reluctance, but applauded his
eventual decision to accept the position when
he considered that the time was right to take on
the challenge. He subsequently earned the ad-
miration of his colleagues and students for his
ability to successfully juggle his large research
group, the administration of the department, and
his family obligations.

In 1993, the Dean of Engineering asked Mike
to head a college-wide committee to develop a
plan for intensifying the college's investment in
bioengineering. The charge and vision have
evolved over the last decade, but Mike has al-
ways played a leadership role. He led the devel-
opment of a new bioengineering curriculum-a
long-standing college goal. At the undergradu-
ate level, the new curriculum comprises several
series of courses that have been tailored to ac-
commodate students from any engineering ma-
jor. Students who complete the bioengineering
option receive recognition on their transcripts.
Early on, Mike recognized the importance of
advising, and he developed a system where each
student participating in this option is assigned a
bioengineering faculty who oversees that part of
the student's curriculum. Thus, the students not
only receive formal recognition for completing
the option but they also obtain sound advice on
planning a career in bioengineering. Mike was
also instrumental in designing a new Bioengi-
neering Seminar for juniors and seniors that gives
students from different engineering disciplines a
venue for interaction. The seminar provides stu-
dents from different engineering disciplines with
a venue for interactions and illustrates the inter-


disciplinary nature of the bioengineering field.
The college and university have asked Mike
to play important roles in leading university-wide
efforts in developing bioengineering and the role
of engineering in the university's "New Life Sci-
ences Initiative." Currently Mike is director of
the Biomedical Engineering Program (BMEP).
The BMEP is responsible for an undergraduate
minor in BME available to all engineering stu-
dents. The BMEP is also the home for an inter-
disciplinary graduate field in BME. This field
draws upon approximately 30 faculty members
from 12 different departments including seven
faculty from Cornell's Weill Medical College in
New York City. Four of the Chemical and
Biomolecular Engineering faculty currently be-
long to the BME field and two new faculty are in
the process of election-with six faculty as mem-
bers, Chemical and Biomolecular Engineering
will have the greatest concentration on BME.
Such a complicated program is an administrative
challenge, but one that is being well met by Mike
and his colleagues.

FAMILY AND AVOCATION
While at the University of Minnesota, Mike met
Karen Beck, a physical therapy student from
Springfield, Minnesota. They married in June of
1972. Mike considers this decision to be the best
and most important one he has ever made. Dur-
ing the final year of Mike's PhD thesis, Karen
worked at a local hospital, and after defending
his PhD thesis in September of 1973, the two of
them headed to Aberdeen Proving Grounds in
Maryland. There, Mike spent three months in the
Army at the Ordinance Officers Basic School
(where he graduated at the top of his class).
Mike and Karen then migrated to Ithaca, arriv-
ing on January 17, 1974. In those days, wives
were not invited to visit during the recruitment
process, so Karen had not had the opportunity to
see Ithaca before Mike accepted the Cornell of-
fer. They both fell in love with the hills, lakes,
woods, and the small city of Ithaca, however.
While that area of the country is often consid-
ered extremely cold, Karen and Mike had left
Minnesota when it was 25 below-so when they
arrived in Ithaca and heard a radio announcer say
"12 degrees Fahrenheit and bitterly cold," they
knew they would be right at home with the
weather there.
Mike and Karen have four children (Andy,
Kristin, Eric, and Kathy) and one very new grand-


Chemical Engineering Education




























Mike (with guide, Mike Williams) and a 45-inch
muskie caught on the Kawartha Lakes in
Canada. The fish was released. V



child (Jeremy). Andy graduated from Cornell
in Mechanical and Aerospace Engineering and
is currently a PhD student in Aerospace Engi-
neering at the University of Texas, Austin.
Andy married a fellow PhD student (Harrey
Jeong) from Korea in 2001 and Jeremy was
born last December. Having a grandchild is a
newfound source of great joy for both Mike
and Karen.
Kristin, who has Down syndrome and lives
at home with Mike and Karen, works at Cornell
in two part-time office jobs. Both Mike and
Karen have been active in various community
groups supporting developmentally delayed
persons.
Eric, who graduated from Cornell with a BA in Religious
Studies and History has begun PhD studies in Medieval Stud-
ies at the University of Notre Dame. He completed a year of
service with Francis Corps between his BS studies and PhD
program. Mike is pleased to have one of his children decide
to attend his alma mater.
The youngest child, Kathy, was adopted from Korea.
Kathy arrived in Ithaca when she was only 4 months old.
Mike and his family are strongly connected to Korea, both
through his family members and through the Korean alumni
of the Shuler group. Kathy, a senior in high school, has de-
cided to attend the University of Dayton, where she intends
to study psychology with a minor in biology and to focus on
animal behavior (especially marine mammals). She is active
in dance and the retreat program of the Diocese of Rochester
and expects to continue similar activities in college.
Mike's family has always been a source of strength and
comfort to him. As is true with any faculty member, his big-


4 Mike and Karen relax in front of a quilt
made by Karen.


The Shuler family in Arizona's White Mountains in
2001. Left to right: (first row): Harrey (Andy's wife),
Kathy, and Kristin; (second row): Eric, Mike, Louis
(Mike's Dad), Andy, and Karen. V


gest challenge has been to balance professional
and family activities through the years, and
since family and professional obligations are
dynamic, the balance has had to be continu-
ously reconfigured.
Mike loves the outdoors in general and fish-
i. ing in particular. He has successfully passed
this interest on to his children. While he en-
joys all types of fishing, he particularly likes
the challenge of muskie fishing, with fly fish-
ing for trout or smallmouth bass running a
close second. Mike got his biggest muskie in
Canada last year while fishing with his son Eric
and a guide-the fish was just shy of 45 inches.
Mike's advice is that everyone should have a second fa-
vorite job as a "fall back" position in case they get discour-
aged with what they are doing. For Mike, his fantasy job is
to be a guide for people who enjoy fishing as a hobby.
For now, however, teaching at Cornell is still his "dream"
job come true.
References

1. Payne, G.F., V. Bringi, C.L. Prince, and M.L. Shuler, Plant Cell and
Tissue Culture in Liquid Systems, Hanser Publ., New York, NY (1992)
2. Shuler, M.L., and F Kargi, Bioprocess Engineering: Basic Concepts
(2nd ed.), Prentice Hall, Englewood Cliffs, NJ (2002)
3. Shuler, M.L., N. Mufti, M. Donaldson, and R. Taticek., "A Bioreactor
Experiment for the Senior Laboratory" Chem. Eng. Ed., 28(1), 24
(1994)
4. Stanlake, G.J., and M.L. Shuler, "Classroom Adaptation of Escheri-
chia coli Single Cell Model, in Computer and Information Science
Applications in Bioprocess Engineering, A.R. Moreira and K.K.
Wallace (eds), Kluwer Academic Publishers, The Netherlands, pp. 395-
400 (1996) C


Spring 2004









S"1department


ChE at...



Rice University


KYRIACOS ZYGOURAKIS
Rice University Houston, TX 77251-1892
When Rice Institute opened its doors to its first class
of students on September 23, 1912, a bachelor's
in chemical engineering was among the degrees
offered, and in 1916, three chemical engineering graduates
(James L. Bramlette, William M. Standish, and Herbert W.
Wilber) received their bachelor's diplomas from President
Lovett. They were the first in a long line of Rice chemical
engineering graduates who went on to have successful pro-
fessional careers.
The arrival on campus of Arthur J. Hartsook in 1921 marked
the true beginning of an independent program in chemical
engineering at Rice. Hartsook was educated at MIT and his
choice of Rice and Houston was not an accident. At MIT he
had gathered information that convinced him that the city
and its vicinity would, in time, grow into one of the foremost
chemical engineering centers in the country and the world.
By the 1940s, Houston had become the petrochemical and
energy capital of the world, and Rice graduates began to fill
many of the positions created by the growing industrial es-
tablishment along the Houston ship channel. For many of
these graduates, Houston's ship channel became the launch-
ing pad for important leadership positions, first in corporate
headquarters in New York or other major U.S. industrial cen-


ters and, subsequently, to key posts all over the world.
The 1950s and 1960s were years of almost exponential
growth, and in 1955, the department became a major player
in the emerging field of nuclear energy, securing a grant from
the Atomic Energy Commission to develop a program in
nuclear engineering, supported by a radiation laboratory and
a 10-watt operating reactor. Chemical engineering also "an-
nexed" a sanitation laboratory program in civil engineer-
ing and began to address the broader problems of envi-
ronmental pollution.
In the 1960s, the department was ranked among the top
seven in the country. Capitalizing on Rice's location, a strong
thermodynamics group led by Tom Leland and Riki
Kobayashi developed new theories and an extensive data-
base of thermophysical properties for the petrochemical in-
dustry. Chemical engineering purchased the first digital
computer ever installed at Rice (an LGP-30) and later
acquired Rice's first solid-state programmable computer
(an IBM 1620) that could be programmed with a (then)
new language called FORTRAN.
Perhaps the most significant example of a bold move into
new areas, however, was an artificial heart project and estab-


Copyright ChE Division of ASEE 2004


Chemical Engineering Education









lishment of a Biomedical Engineering Laboratory in the mid
'60s. In partnership with Baylor's Department of Surgery,
Rice chemical engineers (led by Bill Akers) were instrumen-
tal in developing the first left ventricular heart bypass device
and carried out a large number of pioneering studies that so-
lidified Rice's national reputation in the field of biomedical
engineering. Among their early successes was development
of an implantable artificial lens for the eye that restored
sight to hundreds of patients. During the late 70s and 80s,
however, the Biomedical Engineering lab shifted its re-
search efforts away from medical devices to become one
of the strongest centers of applied cellular engineering
research.
The 80s and early 90s witnessed continued growth. Many
new faculty joined the department to strengthen the areas of
thermodynamics (Chapman, Robert), interfacial phenomena
and petroleum engineering (Miller, Hirasaki), reaction engi-
neering and applied math (Zygourakis), process control
(Badgwell), biochemical engineering (Papoutsakis, San,
Shanks) and biomedical engineering (Glacken, Mikos). New
laboratories were built in 1982 to house the expanding re-
search programs in bioengineering, catalysis, reaction en-
gineering, and thermodynamics. In 1992, the chemical en-
gineering faculty forming the core of Rice's bioengineer-
ing group (Hellums, McIntire, Mikos, San, Shanks, and
Zygourakis) moved their offices and laboratories to a new
research facility (George R. Brown Hall), especially built
to promote interdisciplinary research between biologists
and engineers.
Like the chemical engineering profession, the department


reached a major crossroad as it prepared to enter the 21st cen-
tury. Faculty retirements and the move in 1997 of four faculty
members to a new bioengineering department forced the de-
partment to redefine its research and educational missions.

CHARTING NEW DIRECTIONS:
FROM MOLECULES TO SYSTEMS
The mission of the department is shaped by the same forces
that are redefining the chemical engineering profession. On
one hand, revolutionary advances in nanoscale science and
molecular biology open exciting new avenues for develop-
ing new materials, biological products, and medical thera-
peutics. At the same time, economic and social forces are
driving a transition toward more sustainable and environmen-
tally friendly production methods. Chemical engineers are
uniquely qualified to play leading roles in these revolutions.
For more than a century, we have been very successful in
developing and refining the tools necessary for translating
molecular-level discoveries into new and cost-effective prod-
ucts. To meet the challenges of the new century, however, we
must integrate molecular biology and nanoscale science into
the scientific foundation of our discipline. Such an expanded
knowledge base will enable us to engineer new products by
scaling up processes from the molecular to the system level.
These challenges and needs form the foundation of the
strategic plan formulated in 1998 to guide the future growth
of the department. The plan provided a blueprint for research
directions, faculty and graduate student recruitment, curricu-
lum development, and facility renovation. At its heart was a
dual research and education mission that called for


Spring 2004


Capsule History of the Department

1912 Rice Institute opens the doors to its first class of students. f .
1916 The first three chemical engineering degrees are awarded.
1921 A.J. Hartsook is hired as instructor of industrial chemistry.
1927 Hartsook is promoted to Assistant Professor of Chemical Engineering and assumes the leadership of chemical
engineering, a post he holds until 1956.
1928 Anna Rebecca Lay becomes Rice's first woman graduate in chemical engineering.
1938 Chemical engineering ceases to be a part of the department of chemistry and becomes one of the four Rice
engineering departments.
1941 Rice Chemical Engineering becomes the first chemical engineering department in Texas accredited by ECPD (the
precursor of ABET).
1941 The first MS degrees in chemical engineering are granted to Sam Bethea (thesis title, "Studies on Decolorizing
Clays") and Ervon Eggimann (thesis title, "Performance of an Adiabatic Fractionating Column").
1947 The Chemical Engineering Department receives a mandate from Rice President Houston to start a full-scale graduate program.
1955 Orrin K. Crosser is awarded the first PhD in chemical engineering (thesis title, "Condensing Heat Transfer Within Horizontal Tubes").
1965 The Biomedical Engineering Laboratory is established with the help of a large federal grant. Under the leadership of Bill Akers, this
laboratory develops the first successful left ventricular heart bypass in cooperation with the Department of Surgery at Baylor; Bill Akers
holds a left ventricular bypass device that was implanted in 10 patients in the late 60s.
1967 Fritz Horn and other chemical engineering faculty are instrumental in the formation of the Mathematical Sciences (applied math) department
and Horn becomes acting chair of the new department.
1968 The Department of Environmental Science & Engineering is created, with chemical engineering faculty at its core.
1997 The Department of Bioengineering is created; our chemical engineering faculty transfer their primary appointments to the new department
and Larry McIntire becomes its first chair.
1998 A strategic plan for chemical engineering is approved by the University and its implementation begins with the help of an Advisory and
Development Board.









Conducting world-class research in the areas of advanced
materials and complex fluids, biosystems engineering, and
energy and environmental systems
Educating outstanding undergraduate and graduate
chemical engineers to rise to leadership roles in academia,
industry, law, business, medicine and government
1 Promoting interdisciplinary collaborations and forming
bridges linking Rice innovations to applications in the
chemical, energy, biotechnology and materials industries
The strategic plan also calls for complementing the
department's internal research thrusts with strategic alliances
on pollution control with faculty from the Civil and Environ-
mental Engineering Department, and on biomaterials for tis-
sue engineering with faculty from the Bioengineering De-
partment. Because both these departments grew from pro-
grams initiated within chemical engineering, our faculty mem-
bers have many strong connections through research collabo-
rations, joint participation in university-wide research cen-
ters and joint faculty appointments. Our new undergraduate
curriculum, a very flexible schema that encourages students
to build interdisciplinary skills, also calls on these depart-
ments as teaching resources, thus solidifying these alliances.
Implementation of the strategic plan began in 1998 with
the help of a select Advisory and Development Board whose
members include academic, industrial, and professional ex-
perts. Over the past five years
0 Four new faculty members were hired in the areas of
materials and biosystems.
1 Over 14,000 sq. ft. of laboratory and office space was
renovated and now houses state-of-the-art facilities for
research on complex fluids, catalysis, and nanomaterials.
A fundraising campaign for endowed graduate student
fellowships has raised more than half a million dollars.
The undergraduate curriculum was restructured by introduc-
ingfourfocus or specialization areas, by integrating courses
or updating their content, and by introducing a biology (or
biotechnology) requirement for our BS degree.
The focus areas in biotechnology, environmental engineer-
ing, computational engineering, and materials science are very
popular among our undergraduates.

FACULTY RESEARCH
Hiring new faculty has been a top priority and four new
colleagues joined our department after 2000 to strengthen
the materials and biosystems areas.
Matteo Pasquali joined the department as an assistant pro-
fessor in January 2000 after receiving his PhD from the Uni-
versity of Minnesota in 1999 and completing his postdoc-
toral studies at the same University. His research focuses on
processing flows of microstructured liquids that are ubiqui-
tous in the chemical, polymer processing, coating, food, and
biomedical industries. Current projects include the computa-
tional modeling of process flows with mesoscopic rheologi-
cal properties, the solution of microscopic transport equa-


tions, and the visualization of single DNA molecules in pro-
cess flows. He recently led a team of researchers who dis-
covered that a sulfuric acid-based superacid makes an excel-
lent medium for dispersing single-walled carbon nanotubes
(SWNTs) at concentrations that are useful for industrial pro-
cesses. This enabled them to process the dispersion into the
first continuous fibers of aligned, pristine SWNTs.'1-3] Fibers
like these might be used to make ultralight, and yet ultrastrong,
materials with remarkable electronic, thermal, and mechani-
cal properties.
Michael Wong joined our faculty as an assistant profes-
sor in July 2001 after receiving his PhD from MIT in 2000
and completing a year of post-doctoral studies at UC Santa
Barbara. Mike's research program focuses on designing new
and improved materials for catalytic and encapsulation ap-
plications. His approach is based on the concept that the mac-
roscopic properties of materials can be manipulated and en-
gineered for target applications if their structural features can
be controlled at the nanometer scale. Current efforts focus on
the development of (a) supported nanoparticle-based metal
oxide catalysts for solid acid and oxidation reactions, (b)
metal-supported nanoparticles for the breakdown of environ-
mentally-unfriendly organic compounds, (c) "quantum dots"
for photocatalysis, and (d) nanoparticle-based hollow
microspheres and microshells that can be engineered to en-
capsulate either enzymes to form micro-bioreactors or drug
molecules to form drug delivery devices.
Nikos Mantzaris also joined our department as an assis-
tant professor in July 2001. He received his PhD from the
University of Minnesota in 2000 and completed his two-year
post-doctoral studies on mathematical biology at the same
university. His research lies in the emerging area of biosystems
engineering and it aims at understanding, optimizing, and con-
trolling the behavior of biological systems with the use of
mathematical modeling and dynamical studies. Specific sys-
tems of interest include recombinant E.coli cell populations,
tumor-induced migration of endothelial angiogenetic cells and
astrocytic signal transduction systems in the mammalian
brain. The fundamental question Nikos is trying to answer
for each of these systems is how single-cell events lead to the
complex behavior and patterns exhibited by cell populations
exhibiting heterogeneous phenotypes. The common thread
in these studies is a methodological framework used to ac-
complish these tasks. This framework includes the develop-
ment of (a) simplified models that can capture the essential,
experimentally observed features of the system under con-
sideration, and (b) sophisticated two- and three-dimensional
numerical algorithms that, in combination with the under-
standing gained from the simplified model results, serve as
the basis for studying the asymptotic and transient behavior
of the detailed model.
Paul Laibinis joined our department as an associate pro-
fessor in January 2003. He received his PhD in organic chem-


Chemical Engineering Education









istry from Harvard University in 1991 and was a member of
the MIT faculty from 1993 to 2002. His laboratory employs
methods of self-assembly and chemical modification for gen-
erating interfaces with enhanced properties. These efforts in-
clude generation of molecular and polymeric thin films that
are tailored through chemical synthesis to afford specific
molecular architectures on surfaces as needed to alter and
control surface events. The ap-
proach is general and reveals an
ability to manipulate macroscopic
interfacial events through
nanoscopic changes at surfaces.
Current research projects include
approaches that control the inter-
actions of biological species with
surfaces, both to avoid their ad-
sorption and to direct the adsorp-
tion of specific agents from solu-
tion. Such properties are important
in the area of biodiagnostics,
where selective recognition of tar-
gets in conjunction with the avoid-
ance of nonspecific adsorption Jessica Dunn, an und
events determine the performance cell migr
of many sensor and microarray
technologies.
The research efforts of the new faculty nicely complement
the work of the senior faculty in the areas of materials, en-
ergy systems and biosystems.
Chapman uses tools such as molecular simulation, com-
puter visualization, statistical mechanics, and NMR to dis-
cover how material properties and structure depend on mo-
lecular forces. His current research program focuses on poly-
mer solutions and blends, associating fluids, confined fluids,
and natural gas hydrates.
Hirasaki conducts research in fluid transport through po-
rous media ranging from the microscopic scale intermolecu-
lar forces governing wettability to the megascopic scale nu-
merical reservoir simulators for field-wide modeling. A re-
curring theme throughout this research is the dominance of
interfaces in the determination of fluid transport processes.
Miller's research focuses on interfacial phenomena, es-
pecially those involving surfactants and their applications in
detergency, pharmaceutical and food products, petroleum
production, ground water cleanup, agricultural chemicals, and
personal care products. His current projects include studies
on the dissolution rates of surfactants, foam flow in porous
media, and transport in emulsions.
Robert conducts theoretical, experimental, and com-
puter simulation studies of the properties of matter. His
current efforts focus on magnetic nanoclusters and car-
bon nanotubes, polymer, colloidal systems, ferroelectrics,


ergr
'atio;


and disordered systems.
Zygourakis studies the mechanisms through which cell
migration and proliferation affect the growth of three-dimen-
sional tissues. Experimental data and large-scale simulations
are used to analyze the dynamic behavior of large cell popu-
lations proliferating on 3-D scaffolds and find how tissue
growth rates are modulated by the culture conditions.

INTERDISCIPLINARY
RESEARCH
Over the past two decades, Rice
has established several institutes
and centers to promote interdis-
ciplinary research on nano-tech-
nology, biological sciences and
engineering, information technol-
ogy, and environmental engineer-
ing. Almost all our faculty mem-
bers play key roles in these re-
search efforts.
s Center for Biological and
aduate, works on her Environmental Nano-technol-
raduate, works on her
n project. ogy (CBEN): Established in 2001
as one of six Nanoscale Science
and Engineering Centers funded
by the National Science Foundation, CBEN is the first to fo-
cus on applications of nanotechnology to human health and
the environment. The Center's research activities explore
the wet/dry interface between nanomaterials and aqueous sys-
tems at multiple length scales, including interactions with
solvents, biomol-ecules, cells, whole-organisms, and the
environment. Collaborations with industry, entrepreneurs,
and the Jones Graduate School of Management are inte-
gral to the Center's mission of creating sustainable
nanotechnology.
Many chemical engineering faculty members have formed
strong collaborations with other CBEN researchers. In col-
laboration with Richard Smalley and other chemistry research-
ers, Pasquali is studying the theological properties of sus-
pensions of single walled carbon nanotubes and is working
on the production of fibers from concentrated, strong-acid
solutions of nanotubes. Wong and his collaborators are de-
veloping new and improved materials for catalytic and en-
capsulation applications. Their approach is based upon the
concept that the macroscopic properties of materials may be
manipulated and engineered for target applications if their
structural features can be controlled at the nanometer scale.
Laibinis is exploring methods of self-assembly and chemi-
cal modification for generating interfaces with enhanced prop-
erties. Finally, Mantzaris, Zygourakis, Pasquali, and Wong
are developing transient population balance models and ap-
ply them to design reactors and develop optimal control poli-
cies for the large-scale production of high-quality


Spring 2004









nanoparticles (quantum dots).
E Institute of Biosciences and Bioengineering (IBB): Es-
tablished in 1986 in recognition of the revolutionary advances
in biotechnology and molecular biology, IBB promotes in-
terdisciplinary interactions among scientists, engineers, and
clinicians from Rice and neighboring institutions.
Mantzaris and Zygourakis are collaborating with fac-
ulty from biochemistry and cell biology, bioengineering, and
chemistry, to develop a novel framework that combines ex-
perimental, theoretical, and computational tools to study het-
erogeneous cell populations as complex, and highly inter-
connected systems with interacting components. This sys-
tem-based approach will change the design principles used
to develop effective drugs, tissues with desirable structure,
materials with novel properties, and other bio-based, envi-
ronmentally friendly, and sustainable technologies.
Zygourakis also collaborates with other IBB members on
research focusing on tissue engineering and biomaterials. A
Biotechnology Training Grant from NIH and an IGERT grant
from NSF provide stipend support for graduate students work-
ing in these and related areas.
* Shell Center for Sustainability: Chemical engineering
played a key role in the creation of the Shell Center for
Sustainability. This Center embraces the central theme of
Rice's environmental initiative: that implementation of a strat-
egy for sustainable development requires both new tools and
a transformation in our understanding of society's needs.
Administered through Rice's Environmental and Energy
Systems Institute, the Center has a three-pronged mission of
education, research, and community service to (a) create the
knowledge required to remove current technological barriers
to sustainability and to enable development of novel "sus-
tainable" processes and products, (b) promote an interdisci-
plinary approach to sustainability that integrates research,
education & public policy, and (c) serve as an independent
forum for open discussions on sustainable development is-
sues and policies.
One of the first projects funded by the Center involves
research on gas hydrates that offer a vast, untapped energy
source and may have played crucial roles in past global warm-
ing events. Chapman and Hirasaki are collaborating with
geoscientists to develop mechanistic models describing the
accumulation and dissociation of gas hydrates that exist in
deep ocean sentiments or in Arctic permafrost. Also, Hirasaki
and Miller are working with faculty members from civil/
environmental engineering to develop and test new methods
for bioremediation of aquifers.

UNDERGRADUATE STUDIES

In today's rapidly changing business climate, industrial
sectors from petrochemicals to biotechnology and semicon-
ductor manufacturing offer a multitude of employment op-


portunities for our graduates.
What opens all these career options to our graduates is a
broad education that encompasses both fundamentals and
applications to give students a sound scientific and technical
grounding for further development in a variety of professional
environments.
Over the past five years, our department has awarded 146
BS (ABET accredited) and 18 BA degrees, for an average of
about 33 graduates per year. After a significant decrease be-
tween 2001-03, our enrollments are increasing again and
our current sophomore class has about 30 students. The
percentage of female students in our classes ranges be-
tween 40% and 50%.
Industry employs the majority (about 61%) of our students
who graduate with a bachelor's degree. Of the remaining stu-
dents, about 18% continue their education in graduate schools
to prepare for academic careers and industrial research jobs,
12% attend medical or law school, and 9% take govern-
ment or other jobs.
Courses in mathematics, chemistry, physics, and compu-
tational engineering provide a foundation for the chemical
engineering core, which introduces students to chemical pro-
cess fundamentals, fluid mechanics, heat and mass transfer,
thermodynamics, kinetics, reactor design, process control, and
process design. Chemical engineering curricula place an em-
phasis on chemistry not found in other engineering disciplines.
This background allows chemical engineers to tackle the wide
variety of technical problems arising in the chemical, elec-
tronic, pharmaceutical, and biotechnology industries.
To complete their technical education, students seeking a
BS degree in chemical engineering take course electives in
at least two other engineering disciplines to satisfy a "breadth"
requirement. Or, they can use their electives to create a focus
or specialization area in 1) biotechnology and bioengineer-
ing, 2) computational engineering, 3) environmental science
and engineering, and 4) materials science and engineering.
Figure 1 shows the components of the chemical engineer-
ing BS degree with their credit requirements. A total of 132
credit hours are required for the BS degree. The BA program
is more flexible, making it even easier for a student to pursue
a double major. Chemical engineering specifies 77 semester
hours for the BA degree, including prerequisites and labora-
tory courses. In addition to these requirements, students must
also satisfy the University distribution requirements (24 se-
mester hours) and complete no fewer than 31 semester hours
of free electives for a total of at least 132 semester hours.

GRADUATE STUDIES
The department offers programs of graduate study lead-
ing to both the Master of Science and Doctor of Philosophy
degrees, with the primary emphasis on the latter. A profes-
sional master's degree (MChE), involving only course work,


Chemical Engineering Education









is also offered.
Currently, the depart- Free
ment has about fifty gradu- Electives
7-12 credits
ate students. All are en- 7-2
gaged in research activities
and receive full financial
aid (stipend plus tuition).
During the past five years, Humanities
the department has gradu- Social Sciences
ated an average of 7.5 PhDs 24 credits
each year, making it first in
per-faculty production of Focus Area
PhD graduates among the
eight departments of the
Rice School of Engineer- Biotechnology/ Environmental
ing. The 1995 National Re- Bioengineering Engineering
20 credits 18 credits
search Council report 20 crs 18
ranked our department 16th
in the United States in edu- Figure 1. Components
rating chemical engineer-
ing PhD students.
Graduate education is aimed at developing each student's
ability to conduct independent creative scientific research.
To ensure the versatility of our graduates, our curriculum pro-
vides a solid background in chemical engineering fundamen-
tals (applied mathematics, thermodynamics, transport phe-
nomena, kinetics, and reaction engineering). It also provides
a mastery of engineering tools and ability to set clear goals
of professional development so that our graduates become
productive and successful in their professional careers.
The graduate program is large enough to offer research
topics in several important areas of chemical engineering and
related fields, but small enough to retain an atmosphere where
students and faculty can have extensive personal contact. In
addition, interdisciplinary research projects provide ample
opportunities for students to interact with researchers from other
disciplines, especially through the interdisciplinary institutes.

THE UNIVERSITY AND ITS CITY
Rice is a leading research university distinguished by a
collaborative, interdisciplinary culture and a global perspec-
tive. Only a few miles from downtown Houston, it occupies
an architecturally distinctive, 300-acre campus shaded by
nearly 4,000 trees. State-of-the-art facilities and laboratories,
internationally renowned centers and institutes, and one of
the country's largest endowments support an ideal learning
and living environment. With just 1,600 graduate students
and 2,700 undergraduates, it offers an unusual opportu-
nity to form close relationships with its faculty scholars
and researchers and the option to tailor graduate programs
to specific interests.
Houston is home of the Texas Medical Center, the largest
concentration of medical schools, hospitals, and research fa-


of t


Spring 2004


cilities in the world.
iematics Chemistry Rice has cooperative
Physics programs with the Uni-
credits 25 credits versity of Houston,
Baylor College of Medi-
cine, the University of
Texas Health Science
'ore Engineering Center, and Texas
emical Breadth or Southern University.
ineering Focus Area Houston is one of the
credits 15-20 credits
credits 15-20 credits few U.S. cities with resi-
I dent companies in all
s I four major performing
Starts (drama, ballet, op-
IComputa l Engineering era, and symphony). It
rials Computational Breadth
nce Engineering 15 credits also boasts a museum
edits 18 credits district featuring exhib-
its of national and inter-
he BS curriculum in ChE. national prominence. As
urban as it is, Houston is
also a surprisingly green
city. Houstonians enjoy the outdoors in over 300 municipal
parks and 120 open spaces, while the beach at Galveston Is-
land is only a 45-minute drive away.

FORWARD PATH
The Rice chemical engineering department has grown from
its embryonic beginnings as an adjunct to the chemistry de-
partment to a major research and educational center. To a large
extent, this success has been due to the foresight of its faculty
and an ability to exploit opportunities in emerging areas.
Past achievements set a precedent for the departmental
rejuvenation that is occurring today. Faithful to its tradition,
the department is refocusing its research and teaching efforts
in areas that address the evolving needs of our society. These
efforts are already receiving national recognition.'I3] The be-
lief that chemical engineering is an enabling discipline, how-
ever, continues to shape our mission. Like our predecessors,
we continue to use the same quantitative, systems-based ap-
proach in our research, we adapt our curricula to meet the
evolving needs of our students, and we apply the same com-
mitment to excellence.

ACKNOWLEDGMENTS
Our special thanks go to Bill Akers, Professor Emeritus,
and the staff of the Woodson Research Center, Fondren Li-
brary, Rice University.

REFERENCES
1. "Coming: Superthreads from Nanofibers," New York Times, Science
Section, December 9 (2003)
2. "Acid Route to Nanotube Fibers," Chem. & Eng. News, 81(50), 9
(2003)
3. Davis, V.A., et al., "Phase Behavior and Rheology of SWNTs in
Superacids," Macromolecules, 37, 154 (2004) 0











Dow Award Lecture




MEMBRANE


SCIENCE AND TECHNOLOGY

IN THE 21ST CENTURY


WILLIAM B. KRANTZ
University of Cincinnati Cincinnati, OH 45221-0012
he author's career path in membrane science was any-
thing but direct. Indeed, he embarrassingly admits that
he has no formal training in the subject. His interest
in membranes began during his tenure as a Director for NSF's
(former) Thermodynamics and Mass Transfer Program in
1977-78. This program, which provided the only source of
NSF funding for membrane research, had only two grants in
the field at the time.
The author's efforts to enhance funding for membrane sci-
ence ultimately led to establishing NSF's Separations and
Purification Processes Program. In the process of doing this,
the author also convinced himself that membrane science was
a fertile area for challenging research. The actual impetus
for initiating a research program in membrane science, how-
ever, came from one of his students, who upon completing
his MS degree in the author's geophysics research program
in 1981, asked if he could pursue a PhD in membrane sci-
ence. Indeed, we sometimes gain more inspiration from our
students than we give to them!
In 1990 the author and Professor Richard D. Noble estab-
lished the NSF Industry/University Cooperative Research
Center (I/U CRC) for Separations Using Thin Films that sub-
William B. Krantz received a BA in chemistry
(1961) from Saint Joseph's College (Indiana)
and his BS (1962) and PhD (1968) degrees in
chemical engineering from the University of Illi-
nois-Urbana and the University of California-
Berkeley, respectively. From 1968-1999 he was
Professor of Chemical Engineering and
President's Teaching Scholar at the University
of Colorado. In 1999 he accepted the Rieveschl
Ohio Eminent Scholar Chair at the University
of Cincinnati where he is Director of the NSF I/
U CRC for Membrane Applied Science and
Technology. He is the recipient of a Guggenheim, NSF-NATO and three
Fulbright fellowships, Special Achievement and Outstanding Performance
Awards from NSF, the Innovation in Coal Conversion Award of the Interna-
tional Coal Conference, the George Westinghouse and Dow Chemical En-
gineering Division Lectureship Awards of ASEE, and the Malcolm E. Pruitt
Award of the CCR. He has been named a Fellow of AAAS and ASEE and a
Sigma Xi National Research Lecturer. In addition, he has received27 awards
for teaching excellence and service to student organizations.
Copyright ChE Division of ASEE 2004


sequently became the Center for Membrane Applied Science
and Technology (MAST). In 2000 the author accepted the
Rieveschl Ohio Eminent Scholar Chair in Membrane Tech-
nology at the University of Cincinnati where he established a
partner MAST Center site with a particular focus on bio-ap-
plications of membrane science.

CHRONOLOGY OF MEMBRANE SCIENCE
AND TECHNOLOGY

A membrane is a semipermeable medium that permits the
passage of some molecules, colloidal aggregates, or particles
relative to others. The dawn of membrane science is gener-
ally considered to be in the 18"' century when Abbe' Nollet
used the semipermeable properties of animal bladders to study
reverse osmosis. The practical use of semipermeable media
dates back many centuries before this, however. For example,
the Arawak peoples of the West Indies used microporous stone
filters to purify drinking water as early as 1600 BC, and the
use of membrane filters for clarifying alcoholic beverages
probably predates this by a few thousand more years. An ex-
ample of one of these primitive stone membrane filters as
shown in Figure 1.
It is interesting to note that membrane science developed
much faster than membrane technology. For example, dialy-
sis was studied by Graham in 1861 but it was not until 1944
that the first commercial dialyzer was developed by Kolff
and Beck-think of how many lives could have been saved
if the membrane dialyzer (artificial kidney) had been de-
veloped sooner.
Electrodialysis was reported by Pauli in 1924 but did not
become practical until the 1950s. Ultrafiltration was studied
in the early part of the 20th century but was not commercial-
ized until the 1960s. Gas separations were studied by Gra-
ham in 1864 but the development of the membrane lung oxy-
genator by Kolff and Balzer did not occur until the 1950s;
again consider the impact of this delay on the loss of life.
Although the pH meter was developed by Cremer in 1906,


Chemical Engineering Education









the smart sensor industry did not emerge until the 1960s.
Nelson and Rose's pioneering studies of controlled release
technology were reported in 1955, but this was not commer-
cialized until the 1980s. This delay between scientific dis-
covery and commercial implementation is in stark contrast
to the rapid commercialization that we associate with com-
puter- and information-age technologies. Unfortunately this
delay is very much associated with our inability to address
challenges that have confronted membrane science.

DISADVANTAGES AND ADVANTAGES
OF MEMBRANES
The significant time gap between discovery and commer-
cialization of membrane technology is a result of several per-
ceived disadvantages of membrane technology. Since most
membranes are made from solids, the permeation rates will
be low unless they can be made very thin. Membrane pro-
cesses scale linearly; that is, the required membrane area is
directly proportional to the feed flow rate. Hence, membrane
processes do not have the economies-of-scale that are en-
joyed by separations technologies such as distillation, extrac-
tion, and crystallization.
The attractiveness of membrane technology is also tied to
energy costs since they are very energy efficient relative to
separations processes that require a change of phase; fortu-
nately (but unfortunately for membrane technology) we have
lived during a time of abundant and relatively inexpensive
energy. Concentration polarization and fouling have hindered
the application of membranes to many liquid separations ap-
plications. Most membrane separations employ polymeric
membranes that are subject to compaction under high pres-
sure and swelling in the presence of aggressive organic sol-
vents, both of which compromise performance.
Interestingly, membranes have been successful primarily
in niche markets where they do not have any competition;
consider for example contact lenses, renal dialyzers, artifi-
cial lungs, breathable garments, battery separators, etc. Poly-
meric membranes are subject to UV degradation, hydrolysis,
and attack by polar aprotic solvents as well as a relatively
low maximum use temperature; hence they have limited ap-
plications and relatively short lifetimes. Ceramic membranes
are brittle, friable, difficult to package into modules and rela-
tively expensive. Finally, polymeric membrane manufacture
requires relatively little material; hence, it does not have much
influence on the polymer market and the corresponding de-
velopment of new polymers.
Membranes also enjoy several advantages that will lead
to exciting new applications in the 21' century. They are
attractive from a First Law of Thermodynamics perspective
since they in general require less energy than competitive
separation processes owing to the fact that they do not nec-
essarily require an energy-intensive phase change. They are
also attractive from a Second Law of Thermodynamics de-


Spring 2004


Figure 1. The author and a stone membrane filter of the
type used by the Arawak peoples of the West Indies to
purify drinking water as early as 1600 BC.
sign perspective since they are not subject to the entropy pen-
alty associated with a phase change. The linear scaling law also
has positive implications since it implies that membrane tech-
nology is modular and therefore amenable to easy scale-up.
Membrane separations are also easily adapted to both con-
tinuous and batch-wise operation. In contrast to many com-
petitive technologies, harsh or extreme operating conditions
are not generally required for membrane separations. They
can also be used to separate feeds that exhibit an azeotope
that precludes separation via distillation.
Membranes are particularly amenable to hybrid separa-
tions in which they are combined with some other technol-
ogy such as distillation where they can be used to process a
slip stream to break an azeotope. They can also be
functionalized to permit highly discriminating separations
and can be made from renewable resources such as wood
pulp and cotton delinters. Membrane separations are also
considered to be environmentally friendly, green technolo-
gies. One particular advantage of membranes that has been
realized only recently is that they are the only technology
that can separate, deliver, or discriminate between solutes
on the microchip scale or MEMS (Micro-Electro-Mechani-
cal-Systems) scale.

THE NSF CENTER FOR MEMBRANE
APPLIED SCIENCE AND TECHNOLOGY
The challenges and opportunities for membranes led to the
establishment of the Center for Membrane Applied Science
95









and Technology (MAST) (the only NSF block-funded mem-
brane research center in the U.S.) at the University of Colo-
rado in 1990. In 2001 the MAST Center became a Multi-
University Industry/University Cooperative Research Cen-
ter (I/U CRC) by adding a partner site at the University of
Cincinnati. This permitted the MAST Center to expand its
focus to include biomedical, pharmaceutical and biosensor
applications of membranes since the requisite
multidisciplinary research was facilitated by having the Col-
leges of Engineering, Arts and Sciences, Medicine, and Phar-
macy on the same campus at Cincinnati.
The objectives of the MAST Center are (1) to conduct ba-
sic research and related developmental activities for the use
of membrane technology, (2) to provide timely and effective
technology transfer between the Center and its sponsors, and
(3) to promote education in membrane technology. The MAST
site Co-Directors are Professors Richard D. Noble and Alan
R. Greenberg at the University of Colorado and Professors
Jerry Y.S. Lin and the author at the University of Cincinnati.
The primary source of funding for the MAST Center is its
industrial and governmental lab sponsors, each of whom pays
an annual fee. Each sponsor designates one representative on
the Industrial Advisory Board (IAB) of the Center. Twice each
year, the sponsors are invited to submit topics for possible
MAST Center research projects. These topics can involve
any aspect of membrane science and technology, the under-
taking of which would be appropriate to MS or PhD research
in any field of pure and applied science. The list of topics
along with any supporting information the sponsors choose
to provide is sent to faculty spanning a broad spectrum of
disciplines at the two MAST sites. Interested faculty then
submit pre-proposals that are evaluated by the IAB to de-
termine who will be asked to submit expanded proposals.
On the basis of the latter, the IAB selects the proposals
that will be funded.
Projects are generally funded for three years at a level ad-
equate to fund one graduate student and to provide a modest
budget for supplies, equipment, travel, and some faculty sum-
mer support. The return to industry for its investment in the
Center is rapid technology transfer, access to University fac-
ulty, students, and facilities, a forum to discuss membrane
science and technology, and a proven 20:1 leveraging on their
annual sponsor fee. The benefit to faculty and students is rel-
evant research, participation of sponsors on the research
teams, industry internship opportunities, and access to spon-
sor facilities. In order to protect the intellectual property inter-
ests of the Universities and the sponsors, all MAST Center re-
search is held in confidence until the IAB approves disclosure.
Membrane research at Cincinnati and Colorado impacts not
only the graduate programs at each university, but also has
had a significant positive impact on undergraduate programs
via NSF Research Experiences for Undergraduates (REU)
Summer Site Programs that are tied to the MAST Center sites.


Over 175 undergraduates have received financial support and
invaluable research experience through the NSF REU pro-
grams at the two sites.
Recently the Cincinnati MAST Center site was awarded
an NSF Integrative Graduate Research Education and Re-
search Traineeship (IGERT) grant that focuses on bio-appli-
cations of membrane science and technology. This program
includes a set of prescribed courses in membrane science that
provide the impetus for complementary curricular tracks in
biomedical, pharmaceutical, and biosensor applications of
membranes. Program components include a set of laboratory
rotations, training in entrepreneurship, professional ethics,
communication skills, industry internships, and the opportu-
nity for an international technical cultural experience. This
NSF IGERT program complements the MAST Center in that
the membrane research undertaken by the former will be gen-
erated by the faculty, whereas that undertaken by the latter is
suggested by the MAST sponsors.
Over the past 13 years, the MAST Center has funded 55
research projects that in turn have provided stipend support
for 16 MS, 37 PhD, and 6 postdoctoral students. The fact that
77% of the project teams were multi-investigator and 56%
were interdisciplinary demonstrates that the MAST Center
has successfully promoted the research environment neces-
sary to tackle the difficult challenges facing the advancement
of membrane science and technology. This research has led
to 90 peer-reviewed and 253 presented papers as well as 10
patents. Most impressive is the fact that MAST Center stu-
dents have won 27 regional and national awards for their re-
search accomplishments. MAST Center activities have cata-
lyzed more than $16 million in spin-off funding, primarily as
a result of facilitating the formation of faculty research teams.

EXAMPLES OF SUCCESSFUL
NSF MAST CENTER RESEARCH
It is interesting to follow the chronology of a successful
MAST Center project. In 1993, the U.S. Army Tank and Au-
tomotive Research and Development Center (TARDEC)
joined the MAST Center. TARDEC's motivation to join the
Center was based on problems encountered during the 1991
Desert Storm war in Iraq. Indeed, a major factor that limited
our ability to move through the vast desert into Baghdad was
the supply of water to the ground forces. TARDEC was re-
sponsible for mobile water treatment units, each of which
could produce 100,000 gallons per day using reverse os-
mosis membrane technology.
Approximately one-half of the weight of these units was
associated with equipment and chemicals needed to clean the
membranes of fouling deposits. Owing to the buildup of foul-
ing deposits on the membranes, the mobile units had to be
periodically shut down and de-fouled by circulating chemi-
cal cleaning agents through them. Considerable time and
cleaning agent materials were wasted owing to the uncer-


Chemical Engineering Education









tainty in the amount of time required for cleaning the units.
The author participated in a research team that quickly rec-
ognized that this problem could be addressed by adapting
Ultrasonic Time-Domain Reflectometry (UTDR) to monitor
the condition of the membranes. Ultrasound was an estab-
lished medical technology for imaging centimeter-scale in-
ternal organs and fetuses. Adapting it to image micron-scale
fouling deposits was a challenging task for our MAST Cen-
ter research team-one that ultimately led to a U.S. patent."'
TARDEC projects that this UTDR technology will pro-
vide a savings of as much as 50% in cleaning time and chemi-
cal cleaning agent use. Subsequent
publications describe the use of
UTDR to monitor membrane foul-
ing,"261 membrane compaction," 78
and membrane formation.'91 Current
research involves the use of UTDR
to study membrane swelling, defect
detection, and optimal design of
membrane modules. This provides an
excellent example of how MAST
Center projects address both the
immediate goals of its sponsors and
provide incentives for faculty to pur- Figure 2. Andrew Nei
sue new research thrusts that result in and the author righth
scholarly publications and significant Ceremony for the Coli
Sfund. Year. Dr. Hillier was h
external funding. related to the invent
The MAST Center not only has croscope. Andrew, a
provided stipend support for many received the Undergi
graduate and postdoctoral students, ventor of the Year At
but also has provided financial sup- that emanated from
port and research experience for over search focused on m
tion in polymeric me
175 undergraduate students. It has tion in pymeric me
been said that if faculty are reason-
ably confident the objectives of a research project can be
achieved, it should be given to a PhD student since they must
achieve positive results for their thesis; but if the project is
highly speculative, with a low chance of positive results, it is
appropriate for an undergraduate researcher! Indeed, under-
graduates are delightfully naive; they are not experienced
enough to know what is impossible!
One such example is provided by a freshman who came to
the author to undertake an independent study research project
that involved adapting video-microscopy for imaging the evo-
lution of pore structure during membrane formation via phase
inversion. The student happened to be taking physics while
he was working on this project and proposed using nonuni-
form electric fields to eliminate undesirable macrovoid pore
defects that compromise the mechanical integrity of poly-
meric membranes. The author knew that this idea would not
work since he had published a paper that explored this same
idea,101l but he did not want to discourage a young researcher
by telling him about the earlier paper. He provided some funds


to allow the young man to explore his 'ill-conceived' idea.
What subsequently happened is best described by a beautiful
poem of Edgar Guest:
So he buckled right in
with the trace of a grin
on his face. If he worried, he hid it.
He started to sing
as he tackled the thing
that couldn't be done, and HE DID IT!
Edgar A. Guest It Couldn't Be Done
Indeed, as it turned out, the student was right and the pro-
fessor was wrong-wrong in his science, that is, but right in


ce, James Hillier (left),
), at the 1999 Awards
egiate Inventors of the
honored for his research
on of the electron mi-
freshman at the time,
graduate Collegiate In-
vard for his invention
NSF MAST Center re-
icrovoid defect forma-
mbranes.


his mentoring of a young researcher!
This research led to a U.S. patentE" and
the 1999 U.S. Collegiate Inventor of the
Year award for the young man-and just
incidentally, to a new area of research
for the professor. Figure 2 shows this
young man with a 'hint of a grin' re-
ceiving his award!

THE FUTURE OF
MEMBRANE SCIENCE

Now let us consider the future of
membrane science in this 21st century.
There are several important consider-
ations regarding membranes that need
to be kept in mind when considering
where this field of scientific endeavor
is going:
* Membranes are competitive when
the desired product is the permeating
component.
* Membranes thrive in noncompetitive


markets.
Membrane processes obey a linear scaling law.
Membranes are the only separations technology that
will work on the microscale.
Membranes stand to benefit significantly from the
science of biomimetics.
It might seem surprising that reverse osmosis membrane
technology is now used to produce more than 400 million
gallons per day of fresh water from seawater; yet, it is not
being used to provide fuels-grade ethanol from aqueous bio-
mass solutions. The answer is simple, however-any water
that passes through a membrane via reverse osmosis is a use-
ful product. In contrast, to produce a fuels-grade ethanol prod-
uct from aqueous biomass solutions would require passing
nearly all the water through the membrane since the ethanol
molecule is larger than water. Indeed, successful membrane
technologies usually require that the desired product pref-
erentially permeate.
Any technology would like to be the 'only show on the


Spring 2004









road,' but this is especially true for membranes. Consider, for
example, life-saving medical technologies such as the mem-
brane lung oxygenator and membrane dialyzer, pharmaceu-
tical devices such as controlled release patches, and consumer
products such as contact lenses.
Educators who teach process design are quick to point out
the disadvantage of membrane technology in that it obeys a
linear scaling law. As stated earlier, it is for this reason that
membranes do not have the economies-of-scale that other
separations technologies enjoy. It is important to realize, how-
ever, that a technology damned for large-scale applications
might be blessed for small-scale applications! Indeed, mem-
branes provide the only technology that can separate, deliver,
or discriminate between solutes on the micro- or nano-scales.
It is well known that the human body consists primarily of
water-the second pervasive component in our bodies is
membranes. Indeed, our skin, kidneys, spleen, liver, eye cor-
neas, and every cell in our body involve biological mem-
branes. The science of biomimetics seeks to unravel how liv-
ing systems accomplish their exceptional performance. We
are delighted when we can fabricate a synthetic membrane
that will provide a separation factor of 20:1 for some solute.
Imagine the impact of being able to make synthetic mem-
branes that can mimic ion channel performance in the bilipid
layers of cell walls that can discriminate between sodium and
potassium with a separation factor of over one million!
The aforementioned considerations for membrane technol-
ogy suggest the direction for membrane science applications
in this 21 century. For example, changes in federal regula-
tions may soon require that small incorporated municipali-
ties with a population under 10,000 meet the same high wa-
ter-quality standards that are imposed on larger towns and
cities. Membrane technology may provide the economic al-
ternative for these smaller scale applications.
It is well known that the demand for kidney, liver, spleen,
and other human organ transplants exceeds the supply of
donor organs. There are alternative ways to meet this organ
demand that must keep the gods laughing-the animal whose
organs are most readily transplanted into humans is the pig!
Porcine organ transplants cannot be done directly owing to
problems of potential transmission of viruses as well as sen-
sitive religious considerations for some people, but it is pos-
sible to encapsulate the porcine organ in the shell of a hol-
low-fiber membrane module, a device that resembles a tube-
and-shell heat exchanger, and to have the blood flow through
the fibers. In this way organ functionality is retained by having
the solutes permeate through the walls of the hollow fibers with-
out any direct contact between the porcine organ and the blood.
A very desirable alternative to organ transplants is, of
course, to cure the disease itself. Tremendous progress is now
being made using islet cell transplantation. Islet cells are the
'drill sergeants' of the cell community that can restore healthy
organ functionality. Membranes are used both to isolate the


islet cells and to encapsulate them for transplantation via sur-
gical techniques such as laparoscopy.
We have already alluded to controlled release for deliver-
ing drugs at the precisely required dosage transdermally. This
well-established membrane technology will soon be comple-
mented by smart drug delivery. The latter will involve inter-
facing devices for sensing the need for treatment and deliver-
ing the appropriate drug. Consider, for example, people suf-
fering from heart disease. If they encounter a stressful situa-
tion such as a near-accident while driving, they could find
themselves in a life-threatening situation if they do not re-
ceive one of the drugs that suppresses the adrenaline in their
bloodstream. Smart drug delivery might involve a piezoelec-
tric sensor on the chest that would detect an excited heartbeat
and send an analog signal to a MEM-MEMS device that would
deliver an adrenaline-suppressing drug transdermally into the
bloodstream. The latter would close the feedback loop back
to the piezoelectric sensor that would gradually reduce the
drug delivery as dictated by the heartbeat.
The pharmacy-on-a-chip will use a MEM-MEMS device
that incorporates many microscale membranes into a micro-
chip to deliver a solute. The particular advantages of MEM-
MEMS technology are miniaturization and analog input/out-
put electrical signals. These devices capitalize on the unique
linear scaling characteristic that permits membranes to retain
their ability to deliver, separate, or discriminate between sol-
utes even on the micro- and nano-scales. For this reason
MEM-MEMS devices will be used in separation devices such
as compact portable oxygenators that will greatly improve
the quality of life for chronic obstructive pulmonary disease
(COPD) patients and in smart sensors for the discriminating
detection of solutes such as glucose and insulin levels in the
bloodstream of diabetics.
In the past, profitable applications of membranes such as
water desalination have required making membranes via
large-scale, continuous-production-line processing. Indeed,
heretofore most commercial membranes have been manu-
factured either via phase-inversion or interfacial polymer-
ization. The application of self-assembly, molecular scaf-
folding, and other templating techniques for making highly
functionalized membranes has been an academic curiosity.
The exciting potential for MEM-MEMS technology, how-
ever, has dictated a new paradigm for membrane manufac-
ture in the 21"s century-namely making commercial mem-
branes by the millimeter rather than the mile. Consider for
example that a single 8.5 x 11-inch sheet could provide 2
million micromembranes having a 200-micron diameter such
as might be incorporated into a MEM-MEMS device. A par-
ticular advantage of these micromembranes is that the total
force on their area for typical operating pressures will be quite
small, thereby obviating the need for a thick support layer
that reduces the permeation rate.
Making membranes by the millimeter rather than the mile


Chemical Engineering Education










for MEM-MEMS and other high-tech applications opens the
door to an array of techniques for introducing unique struc-
ture and functionality. The MAST Center has developed a
program that is currently focusing on using Langmuir-
Blodgett (L-B) technology to address the MEM-MEMS chal-
lenge. The L-B technique involves spreading one or more
surface-active components at the water/air interface to form
a monomolecular film. This film is subsequently compressed
by a moving barrier in the interface. This compression causes
ordering of the surface-active molecules and possibly a two-
dimensional phase transition. The ordered or phase-separated
film can then be collected on a solid support by drawing the
latter through the interface. Composite membranes can be
fabricated by employing a microporous membrane as the solid
support on which the L-B layers are collected. The thickness
of these composite membranes can be controlled by adding
more L-B films. The functionality of these composite mem-
branes can be engineered by changing the properties of the
monomolecular film, either by employing different surface-
active molecules or by changing the surface state of the film.
Microporous L-B films can be created by first using a mono-
layer within which a two-dimensional phase separation has
occurred and then dissolving the dispersed phase component
from the supported L-B film.
It is particularly desirable to make fluorocarbon membranes
using this L-B technique. Fluorocarbon membranes are at-
tractive owing to their excellent solvent resistance and ap-
proved use in the human body, but many fluorocarbon poly-
mers are not amenable to membrane fabrication via conven-
tional phase-inversion. Blends of surface-active alkyl- and
fluoro-trichlorosilanes will phase-separate under interfacial
compression to form dispersed two-dimensional micro-do-
mains. Moreover, the trichlorosilanes will spontaneously hy-
drolyze in water and subsequently polymerize under com-
pression to form a robust film that can be collected on a po-
rous support. By appropriately choosing the concentration
and chain length of the alkyl- and fluoro-trichlorosilanes, it
is possible to make the former the dispersed phase compo-
nent. The latter then can be easily dissolved from the sup-
ported L-B film using common solvents such as benzene.
Challenges in applying the L-B method for fabricating com-
posite membranes relate to achieving the desired functional-
ity, pore size, and mechanical integrity. The NSF MAST Cen-
ter is exploring the use of chain length, surface compression
speed, substrate pH, temperature, and DC electric fields to
decrease the dispersed phase domain size in these L-B films.
We are optimistic that this research will lead to an enabling
technology for manufacturing highly functionalized membranes
that will address the demands of 21 t century applications!

SUMMARY

Membrane technology has unique advantages that position
it to address exciting challenges in the 21 century that relate


to environmental, biomedical, biosensor, and pharmaceuti-
cal applications. In particular, it is the only separations tech-
nology that retains its functionality at the micro- and nano-
scales and thereby has unbounded potential for use in MEM-
MEMS and related applications.

ACKNOWLEDGMENTS
The author is grateful for the opportunity to have served as
Program Director at NSF, which excited his interest in mem-
brane science research. He is also grateful to his students for
both providing the impetus to get into this field of research
and for reminding him that the professor is not always right!
He is particularly grateful to Professor Alan R. Greenberg at
the University of Colorado, with whom he has shared a co-
operative membrane research program for nearly 20 years.
The author also acknowledges the following NSF support:
for the MAST Center via grants CDR-8816893, ECD-
9103095, EEC-9632722, EEC-0086182 and EEC-0120823;
for our REU programs via grants EEC-9300435, EEC-
9531361, EEC-9820477 and EEC-0139438; and for the
IGERT program via grant EEC-0333377.

REFERENCES
1. Bond, L.J., G.-Y. Chai, A.R. Greenberg, and W.B. Krantz, "Method and Ap-
paratus for Determining the State of Fouling/Cleaning of Membrane Mod-
ules," U.S. Patent No. 6,161,435, issued December 19, 2000
2. Bond, L.J., A.R. Greenberg, A.P. Mairal, G. Loest, J.H. Brewster, and W.B.
Krantz, "Real-Time Nondestructive Characterization of Membrane Compac-
tion and Fouling," in Review of Progress in Quantitative Nondestructive Evalu-
ation. D.O. Thompson and D.E. Chimenti (eds.). Plenum Press, New York,
1167(1995)
3. Mairal, A.P., A.R. Greenberg, W.B. Krantz, and L.J. Bond, "Real-Time Mea-
surement of Inorganic Membrane Fouling in RO Desalination Using Ultra-
sonic Time-Domain Reflectometry," J. Membrane Sci., 159, 185 (1999)
4. Mairal, A.P., A.R. Greenberg, and W.B. Krantz, "Investigation of Membrane
Fouling and Cleaning Using Ultrasonic Time-Domain Reflectometry," De-
salination. 130(1), 45 (2000)
5. Greenberg, A.R., and W.B. Krantz, "UTDR-Based Sensors for Real-Time
Quantification of Membrane Module Fouling and Cleaning," Fluid/Particle
Sep. J., 15(1) 43(2003)
6. Zhang, Zh.-X., A.R. Greenberg, W.B. Krantz, and G.-Y. Chai, "Study of Mem-
brane Fouling and Cleaning in Spiral Wound Modules Using Ultrasonic Time-
Domain Reflectometry." in New Insights into Membrane Science and Tech-
nology: Polymeric, Inorganic and Biofunctional Membranes. A.A. Butterfield
and D. Bhattacharyya (eds.), Elsevier, Amsterdam, 65 (2003)
7. Peterson R.A., A.R. Greenberg, L.J. Bond, and W.B. Krantz, "Use of Ultra-
sonic TDR for Real-time Noninvasive Measurement of Compressive Strain
During Membrane Compaction," Desalination, 116, 115 (1998)
8. Aerts, P., A.R. Greenberg, R. Leysen, W.B. Krantz, V.E. Reinsch, and P.A.
Jacobs, "The Influence of Filler Concentration on the Compaction and Filtra-
tion Properties of Zirfon Composite Ultrafiltration Membranes," Sep. and
Purification Tech.. 22-23. 663 (2001)
9. Kools. W.F.C.. S. Konagurthu. A.R. Greenberg, L.J. Bond, W.B. Krantz, T.
van den Boomgaard, and H. Strathmann. "Use of Ultrasonic Time-Domain
Reflectometry for Real-Time Measurement of Thickness Changes During
Evaporative Casting of Polymeric Films," J. Appl. Polym. Sci., 69,2013 (1998)
10. Shojaie, S.S., A.R. Greenberg, and W.B. Krantz, "Use of an Electric Field to
Alter Membrane Morphology in a Polysulfone-Polyvinylpyrrolidone Blend,"
J. ofMemb. Sci., 79. 115 (1993)
11. Greenberg, A.R., A. Neice and P. Todd "Apparatus and Method for Control-
ling Pore Size in Polymeric Membranes and Thin Films," U.S. Patent No.
6,479,007 BI, issued November 12, 2002 1


Spring 2004










laboratory


AGITATION AND AERATION


An Automated Didactic Experiment





ALBERTO C. BADINO, JR., PAULO I.F. DE ALMEIDA, ANTONIO J.G. CRUZ
Federal University of Sao Carlos Sao Carlos, SP, Brazil


S ince its foundation, the Federal University of Sao
Carlos' Chemical Engineering Department has tradi-
tionally conducted instructional experiments to
complement the theoretical knowledge passed on in the class-
room. This practice is part of a successful approach to pro-
vide students with a good practical background, preparing them
to become competent and competitive professionals in the
chemical industry. Our undergraduate laboratory follows the
principle of designing instructional kits for short experiments
to be prepared and conducted by the students themselves.11'
Recent advances in equipment design and techniques have
facilitated the acquisition of on-line information concerning
chemical processes. This reality is undoubtedly becoming part
of the daily routine of chemical engineers in research labora-
tories and industry. Therefore, undergraduate courses should
provide opportunities for students to get first-hand practice
in the techniques involved in automated chemical processes.
During the establishment of a "Special Training Program"
for undergraduate students (called "The Factory of the Fu-
ture"12]), a set of automated experiments was implemented
with the financial support of CAPES, a Brazilian education
funding agency. One of the kits in this program is an Agita-
tion and Aeration Experiment, which was used in a case study
of an important biochemical process, "Aerobic Cultivation
of Baker's Yeast." The supervisory system was designed by
the authors, and the controllers were programmed by a spe-
cialized engineering firm, using commercial software.
Many processes involved in the manufacture of important
chemical products, such as the hydrogenation of oils in the
production of margarine or the aeration of fermentation broths
in antibiotic production, make use of agitation and aeration
tanks to promote efficient mass transfer from the gas phase
to the liquid phase. Despite its obvious importance, teaching
about the operation of agitation and aeration is usually rel-


egated to a secondary role in chemical engineering courses.
This paper describes automated equipment developed to
run teaching experiments on these unit operations in the labo-
ratory. The purpose of the experiment is to show students a
real chemical process. The main adjusted parameters, such
as the stirrer speed (N) and gas (air) flow rate (Q), can be
monitored. Simultaneously, phenomena involving changes in
dissolved oxygen concentration (C) and gassed (P ) and
ungassed (P0) power consumption can be viewed by the su-
pervisory computer. The continuous acquisition and storage
of experimental data allows the volumetric oxygen transfer

Alberto C. Badino, Jr., is Associate Professor
of Chemical Engineering at the Federal Univer-
sity of Sao Carlos. He received his Doctorate in
Biochemical Engineering from the State Univer-
sity of Sao Paulo (Brazil) in 1997. His research
interests are in mass transfer, hydrodynamics,
and rheology of fermentation broths in conven-
tional and nonconventional bioreactors.



Paulo Ignacio F Almeida is Associate Profes-
sor of Chemical Engineering at the Federal Uni-
versity of Sao Carlos. He received a Doctorate in
Mechanical Engineering in Thermal and Fluids
Area from the State University of Campinas (Bra-
zil) in 1992. His research interests are in plant-
wide control, automation, and control of batch in-
dustrial chemical and biochemical processes.



Antonio J.G. Cruz is Associate Professor of
Chemical Engineering at the Federal University
of Sao Carlos. He received his Doctorate in Mod-
eling and Simulation of Biochemical Processes
from the Federal University of Sao Calos (Bra-
zil) in 2000. His research interests are in model-
ing, simulation, control, and instrumentation of
bioprocess.


Copyright ChE Division of ASEE 2004


Chemical Engineering Education


100










[We have] traditionally conducted instructional experiments to complement the
theoretical knowledge passed on in the classroom.... Our undergraduate laboratory
follows the principle of designing instructional kits for short experiments
to be prepared and conducted by the students themselves.


coefficient, (kLa) to be accurately determined, facilitating the
students' analyses of the data and writing of reports.

THEORY
Owing to their versatility, conventional agitation and aera-
tion tank fermentors are still the ones most frequently used
on bench, pilot, and industrial scale in the fermentation in-
dustry.31 In fermentation processes involving the cultivation
of aerobic and facultative aerobic microorganisms, oxygen
is an essential element in the supply of energy for cell me-
tabolism and, hence, a key element in the syntheses of biom-
ass and products.
A convenient way of evaluating oxygen transfer in aerobic
cultures is to measure the volumetric oxygen transfer coeffi-
cient (kLa). The gassed power consumption per unit volume
of broth (P /V) and the volumetric oxygen transfer coeffi-
cient are still the most widely used criteria in the design and
scale up of conventional stirred and aerated tank fermentors.4'"
Various equations correlating the volumetric oxygen trans-
fer coefficient (kLa) to other quantities have been published."5'
Cooper, et al.,'61 originally proposed the simplest, correlating
kLa to the gassed power consumption per unit volume of broth
(Pg/V), the superficial gas velocity (v), and the geometry of
the vessel (D)

kL C P / V) (Vs) (1)

where
4Q (2)
vs 2 (2)
7tD2

Although initially developed for fluids different from fer-
mentation broths, the relation given in Eq. (1) has been widely
used in fermentation systems.'J1 Values reported in the litera-
ture for the constants ac and p and the proportionality con-
stant C, vary considerably with the geometry of the system,
the range of variables covered, and the experimental method
used. Therefore, specific correlations for kLa for use in sub-
sequent monitoring, control, and scale up of aerobic fermen-
tations should be obtained from bench-scale experiments.
Similarly to the volumetric oxygen transfer coefficient (kLa),
the gassed power consumption (Pg) can be correlated to the
system's geometric parameters (impeller diameter, D), to op-
eration variables (impeller speed, N, and gas flow rate, Q),
and to the physical properties of the fermentation broth (den-


sity, p, and viscosity, p.) that are implicit in the ungassed power
consumption (P(). The power requirements in gassed and
ungassed systems can be represented by the traditional equa-
tion proposed by Michel and Miller'7'


Pg = C2 0.5 (3)

where the constants C2 and 0 also depend on the system's
geometry and on the range of operating variables used.

MATERIALS AND METHODS
Microorganism
Fermentations were conducted using Saccharomyces
cerevisiae in the form of commercial Baker's yeast
(Fleischmann).


Culture Media
To activate and ferment the inocula, a culture medium con-
taining (in kg-m-3)
glucose (15.0)
KHPO4(5.0)
MgSO,47H,O (0.5)
yeast extract (3.0)
(NH4)2SO4 (4.5)
antifoam (1.0 Lm-3)
was used. The initial pH was adjusted to 4.6 with the 4M
HSO4 or 3M NH4OH.


Supervisory and Data Acquisition System
The experiment was monitored and controlled in an auto-
mated teaching system, which uses a Programmable Logic
Controller (PLC) coupled to a supervisory system that pro-
vides the man/machine interface (MMI). The main purpose
of this interface is to allow the user to interact with the pro-
grammable controller. The system was implemented in a Su-
pervisory Control and Data Acquisition Architecture where
the PLC performs control functions while monitored and su-
pervised by a microcomputer.
The computer screen is divided into two windows. The first
one shows an animated view of the process with continuous
variable updates. The on-line values are displayed in boxes
next to the respective variables during the process. The sec-


Spring 2004









ond window presents a graphical display of the time
course of the process variables. These windows can
be seen in Figures 1 and 2.


Experimental Apparatus
The experiment was carried out in an agitated and
aerated cylindrical vessel equipped with two six-
flat-blade turbines and an air sparger. The stirrer
speed (N) was controlled by a frequency inverter
(0.5 HP, Siemens, Germany) coupled to the AC 0.4
kW motor. Stirrer speed was measured directly by
an optical tachometer (Monarch Instruments,
Amherst, NH) on the agitator shaft. The electric sig-
nal was used to monitor the stirrer speed.
The air flow rate (Q) through the vessel was moni-
tored by a mass flow meter (0 to 20 SLPM, KURZ
Instruments, Inc., Monterey, CA). The temperature
was kept at 300C by circulating water between a
thermostatic bath and a water jacket. Apolarographic
electrode (submersible type, IC Control, Canada)
connected to a dissolved oxygen transmitter (model
855, Microprocessor Analyzer, CI Control, Canada)
was used to measure the dissolved oxygen concen-
tration (CE). Figure 3 shows the dimensions of the
vessel and Figure 4 illustrates a sketch of the ex-
perimental system. All the analog signals between
the physical system and the PLCs (GE Fanuc, se-
ries 90-30, GE Fanuc Automation North America,
Inc.) were provided as current signals (4 to 20 mA).
The power consumption (P) was monitored and
controlled through a frequency inverter, which con-
trolled the stirrer speed (N) by modulating the elec-
tric power transferred to the AC motor. The volu-
metric oxygen transfer coefficient (kLa) values ob-
tained at different stirrer speeds (N) and air flow
rates (Q) were correlated with the system's operat-
ing parameters and physical dimensions. Correla-
tions of this type are useful for the design, control,
and scale up of fermentors.


Determination of ka
The volumetric oxygen transfer coefficient (kLa)
was determined using the method proposed by
Mignone and Ertola,181 based on a step change in
stirrer speed during cultivation.
Assuming mixing is complete, the mass balance
for dissolved oxygen in the liquid phase during aero-
bic batch cultivation can be expressed as

dC kLa(- QoC (4)
dt OTR -
OTR OUR


where OTR is the oxygen transfer rate and OUR is the oxygen uptake
rate.
Initially, the culture is maintained under a constant air flow rate (Q)
and stirrer speed (N,). This situation corresponds to the steady state I, in
period I shown in Figure 5. During this phase, we have (kLa), and the
dissolved oxygen concentration (C,).
The stirrer speed is changed to N, where N, > N,. After a short time, a
transient state is produced (period II in Figure 5) during which the con-
centration of oxygen varies until the new steady state II is reached (pe-
riod III). This new period is characterized by another volumetric oxygen
transfer coefficient (kLa), and by another oxygen concentration (C,).
Each of these periods in Figure 5 can be described by the following
equations:


Figure 1. Representative screen of instrumentation created by
the supervisory system, showing updated signals (variables) in
real time.


A .... .... ... ...i.. .. -

7'. 1I7- 0. 1 ., "', :. 7.1 .6 7. . . ",
16:35:15 16-37:51 16:40:27 16:43:03 16 45:39 16:4 15 16:50:51 16:53 27 16:5603 16:8:39
-I.' -- _:'_ ....... M. . '_ -_ I ........Ji B I .-E u -a
w ^- ------ --- -- A, S" V .
.*... .... C .-----
-- - ------ ------ b h C-
-,. .- "" ' I |.B... 1 - ............ ..i

Figure 2. Graphic display showing the time course of
process variables.


Chemical Engineering Education










































Figure 4. Experimental apparatus: (1) agitation and aeration
tank; (2) stirrer motor; (3) thermostatic bath; (4) photo tachom-
eter; (5) dissolved oxygen electrode; (6) dissolved oxygen ana-
lyzer; (7) mass flow meter; (8) frequency inverter; (9) supervisory
computer (InduSoft Studio 3.0 running on Windows 95,
Microsoft); (10) Programmable Logical Controllers.


Pe"aIdiB


W I


Q
----C, -. . ...
, I ..... J -.- -. -,.
Ct-C _ _


-- *1*-----


Period I (steady state)



Period II (transient phase)

C= (kLa)(C -C)-Qo, C (6)

Period III (new steady state)

dC (kLa)2C -C2)QoCx= 0 (7)

In each case, the oxygen concentration was kept above
the critical value. For short periods of time, it may be
assumed that the oxygen uptake rate ( Qo, C,) remained
constant.
From Eq. (7)

QoCx = (kLa)2(C* -2) (8)

Substituting Eq. (8) into Eq. (6) gives
dC
d = (kL2(C2 -C) (9)

Eq. (9) can be integrated from t = 0, C = C, to t = t, C =
C,, giving

C C2 -(C2 C1)e-(kLa) t (10)
Defining a dimensionless concentration


C2 C1
Eq. (10) can be rewritten in dimensionless form

C=1-e-(kLa)2t (12)


six flat blade
turbine
H = 0.48 m
S HL= 0.325m
SHL L D- 0.28 m
W 1E] Di= 0.12 m
H Hi= 0.12 m
SI I H,= 0.12m
Di L = 0.0285 m
Dt W- 0.024 m

Figure 3. Dimensions of the vessel.


Spring 2004


I I __ J*
Figure 5. Dissolved oxygen concentration profile (CE) after steps in
stirrer speed (N).


i. : I 1


~~
I*~Aiil~C1C*


.i
~b~~II









Equation (12) describes the kinetics of the liquid phase.
This equation, however, does not take into account the elec-
trode kinetics and the kinetics of the stagnant film that cov-
ers the membrane of the electrode. These two phenomena
are described by the first-order equations

dCF C-CF
(13)
dt TF

dCE CF-CE
(14)
dt TE

where CF and CE describe, respectively, the dimensionless
diffusion film and electrode responses; TF and TE are the
response times for the liquid film and electrode diffusion re-
sistance.
The composition of the gas phase is assumed to remain
constant, so it is unnecessary to consider its kinetics.

The value of CE can be derived from Eqs. (12), (13), and


-kat) +


(F -TE)


(15)


The area (al) below the electrode response curve, (1- CE)
versus time, is given by


Thus, this method provides a simple procedure to evaluate
(kLa)2 after a step change in the stirrer speed (N).
It is evident that, by returning the stirrer speed to its origi-
nal value N,, the value of (kLa), can be obtained by means of
a similar procedure.


Experimental Procedure
The experiment was carried out in the apparatus illustrated
in Figure 4, with 20-L (0.020 m3) of culture medium in the
tank.
The air flow rate (Q) was initially set to 12.0 SL-min-
(2.0-104 Sm3-s-l) and stirrer speed (N) to 400 rpm (6.67 s-l),
until the dissolved oxygen concentration given by the elec-
trode (CE) attained the saturation value (equilibrium). The
electrode response time (TE) was obtained by placing the elec-
trode in the tank after equilibration in an 0.1 M sodium sulfite
solution (Na2SO,) (CE = 0).


0 20000 40000 60000 80000
( p> 2 N D i3 /Q 0.560V '3 ( 2 -1 3 ( 3 -1 -0.56)
(P0 N Di3/Q 0.)39) (W2.s .m .(m3.s )-o5 )

Figure 6. Experimental correlation between gassed (P)
and ungassed (P) power requirements.


a= l-CE)dt (kLa)+ + ETF
S(k a)2


The response times tF and TE are determined by a step
change in oxygen concentration in the medium, carried out
before inoculation (no cells present). The area (alE) below
the response curve versus time is then given by


alE = (1-CE)dt= E + F
*0


The difference between at and alE then yields


1
a1 -alE -(kLa)2


Figure 7. Electrode response in a step assay at 400 rpm
(6.67 s-1) in culture medium without cells.


Chemical Engineering Education










The tank was inoculated by adding 0.40 kg of commercial
Baker's yeast (40% on a dry basis), giving an initial cell con-
centration (Cx) of approximately 8 kg-m 3.
After inoculation, the air flow rate was kept at 12.0 SL-min-',
the stirrer speed was altered to approximately 250 rpm (N1),
and the first steady state was produced, with CE = C E. The
stirrer speed was raised to approximately 400 rpm (N,) and
the second steady state was reached, with CE = CE. The stir-
rer speed was increased again to approximately 550 rpm (N,),
producing the third steady state, with CE = CE3. After the third
steady state was reached, the stirrer speed was returned to
approximately 400 rpm (N2) and 250 rpm (N1), respectively,
as illustrated in Figure 8.
This experimental procedure was repeated at an air flow


TABLE 1
Power Requirements for Various Conditions of
Agitation (N) and Aeration (Q)


N(s') Q104(Sm's') P,(W) P,(W)


P(JND' [ S ,m3(m3s, -10.56
Q 0.56


4.15 2.0 53.9 2,228.4
4.18 3.0 51.4 39.7 1,791.5
6.81 2.0 90.8 13,389.7
6.85 3.0 98.3 78.7 10,743.0
9.35 2.0 171.0 71,794.9
9.37 3.0 193.5 161.8 56,850.8


mIeDkseragmlw nkratrl (<,)



0 i ..- Q... ... r
16:51 16315 16 0 ia 081:2) 114.o 0810 1 I 1 7
165:15 16:37:51 16:40:27 16:43:03 1:45:39 1:4:1 16:50:51
16:35:15 16:37:51 16:40:27 16:43:03 16:45:39 16:48:15 16:50:51


I ii i.I~'~ I


LJ] irpeiasuie IT|

I M- ,.'.
I POW M.ll.I .


o .." "o ..0
ri .0S 188


I raromr Dt fl
*L -
~ C 3OLL "' r60 hC
STIr lul cl' | )i
Conve
earn me


Figure 8. Supervisory screen for viewing the changes oc


rate (Q) of 18.0 SL-min-1 (3.0-10- Sm3.s- ) so that six values
of volumetric oxygen transfer coefficient (kLa) could be ob-
tained. After the CE curves were measured, the air flow rate
(Q) was turned off and stirrer speed (N) adjusted to approxi-
mately 250, 400, and 500 rpm, respectively. For each agita-
tion and aeration condition, the system recorded the gassed
and ungassed power consumption (Pg and Pd).
The total duration of the experiment, which involved three
stirrer speeds and two air flow rates, was approximately 2.5
hours.

RESULTS AND DISCUSSION
As described previously, the experiment is totally moni-
tored and controlled through a microcomputer with a soft-
ware supervisor. The students may alter any given variable
and view its effect over time. Experimental data are recorded
at 2-second intervals in the form of a binary file that is subse-
quently transformed into an ASCII file and transferred to a
computer worksheet such as Excel.
Figure 1 illustrates the main screen of the supervisor soft-
ware and shows the variables, which can be manipulated by
using the mouse and keyboard, such as the stirrer speed (N)
and the air flow rate (Q). The response variables, which vary
according to changes in the manipulated variables, appear in
their respective fields, as do the power requirements (P) and
the dissolved oxygen concentration given by electrode (elec-
trode signal, CE).
Power Requirements Table 1 lists the values of power
input in gassed (P ) and ungassed (Po) broth, monitored and
controlled through a frequency inverter at various stirrer
speeds (N) and air flow rates
(Q). As can be seen, the gassed
power requirements (P ) are
lower than the ungassed values
(Po) under the same agitation
_condition (N), which is due to
-- -- the fact that the introduction of
Sgas into the fluid decreases its
-____. apparent density (p). It can also
: i be observed that an increase in
stirrer speed (N) raises the
power requirements (Po and P ),
-- indicating the strong influence
of viscous forces in the opera-
o8:l .O tion of this important unit.
16:53:27 16:56:03 16:58:39
H Gassed power consumption
Aw) ail .: -T values (P ) were correlated with
MGM.- the ungassed power consump-
rsion
._ E1tion (Po), stirrer speed (N), air
Slow rate (Q), and the impeller
diameter (D) through a well-
'curring during a run. known equation proposed by


Spring 2004


I1] van abile = I m"









Michel and Miller[71


(2 N D 0.39 Q = 18.0 SL.minm (3.0 x 10 Sm'.s')
P = 2.46 N Di R 2.= 0.995 (19)
2.46 0.56 0.995 (19) 10 :251 -411 rpm (4.18 6.85 s')
:411-562 rpm (6.85- 9.37 s')
The experimental data and fitting equation are illustrated in 0.8 411 562 rpm (6.85 9.37 s
Figure 6. The high correlation coefficient indicates that the
equation represents closely the data obtained from the ex- i- 0.6
perimental system.
Oxygen Transfer Similar values of electrode response 0.4
time ( E) were obtained in triplicate assays at a stirrer speed
of 400 rpm (6.67 s -). The resistance of the liquid film cover- 0.2
ing the membrane of the electrode was considered negli-
gible in the range of stirrer speeds studied. Thus, TF = 0, 00 50. 100 -150 200 250
and from Eq. (17), aIE = TE. Figure 7 illustrates a typi- t(s)
cal electrode response curve obtained in the culture me-
dium at 400 rpm (6.67 s-'). Figure 9. Electrode response at a constant air flow rate
(Q = 3.0 x 104Sm3s- ), after the following step changes in
Figure 8 shows the supervisory window where the students stirrer speed
stirrer speed:
can view changes occurring in the dissolved oxygen concen- () 251-411 rpm (4.18 6.85 s-1); ka = 0.0346 s-1;
tration given by the electrode (CE) after step changes in the (0) 411-562 rpm (6.85 9.37 s-), kLa = 0.0469 s-';
stirrer speed (N), with the air flow rate (Q) maintained con- (-) continuous lines.
stant, as described in the methodology of Mignone and
Ertola1l8 for kLa determination. In the actual window, the vari- TABLE 2
able profiles appear in color, but to facilitate viewing here, T
Results of ka Measurements at Various
the profiles of Q, N, and CE are illustrated, respectively, with Air Flow Rates (Q) and Stirrer Speeds (N)
dashed, dotted, and solid lines, and temperature, air flow rate,
and power requirement were omitted. N(s-') Q-104(Sms ') P/V(Wm ) v- 103(ms ') kLa(s-')
Figure 9 shows typical curves of the electrode response 4.15 2.0 2,695 3.25 0.0187
after the stirrer speed is subjected to sudden changes. The 4.18 3.0 1,985 4.87 0.0201
symbols (0 and ]) represent experimental data acquired and 6.81 2.0 4,540 3.25 0.0253
stored by software, which are linked by smooth lines. Ac-
6.85 3.0 3,935 4,87 0.0346
cording to the methodology proposed by Mignone and
Ertola,181 the area below the response curve obtained by plot- 9.35 2.0 8,550 3.25 0.0331
ting vs time corresponds to 9.37 3.0 8,090 4.87 0.0469


a = (l-CE)dt= +TE+F (20)
(kLa)2 (20) 0.06
S2 + 10%
and given zF = 0 and TE = 14.2 s, kLa values in the second 005 0%
steady state (kLa)2 can be determined by Eq. (21) -
0.04 I
Sl-a1E -= (21) .
(kLa)2 003
Table 2 shows the results of kLa measurements at different 002
values of air flow rate (Q) and stirrer speed (N), as well as
the variables of gassed power consumption per unit volume o.o 0 -
of broth (P /V) and superficial gas velocity (v) calculated
for the experimental conditions used. 0oo0
000 0.01 0.02 0.03 004 0.05 006
Experimental data for kLa, determined by the proposed kLa, (s')
method, ranged from 0.0187 to 0.0469 s-' (67.3 168.8 h-'), ___a
in complete agreement with the various agitation and aera- Figure 10. Comparison of the experimental values of kLa
tion conditions tested. Therefore, the results confirm the compared with the values of ka calculated by the
method's reliability and consistency. equation (Eq. 1) proposed by Cooper, et al.'61


Chemical Engineering Education










As mentioned earlier, the most classical relation for the
volumetric oxygen transfer coefficient (kLa) in agitated and
aerated tanks, proposed by Cooper, et al.,161 is given by Eq.
(1). This equation was fitted to the experimental values and
the parameters were estimated through nonlinear regression.[91
The criterion for the best fit and parameter optimization was
the sum of squares of residuals (SSR). The fitted correlation


kLa=0.046(Pg /V)053 (s)0.89


CF
Cx
Di
D
kLa
N
OTR
OUR
P
Po
P
Q


R2 = 0.98 (22)


The parameters estimated by the nonlinear regression are
unequivocally within the range of values mentioned in the
literature."I1
Figure 10 illustrates the congruence between the experi-
mental kLa values and the data calculated by Eq. (22). The
correlation coefficient (R2), as well as Figure 10, demonstrate
that a good fit was obtained, indicating that Cooper's correla-
tion (Eq. 1) can be used to estimate, with good precision, the
volumetric oxygen transfer coefficient (kLa) in these systems.


CONCLUSIONS

* The experiment can be completed rapidly, i.e., in about 2
1/2 hours, which is appropriate for an undergraduate ex-
perimental class.
All the transient phenomena can be monitored and con-
trolled on-line by the students, and the data can be stored
for subsequent manipulation and calculation. This pro-
cedure, which is typical of an industrial supervision pro-
cess, puts students in direct contact with a reality they
will face in their careers as process engineers.
The easy handling of data facilitates data processing in
optimization programs for the industrial operation pro-
cesses the students are learning to deal with.
Student training in automated laboratory facilities that
simulate industrial processes offers opportunities for ac-
quiring and enhancing their professional skills in today's
competitive industrial environment.

NOMENCLATURE
C, constant ofEq. (1)
C2 constant of Eq. (3)
C dissolved oxygen concentration in the broth in steady
state (kgO,m 3)
C' dissolved oxygen saturation concentration in the
broth (kgO m-3)
CE dissolved oxygen concentration given by the
electrode or electrode signal (kgO,m-3)
CF dissolved oxygen concentration on stagnant film
(kgO2m-3)
CE dimensionless electrode signal (-)


Qo,
Qo, Cx
R2


dimensionless oxygen concentration on diffusion film (-)
cell concentration (kg-m-3)
impeller diameter (m)
tank diameter (m)
volumetric oxygen transfer coefficient (s-')
stirrer speed (s or rpm)
oxygen transfer rate (kgO,m-'s')
oxygen uptake rate (kgO2m-3s1)
power consumption (W)
ungassed power consumption (W)
gassed power consumption (W)
air flow rate (Sm's or SL-min-', where S indicates
standard condition: 0C and 1 atm)
specific oxygen uptake rate (kgO2kg- s')
global oxygen uptake rate (kgOm 3s 2)
correlation coefficient (-)


t time (s)
vs superficial air velocity (ms ')
V broth volume (m3)
Greek Letters
a constant of Eq. (1)
p constant of Eq. (1)
aI area below the response curve obtained by plotting
1 CE VS time during cell growth (s)
alE area below the response curve obtained by plotting
1 CE vs time after step assay (s)
0 constant of Eq. (3)
TE electrode response time (s)
IF film response time (s)

REFERENCES
1. Badino, A.C., and C.O. Hokka, "Laboratory Experiment in Biochemi-
cal Engineering: Ethanol Fermentation," Chem. Eng. Ed., 33(1), 54
(1999)
2. De Almeida, P.I.F., "The Factory of the Future: A Special Training
Program," Int. Conf. on Eng. Ed ICEE-98, Rio de Janeiro, Brazil;
CD Rom (1998)
3. Royce, P.N., "A Need to Refocus Research in the Operation of Fer-
menters?" Trends in Biotech., 10, 223 (1992)
4. Einsele,A., "Scaling-up Bioreactors," Process Biochem., 13,13(1978)
5. Badino, A.C., M.C.R. Facciotti, and W. Schmidell, "Volumetric Oxy-
gen Transfer Coefficients (kLa) in Batch Cultivations Involving Non-
Newtonian Broths," Biochem. Eng. J., 8, 111 (2001)
6. Cooper, C.M., G.A. Fernstrom, and S.A. Miller, "Performance ofAgi-
tated Gas-Liquid Contactors," Ind. Eng. Chem., 36(6), 504 (1944)
7. Michel, B.J., and S.A. Miller, "Power Requirements of Gas-Liquid
Agitated Systems," AIChE J., 8(2), 262 (1962)
8. Mignone, C.E, and R.J. Ertola, "Measurement of Oxygen Transfer
Coefficient under Growth Conditions by Dynamic Model Moment
Analysis," J. Chem. Tech. Biotech., 34B, 121 (1984)
9. Marquardt, D.W., "An Algorithm for Least Square Estimation of Non
Linear Parameters," J. Soc. Ind. Appl. Math., 11, 431 (1963)
10. Kawase, Y., and M. Moo-Young, "Volumetric Mass Transfer Coeffi-
cients in Aerated Stirred Tank Reactors with Newtonian and Non-New-
tonian Media," Chem. Eng. Res. Des., 66, 284 (1988) O


Spring 2004










[e, curriculum


INTEGRATING BIOLOGY AND ChE


AT THE LOWER LEVELS




KATHRYN A. HOLLAR,* STEPHANIE H. FARRELL, GREGORY B. HECHT, PATRICIA MOST
Rowan University Glassboro, NJ 08028


Instilling a working knowledge of biological principles
in students and developing their ability to apply engi-
neering principles to biological systems (and vice versa)
is recognized nationwide as a goal for chemical engineering
programs."'- Many schools offer specialized bio-focused cur-
ricula or courses at the senior or graduate level,t6-81 and there
is a significant movement to change chemical engineering
department names to reflect faculty expertise in bio-focused
engineering. Integration of biology and chemical engineer-
ing at the lower levels, however, is difficult in an already
overloaded curriculum.
We have developed an integrated, collaborative approach
between engineering and biology faculty to introduce chemi-
cal engineering students to the application of engineering
principles in biological systems at the lower levels. Through
specially designed courses and active learning modules that
can be easily adapted to any course, students are exposed to
this newest pillar of the chemical engineering curriculum.
The systematic implementation of this philosophy exposes
students to key areas of collaboration between biologists and
chemical engineers in the early stages in their undergraduate
education. This strategy also enables faculty to build increas-
ing detail and technical content into problems and projects
that address the interface between biology and engineering
as students progress through the curriculum because students
develop a cumulative knowledge of biological principles.
Revisions to the chemical engineering curriculum include
a "Biological Systems & Applications" (BS&A) course de-
signed to introduce students to a variety of biological prin-
ciples that are directly relevant to chemical engineering. Ad-
ditionally, several laboratory modules and projects that can
be easily incorporated at the freshman and sophomore levels
have been developed. These modules include reverse engi-
neering of the human body, reverse engineering of the beer-

*Current address: Division of Engineering & Applied Sciences, Harvard
University, 20 Oxford Street, Cambridge, MA 02139


making process, and design of a microbial fuel cell.
These modules in the freshman year expose students to
chemical engineering principles as they apply to living sys-
tems. The BS&A course, specifically designed for sophomore
chemical engineering students and taught by faculty in biol-
ogy, introduces students to a wide variety of topics, from
prokaryotic and eukaryotic regulatory systems to food mi-
crobiology. A sophomore-level engineering project on mi-
crobial fuel-cell design reinforces concepts in microbial
growth and nutrition that are covered in the BS&A course.
This collaborative approach to integrating biology and
chemical engineering helps prepare students for industrially
sponsored projects at the junior and senior level, and for ca-
reers in the food, biotechnology, and pharmaceutical indus-
tries. The projects, courses, and activities that we describe in
this paper address key areas in which chemical engineering


Kathryn A. Hollar received her BS in Chemical Engineering and English
from North Carolina State University and her PhD in Chemical Engineering
from Cornell University Her research expertise is in the field of recombinant
protein production. Her current focus is on developing laboratory experi-
ments and course activities in food processing, biochemical engineering,
and green engineering, particularly at the freshman and sophomore levels.
Stephanie H. Farrell received her BS in 1986 from the University of Penn-
sylvania, her MS in 1992 from Stevens Institute of Technology, and her PhD
in 1996 from NJIT Prior to joining Rowan in 1998, she was a faculty mem-
ber at Louisiana Tech University. Her research expertise is in the field of
drug delivery and controlled release, and she is currently focusing efforts on
developing laboratory experiments related to membrane separations, bio-
chemical engineering, and biomedical systems for all level students.
Gregory B. Hecht has extensive research experience in prokaryotic genet-
ics and molecular biology. With Dr. Mosto, he has developed a new course
for chemical engineering students, "Biological Systems & Applications." He
is the creator and coordinator of the Rowan University Student Research
Symposium, an annual forum at which Rowan students from all of the STEM
disciplines present the results of their independent research.
Patricia Mosto has extensive environmental science experience. She has
been actively involved with field and laboratory projects related to water
quality and water pollution issues for the last thirty years. She worked with
the Department of Water and Power and the Department of Sanitation in
Los Angeles for ten years. During her ten years at Rowan, she has super-
vised 41 independent undergraduate projects, taking many students to na-
tional and international conferences.
Copyright ChE Division of ASEE 2004


Chemical Engineering Education











We have developed an integrated, collaborative approach between engineering and
biology faculty to introduce chemical engineering students to the application
of engineering principles in biological systems at the lower levels.


and biology have a strong connection, such as bioprocess en-
gineering (biochemical reaction engineering for production
of commodities and waste treatment), bioseparations,
biocatalysis, and metabolic engineering. We will discuss the
implementation and impact of these modifications in the en-
gineering curriculum.

EXPERIMENTS AT THE FRESHMAN LEVEL
A two-semester freshman clinic sequence introduces all
freshman engineering students to engineering at Rowan Uni-
versity. In the clinic we immediately establish a hands-on,
active learning environment for the reason explained by sci-
entist and statesman Benjamin Franklin: "Tell me and Ifor-
get. Show me and I may remember. Involve me and I under-
stand. Multidisciplinary engineering experiments using en-
gineering measurements are the common theme of the first
semester of the clinic, while in the second semester, students
reverse engineer a product or process. In the following para-
graphs, we describe two experiments that have been incor-
porated in the first and second semesters of this unique course
sequence. The experiments, of varying lengths (from one
three-hour lab to a semester project), illustrate various meth-
ods for integrating biological concepts at the lower levels.
Biomedical Experiment The human body is an exquis-
ite combination of interacting systems that can be analyzed
through the application of chemical engineering principles.
Familiar examples include fluid flow of blood through arter-
ies and veins, mass transfer in the lungs, pumping of the heart,
and chemical reactions in cells. Biomedical topics in chemi-
cal engineering are explored in many curricula through ad-
vanced-level elective courses, and they are sometimes worked
into homework problems in core courses. Freshman and
sophomore chemical engineering students are rarely exposed

TABLE 1
Biological Principles and Topics
Related to Beer Production
Principles/Topic Where Applicable
Germination and enzyme production Malting
Enzymatic reactions Starch hydrolysis to sugars during
mashing; protein breakdown to amino
acids during mashing
Yeast growth curve Fermentation process
Fermentation Fermentation
Fermentation monitoring Fermentation
Disinfection and contamination Fermentation and sampling


Spring 2004


to real biomedical applications of their discipline, however,
and are unaware of the analogies between physiologic sys-
tems and chemical engineering operations.
We developed a freshman-level experiment that is used to
introduce students to a wide range of chemical engineering
principles through their application to physiological processes.
The details of the experimental procedure and analysis are
provided elsewhere.191 Students take measurements of physi-
ologic variables both at rest and during exercise, and then
perform engineering calculations that involve basic principles
of mass and energy balances, fluid flow, chemical reactions,
energy expenditure, mechanical work, and efficiency.
During the three-hour experiment, students measure volu-
metric breathing rate and heart rate both at rest and during
exercise on a bicycle ergometer. They also measure blood
pressure at different elevations in the body using a blood pres-
sure cuff (sphygmomanometer). Students use their physi-
ologic data for breathing rate and heart rate to estimate their
rate of oxygen consumption, blood flow rate, and rate of en-
ergy expenditure. The blood pressure measurements are used
to calculate hydrostatic pressure differences to compare with
expected values. This experiment provides an initial expo-
sure to a variety of principles from engineering, physiology,
and cellular metabolism, and provides motivation and frame-
work for future courses on related topics.
Reverse Engineering of Beer Production The theme of
the second semester is the reverse engineering of a commer-
cial product or process. Previous reverse engineering projects
have involved products such as automatic coffee makers, 10"12
hair dryers, and electric toothbrushes.t 3 One of these semes-
ter-long modules, an investigation of the beer production pro-
cess, incorporates the biology and reverse engineering of a
biochemical process into our freshman clinic. A detailed struc-
ture of the course has been previously described.14-151 In this
paper we describe the integration of principles from biology
and engineering into this introductory, multidisciplinary engi-
neering course. The principles are summarized in Table 1.
Near the beginning of the laboratory-intensive part of the
project, a biology professor gives a guest lecture to provide an
overview of the biological processes involved in the produc-
tion of beer. The addition of this guest lecture emphasizes not
only the interdisciplinary aspects of biochemical processes,
but also illustrates the collaboration between engineering
and biology faculty and the importance of
multidisciplinary teamwork. Once this foundation is laid,
students work in teams to investigate and reverse engi-
neer the major steps of beer production.
109









The students' investigation focuses on three of the major steps
of the brewing process: mashing, boiling, and fermentation. The
brewing process is shown in Figure 1. Mashing is the first ma-
jor step in the brewing process. Using the raw materials of malted
barley and water, this process produces a nutritionally com-
plete wort for fermentation. Students mash both malted and
unmalted barley and then compare the worts obtained from each
type. Analyses of the total extract and concentrations of fer-
mentable sugars using an enzyme test kit reveal that only the
malted barley produces a wort containing fermentable sugars,
as shown in Figure 2.
After mashing, the wort is boiled for stabilization and chilled
rapidly to avoid contamination. Yeast is added and the fermen-
tation process takes place over the next 8 to 14 days, with most
vigorous fermentation occurring within the first three days. Stu-
dents may also perform experiments to determine yeast vi-
ability and activity.
The fermentation process provides an impressive visual show
of biological systems in action. As the fermentation proceeds,
students can observe changes in color and turbidity, the forma-
tion of bubbles, and eventually the settling of yeast. The fer-
mentation pathway and yeast growth curves are followed ana-
lytically as sugars are consumed to produce more yeast, alco-
hol, and carbon dioxide; analytical measurements include yeast,
sugar, alcohol concentrations, and pH. Students also learn about
disinfection techniques and contamination issues as they clean
and sterilize the glassware and other supplies used for fermen-
tation and subsequent sampling. At the end of the experiment,
engineering problems such as disposal of wastewater, organic
wastes, and biochemical oxygen demand may be addressed.

EXPERIENCES AT THE
SOPHOMORE LEVEL

The two freshman-level modules discussed above intro-
duce the interplay between biology and engineering using
familiar systems. These modules also expose
students to two different areas in which en-
gineering principles can be applied to bio- Malted Ba
logical systems: 1) using engineering prin-
ciples for analysis in biomedical applications
(reverse engineering of human body) and 2)
using engineering principles for manipulation
of microbial cultures to generate products
for human use (beer-making process). The
experiments illustrate basic principles and
excite students about careers for which a Mash
Tun
chemical engineering degree can prepare
them.
Before tackling in-depth analysis and ma-
nipulation of biological systems using engi-
neering principles, however, students must
have a firm grounding in biological principles Figure 1.
that only a course can provide. A chemical


engineer who is well-versed in biological, biochemical, and
microbiological applications is an attractive candidate for
employment in the pharmaceutical, biotechnology, and food
manufacturing industries. To meet the anticipated growing
demand for biology-literate engineers in these industries, bio-
logical sciences and chemical engineering faculty worked
closely together to develop a course that prepares chemical
engineering students for these careers. The 4-credit, lab-in-
tensive course is open only to fall-semester sophomore chemi-
cal engineering majors who have completed Advanced Chem-
istry I. Concurrently with this specially designed biology
course, students also are enrolled in a multidiciplinary engi-
neering course that has a biological component.

Biological Systems andApplications Course The BS&A
course was designed to meet the following four objectives:
1. To provide engineering students with a basic understanding of
fundamental biological principles and a working vocabulary
that would enable them to expand their knowledge base
during their academic and professional careers.
2. To convey to the students an appreciation of the wide variety
of engineering applications that are related to the fields of
biochemistry, cell biology, genetics, general microbiology,
and environmental microbiology.
3. To provide laboratory experiences that teach "hands-on"
mechanical skills such as micropipetting and culturing
techniques.
4. To provide additional laboratory experiences that collectively
instill in the students a general "biology common sense" that
can be applied to work in any microbiology laboratory or
project.
Beyond the basics of cell and membrane biology, highlights
in the lecture portion of the course include units about prokary-
otic and eukaryotic regulatory systems, biotechnology,
genomics, microbial growth and nutrition, the physiological
diversity of microbes, environmental microbiology, indus-
trial microbiology applications and concerns, and food mi-


Heat
Exchanger

Schematic representation of the brewing process showing
major process steps.


Chemical Engineering Education


irley









crobiology. The laboratory exercises in the course are all de-
voted to skill development and/or directly connected to lec-
ture topics.116' The BS&A course has benefited from using
the "project-based" learning approach1'7 and from strong
interactivity between chemical engineering and biological sci-
ences faculties. Extensive assessment data demonstrating the
effectiveness of the course have been presented elsewhere."'61
Microbial Fuel Cell Semester Project Chemical engi-
neering students who are taking the BS&A course also take
Sophomore Clinic I, a multidisciplinary engineering design-
and-practice course that is integrated with technical commu-
nications. It combines a 1-credit multidisciplinary engineer-
ing laboratory with the 3-credit college composition and rheto-
ric requirement and is co-taught by engineering faculty and
composition/rhetoric faculty.118' The 1-credit lab for the course
includes a semester-long project in which student teams de-
sign and create a microbial fuel cell(MFC) that powers a



TABLE 2
Biological Principles and Topics
Related to Microbial Fuel-Cell Design
Principle/Topic Where Applicable
Reaction stoichiometry/yield Calculating theoretical yield of electrons based on
initial glucose concentrations
Metabolic pathways Investigating pathways for metabolism of glucose
Yeast growth curve, doubling Determining "feeding time" for cell to
time, fermentation continuously produce current
Fermentation monitoring Current and glucose monitoring


Glucose e


Potassium
Ferricsunide
solution

4 Fe F '
e- 7
Fe(ClW


cels Electrodes




Cation exchange
membrane (Nafiona)

Figure 3. Microbial fuel cell schematic, adapted from Ref. 20.


Lego Mindstorms robot. The project combines mechanical,
chemical, civil/environmental, and electrical/computer engi-
neering skills. Students determine how changing certain fuel
cell parameters and conditions affect voltage and current, then
construct a Lego Mindstorms robot that derives its energy
from a MFC stack. The project reinforces concepts from ear-
lier courses such as chemistry, biology, and physics.
Fuel cell technology and alternative energy sources such
as biofuels and photovoltaics are developing technologies that
are exciting to students. MFCs operate on the same principles
as the more widely used (and more powerful) hydrogen fuel
cells. Rather than a nonrenewable source such as natural gas,
however, MFCs use biomass as the substrate and microor-
ganisms as the catalyst. While MFCs in which various types
of substrates and waste products are converted to energy by a
range of microorganisms, this project focused on yeast as the
catalyst and glucose as the primary substrate.
The project was inspired by the University of South
Florida's research on the "Gastrobot," (a self-sustaining, semi-
autonomous robot'19,201) and educational materials available
from the National Centre for Biotechnology Education at the
University of Reading.[21'241 This combination of readily avail-
able educational kits and supplies (see reading.ac.uk> for supplies) and accessible literature (see
) that de-
scribes cutting-edge research makes the project feasible yet
stimulating for the students.
A microbial fuel cell takes advantage of the metabolic reac-
tions of microbes to generate electricity. Organisms carry out


the following respiration reaction:

C6H1206 +602 ->6CO2 +6H20 (1)
to draw energy from food or carbohydrates.[21' The
above reaction can be broken down as shown in
Eqs. (2) and (3)

C6H1206+6H20->6CO2+24H++24e- (2)
602+24H++24e- -12H20 (3)

which follows the activity of electrons.
A redox mediator (methylene blue) can traverse
the cell membrane and scavenge electrons from
intermediates in which the electrons are stored.
The mediator can then present these electrons to
an electrode, and if an electron sink is provided
(potassium ferricyanide), the circuit is completed.
This process is shown in Figure 3. Voltage and
current can be monitored using a multimeter. A
single microbial fuel cell is capable of producing
approximately 0.5 V.
The microbial fuel cell project introduces and
reinforces several key concepts (summarized in
Table 2): stoichiometry, cells as biocatalysts, cell


Spring 2004









metabolism, and modeling of the system to enhance design
and performance. Student teams are asked to design a micro-
bial fuel cell based on optimization experiments performed
with a prototype. They investigate the effect of glucose and
yeast concentration on voltage and current produced by the
cell. As the yeast consumes the glucose, the current produced
drops until no glucose remains and the cell is unable to pro-
duce current. By varying the initial amount of glucose and
yeast and plotting the concentrations of these two variables,
students are introduced to the kinetics of batch fermentation.

IMPACT ON THE CURRICULUM
The combination of experiments and modules at the fresh-
man and sophomore level and a Biological Systems & Ap-
plications course specifically designed for chemical engineers
helps prepare students for research and industrial projects at
the junior and senior levels. As part of the clinic sequence at
Rowan, students participate in sponsored research projects
during their junior and senior years. Each semester, students
work in multidisciplinary teams as part of a 2-credit course.
Project funding is provided through government or indus-
trial grants or sponsorships. Located in southern New Jersey,
Rowan interacts with many companies in the pharmaceutical
and food industries through its junior/senior clinic sequence.
Accordingly, the chemical engineering faculty's expertise
reflects this bio-intensive regional interest; over half of our
faculty have training in biomedical, bioprocess, and biotech-
nology fields. This research interest is reflected in the types
of projects offered in the junior/senior clinic, from experi-
ment development in drug delivery to the effect of packaging
conditions on product spoilage.
Students often cite a potential career in biochemical engi-
neering as the motivator for pursuing a chemical engineering
degree. This interest in the interplay between biology and
engineering is apparent in the student demand for bio-ori-
ented research projects at the junior and senior levels. One
measure of student interest in bio-related projects is their par-
ticipation in the University's Science, Technology, Engineer-
ing, and Math (STEM) Symposium. As shown in Figure 4,
the percentage of bio-related engineering projects presented
at the symposium has increased from 10% in 1998 to al-
most 50% in 2002.
In some cases, such as the Food Microbiology laboratory
exercise, the lab component of the BS&A course has directly
benefited students working on research projects. In an indus-
trially sponsored clinic project in which the effect of packag-
ing on food spoilage was studied, a (junior) student who had
taken the BS&A course became the leader of the group and
taught the other members (two seniors) the techniques nec-
essary for determining bacterial counts. This student's expe-
rience in the course helped the group hit the ground running.
As the beginning cadre of students who have been exposed
to these innovations in the curriculum progress, we expect to


see similar results in future junior/senior clinic projects, as
well as in engineering courses that have a biotechnology or
bioengineering component.
Like many chemical engineering schools, Rowan offers
several bio-related technical electives at the senior and gradu-
ate-student level on a rotating basis. Students may enroll in
courses on drug delivery, biomedical engineering, or bio-
chemical engineering. Other courses at this level, such as food
engineering, polymer engineering, and membrane process
technology, have increasing applications in the biotechnol-
ogy and pharmaceutical industries. Preparing students with a
course in biological systems and bio-related modules during
their freshman and sophomore years aids in developing ma-
terial for these courses that explore the link between biology
and chemical engineering at a much deeper level. Addition-
ally, we anticipate being able to cover more material and ap-
plications in the time that is traditionally spent in an intro-
duction to biological principles.


Figure 2. Fermentable sugars in the wort from
malted and unmalted barley: M=Maltose;
S=Sucrose; G=Glucose


Figure 4. Number of bio-oriented abstracts
and total abstracts submitted by engineering
students at Rowan's STEM Symposium.


Chemical Engineering Education












































ACKNOWLEDGMENTS

Funding for development of the biomedical experiment in
the freshman year was supported by the National Science
Foundation Division of Undergraduate Education (NSF DUE
#0088437). Nafion membrane for the microbial fuel cell
project was generously donated by DuPont.


REFERENCES

1. Baum, R.M., "The Engineering Approach to Molecular Biology,"
Chem. Eng. News, 76(13), 38 (1998)
2. Breslow, R., "Into the Future," Chem. Eng. News, 78(47), 5 (2000)
3. Newell, J.A., H.L. Newell, and K.D. Dahm, "Rubric Development
and Inter-Rater Reliability Issues in Assessing Learning Outcomes,"
Chem. Eng. Ed., 36(3) (2002)
4. Rawls, R.L., "Biochem Meets Engineering," Chem. Eng. News, 77(35),
53(1999)
5. Westmoreland, P.R., "Chemistry and Life Sciences in a New Vision of
Chemical Engineering," Proc. Ann. Meet. of AIChE, Los Angeles
(2000)
6. Lauffenburger, D.A., "A Course in Cellular Bioengineering," Chem.
Eng. Ed., 23(4), 208 (1989)
7. Oerther, D.B., "Introducing Molecular Biology to Environmental En-
gineers Through Development of a New Course," Chem. Eng. Ed.,
36(4), 258 (2002)
8. Worden, M., and D. Briedis, "Training in Multidisciplinarianism," in
Proc. 2000 ASEE Ann. Conf., St. Louis, MO (2000)
9. Farrell, S., R. Hesketh, and E.C. Chaloupka, "Exercise in Chemical
Engineering for Freshmen," in Proc. ASEEAnn. Conf., Albuquerque,
NM (2001)
10. Hesketh, R., and C.S. Slater, "Demonstration of Chemical Engineer-
ing Principles to a Multidisciplinary Engineering Audience," in Proc.


Spring 2004


ASEEAnn, Conf, Seattle, WA (1997)
11. Marchese, A.J., et al., "Design in the Rowan University Freshman
Clinic," in Proc. ASEEAnn. Conf, Seattle, WA (1997)
12. Hesketh, R.P., et al., "Multidisciplinary Experimental Experiences in
the Freshman Clinic at Rowan University," in Proc. ASEEAnn. Conf.,
Seattle, WA (1997)
13. Ramachandran, R.P., J.L. Schmalzel, and S. Mandayam, "Engineer-
ing Principles of an Electric Toothbrush," in Proc ASEE Ann. Conf.,
Charlotte, NC (1999)
14. Farrell, S., J.A. Newell, and M.J. Savelski, "Introducing Chemical
Engineering Students to Product Design Through the Investigation of
Commercial Beer," Chem. Eng. Ed., 36(2), (2002)
15. Farrell, S., et al., "Introducing Freshmen to Reverse Process Engi-
neering and Design Through Investigation of the Brewing Process,"
Int. J. Eng. Ed., 17(6), (2001)
16. Hecht, G.B., P. Mosto, and C.S. Slater, "Effectiveness of an Applied
Microbiology Course Specifically Designed for Chemical Engineer-
ing Majors," Microbio. Ed., accepted for publication (2002)
17. Mosto, P., "Project Based Learning: Reflections," The Communique,
5,2(2001)
18. Hollar, K., et al., "Bugbots! A Multidisciplinary Design Project for
Engineering Students," in Proc. ASEEAnn Conf, Montreal (2002)
19. Wilkinson, S., 'Gastrobots'-Benefits and Challenges of Microbial
Fuel Cells in Food Powered Robot Applications," Autonomous Ro-
bots, 9, 99 (2000)
20. Wilkinson, S. 'Gastronome'-A Pioneering Food Powered Mobile
Robot," in lASTED Internat. Conf, Honolulu (2000)
21. Bennetto, H.P., et al., Operation ofMicrobial Fuel Cells, King's Col-
lege, London, Great Britain (1987)
22. Bennetto, H.P., "Electricity Generation by Micro-Organisms," Biotech.
Ed., 1(4), 163 (1990)
23. Bennetto, H.P., "Microbial Fuel Cells," Life Chem. Repts., 2,363 (1984)
24. Allen, R.M., H.P. Bennetto, "Microbial Fuel Cells: Electricity Pro-
duced from Carbohydrates," App. Biochem. Biotech., 39/40, (1993) p

113


CALL FOR PAPERS


Fall 2004 Graduate Education Issue of


Chemical Engineering Education


We invite articles on graduate education and research for our Fall 2004 issue. If you are interested in contributing,
please send us your name, the subject of the contribution, and the tentative date of submission.


Deadline is June 1, 2004

Respond to

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

Phone and Fax: 352-392-0861
e-mail: cee@che.ufl.edu









Random Thoughts...



WE HOLD THESE TRUTHS

TO BE SELF-EVIDENT


In 19911 wrote a column listing several widely accepted academic mythsfor which I've never
seen a shred of evidence. I glanced over it today to see if anything has changed in the interven-
ing 13 years. It hasn't. So, I'll re-run it. Tune in again in 13 years.
Rich


RICHARD M. FIELDER
North Carolina State University Raleigh, NC 27695-7905


Being engineering professors, we all know about the
need to make assumptions... and we also know that
if the assumptions are invalid, the results can be
worthless. We learn early in our careers to check our results
(Does the model fit the data? Does the algorithm converge?
Does the product meet quality specifications?) and if they
are not satisfactory, to question our assumptions (Is the solu-
tion ideal? Is the reactor isothermal? Is flow laminar?), and
we try to develop the same critical, questioning mentality in
our students.
When it comes to education, however, our mentality
changes. We generally do whatever it is we do without much
critical evaluation of how well or how poorly it is working,
and we accept without question what Armando Rugarcia'l1
calls academic myths assumptions that have never been
shown to have any basis in reality and often defy common
sense. Here are some of them.

MYTHS ABOUT ...
FACULTY RECRUITMENT
People who (1) don't have Ph.D.s or (2) have spent
their careers in industry and have no research publica-
tions, are not qualified to be engineering professors.
When filling faculty vacancies, an engineering depart-
ment benefits most by selecting the candidates in the
hottest and years is more important than whether they
know enough engineering to teach the core courses and
to change research areas if their present one goes out of
fashion.
The best way to handle required courses that no one wants
to teach, such as the unit operations laboratory or the


capstone design course, is to rotate the faculty so that no
one gets stuck with them too often. An inferior solution
is to fill a vacant faculty position with someone who has
the desire to teach these courses and the expertise to teach
them well.
When selecting a department head, the faculty benefits
most by choosing the candidate with the strongest re-
search record, regardless of administrative experience
or ability. How he or she runs the department in the next
five to ten years is less important that what he or she
does in research after that.

MYTHS ABOUT ...
RESEARCH AND TEACHING
Excellence in research and excellence in teaching are
highly correlated.
Requiring EVERY faculty member to build up a strong
research program as a condition for promotion and ten-
ure is in the students' (professors', department's) best in-
terests.
Excusing new professors from teaching responsibilities

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


Copyright ChE Division of ASEE 2004


Chemical Engineering Education











so they can write proposals is a good thing to do. Excus-
ing them from research responsibilities so they can de-
velop a couple of good courses makes no sense.

Professors who are excellent at research and mediocre-
to-adequate at teaching deserve tenure. Professors who
are excellent at teaching and mediocre-to-adequate at
research don't.

MYTHS ABOUT...
CURRICULUM DESIGN AND PEDAGOGY
Our graduates routinely say they never use 90% of what
we taught them. Since we're engineering professors, 90%
of what they're doing must not be engineering.
It makes sense educationally to teach students a gener-
alized theory (e.g., transport theory) before teaching them
anything about the specific phenomena and devices that
the theory was invented to describe (e.g., unit operations).
Tensor calculus, quantum chemistry, and statistical me-
chanics should be taught to every chemical engineering
undergraduate; statistical process control, project man-
agement, and technical writing they can pick up on their
own-there's no room for those subjects in our crowded
curriculum.
The best thing to do with ethics, safety, environmental
science, and all those other important things ABET says
we have to teach, is stick them all in the capstone design
course.
I accomplish something useful when I spendfifty min-
utes in class writing detailed derivations on the chalk-
board for the students to copy.
We can't teach students to think critically or creatively-
they can either do it or they can't.
Students who complain that our lectures have nothing to
do with the real world don't know anything about the
real world-and we do.
If I have covered the syllabus, I have done my job suc-
cessfully.

MYTHS ABOUT ...
EVALUATION OF STUDENTS (GRADING)
How well our students will do as engineers correlates
highly with (a) their undergraduate GPA; (b) their abil-
ity to solve problems with unfamiliar twists on 50-minute


exams; (c) anything else that we typically use to evalu-
ate them.

An average score of 40 on myfinal exam proves (a) I set
high standards; (b) they didn't understand the material.
There is no possibility that it proves (c) the test was lousy.
An average score of 85 on yourfinal exam proves (a) it
was a trivial test; (b) you're a soft grader; (c) there was
widespread cheating. There is no possibility that the re-
sult proves (d) they learned the material.

MYTHS ABOUT ...
EVALUATION OF TEACHING

All methods of evaluating teaching are unreliable, and
student evaluations are the most unreliable of all.

If you get consistently outstanding student evaluations,
it must be because you are (a) an easy grader; (b) an
"entertainer It is certainly not because you are (c) an
outstanding teacher.
If I get rotten student evaluations it is because (a) the
students are ignorant and lazy; (b) I don't water down
the material for them; (c) they don't understand what I
am doing for them now but in later years they'll come
back and thank me. It is definitely not because (d) I am
doing a rotten teaching job.

I could go on, but you get the idea.
When I classify these points as myths I am not saying there's
nothing to them; it's just that as far as I know they've never
been scientifically or even empirically validated. (Mention-
ing someone who is great at both teaching and research, for
instance, doesn't quite do it.) If you can justify one or an-
other of these assumptions, let me know and I'll set the record
straight.* If, on the other hand, you conclude that the assump-
tions might be faulty, then how about considering whether
some alternative assumptions might lead to better ways of
doing things? Couldn't hurt.
References
1. Rugarcia, A. "The Link Between Teaching and Research: Myth or
Possibility?" Engineering Ed., 81, 20 (1991)
2. McKeachie, W.J., Teaching Tips: A Guidebookfor the Beginning Col-
lege Teacher, 8th ed., Toronto, D.C. Heath & Co. (1996)
Before you attempt it, though, you might want to check out the literature:
McKeachie 2' provides invaluable summaries of the research on most of
the topics in question, and Rugarcia'" makes some interesting points
specifically on the research/teaching dichotomy. O


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


Spring 2004









[e, M classroom


IMPROVING COHERENCE IN

TECHNICAL WRITING




G.K. SURAISHKUMAR
Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India


t is a well-recognized fact that effective technical com-
munication is a skill that graduate/undergraduate students
in engineering disciplines need to develop, and many
good suggestions have been made[1-6] to improve it. Although
good communication involves good skills in writing, speak-
ing, reading, and listening ( ac-
cessed on 3 Feb 2003), we usually concentrate on develop-
ing only the speaking (presentation) and writing skills in stu-
dents. Between the two, it is usually more difficult to de-
velop good writing skills, probably because it requires higher
clarity and rigor in the thought processes. Furthermore, there
is a general notion that one learns to write well in the same
way that one learns to ride a bicycle, to play a musical instru-
ment,171 or to swim-in an experiential manner.
When graduate students write their first manuscripts, spell-
ing and grammatical aspects are addressed first, either by a
person comfortable with the language or by a word proces-
sor. A grammatically correct document may not always read
well, however. Often, advisers know that the manuscript is
not written well, but cannot clearly explain why. They tend
to talk about "clarity" and "style" (as distinct from that de-
scribed in the Chicago or the American Chemical Society
style manuals), and further confuse the student. Ultimately,
the adviser might say something like, "It's all in there-you
just need to communicate it better," and without further di-
rection, ask the student to rewrite the paper.
Students typically rewrite their first manuscript many times,
and the writing improves intuitively with each progression.
Throughout this process, it is helpful if the student reads well-
written scientific literature.18' When the adviser finally ac-
cepts the manuscript, the student knows that the final draft is
written better, but is usually not aware of why the paper is
improved. These same students later become professors/ad-
visers, attesting to the fact that university teaching is prob-
ably the only skilled profession for which there is no formal


training ( accessed
on 3 Feb 2003), and the cycle continues.
This article presents a reasonably structured approach that
faculty members can use to improve the writing skills of stu-
dents. Alternatively, it also provides a direction for the not-
so-experienced writers of scientific material to consciously
improve their writing skills.

COHERENCE
Good communication, in either written or oral form, re-
sults from a solid knowledge of the subject, a clear aware-
ness of the aspects that need to be communicated (and those
that need to be left out), clear thinking, and good organiza-
tion, assuming that the language (grammar, spelling, and pro-
nunciation) are also addressed. For the most appropriate or-
ganization, one should be aware of how the reader/listener
will perceive the information. In other words, one needs the
ability to "tell a story," or the ability to coherently present
relevant material. In fact, many advisers ask the question,
"What is your story?" when they discuss a student's research
work. The above requirements are even more critical in writ-
ten communication because the communicator is not present
when the receiver reads the document, unlike an oral presen-
tation when the communicator can draw on nonverbal com-
munication cues to bolster the material.

G.K. Suraishkumar is currently Associate Pro-
fessorin the Chemical Engineering Department
at Indian Institute of Technology, Bombay. He
received his B. Tech from the Indian Institute of
STechnology, Madras, and his PhD from Drexel
University in Philadelphia. His current research
interest is in the area of free radical based im-
provements in the productivity of bioreactors.
He can be reached at iitb.ac.in>.
#


@ Copyright ChE Division of ASEE 2004


Chemical Engineering Education









A fairy tale is a good example of coherent presentation.
Children can listen to a fairy tale being read and completely
understand the story, even if they are sleepy at the time. If
engineering students could embrace the same simplicity and
coherence embodied in a fairy tale when they present scien-
tific material, their message would be more
easily understood.
This idea led to an exercise for students for
taking a M.Tech (graduate level) communi- clear awar
cation skills course at the Indian Institute of
Technology, Bombay. They were required to (an ose
narrate a fairy tale in front of the class. Many org
of them could not coherently narrate even a
simple fairy tale such as Cinderella the first
time they tried. One example: "Cinderella is that girl who
wore a shoe. The Prince found her with the shoe. She went to
the ball in a chariot made from a pumpkin. There was a fairy
that helped her. While running away from the ball she lost
one of her shoes. Cinderella had a wicked stepmother and
two stepsisters...." A chronological organization of infor-
mation, a crucial requirement for delivery of the tale, was
completely absent.
Faculty members quite often find the same type of disor-
ganization in a first draft of a report or manuscript written by
a student. The presentation is a recital of mere facts, written
in the order of recall without bothering about the relationship
between them.
There is an important distinction between a fairy tale and a
scientific narrative, however-the concept of external and
internal times.9,'01 External time refers to the time taken for
the actual presentation and internal time refers to the dura-
tion of the sequence of events in the presentation. For ex-
ample, if narrating the Cinderella story takes fifteen minutes,
then the external time is fifteen minutes; the internal time is
the days or months or years over which the story is set. While
external time is relevant for both types of presentations, the
internal time is normally absent in a scientific narrative (ex-
cept, perhaps, in a background/introduction section). Instead
of a chronological detailing of internal time, a logical sequence
of scientific information (facts, graphs, tables, derived in-
formation, discussion, etc.) is present. Students usually
develop this ability to present material in a logical se-
quence through experience.

TOOLS FOR COHERENCE
Students can be encouraged to develop a logical sequence
of presentation, as described in a later section. Here we will
consider the tools that can be used to improve coherence (see
ac-
cessed 3 Feb 2003) and communication of the material after
it is logically sequenced.
To appreciate the use of tools, consider this well-written
passage by Bird,E"1 which has been slightly adapted:


In educational circles today we hear a great deal about
teaching and research. However, we hear very little about
the activity of book-writing, which ought to be included as a
third principal activity of a university teacher since it is
concerned directly with the production, evaluation,
organization, and dissemination of new knowledge.

Good communication, in either written or oral
n, results from a solid knowledge of the subject, a
*ness of the aspects that need to be communicated
that need to be left out), clear thinking, and good
anization, assuming that the language (grammar,
spelling, and pronunciation) are also addressed.

Therefore, I thought it might be useful to use this lecture to
focus attention on the "rites, rewards, and responsibilities"
of book authorship. Since I have had the pleasure and good
fortune to coauthor several books, perhaps I can offer some
appropriate words of encouragement to aspiring writers and
even a few helpful suggestions regarding the art of writing.
Maybe I can help others avoid some of the mistakes I've
made. From time to time I will cite specific personal
experiences in order to avoid discussing the problems of
authorship in the abstract.
What Kind of Books do Chemical Engineers Need?
A library of professional volumes includes various classes
of books: (i) edited volumes to present very recent develop-
ments by teams of experts; (ii) research monographs to
catalog and evaluate the research published in the preceding
5-10 years; (iii) treatises to give authoritative, encyclopedic
coverage to one particular topic; (iv) textbooks to set forth
the basic ideas in the field in a form suitable for students;
(v) handbooks to summarize standard results of widespread
use; and (vi) design manuals to provide up-to-date proce-
dures for practicing engineers. Each of these categories has
a different audience, and each requires special organiza-
tional talents. Generally speaking, there is a flow of
information from (i) toward (vi) in the above listing-that
is, from innovative, exploratory, and (sometimes) impracti-
cal ideas of the researcher all the way to the time-tested and
reliable tools of the practitioner. Along the way many ideas
and methods are inevitably discarded, and only the most
useful material survives to the arena of industrial practice.
But without this constant exploration of new ideas and
subsequent filtration, a profession can stagnate and atrophy.

Repetition is an important tool for improving coherence.
In the above example, Bird repeats the word books(s" in a
few places to build coherence.
If repetition becomes boring, synonymy can be used; e.g.,
Bird uses the word "volumes" to avoid a tiresome repetition
of the word "books" in the first sentence of the last para-
graph. Similarly, antonymy (using an opposite word) can
improve coherence; e.g., see the use of "impractical ideas"
and "reliable tools" in the same sentence in the last para-
graph.


Spring 2004









* The pronoun is commonly used to improve coherence
between sentences; e.g., the pronoun "it" is used to refer to
"book-writing" in the first paragraph. Also, parallelism,
which refers to the use of the same sentence structure in sub-
sequent sentences, improves coherence.
* A tool that is commonly used by engineers is enumera-
tion, which refers to the use of specific markers of sequence
to achieve a connection between thoughts. A good example
of enumeration appears in the second paragraph in which Bird
uses enumeration to link the various classes of books.
* A tool that students learn easily is transition. Transitions
are conjunctions or conjunctive adverbs that link sentences
with specific logical relationships. They can be subcategorized
according to their meaning (see engl_126/bookl26.htm> accessed on 3 Feb 2003) as follows:

Identity Indicates sameness (that is, in other words)
Opposition Indicates contrast (but, yet, however,
nevertheless, still, though, although, whereas, in
contrast, rather)
Addition Indicates continuation (and, too, also,
furthermore, moreover, in addition, besides, in the
same way, again, another, similarly, a similar, the
same)
Cause and effect (therefore, so, consequently, as a
consequence, thus, as a result, hence, it follows that,
because, since, for)
Indefinites Indicates a logical connection of an
unspecified type (in fact, indeed, now)
Concession Indicates a willingness to consider the
other side (admittedly, I admit, true, I grant, of course,
naturally, some believe, it has been claimed that, once
it was believed, there are those who would say)
Exemplification Indicates a shift from a more general
or abstract idea to a more specific or concrete idea (for
example, for instance, after all, an illustration of, even,
indeed, in fact, it is true, of course, specifically, to be
specific, that is, to illustrate, truly)

It is easy to identify the use of transitions in the Bird ex-
ample. These coherence tools can be used to improve scien-
tific writing. While it is unlikely that Bird was aware of the
coherence tools he used while writing, a faculty member
should encourage a novice writer to use the tools consciously
until a time when they become a subconscious part of the
writing process.

STRUCTURED APPROACH
FOR COHERENCE
Coherence tools can only help improve something that was
initially reasonably well written. The following structured ap-


proach is one way by which well-written drafts can be
achieved. It is neither a panacea nor the only way, however,
since there are innumerable factors that contribute to good
writing, including the writer's own personality.

The Preliminaries
1. The writer must have the requisite knowledge/informa-
tion before beginning to write. This is an absolute pre-
requisite.
2. If a manuscript is for journal publication, a thesis, or a
report, a substantial number of discussion aspects (say,
50%) should be clear to the student before writing be-
gins. The student should be encouraged to analyze sci-
entific material and note the salient points for discus-
sion, with clarity, before writing anything. Many first
manuscript drafts are poor in the discussion of data/simu-
lations.
3. The student should be relaxed; (s)he should be encour-
aged to drink a glass of water or to take a few deep
breaths, taking care to exhale more slowly than during
inhalation. Then the student should take a few blank
sheets, a pencil and an eraser (or a word processor), and
sit where (s)he will not be disturbed.

The Questions
Now the student should ask himself/herself the following
questions and incorporate the suggestions that follow them.
To illustrate, I will present an example of my own thought
processes while writing a paper a few years ago.
What is the main idea that I need to communicate?
For example, we had just discovered that induced
free radicals could improve the productivity of cells
in bioreactors, and we were very excited about it.
Therefore, the main idea that we wished to communi-
cate was "induced free radicals can be used as a novel
means to improve bioreactor productivity." Consider-
able thought may be required for first-time writers to
recognize the main idea to be communicated, but that
is the absolute starting place.
How do I communicate the main idea?
This is fairly simple for engineers/scientists who are
normally bound by the required format of a journal,
the university, or a funding agency. Typically, we are
required to communicate the main idea in various
sections, such as Introduction, Materials and Meth-
ods, Mathematical Model, Results and Discussion,
Conclusions, Nomenclature, References, Appendices,
etc. Also, we rarely use anything except a linear
presentation of information, which makes this aspect
very simple.
Taking one section at a time, ask the question, "What do I
want to communicate in this section?"


Chemical Engineering Education


118









Jotting It down
Write down the points as they occur to you.
For example, what do I want to communicate in the Intro-
duction of the manuscript on induced free radicals? We were
excited about the novelty of the strategy-therefore I needed
to communicate that. I also wanted to communicate the vari-
ous contributions we have made in this work. In addition, I
wanted to tell the readers what motivated us to do the work,
and since the typical reader of this journal was unlikely to
know much about free radicals, I needed to give the relevant
background on free radicals. Also, to provide focus, I
wanted to present the overall aim and objectives of the work.
If the information is given in the order above, the readers
(who will most probably not be familiar with the work) will
find it difficult to understand. But if the same information is
presented in a logical sequence, the reader's understanding
and the paper's readability would improve significantly.
Ordering
Put yourself in the reader's position and then logically or-
der the aspects written in bold in the above section. One pos-
sible list would be
1. Relevant background on free radicals
2. Motivation for the work
3. Novelty of the strategy
4. Overall aim and objectives, along with contributions
Jotting Down: Paragraph Level
Addressing the items in the list above, I first determined
what it was that I wanted to communicate to the reader and
noted the points down on a piece of paper as I thought of
them. They were that: 1) Free radicals can be expected to
improve bioreactor productivity; 2) They mediate cell pro-
cesses such as cancer, apoptosis, etc.; and 3) They are sus-
pected mediators of the effects of temperature, osmolarity,
and nutrient levels (important bioreactor variables) on cells.
Logical Ordering: Paragraph Level
If I had written the points down in the order listed above,
an intelligent lay-person (the common reader) would have
found my message difficult to understand. Therefore, putting
myself in the reader's position showed me that I needed to
reorder the points for better understanding. One possibility
was: 1) Free radicals are known to mediate a number of cell
processes, including apoptosis and cancer (citing references);
2) Free radicals are suspected of being mediators of the ef-
fects of temperature, osmolarity, and nutrient levels on cells
(citing references), which are also important bioreactor envi-
ronment variables; and 3) Free radicals can be expected to
play a significant role in determining bioreactor productiv-
ity. Note that in the process, I also improved the accuracy of
the information.
Linking Sentences: Paragraph Level
Next, the sentences should be linked to improve coherence,


Spring 2004


using the various linking tools mentioned earlier:
Free radicals are known to mediate a number of
significant cell processes, including apoptosis and cancer
(Feig and Loeb, 1994; Feig, et al., 1994: Okuno, et al.,
1998; Reid and Loeb, 1993). Further, free radicals are
suspected of being mediators of the effects of temperature,
osmolarity, and nutrient levels on cells (Nagarathnamma, et
al., 1997; Osbourn, et al., 1990), which are also important
bioreactor environment variables. Therefore, free radicals
can be expected to play a significant role in determining
bioreactor productivity.
Note that in addition to the transition tools indicated in bold,
I also used repetition when I framed the paragraph.
Now, the first paragraph, which communicated (and not
merely presented) the background, was ready.

Second Paragraph
Similarly (jotting down thoughts, ordering sentences, and
linking them) motivation could be communicated in the fol-
lowing paragraph:
Xanthan gum is secreted by Xanthomonas campestris
when it attacks plants (Chamnongpol, et al., 1995). The
extent of xanthan gum secretion (mucoidy) is directly related
to the pathogenicity of the organism on plants, which it
attacks (Throne, et al., 1987; Weiss, et al., 1994). Pathoge-
nicity is related to the induced oxy free radicals (Sutherland,
1991). From an industrial viewpoint, Xanthomonas
campestris is employed for commercial bioproduction of
xanthan gum, which has wide applications in food, pharma-
ceuticals, oil, and other industries (Lee, 1996). If the
relationship between free radical induction and gum
production is better understood, free radical induction may
be employed as a means to improve xanthan gum productiv-
ity. In addition, a better understanding will help to improve
cultivation strategies where oxygen is provided in situ
through the liquid-phase oxygen supply strategy (Sriram, et
al., 1998).
When the two "completed" paragraphs are read one after
the other, the reader will notice an abrupt jump in ideas be-
tween them. The paragraphs do not seem to be linked. The
first talks about free radicals and the second talks about
xanthan gum. The reader, who is subconsciously expecting a
link, will experience discomfort when (s)he does not find one,
and this leads to a loss in communication.

Linking Paragraphs
In the example above, the relationship between free radi-
cals and xanthan gum, especially from a production view-
point, was unknown in the literature at that time, and there-
fore known information could not be used to link the two
paragraphs. Given this constraint, how could the ideas be
linked?
The third sentence in the second paragraph talks about free
radicals and could thus be used as a connecting sentence.
Bringing this sentence to the beginning of the paragraph and
119









suitably modifying it serves the purpose of linking both para-
graphs.
Free radicals are known to mediate a number of
significant cell processes, including apoptosis and cancer
(Feig and Loeb, 1994; Feig, et al., 1994: Okuno, et al.,
1998; Reid and Loeb, 1993). Further, free radicals are
suspected of being mediators of the effects of temperature,
osmolarity, and nutrient levels on cells (Nagarathnamma, et
al., 1997; Osboum, et al., 1990), which are also important
bioreactor environment variables. Therefore, free radicals
can be expected to play a significant role in determining
bioreactor productivity.
Oxy free radicals and oxidative stress are important
aspects of plant defense mechanisms against invading
microorganisms (Chamnongpol, et al., 1995;
Sutherland, 1991) such as Xanthomonas campestris, a
plant pathogenic bacterium. Xanthan gum is secreted by
Xanthomonas campestris during its attack, and the extent
of xanthan gum secretion (mucoidy) is directly related to
the pathogenicity (Throne, et al., 1987; Weiss, et al., 1994).
From an industrial viewpoint, Xanthomonas campestris is
employed for commercial bioproduction of xanthan gum,
which has wide applications in food, pharmaceuticals, oil,
and other industries (Lee, 1996). If the relationship between
free radical induction and gum production is better
understood, free radical induction may be employed as a
means to improve xanthan gum productivity. In addition, a
better understanding will help to improve cultivation
strategies where oxygen is provided in situ through the
liquid-phase oxygen supply strategy (Sriram, et al., 1998).
In a similar fashion, I could proceed to compose other para-
graphs, linking them to produce a coherent document.
Summarizing the steps for improving coherence, the most
important factor is that the student needs to be knowledge-
able in the area and aware of the aspects that need to be com-
municated. Then the student needs to
Write down the points that (s)he needs to communicate
in each section, as they come to mind.
Order them logically
Improve coherence (by using tools)
Link paragraphs
Link sections/chapters, when needed

EFFECTIVENESS OF
THE STRUCTURED APPROACH
The structured approach was presented to students taking
a communication skills course at IIT Bombay. In addition,
they were given exercises to practice writing (and presenta-
tion). Their writing (and presentation skills) improved sig-
nificantly, and they ultimately expressed gratitude that such
a course was offered to them. Some students who were com-
fortable with the language were initially skeptical about the
utility of the course, but they learned that effective commu-
nication does not necessarily arise from an ability to write


correct grammar. At the end, these students also felt they sig-
nificantly benefited from the course.
Many faculty members (15 out of 28) expressed apprecia-
tion for, and satisfaction with, the improvement in communi-
cation that they observed in the M.Tech seminar course, where
students worked on a research area and presented a critical
evaluation of the literature through both a written and an oral
presentation. After being asked for their input, six other
faculty members said the course was useful. No negative
comments were received from the faculty or the students,
with the exception of certain individual preferences on
presentation style.
To summarize, through clear thinking and better organiza-
tion of information that is based on a sensitivity to the reader's
needs, better writing can be achieved. Very often, bad writ-
ing results from muddled thinking or an inability to perceive
the reader's needs. Further, writing is a skill-as is swim-
ming. One cannot expect a person who does not know how
to swim to learn from a set of verbal/written instructions
alone-a lot of practice is required. Similarly, good writing
requires a lot of practice. The structured approach given in
this article cannot obviate that requirement, but it can pro-
vide direction for practice.

ACKNOWLEDGMENT
I thank my colleagues, Professors Preeti Aghalayam,
Hemant Nanavati, Kartic Khilar, and Santosh Noronha, as
well as my graduate student, Susmita Sahoo, for their input.

REFERENCES
1. Woods, D.R., R.M. Felder, A. Rugarcia, and J.E. Stice, "The Future of
Engineering Education: Part 3. Developing Critical Skills," Chem. Eng.
Ed., 34, 108 (2000)
2. Nirdosh, I., "Making Successful Oral Presentations: A Guide," Chem.
Eng. Ed., 31, 52 (1997)
3. Newell, J.A., D.K. Ludlow, and S.P.K. Sternberg, "Development of
Oral and Written Communication Skills Across an Integrated Labora-
tory Sequence," Chem. Eng. Ed., 31, 116 (1997)
4. Prausnitz, M.R., and M.J. Bradley, "Effective Communication for Pro-
fessional Engineering Beyond Problem Sets and Lab Reports," Chem.
Eng. Ed., 34, 234 (2000)
5. Sureshkumar, G.K., and K.C. Khilar, "On Improving 'Thought With
Hands'," Chem. Eng. Ed., 36, 292 (2002)
6. Bendrich, G., "Just a Communications Course? Or Training for Life
after the University," Chem. Eng. Ed., 32, 84 (1998)
7. Van Ness, H.C., and M.M. Abbott, "Technical Prose: English or
Techlish?" Chem. Eng. Ed., 11, 154 (1977)
8. Haile, J.M., "Easy Writing Makes Hard Reading," Chem. Eng. Ed.,
28, 278(1994)
9. Chatman, S., Coming to Terms: The Rhetoric of Narrative in Fiction
and Film, Comell University Press, Ithaca, NY (1990)
10. Abbott, H.P., The Cambridge Introduction to Narrative, Cambridge
University Press, Cambridge, UK (2002)
11. Bird, R.B., "Book Writing and Chemical Engineering Education: Rites,
Rewards, and Responsibilities," Chem. Eng. Ed., 17, 184 (1983)
12. ManjulaRao, Y., and G.K. Suraishkumar, "Improvement in Bioreactor
Productivities Using Free Radicals: HOCl-Induced Overproduction
of Xanthan Gum from Xanthomonas campestris and Its Mechanism,"
Biotechnol. Bioeng., 72, 62 (2001) O


Chemical Engineering Education












r teaching tips


LLOYD WHILE
California State University, Long Beach Long Beach, CA 90840-5003


To encourage development of collaboration skills,
many classes are split into teams of students to work
on projects. Common methods of selecting the team
composition are

1) To let students form their own groups,
2) To have the instructor assign team members,
or
3) To randomly choose members.

A disadvantage to student-selected groups is that of-
ten the better students choose each other and the weak-
est students end up together; this approach can also
hinder achieving racial and gender diversity within a
group. Instructor-selection may lead to complaints of
favoritism, and random-selection leaves too much to
chance and is not a method that would be used in the
"real world."
For several years we have used a novel approach for
forming teams. We require students to bring a one-page
resume to class with them. The students' names are cov-
ered with Post-It notes and the resumes are laid out on
a table so all can be viewed. "Team selectors" then study
them and pick up the resumes of the students to be on
their team. They do not know the students' identities at
that point and make their selection based solely on the
skills displayed in the resume.
The team selectors could be chosen by a random
method, but we prefer to have the instructor identify
those with the best grades in pre-requisite subjects to


ensure there is at least one strong student per group. These
selectors then randomly draw a number from a hat that
will serve both as their team number and the order of
selection. For example, selector #1 will form team #1
and will pick up the first resume; selector #2 will form
team #2 and will have second choice; the last team se-
lector will have last choice, but will immediately select
again (i.e., pick up two resumes), and then the selection
sequence reverses. This process continues until all the
students have been selected and teams of the desired size
are formed. Identities are not revealed until all resumes
have been picked up, at which point the Post-Its are re-
moved from each resume.
An advantage of this method is that students have taken
an active role in forming their team, which is similar to
the approach companies use to select job applicants. We
take some time to discuss the reasons students had for
making their selections. This is an excellent way to il-
lustrate the importance of a good resume.
Many students have never attempted to prepare a re-
sume prior to this requirement, so we provide some sug-
gestions to help them. A good starting point is the "Re-
sume Wizard" in Microsoft Word, which takes a step-
by-step approach to creating a resume. Also, most col-
leges have career centers that provide helpful informa-
tion on this topic. An added bonus is that the self-analy-
sis needed to construct a resume makes students aware
of shortcomings in their skills and experience and pro-
vides incentive for them to work on filling those gaps as
they progress through their studies.


This new feature in CEE, "Teaching Tips," is intended to be a forum for sharing innova-
tive teaching practices with others in the profession. These short (600 words or less) con-
tributions should clearly state the innovation, summarize the evidence for its success, and
offer guidelines for its implementation. Contributions will be reviewed for originality, gen-
eral interest, and appropriateness. They should be sent to Professor Phillip C. Wankat,
Chemical Engineering Department, 480 Stadium Mall Drive, Purdue University, West
Lafayette, IN 47907-2100, or .
^ ____________________ __________________


Spring 2004









rim" 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 fourteen double-spaced pages and should
be accompanied by the originals of any figures or photographs. Please submit them to Professor
James O. Wilkes (e-mail: wilkes@umich.edu), Chemical Engineering Department, University
of Michigan, Ann Arbor, MI 48109-2136.




A SIMPLE OPEN-ENDED

VAPOR DIFFUSION EXPERIMENT


DAVID WHITMIRE, WAYNE BLAYLOCK*
Tennessee Technological University Cookeville, TN 38505


he value of problem-based learning in undergraduate
chemical engineering instruction is well established.'1
But even with publication of a simple mass-transfer
experiment,[21 de Nevers' statement that, "A permanent prob-
lem in engineering education is to find simple, portable, low-
cost classroom demonstrations of the principles we present
to our students," remains true, and the problem appears par-
ticularly acute with respect to mass-transfer instruction.[31
In an attempt to enhance learning in a lecture course on
mass transfer, we developed a laboratory procedure that pro-
vides an opportunity for students' experimental estimation
of vapor-phase diffusion coefficients. They can also cre-
atively expand their analyses and receive credit for their
creative efforts.
Therefore, inspired by a schematic from Bird, et al.,141 we
developed a mass transfer experiment requiring a simple and
inexpensive laboratory procedure. Simplicity and very low
cost of the laboratory procedure notwithstanding, the great-
est value of the experiment lies in the breadth and depth of
analyses to which students can subject the resulting data. The
overall objective of the present work was to develop a simple
and inexpensive mass transfer experiment suitable for open-
ended analysis by undergraduates in a unit operations course
on mass transfer. Specific aims of the development were sim-
plicity and low cost, use of common equipment and supplies

* Address: ChE Process Design Group, BWXT Y-12, LLC, Oak Ridge, TN
37831


that require no expertise, and certain yield of usable data suit-
able for open-ended analysis by undergraduate students.

MATERIALS AND METHODS
Presentation to the Class
Diffusion is the initial topic in the third lecture course in
our three-course unit operations sequence. To reinforce the
topic effectively, students are asked to perform a class ex-
periment on diffusion. During the last 20 minutes of a lec-
ture, the assignment is distributed to the class along with Lee
and Wilke's paper15" that details an elaborate version of the
same experiment. The assignment is presented as a brief dis-

David Whitmire is Associate Professor of
Chemical Engineering at Tennessee Tech
University. He received his BS from Clemson
University in 1971, his MS from Virginia Tech
University in 1978, and his PhD from Auburn
University in 1988, all in chemical engineer-
ing. Research interests include biochemistry
of alcohol metabolism and human learning.




Wayne Blaylock received his BS in Chemi-
cal Engineering from Tennessee Technologi-
cal University in 2003. He is currently em-
ployed by BWXT Y-12 as a design engineer
at the Y- 12 National Security Complex in Oak
Ridge, TN.


Copyright ChE Division of ASEE 2004


Chemical Engineering Education









cussion of the background of the experiment and the work-
ing equations detailed below. Students are told that the "mini-
mum base-case" analysis is to use the experimental data with
the working equations to estimate the vapor diffusion coeffi-
cient for each of three pure liquids evaporating from gradu-
ated cylinders and to submit an individually prepared written
report of the results.
After discussing the assignment, the class reconvenes in a
lab room to become familiar with the experimental setup,
which consists of an ordinary box fan, nine 100-ml gradu-
ated cylinders, hexane, acetone, distilled water, and a ther-
mometer (see Figure 1). Students are told to use the "low"
fan speed and to measure the volumetric disappearance of
the three liquids from each of three graduated cylinders. Three
cylinders of each liquid are included as an additional vari-
able for students to consider; the students are given no
instruction on the purpose of having more than one cylin-
der for each liquid. Optionally, an electronic balance can
also be used for another measure of disappearance of the
three liquids.
Students then organize themselves for data collection and
e-mail data reporting over the ensuing two-to-three-day pe-
riod required for significant evaporation of water (hexane and
acetone evaporate more quickly than the water).
An Open-Ended Assignment
Since the assignment is open-ended, students are told that
they may creatively expand their analysis of the experimen-
tal data and receive unlimited additional credit for their extra
work. Since no particular "correct" answer is sought even in
the base case, the assignment is open-ended relative to tech-
nical scope and to grading. Student ideas for analyses be-
yond the base case are discussed with individual students only
if discussion is initiated by the student.
Working Equations
Equation (1) is used to compute the experimental diffusion
flux of each species:

NA= WA (1)
NA W



Figure 1. Sche-
matic top-view of 0 0 0 Hexane
experimental ap-
paratus. A com- F Air Flow
mon boxfan on low A 1 Q Acetone
speed directs a N
stream of air over Air Flow
the tops of gradu- 0 0 0 Water
ated cylinders con-
taining hexane, ac-
etone, and water. Three cylinders of each liquid were used
to provide additional opportunities for student analysis.


where
NA molar flux ofA (mole/hr-cm2)
WA mass of component A vaporized (grams)
MA molecular weight of A
0 elapsed time (hours)
S cross-sectional area of graduated cylinder (cm2)

Based on Figure 2, a diffusion coefficient is defined using
the integrated form of Fick's law:

NA= DPAp (2)
RTpfx

where
D diffusion coefficient for A (cm2/s)
P atmospheric pressure (atm)
Ap difference between the partial pressure of A at the liquid-
gas interface and the partial pressure of A in air at the
mouth of the cylinder (atm)
R universal gas constant (cm3 atm/mole/K)
T ambient temperature (K)
p, log-mean partial pressure difference (atm)
x effective length of the diffusion path (cm)

Assuming Raoult's-law behavior, the vapor pressure of liq-
uid A, PA' at the liquid-gas interface was substituted for the
partial pressure of A, PA, at the same location. Thus the log-
mean partial pressure difference, p, becomes

(P-o)-(P-PA) (3)
Pf = (3)
SP-PA

where
PA vapor pressure of A (atm)
po partial pressure of A in air above the cylinder (atm)
P total pressure of gas above the cylinder (atm)

Because an average molar diffusive flux for the period of
observation was used, the average length, xa, cm, of the dif-
fusion path over the period of observation was also used:


AIR FLOW AIR FLOW


T t
Xa X



L 7


Figure 2. (Left)
Schematic of air
stream passing over
the top of a gradu-
ated cylinder con-
taining vaporizing
liquid (cross-
hatched region).
(Right) Schematic of
graduated cylinder
with turbulence indi-


cated at the cylinder top and a curved meniscus at the liq-
uid surface. xais the apparent distance of the diffusion path.
dxc is a correction to the apparent diffusion path due to
putative turbulence; dxs is a correction to the apparent dif-
fusion path due to curvature of the meniscus.


Spring 2004









Xe +Xe
Xa = ,n n (4)
2
where

X0., apparent path length at time On+l (cm)
xe, length at time On (cm)

Because of putative turbulence at the top of the cylinder
and the effect of a curved meniscus at the liquid surface (see
Figure 2), the actual diffusion path length, x, is somewhat
less than the apparent diffusion path length, Xa, that can be
physically measured along the vertical wall of the cylinder.
Thus the actual diffusion path length, x, is obtained by cor-
recting the apparent diffusion path length, xa:
X=Xa -Axs-Axe =Xa -AX (5)
where
Ax, distance correction due to the meniscus (cm)
Ax, distance correction due to turbulence (cm)
Ax total distance correction (cm)

Equation (2) can be rewritten to yield a value for an appar-
ent diffusion coefficient based on an apparent diffusion path:

Da = NARTpfxa (6)
PAp

where
Da apparent vapor diffusion coefficient (cm2/sec)

Rearrangement of Eq. (6) gives

NA = DaP (7)
RTpfxa
Equating the expression for NA in Eq. (7) with that in Eq. (2)
and noting that x = xa Ax,

NA= DaPAp DPAp
RTpfxa RTpf(xa -Ax) (

Rearranging Eq. (8) leads to a useful expression relating the
apparent diffusion coefficient (Da), the apparent diffusion path
length (Xa), the actual diffusion coefficient (D), and the total
path length correction (x),
1 -Ax 1 1
-=-.- +-- (9)
Da D xa D

The value of Eq. (9) in finding the actual diffusion coeffi-
cient should now be clear. Measurements at different times
of liquid remaining in the cylinder, each associated with a
measured x. value, allows computation of D values from Eq.
(6). According to Eq. (9), a plot of 1/Da versus l/xa should
yield a straight line with an intercept of 1/D and slope -Ax/
D. Using this technique, a value for the actual diffusion
coefficient along with the total path length correction (Ax)
can be obtained.


RESULTS AND DISCUSSION
Students' Extra Work
Is it at steady state? The working equations were based
on steady-state diffusion, but many students never questioned
the assumption of steady state relative to the experimental
data. Several other students rightfully questioned the assump-
tion of steady state using Eq. (10) from Lee and Wilke,5" which
provides an estimate of the approach of the diffusive flux to
the steady-state diffusive flux of the same system:

(NA) 0=
(NA)e=-
Dt20 4Dir2e 9Dit'2 16DI62 25 D20
1-2e x0 +2e x. -2e x; +2e x; -2e x;
(10)


where
(NA)O= flux of species A at any time 0
(NA)O= flux of species A at infinite time
xo length of diffusion path
0 time


The estimated times required to achieve (NA)0=0/(N A) =
0.999 were 4 minutes, 4 minutes, and 2 minutes for acetone,
hexane, and water, respectively. Data acquired during the
unsteady period should be excluded from analysis.
Turbulence and Ax Most students were able to use the
experimental data along with the working equations to com-
pute an actual diffusion coefficient according to Eq. (9). An
example of the type of plot dictated by Eq. (9) is shown for
hexane in Figure 3. For all three liquids, however, the plots
consist of two definite linear regions, i.e., a value for 1/x


I -a.(-'0) I
Figure 3. Plot of the inverse diffusion path abscissaa) ver-
sus the inverse of the apparent vapor diffusion coefficient
ordinatee) for hexane from a single graduated cylinder. Data
left of abscissa values 1/x. = 0.125 are used with Eq. (9) as
a means to obtain a vapor diffusion coefficient value cor-
rected for turbulence at the top of the graduated cylinder
and a curved meniscus at the liquid surface.


Chemical Engineering Education









exists that indicates transition between the two linear regions.
A plurality of students simply ignored this inconsistency while
others pursued its potential meaning. Students who pursued
it quickly recognized that 1/x was related to Ax, and ulti-
mately realized that some part of the Ax region estimated
using Eq. (9) and Figure 3 actually included data from within
the putative turbulent region at the top of the
cylinder, i.e., during early times, for small xa;
mass transfer in the turbulent region is not due
to simple diffusion alone and must be excluded The oj
in estimating a vapor diffusion coefficient. nati
Thus students were faced with using an it- ass
erative approach of estimating Ax, using Eq. motiv
(9) and Figure 3 and then repeating the esti- studei
mation procedure after excluding data from extra w
within the estimated turbulent region. Iteration the
continued until Ax estimates stabilized at 6.51 and
cm, 8.68 cm, and 4.53 cm for acetone, hexane, seven
and water, respectively, compared with the dis- t_-a
tance from the initial liquid level to the cylin- cr
der top of 5.39 cm and the total height of each
cylinder of 23.7 cm. Each cylinder's inside p
diameter was 2.8 cm.
their
Although it may appear that the turbulent
region varied with the liquid species, we sug- engine
gest that a more likely cause of variation among
the heights of the turbulent regions was cylin-


der placement relative to the fan and other cylinders. In an
attempt to eliminate the turbulent region in the tube above
the liquid, a variation of this experiment could possibly use
wire gauze to fill the space between the cylinder top and the
initial liquid level in the cylinder.
Ambient Partial Pressure: Does p, = 0? The partial pres-
sure of the diffusing species in air supplied by the fan pass-
ing over the cylinders appears in Eq. (3), which provides an
estimate of the diffusion driving force. Clearly, the partial
pressure for organic species that vaporize relatively slowly
may realistically be assumed as zero, but since water was
one of the three diffusing species used in this assignment and
since water can be present in ambient air due to humidity, the
assumption of zero partial pressure of water was investigated
(but by fewer than five students). Several students attempted
to include a partial pressure value obtained from humidity
data archived by the National Weather Service. Unfortunately,
these humidity values were not realistic for room air in the
air-conditioned lab. One student, however, included the wa-
ter partial pressure as an adjustable parameter in his model in
an Excel spreadsheet. He used Excel's Solver tool to perform
a nonlinear least squares minimization of error on water mo-
lar flux to compute a "most likely" estimate of water partial
pressure. The relative humidity thus obtained was 68%.
Pedagogical Impact
In the first year this assignment was used (Year 1, N=3), 14


students received grades of 100 or greater, while only two
students did not complete the minimum base-case. The Year-
1 average grade was 97.0, with standard deviation of 12.59.
In Year 2 (N=19), grades bifurcated into a high grade group
(n=12, avg grade = 105.0, std dev = 5.0) and a low grade
group (n=5, avg grade = 60.5, std dev = 8.5), with two stu-
dents not submitting a report.


In end-of-semester surveys, students were
asked to rate their level of learning on course
topics including diffusion. The levels were
modeled on Bateson's three levels of learn-
ing.[61 In Year 1, of students responding, six
students (24%) indicated they had mastered
aspects of diffusion, 17 students (68%) indi-
cated they had some facility with diffusion,
and only two students (8%) indicated they
were merely familiar with diffusion. In Year
2, of students responding, six students (46%)
indicated they had mastered aspects of diffu-
sion, five students (38%) indicated they had
some facility with diffusion, and again, only
two students (15%) indicated they were
merely familiar with diffusion.
Based on instructor evaluation of student
work and students' self-assessment of learn-
ing, we were very satisfied with the utility of
this assignment to enhance student learning


of mass transfer. The open-ended nature of the assignment
motivated many students to pursue extra work beyond the
minimum and sparked several students to a level of creativ-
ity not previously exhibited in their chemical engineering
work. Students' grades were determined from the same in-
strument from which they learned and they could receive
any amount of credit for which they were willing to work;
they faced no artificially imposed upper limit on credit
they could accrue.

REFERENCES
1. Woods, Donald R., "Problem-Based Learning for Large Classes in
Chemical Engineering," in Bringing Problem-Based Learning to
Higher Education: Theory and Practice, Number 68, 91-99, L.
Wilkerson and W.H. Gijselaers, eds., in the series New Directions for
Teaching and Learning, R.J. Menges, editor-in-chief, Jossey-Bass Pub-
lishers, San Francisco, CA (1996)
2. Rodriquez, J.M., V. Henriquez, and A. Macias-Machin, "A Simple
Experiment for Mass Transfer," Chem. Eng. Ed., 32, 142 (1998)
3. de Nevers, N., "A Simple Heat of Crystallization Experiment," Chem.
Eng. Ed., 25(3), 154 (1991)
4. Bird, R.B., W.E. Stewart, and E.N. Lighfoot, Transport Phenomena,
John Wiley and Sons, New York, NY, 523 (1960)
5. Lee. C.Y., and C.R. Wilke, "Measurements of Vapor Diffusion Coeffi-
cient," IndEng. Chem., 46, 2361 (1954)
6. Perry's Chemical Engineers Handbook, 4th ed., R. Perry, C. Chilton,
and S. Kirkpatrick, eds., McGraw-Hill Book Co., New York, NY (1969)
7. Bateson, G., Steps Toward an Ecology of the Mind, Paladin, London,
England (1973) 1


Spring 2004


pen-ended
re of the
chemict
erig-ed mw
ts1Japwww-

aizdwum :-







Ibitedin
chemical
Serngwek.









e, M classroom


COMPUTER EVALUATION OF EXCHANGE

FACTORS IN THERMAL RADIATION



REDHOUANE HENDA
Laurentian University Sudbury, Ontario, Canada P3E 2C6


Computer software packages that supplement tradi-
tional classroom presentations can provide
opportunities in reflection and action to students.' 4]
A key factor for successful development and use of computer
packages in engineering education is identification of activities
that cannot be accomplished by other means, e.g., analytically.
Such is the case for most thermal radiation problems.
Thermal radiation is offered as part of heat transfer courses
in all chemical engineering programs, although not as inten-
sively as in their mechanical engineering counterparts. At least
two aspects are very peculiar to thermal radiation transfer
compared to other modes of transport. First, the mechanism
of thermal radiation has no analogy in momentum and mass
transport-a very useful tool in the instruction of the differ-
ent modes of transport. Second, the equations of heat transfer
by radiation are strongly nonlinear, making it difficult to solve
radiation problems using analytical or even numerical tech-
niques, especially for complex geometries and/or real surfaces.
In evaluating any thermal radiation exchange, it is funda-
mental to calculate the view or exchange factors. Generally,
the instruction of thermal radiation to chemical engineers re-
volves around a thorough understanding of the properties of
the view factor and of the tools to evaluate it. For simple
geometries, view factors have been evaluated and tabulated
in catalogues.[561 For moderately involved geometries, how-
ever, view factors are exceedingly difficult to evaluate ana-
lytically. Different numerical techniques, such as the finite
element method, area and line integral methods, and the Rom-
berg formula, were used to compute the view factor.17-9 More
often than not, statistical approaches, such as the Monte Carlo
(MC) technique, are the only alternative.
An advantage of the MC technique is its ability to tackle
the most complicated problems with relative ease. It also al-
lows for a better analysis of the design parameters that affect
distribution of thermal patterns arising from thermal radia-
tion exchange in realistic geometries in an effort, for instance,
to optimize these parameters. Its disadvantage lies in the is-
sue of accuracy of the results, although statistical errors can


be resolved through an increase in sample size and an ad-
equate choice of a random number generator.
This paper describes a computer program that serves as an
educational tool in the instruction of the notion of view fac-
tor in a nontrivial geometry using the MC technique. The
concept of exchange factor, which induces the effect of mul-
tiple reflections, is also considered in detail. Thermal radia-
tion occurs between discrete sources (both a single source
and a linear succession of sources are considered) and the
walls of a rectangular enclosure containing an obstructing
surface. All surfaces are separated by a transparent medium
and are assumed to be isothermal with uniform properties.
The surfaces can be black, diffuse-grey, or diffuse-nongray.
The discrete sources are black bodies emitting at a high tem-
perature (T = 2700K). In order to make the problem tractable
at the undergraduate level, the surfaces are assumed to be
cool enough so their emission can be neglected.

MODELING
The enclosure considered is a box with six sides and ob-
structing surface as sketched in Figure la. Thermal radiation
is emitted by discrete isotropic sources in the form of energy
bundles that travel along straight paths through the enclosure
until they are absorbed by a surface or terminated after a pre-
set, high-enough number of reflections on the surfaces. Solv-
ing exchanged thermal radiation using the MC technique ne-
cessitates tracing the history of randomly sampled photons
from their point of emission to their point of absorption or

Redhouane Henda is Associate Professor in
the School of Engineering at Laurentian Uni-
versity (Canada). He received his MSc and PhD
degrees in chemical engineering from I'lnstitut
de Genie Chimique and LAAS-CNRS (France),
and is a recipient of the Alexander von Humboldt
foundation (Germany). Presently, he teaches
courses related to transport phenomena, nu-
merical analysis, and materials engineering. His
current research interests include dynamical
chemical systems, advanced materials, and
engineering education.
@ Copyright ChE Division of ASEE 2004


Chemical Engineering Education









attenuation, after simple or multiple reflections. A photon
history is tracked using a ray-tracing approach101 (as shown
in Figure lb) relative to enclosure and source systems of co-
ordinates. In order to follow the history of energy bundles in
a statistically sound way, the position and properties of
bundles are chosen according to probability distributions. The
fraction of energy emitted over wavelengths between 0 and X
is expressed by"1


JEddk
R(X)= P(X)dk=- (1)
0 f EXdX
o

Equation (1) is then inverted to correctly model a specific
bundle property. In case of a single source, the position of the
latter in the enclosure is rather deterministic. For the case of
a uniform linear succession of discrete sources, based on Eq.


41 5
5


z


:y ------------
Z ,
/ ------------------- 2


6 / ,
6Sources

4 ---I --


Surface i


Arbitrary system of
coordinates

Figure 1. Schematics of enclosure geometry (a) and vector
description of an energy bundle (b). The dimensionless
depth and height of the enclosure are Y/S = 1 and Z/X = 1
(with X = 20 cm), respectively, and the dimensions of the
obstructing surface are 10 x 10 cm2.


(1) a source position (x, yS, z) is chosen randomly along a
linear segment of coordinates (x sg, ysg.,, Zsg.) and xsg2, Ysg,2
zsg,2), as follows:
xs = (1 Rs)xsg,l + RsXsg,2

Ys = (1 Rs)ysg, + RsYsg,2 (2)
Zs = (1 Rs)zsg,l + RsZsg,2

where R is a number between 0 and 1 picked at random.
For the nongray case, the wavelength of emission from
a source is determined according to the polynomial de-
scription proposed by Haji-SheikhE121 and given in the
Appendix (Part 1).
The direction of emission or reflection of an energy bundle
is defined by the direction vector R., whose coordinates are
expressed in terms of the azimuthal, The two angles are determined by


es = cos-'(1- 2 Re)


0surf =sin-l(Ro)


-Ps,surf = 2 R (4)

0s is valid for bundle emission from a source (solid angle =
4'rr), and O6 is used for a bundle reflected from a surface.
The expression for the azimuthal angle is valid for both a
source and a surface.
The position of a bundle in space is defined in terms of its
origin and the direction vector, Ru, by
x = xs + xt y = ys + t z=z + zut (5)
where t is a positive value.
For a bundle to hit one of the planar surfaces defined by
Ax + By + Cz + D = 0 (6)
where A, B, C, and D are constants, the following equation
must be fulfilled:

A(xs + xut)+ B(ys + yut)+C(z, + zut)+ D = 0 (7)
Equation (7) is used to solve for t, which is then substituted
into Eq. (5) to compute the bundle position. A bundle inter-
sects a surface if its coordinates are within the bounds of the
specified surface. Similarly, further planar surfaces with
differing shapes can also be considered if their geometries
are well defined.
Among the bundles impinging on a surface, a fraction a is
absorbed and the rest of the bundles are reflected from the
surface. Absorption of a bundle is determined by comparing
the surface absorptivity at a given wavelength, acx, with a ran-
domly generated number, R The condition of absorption of
an energy bundle by a surface is given by

Ra < a (8)
In this case, a new bundle is generated and its history is traced
using Eqs. (2) through (8). If a bundle is reflected on a sur-


Spring 2004










face, a new direction for the bundle in the enclosure is chosen anc
bundle history is traced using Eqs. (3) through (8).
In the present work, the exchange factor is defined as the fractic
diffuse radiation emitted by a discrete source, s, that is absorbed
surface, i, either directly or after multiple reflections on the surfaces

number of photons leaving s i (direct & indirect)
dFsi =
total number of photons leaving s


The total exchange factor Fs-i between a source s and a surface,
calculated by integrating Eq. (9) over a large number of energy bun
emitted by source s:

F0 ~~NA,.j
\ NB JNB>>I

Finally, the total exchange factor Fs-, between a linear successio
sources and any surface, i, in the enclosure is obtained by summing u
factors F- given by Eq. (10) over the number of all sources making ul
linear surface.

RESULTS AND DISCUSSION

In the following paragraphs, three cases with increasing complexity,
presented and solved using the computer program. The latter serves z
exercise in an undergraduate course in heat and mass transfer. The les
learned through each solved case include closed- and open-ended q
tions and require students to use thinking skills, in accordance with Blo<
Taxonomy.[4] The dimensions of the system geometry and the value
many other parameters can be changed at will by the students through
input file (see description of the code usage in the Appendix, Part 2).
parameters pertain to enclosure, blocking surface, and heating wir
mensions; to blocking surface and heating wire locations; and to the p
erties of the surfaces.
In Case 1, a black enclosure without the obstructing surface is cor
ered, and a test problem is presented. The test problem also has an ana
cal solution so students can verify the computations for themselves. (
2 concerns radiative exchange in a diffuse-gray enclosure contain
diffuse-gray obstructing surface. Unless otherwise stated, all surface:
assumed to have the same absorptivity, o, of 0.45. Case 3 is simil;
Case 2, but all surfaces are treated as nongray, i.e., with wavelength
pendent absorptivities. In the present study, all calculation results for C
2 and 3 correspond to a number of energy bundles per single source
106. Moreover, thermal radiation is emitted by a linear successio
sources, with a sample size of 500 sources, mimicking a heating wire.
sources are considered to lie between segment ends of dimensionless
ordinates (xs/X=0.5, y g/X=0.25) and (xsg/X=0.5, yg /X=0.75).
Case 1

The Monte Carlo program has been tested for the case of an enclo
whose six walls are black. For the sake of simplification, in the exar
presented here heat radiation is emitted from a single source place
position (x/X=0.5, ys/X=0.5, zs/X=0.5) coinciding with the center ol
black enclosure. Results for this example are shown in Figures 2 anc
Figure 2 illustrates the calculated view factors between the source

128


I the the six surfaces of the enclosure as a function of
the size of the sample of energy bundles. By so
in of doing, the students get a better feel for the Monte
by a Carlo method by observing its accuracy as a func-
tion of sample size. As the total number of emit-
ted bundles increases, the calculated values of the
view factors tend directly to or oscillate about the
exact solution corresponding to Fs-1 = FS2 = F-3=
(9) Fs~4= Fs5 = Fs-6 = 1/6 for surface 1 through surface
6, respectively. One point worthy of notice in Fig-
i, is ure 2 is that accuracy in the calculated view fac-
dles tors is improved only for a bundle sample size of
105 and larger. For the latter bundle sample size,
the computed view factors are: Fs_ = 0.1662, Fs_2
(10) =0.1664, F_ =0.1667, F = 0.1670, F =0.1667,
F6 = 0.1669, for surface 1 through surface 6, re-
n of


o180

0.175

0.170 -

" 0.165

1 0.160

0.155

0.150 -


0 1 2 3 4
Surface, i


6 7


Figure 2. Calculated view factors of the differ-
ent black surfaces as a function of the number
of emitted bundles (in the inset). The dotted line
represents the analytical solution.


Figure 3. Contour plot of the view factor over
the black surface of the bottom wall.


Chemical Engineering Education


0



Q 9

o
V
1,000
o 10,000
v 20.000
V 30,000
50,000
O 80,000 0
90.000
0 100,000










spectively, with a maximum relative error of 0.3%. In Figure
3, a typical contour plot depicting the distribution of the lo-
cal view factor is shown for surface 3 (bottom wall). For
all surfaces, the view factor reaches its largest value
around the center of a surface, which is closest to the
source, and decreases toward surface edges.
Case 2

In this case, the walls of the enclosure and the obstructing
surface are considered diffuse and gray. The center of the
obstructing surface, denoted surface 0, coincides with the
center of the enclosure. A linear succession of sources is
located between the bottom wall of the enclosure and the
obstructing surface.
Figure 4 shows the dependence of the calculated exchange
factors of the surfaces on the height, zsg/X, of the linear source
when all surfaces are assumed to have an absorptivity of 0.45.
As the linear source is moved upward from the bottom wall,
the magnitude of the exchange factor increases from -0.09
to ~0.22 for the obstructing surface, and decreases from ~0.34
to ~0.20 for the bottom wall. This result is expected since the
exchange factor is inversely proportional to the square of the
distance between two interacting bodies. For the remaining
surfaces, i.e., the lateral surfaces and surface 1 (top wall), the
value of the calculated exchange factor slightly increases as
the linear source is raised to a dimensionless height of 0.3,
then the exchange factor slightly decreases with increasing
the height of the linear source (see Figure 4). The results also
show that the exchange factors of surface 2 and surface 4 are
equal in magnitude (as well as those of surfaces 5 and 6).
This is because of the symmetry they exhibit relative to
the obstructing surface and linear source (the same goes
for surfaces 5 and 6).
The distributions of the calculated exchange factor, corre-
sponding to a height of the linear source of 0.1, on the ob-


000- I I
0.05 0.10 0.15 0.20 0.25 030 035 0.40 045
Source dimensionless height, zsg/X

Figure 4. Dependence of exchange factors on the
height of the linear source for the different surfaces
(in the inset).


structing surface and surfaces 1 and 2, are presented in Fig-
ures 5a,b,c, respectively. It is interesting to note that the ex-
change factor was found to be at its maximum value near the
center of the top surface, where the central areas do not face
the source directly, as can be seen in Figure 5b. One of the
reasons for this distribution pattern is that multiple bundle
reflections on the surfaces were needed for the exchange fac-
tors to fulfill the summation relation. For instance, the sum
of the calculated exchange factors for all surfaces is higher


00 02 04 06
Dimensionless position x/X


00 02 04 06
Dimensionless position, x/X


0.8 10


08 10


0 0005 0005
0o

00010 00010
06





02 0 0020 002 00020
W0015 0 015
J\U \ \ /} \


o0 02 04 06
Dimensionless position, x/X


08 10


Figure 5. Contour plots of the exchange factor over
surface 0 (a), surface 1 (b), and surface 2 (c).


Spring 2004


129










than 99.95% only after four, six, and twelve reflections be-
fore terminating a bundle for a surface absorptivity, a, of 0.70,
0.45, and 0.20, respectively. Figure 6 illustrates the effect of
surface absorptivity on the calculated exchange factors of the
different surfaces. A decrease in surface absorptivity induces
an increase (decrease) in the exchange factor for the top wall
(bottom wall). A low surface absorptivity implies that a bundle
has a good chance of being reflected upon impinging on a
surface rather than being immediately absorbed. That is, a
bundle can travel a long distance within the enclosure and
reach remote areas before it is terminated.

Case 3

In this case, the walls of the enclosure and the obstructing
surface are considered nongray and their spectral absorptiv-
ity is assumed to follow the simple distribution shown in Fig-
ure 7a (this is a good approximation for some important met-
als."31) The MC program has a routine to account for the wave-
length dependence of surface absorptivity. Thermal radiation
emission is due to a linear succession of sources with the
same characteristics as those corresponding to the results of
Case 2, and with z /X = 0.1.
sg
Figure 7b illustrates the calculated exchange factors of the
different surfaces in the enclosure as a function of the split-
ting (cut-off) wavelength Xs. The magnitudes of the exchange
factors vary sensibly as Xsp is increased for the top wall
(Fs, varies from ~0.08 to ~0.04) and bottom wall (F varies
from -0.26 to -0.33), but only slightly for the obstructing
surface and lateral walls of the enclosure. As can be seen in
Figure 7b, the calculated exchange factors are insensitive to
hsp for values of the latter higher than 4.0 pim and become
comparable in magnitude to exchange factors corresponding
to the gray-body case (see Figure 6, case with o = 0.45).

CONCLUSIONS

This paper has discussed calculation of the view/exchange

0.40
V
035

030
Surface 0
O Surface 1
S025 Surace 2
v Surface 3
I Surf=e 4
020 C t a Surfeuno
Surface
00 urface



0
0.05 -
0
0.00-1
01 02 0.3 0.4 05 06 07 08
Surface absorptilvity, a

Figure 6. Calculated exchange factors as a function of sur-
face absorptivity for Z /X = 0.1. The different surfaces are
indicated in the inset.


factor in an enclosure, with three problems related to differ-
ent aspects of thermal radiation, using the MC method. The
methodology has been presented in a simple and intuitive
way so that no additional background in statistics is required.
It provides new mathematical concepts that, generally, are
not taught in courses on thermal radiation. The computer pro-
gram is modular, very robust, and easy to use as an educa-
tional tool in the analysis of radiation problems. In the fu-
ture, the author intends to make an executable JAVA version
of the package available on-line to potential users.

ACKNOWLEDGMENT

The author thanks the undergraduate students who helped
in implementing and testing the computer program.


NOMENCLATURE

A,B,C,D constants defining a plane
D,, coefficients of inverse probability function, pim.K
E emissive power, W/m2/tLm
F view factor; exchange factor
NB,NA number of emitted bundles; number of absorbed
bundles
R cumulative distribution function; random number in
range 0-1; direction unit vector


Wavelength, (pm)


0.35


0.30

0.25

S0.20

, 0.15

010

005-

000


1.0
0 2.0
3.0
V 4.0











0 1 2 3 4 5 6 7


ISurface, i
Figure 7. Distribution of the spectral absorptivity for Case
3 (a), and variation of exchange factors of the different sur-
faces with the splitting wavelength in the inset (b).


Chemical Engineering Education










T,t absolute temperature of emitting source, K; length of
bundle
X,Y,Z dimensions of enclosure, m
x,y,z Cartesian system of coordinates; bundle Cartesian
coordinates, m
Greek Symbols
a surface absorptivity
0 angle to normal; polar angle
yp azimuthal angle
X wavelength of emission, tm
Subscripts


i,s
sp,sg
surf
u
s-i
sg-i
h
1,2


surface i, source s
splitting, segment
surface
unit
from source s to surface i
from segment to surface i
at a given wavelength
end-points of linear source


REFERENCES
1. Camahan, B. (Ed.) Computers in Chemical Engineering Education
CACHE, Austin, TX (1996)
2. White, S.R., and G.M. Bodner, "Evaluation of Computer Simulation
Experiments in a Senior-Level Capstone ChE Course," Chem. Eng.
Ed., 33(1), 34 (1999)


3. Thompson, K.E., "Teaching PDEs to Undergraduates: Overcoming
Conceptual and Computational Barriers," Chem. Eng. Ed., 34(2), 146
(2000)
4. Bloom, B.S., (Ed.), Taxonomy of Educational Objectives: The Classi-
fication of Educational Goals. Handbook 1: Cognitive Domain, David
McKay, Co., New York, NY (1956)
5. Feingold, A., "Radiant-Interchange Configuration Factors Between
Various Selected Plane Surfaces," Proc. R. Soc. London, 292,51 (1966)
6. Howell, J.R., A Catalog ofRadiation Configuration Factors, McGraw-
Hill, New York, NY (1982)
7. Chung, T.J., and J.Y. Kim "Radiation View Factors by Finite Elements,"
ASME J. Heat Trans., 104, 792 (1982)
8. Ambirajan, A., and S.P. Venkanteshan, "Accurate Determination of
Diffuse View Factors Between Planar Surfaces," Int. J. Heat Mass
Trans., 36, 2203 (1993)
9. Rammohan, V., and V.M.K. Sastri, "Efficient Evaluation of Diffuse
View Factors for Radiation," Int. J. Heat Mass Trans., 39, 1281 (1996)
10. Haines, E., "Essential Ray Tracing Algorithms," in An Introduction to
Ray Tracing, A. Glassner, ed., Academic Pres, New York, NY; pp 33-
77(1989)
11. Modest, M.F, Radiative Heat Transfer McGraw-Hill, New York, NY
(1993)
12. Haji-Sheikh, A., "Monte Carlo Methods," in Handbook of Numerical
Heat Transfer, W.J. Minkowycz, E.M. Sparrow, G.E. Schneider, and
R.H. Pletcher, eds., John Wiley & Sons, New York NY; pp 673-722
(1988)
13. Brewster, M.Q., Thermal Radiative Transfer and Properties, John
Wiley & Sons, New York, NY; Chapter 2 (1992) O


- APPENDIX


Part 1. For the nongray case, the wavelength of emission from a
source was determined as follows:1"2

XT=D, +D2R8 +D3R1/4 +D4R +D5R/2 when0.0
XT=DI +D2R+ +D3R+ +D4R + D5R4 when0.1< R< <0.9

XT= 15.2886 x 10'0/ [DIR +D2R 2+ D3R3 +D4R4 1/3

when 0.9 < R <1.0

with R, = 1 R,, and T= temperature of emitting source.
The coefficients for XT = f-'(R,) given above are summarized
in the table below (XT in lLm.K)
Range ofR, D, D, D, D, D,
0.0-0.1 503.274 230.243 5863.85 -10759.6 8723.14
0.1-0.4 1560.84 7603.61 -15540.1 31257.7 -20844.8
0.4- 0.7 2846.63 -4430.38 27936.0 -41041.9 25960.9
0.7-0.9 345197 -1828567 3674856 -3284391 1108939
0.9-0.99 1.2 9.476 -44.84 156.9
0.99- 1.0 1.10064 16.8148 -183.445 890.699


Part 2. Flowchart for MC calculation of radiation view/exchange
factors. Data pertaining to the geometry and dimensions of the
enclosure and obstructing surface, optical properties and tempera-
tures of the surfaces in the enclosure, heating source location and
temperature, number of bundles per photon, physical constants,
and preset calculation tolerances are needed as input to the com-
puter program. All data pertaining to length and position are ex-
pressed in absolute quantities. The input file interfaces between
the user and the computer package, and several output files are


obtained upon execution of the code. The program was written in
C++ using Microsoft Visual Studio, and the output files, in .txt
format, are readily transferable to SigmaPlot, a statistics and graph-
ics software from SPSS, Inc., for further processing.


Inpu data


Spring 2004









1 laboratory


EXPERIMENTS AND

OTHER LEARNING ACTIVITIES

USING NATURAL DYE MATERIALS


VERONICA A. BURROWS
Arizona State University Tempe, AZ 85287-6006
ver-increasing concerns and constraints regarding stu-
dent safety and health, waste disposal, and liability
can significantly hamper development of interesting
and relevant laboratory experiences in a chemical engineer-
ing program. Using chemical reagents is likely the most prob-
lematic of the possible safety and health concerns faced by
chemical educators. While placing the highest priority on the
safety of our students, we still wish to provide learning expe-
riences that prepare them for their careers, meet our learning
objectives, and are (we hope) inherently interesting. Ap-
proaches that can be taken to provide especially safe labora-
tory experiences are many; some of them include
*Avoiding chemical reagents (other than water and air)
altogether
Using microquantities of chemical reagents
Investing in ultrasafe equipment, facilities, and
specialized training
Using simulations and/or remote laboratory experi-
ences
Each of these approaches has some significant advantages,
but each also has its disadvantages: avoiding chemical re-
agents altogether is impractical when learning objectives in-
volve chemical reaction kinetics, for example, and using only
microquantities is impractical when learning objectives in-
volve experience with pilot or industrial-scale unit operations.
In recent years, an additional personal concern of the au-
thor has been that chemical engineering students have little
exposure to or opportunity to work with materials in "raw"
form. Laboratory materials nearly always comprise highly
processed and purified reagents. This, despite the fact that a
very large fraction of the top production chemicals produced
in the United States11 are commodities produced from basic
raw materials, including petroleum, mined goods, agricul-


tural products, surface and ocean waters, and air.
A description of an experimental approach designed to sup-
port a particular chemical engineering technical topic, but
that can be adapted to other chemical topics and other uses,
is presented in this paper. The hope is that it can serve as a
guide to others contemplating incorporating raw and natural
materials into chemical engineering laboratory experiences.

MOTIVATION
In addition to the learning goals common to any laboratory
experiment in any chemical engineering laboratory course,
an additional cognitive21 learning goal desired by the author
was that students achieve knowledge of some differences
between raw and highly processed materials when incorpo-
rated into a chemical process. Some additional elements
motivating the use of natural dyestuffs include
Safety Natural dyestuffs have been in use for many
centuries; many of them are extremely well character-
ized, and the ones we chose have no safety warnings
associated with them. We insist on the best laboratory
and safety practice at all times, but the safety of these
materials was a valuable aspect of their performance.
Environmental Friendliness As a result of their low
hazard, disposal of wastes generated when using


Copyright ChE Division of ASEE 2004


Chemical Engineering Education


Veronica Burrows is Associate Professor of En-
gineering in Chemical and Materials Engineer-
ing at Arizona State University. She received her
BS from Drexel University and her PhD from
Princeton. Her research interests include the ap-
plied surface chemistry of semiconductors and
thin film sensors. Her educational research inter-
ests include "girl-friendly" science and technol-
ogy experiences for K-12 students, criterion-
based assessment techniques, and technology-
enhanced learning for visually impaired students.










natural dyestuffs is simple-dye extracts pose no
problems to municipal water treatment systems, and
spent dyestuffs can be disposed of in general waste.
(The author prefers to collect the dyestuffs for garden
composting.)
Cost Commercially available dyestuffs range in price
from very expensive (indigo and carmine, up to $50/
lb) to modest logwoodd, $8/lb), while some usable
materials are nearly free (onion skins).
Aesthetics Natural dyestuffs and their extracts are low-
odor and generally pleasant-smelling, and the colors of
the extracts can be quite vivid, which is of additional
appeal to students.
Topical Richness and Complexity Students often
become bored when working with materials and
processes that are extremely well-characterized-
"Why should I bother studying this when I know the
answer already?" is a common complaint. Although
there is a very large amount of information available
on the production of natural dyes, it is mostly in the
form of recipes and "folk wisdom." This makes an
engineering approach to their processing both interest-
ing and useful. I have had more than a few students
speculate that application of chemical engineering
principles could possibly lead to profitable commercial
production of natural dyes.

NATURAL DYESTUFFS

Forty-seven plants are listed in Dye Plants and Dyeing."
A selection of plants for each major color group is given in
Table 1. In addition to plant sources, there are some animal
sources (cochineal, for example, is extracted from the cara-
paces of scale insects that live on cactus plants) and mineral
sources (for example, ochre, from a type of iron ore). A vari-
ety of sources for some of the more readily available dye-
stuffs can be found with minimal search on the World Wide
Web. Prices for logwood chips, a moderately expensive ma-
terial, are about $30 per pound retail; fifty grams of chips is
enough for 15 or more small-volume (50 ml or less) experi-


mental extractions. Less expensive dyestuffs include alkanet
root (about $4 per pound), madder root (about $4-$10 per
pound), and osage orange sawdust (about $10 per pound). In
addition to a wealth of material available through the World
Wide Web, there are many text references available detailing
dye plants,31 dyeing recipes and approaches,[14'5 and world
dyestuff production.16'

Experiment for Using Statistical Design
of Experiments to Screen Processing Factors
This laboratory experiment is conducted in the junior-level
chemical engineering laboratory course at Arizona State Uni-
versity. The course is designed to meet six learning objec-
tives:
1. Students will be able to design experimental runs to
achieve a specified experimental goal and perform
the experiments in a safe, ethical, and professional
manner
2. Students will be able to record and analyze experi-
mental data, and to interpret analyzed results using
appropriate theory and models.
3. Students will be able to effectively communicate all
aspects of their experimental work and analyses in
oral and written form.
4. Students will be able to make appropriate use of a
computer in data analysis and in oral and written
communication of their experimental work.
5. Students will be able to perform technical work in
teams.
6. Students will have demonstrated the above skills in
the context of chemical engineering knowledge
acquired in earlier courses.
The natural dye extraction experiment is implemented to
introduce students to statistical design of experiments (DoE).
The experiment is framed in the context of performing DoE
on a variety of processing factors for dye extraction as if for
eventual development of a full commercial process.
Students are provided with the following context statement


TABLE 1
Some Plant Sources of Natural Dyes for Various Colorss""s


Greens
Ivy Berries
Wallflower
Lily of the Valley
Plantain Roots
Nettle


Puroles
Logwood
Blackberry
Alkanet
Grapes
Red Cedar Root


Oranges
Annatto
Coreopsis
Silver Birch
Turmeric


Browns
Cutch
Black Walnut Hulls
Elderberry
Henna
Onion
Yew
Sumac Leaves
Juniper Berries
Coffee Grounds


Yellows
Black Oak Bark
Fustic
Osage Orange
Turmeric
Saffron
Weld


Blues
Logwood
Indigo
Woad
Red Cabbage


Reds
Madder
Brazilwood
Bloodroot
Munjeet
Pokeweed
Strawberries
Cherries
Raspberries
Beets


Spring 2004


Greys/Blacks
Black Walnut
Logwood
Mountain Laurel
Sumac
Iris Root









and experiment-specific learning objectives:
"Natural" or naturally based consumer products
currently have a limited, but growing (and devoted!),
customer base. This is especially true for clothing and
in cosmetics and personal care products, where
"organically grown" fibers or a "naturally derived"
specialty chemical (to be used in a cosmetic, a
shampoo, fragrance, or nutraceutical), having the
same functionality and performance as a "non-
natural" fiber or chemical might sell for ten-or-more
times the price.
A challenge in using "natural" feedstocks in
chemical processes is that they have properties that
can be quite variable-consider for example, indigo
(the leaves, stems, and flowers of which contain the
familiar blue dye of blue jeans). The productivity
(bushels of plant matter/acre) of land planted in
indigo will depend on all the usual agricultural
factors (climate factors, water factors, soil-type
factors, farming-practice factors). The quality of the
plants themselves will also depend on most of these
agriculturalfactors -the 'quality' including concen-
tration of the dye chemicals of interest, their ease of
processing, and the types and kinds of impurities. The
quality and quantity of extracted dye will depend on
the plant quality as well as on a host of processing
factors-extraction pH, oxygen concentration in
extraction bath, extraction time, temperature, degree
and types of mixing, extraction co-solvents, etc.
For this experiment, you will be assigned a particu-
lar dyestuff brazilwoodd, logwood, or madder) and
perform a statistically designed ("DoE") experiment
with the intent of testing possible dye processing
factors for significance. You should perform this
experiment in the context of the development of a
commercially viable process for extracting natural
dyes, with a view to tapping into the growing market
for 'natural' materials. You will generate an appropri-
ate statistical hypothesis, design the experiments
required to test the hypothesis, conduct the experi-
ments, analyze the results, and either falsify or
support the proposed hypothesis.
The learning objectives for this experiment are that
you
1. Achieve and demonstrate Comprehension Level
of Learning for generation and testing of DoE
screening-type hypotheses.
2. Achieve and demonstrate Application (Analysis)
Level of Learning for planning experiments, in
particular, full-factorial DoE screening experiments.
3. Achieve and demonstrate Application (Analysis)
Level of Learning when applying principles from


earlier ChE and core courses to analysis of experi-
mental data.
Pre-Lab Expectations Students are assigned a particular
dyestuff brazilwoodd, logwood, or madder) and are required
to do some background research on natural dyes in general
and on their assigned dyestuff in particular. From this back-
ground research they generate a set of possible "measures of
goodness" or outcomes that can be used to characterize their
experimental outcome. From this set, they choose one (nor-
mally, dye concentration or strength). They also generate a
large set of possible processing factors influencing the out-
put variable. They narrow these down to four to be tested in a
full-factorial two-level screening type experiment and develop
the details of the experimental runs in consultation with the
instructor (issues: safety, practicality, available processing,
analytical equipment, etc.). After being signed off by the in-
structor, they conduct the experiment.
Conduct of the Experiment Supplies must be provided
to allow students to test processing factors:
Temperature: hotplates and/or temperature baths,
temperature measurement
Extraction time: timers
Agitation: Stirrers, shakers, or agitators
Solvent composition: reagents and chemicals,
balances and/or graduated cylinders, pipettes, etc.
Pre-extraction dyestuff soaking: reagents, timers, etc.
Dyestuffparticle size: mortar and pestle, grinder, size
screens, etc.
It is important to either provide students with a compre-
hensive list of available equipment and supplies in the labo-
ratory or require them to specify required supplies at a pre-


TABLE 2
Typical Student Choices for High and Low Factor Levels
for Statistical DoE Experiment


Factor Temperature Solvent" Pre-soak time Agitation particle size
Typical 1000C ethanol 24 20-50 high as
'high' hours minutes speed received
factor 20C
value
Typical 20C water none 1-5 none reduced


'low'
factor


"Students were permitted to choose water, ethanol, or with justification from
the literature, to choose other "safe" solvents or solvent systems (e.g.,
aqueous acetic acid, saline solution)
bAs-received particle size varied significantly for different dyestuffs: madder
was received as roots with an average root piece of 4x40 mm, while
brazilwood was received as roughly 300 pim particles. Reduction was
accomplished via laboratory blender or by mortar and pestle.


Chemical Engineering Education


minutes


Process


Extraction


Dyestuff









lab meeting to assure that necessary supplies are adequate
for the testing of their chosen factors. Table 2 gives typical
low and high values chosen by students in the conduct of
their DoE experiments.
Dvebath Analysis There are several approaches that can
be taken to analyze dye concentration-a colorimetric ap-
proach, where samples of extract from each run are com-
pared to a wide set of known standards; a spectroscopic ap-
proach where samples of extract from each run are analyzed
using UV-vis spectroscopy and Beer's law; or an analytical
chemistry approach where extracts are further treated to pu-
rify (and weigh) the pure dye. In both colorimetric and spec-
troscopic approaches, a significant challenge is that the range
of dye concentrations in the extracts might span three or even
four orders of magnitude. In this case, neither the eye nor the
spectrometer will have enough dynamic range to fully ac-
commodate all samples, so a stronger sample might have to
be pre-diluted to fit into a narrower 'instrumental' range.
Data Analysis The experimental data set will comprise a
matrix of data that is analyzed using DoE algorithms, nor-
mally via computer programs such as Minitab, JMP, Design-
Expert, Fusion Pro, or add-on statistical software for MS-
Excel such as XLSTAT.171 Depending upon the students' sta-
tistics background, required analysis can be as simple as a
significance test for each of the factors, or as complex as re-
sponse surface analysis. In this course, students were only
required to complete a significance test.

LEARNING EFFECTIVENESS
Students tended to rather easily develop a competent DoE
experimental design, but they often exhibited confusion be-
tween experimental design in the statistical sense and experi-
mental planning-they often did not adequately imagine the
actions they would be taking in the laboratory, and as a result
they underestimated the time required or did not adequately


Obj. 1 Obj. 2 Obj. 3 Obj. 4 Obj. 5 Obj. 6

Figure 1. Student ratings of the effectiveness of the dye-
extraction experiment in meeting each of the course
learning objectives (listed in the text).


plan for the complexity of simultaneous runs. In this imple-
mentation, we used colorimetric analysis of dyebaths to as-
certain extraction effectiveness. We plan on implementing
spectrophotometric analysis in the next implementation,
both as additional training for students in modern experi-
mental methods and to enhance the precision of the ex-
perimental results.
Students were asked to rate the effectiveness of the dye
extraction experiment and report-writing experience on their
achievement of each of the course's learning objectives. Re-
sults are shown in Figure 1 for the Spring 2003 semester, the
first implementation of this experiment.
On average, students rated the dye extraction experiment
as a very effective activity toward the achievement of three
of the course learning objectives (ability to design and per-
form experiments, ability to record and analyze data, and
working in teams), somewhat-to-very effective for two ob-
jectives (ability to communicate effectively and apply skills
in a chemical engineering context), and somewhat effective
for one (ability to effectively use computers). These data in-
dicate that the students viewed the learning experience as
integrative within the course, which was certainly the inten-
tion of the instructor.
Positive comments about this experience included
I think the lab was the most important assignment of
the course because it encompasses all of the learning
objectives and seemed to be the climax of the course...I
think more emphasis should be placed on it.
Out of the tasks completed in this course, the design of
experiments was the most useful in achieving the
course learning objectives.
Student suggestions for how to improve the experience mostly
referred to a wish for more training in appropriate software
for statistical analysis of their results. For example,
...require students to have more computer interaction
...have an in-class tutorial on how to use available
software

MASS TRANSFER EXPERIMENTS
These are described for sophomore- or junior-level work
after students have been introduced to simple mass transfer
concepts. Simple estimation of an overall mass transfer coef-
ficient can be obtained by looking at the driving force for
extraction. If C represents the concentration of the dye in the
extract, and Cq represents the equilibrium (or saturation)
concentration of dye in the solvent, then assuming that the
mass transfer coefficient is dominated by resistance at the
dyestuff/solvent interface, the relation

aC =k(Ce -C) (1)
atContinued o page 141.
Continued on page 141.


Spring 2004










Mmcurriculum


INCORPORATING NONIDEAL REACTORS


IN A JUNIOR-LEVEL COURSE

Using Computational Fluid Dynamics (CFD)



BENJAMIN J. LAWRENCE, JASON D. BEEN, SUNDARARAJAN V. MADIHALLY, RANDY S. LEWIS
Oklahoma State University Stillwater, OK 74078


Within the last few decades, there has been a sig-
nificant increase in technological and computa-
tional tools that are available to students-using
simulation packages to design process flow diagrams is now
the norm in academia. Despite the advancement of computa-
tional tools, however, there is still growth potential for imple-
mentation of these tools in the chemical engineering under-
graduate curriculum. Such tools can be beneficial for teach-
ing advanced topics to undergraduate students in beginning
courses. For example, analysis of nonideal reactors can be
implemented in a junior-level chemical reaction engineering
course using computational fluid dynamics (CFD).
This paper describes the incorporation of CFD in a junior-
level chemical reaction engineering course at Oklahoma State
University in which students use CFD to predict the single-
reaction conversion of a species in a nonideal reactor. The
last few weeks of the course are dedicated to teaching the
students about nonideal reactors-specifically, obtaining and
using residence time distribution (RTD) functions to assess
the reactor. The students are assigned to teams of two-to-four
students in which each team completes one CFD project to
obtain the RTD for a given geometry and flow conditions.
The students use the RTD, along with a given chemical reac-
tion, to predict the conversion of a reactant at the exit of the
reactor. Two classes of CFD demonstrations are provided on
how to predict an RTD. An example for predicting the RTD
in a straight tube, followed by predicting the conversion of a
reactant, is described below.
In many undergraduate chemical reaction engineering
courses, the content includes in-depth analyses of ideal reac-
tors, such as batch, semi-batch, plug-flow (PFR), and con-
tinuous-stirred tanks (CSTR), for single or multiple isother-
mal or nonisothermal reactions. For the batch, semi-batch,
and CSTR, the ideal assumptions include instantaneous and
complete mixing (no spatial variations in concentration, tem-
perature, or reaction rate). The ideal assumptions for a PFR


include no radial variation in temperature and concentration
and no mixing in the axial direction (spatial variations occur
only in the axial direction). Nonideal reactor analysis is a
subject that is covered near the end of most textbooks'1" and
may not always be covered in a beginning course. Since stu-
dents enrolled in transport phenomena or fluid mechanics
courses at the undergraduate level learn about nonideal flow
conditions, implementation of the nonideal flow concepts in
reactor design would be beneficial for connecting the two
courses and exposing students to reactor design that is more
applicable to industrial settings.

NONIDEAL REACTORS
Several methods, ranging from combinations of hypotheti-
cal ideal reactors to detailed mathematical modeling, are used
to characterize nonideal reactors.[2-41 One method involves
incorporating the residence time distribution (RTD) with mod-
els to predict the conversion of a species following passage
through a chemical reactor. The RTD is a characteristic of
the mixing that occurs in a chemical reactor and is informa-
tive about how long species reside in the reactor. The RTD of

Benjamin J. Lawrence is a senior at Oklahoma State University. He is
currently president of ChemKidz, a group that performs science demon-
strations for elementary school children. His future plans include graduate
study in chemical engineering.
Jason D. Beene is a seniorat Oklahoma State University and will graduate
with a BS in Chemical Engineering with a Biomedical Engineering Option.
Currently, he serves as the President of the College of Engineering, Archi-
tecture, and Technology Student Council. He will be working with ExxonMobil
Development Company in Houston, Texas, after graduation.
Sundararajan V Madihally is Assistant Professor in the School of Chemi-
cal Engineering at Oklahoma State University. He received his BE from
Bangalore University and his PhD from Wayne State University, both in
chemical engineering. His research interests include tissue engineering and
the development of therapies for traumatic conditions.
Randy S. Lewis is Associate Professor in the School of Chemical Engi-
neering at Oklahoma State University. He received his BS and PhD de-
grees in Chemical Engineering from Brigham Young University and the Mas-
sachusetts Institute of Technology, respectively. His research interests in-
clude biomaterials development and the utilization of renewable resources
for the production of chemicals.


Copyright ChE Division of ASEE 2004


Chemical Engineering Education









a reactor is not dependent on the chemical reaction. A major
focus of this paper describes how to obtain an RTD and then
apply the RTD to a model that combines a chemical reaction
to predict the reactant conversion.
Experimentally, the RTD is determined by injecting a tracer
into the reactor and measuring the tracer concentration in the
exit stream as a function of time. The tracer should be a
nonreactive species, be easily detectable, should have physi-
cal properties similar to those of the reacting mixture, and
should not absorb on the walls or other surfaces in the reac-
tor. These requirements are necessary so that the tracer's be-
havior reflects the flow properties of the reactants and prod-
ucts. In reaction engineering textbooks, the RTD is usually
given in homework problems and examples.12' The advan-
tage of incorporating CFD into the curriculum is that stu-
dents can simulate an RTD for any reactor geometry rather
than performing time-intensive experiments or having the
RTD provided. Students can then model the conversion oc-
curring in any nonideal reactor geometry using the simulated
RTD. We introduced the CFD program (CFX, AEA Technolo-
gies, Pittsburgh, PA) in the transport phenomena course, which
is taught in the semester prior to the reaction engineering course.
To demonstrate the use of CFD in analyzing nonideal reac-
tors as part of a junior-level chemical reaction engineering
course, a tubular reactor with laminar flow is provided as an
example. The RTD is obtained for the tubular reactor by ap-
plying a step change in the concentration of a tracer flowing
into the reactor (represented as CO) and evaluating the tem-
poral mixing cup tracer concentration at the reactor outlet.
As stated before, the RTD is independent of the chemical
reaction so the chemical reaction is not integrated into the
CFD modeling. The dimensionless mixing cup (or volumet-
ric flow-averaged) tracer concentration (Cmix/Co) as a func-
tion of time is defined for a cylindrical reactor exit as
R
S[C(r) / Co]V(r)rdr
CO (t) 0 R- (1)
SoJ V(r)rdr
0

where [C(r)/C0] is the dimensionless tracer concentration and
V(r) is the axial velocity for a given radial position, r, at the
reactor exit. R is the radius of the tube. When performing
experiments, Cmix(t) is the measured tracer concentration at a
given time when collecting samples at the reactor exit. In the
absence of experiments, CFD can be used to predict C(r) and
V(r) as a function of time, then (C ix/Co) can be obtained as
a function of time from Eq. (1).
Once (Cmix/C) is determined, the RTD function [E(t)] is
determined for use in models (described later) to predict the
conversion of a species for a given reactor geometry. For a
positive step change in the tracer concentration, E(t) is evalu-
ated from (Cmix/Co) according to


S d[Cmix / C]
E(t) =


E(t) has several characteristics, such as the integral from
t = 0 to t represents the fraction of effluent that has been
in the reactor for less than time, t, and the integral of E(t)
from t = 0 to o is one.

METHOD FOR
OBTAINING RTD FUNCTION USING CFD
Example of a Tubular Reactor
Geometry The tubular reactor is 30 cm long with a 2.5-
cm internal diameter. Laminar flow through the reactor pro-
vides a parabolic velocity profile. For this example

V(cm / sec) = Vax 1 r2(3)

with V... = 2 cm/s. The reactor geometry was created in CFX
with the required dimensions; first, a 2.5-cm-diameter circle
was drawn, then the circle was extruded 30 cm and properly
trimmed to obtain a solid vessel (using CFX terminology, the
solid represents a fluid domain-in this case the fluid do-
main is the liquid inside the reactor).
Fluid Domain A fluid domain must be established that
describes the fluids that will exist within the geometry dur-
ing the simulation. For the RTD predictions, two fluids are
selected when creating the fluid domain-water and tracer.
The water properties at room temperature and pressure are
already in the CFX database. The fluids editor is used to cre-
ate a new fluid called "tracer" that has the same physical prop-
erties as water, with the addition of an aqueous tracer diffu-
sivity of 2 x 10-5 cm2s-'. In the domain options, a transient
problem with 0.1-second timesteps and a total simulation time
of 40 seconds is specified. The specified time step and the
maximum number of iterations for a time step is important
for an accurate solution; decreasing the time step until the
solution no longer changes demonstrates an appropriate time
step. The reference pressure is atmospheric. Isothermal lami-
nar flow is also specified.
Boundary Conditions and Initial Values The inlet and
outlet boundary conditions in CFX may be described using
the normal speed, pressure, Cartesian velocity components,
or the mass flow rate. The inlet boundary was specified using
the 1-D velocity profile according to Eq. (3). The velocity
equation can be input into CFX using an expression editor.
The tracer mass fraction at the inlet was specified at 10% (Co
= 2170 mol/m3). Any tracer mass fraction could be specified,
however, since the tracer has the same properties as the wa-
ter. The outlet boundary condition was specified as atmospheric
pressure. For the initial values within the reactor at t = 0, the
velocity components were chosen according to Eq. (3) and the
mass fraction of tracer was specified as zero.


Spring 2004









Mesh Parameters and Solving Prior to using CFX, a mesh was
automatically generated. To assess the accuracy of the solution, how-
ever, the mesh should be refined until the solution no longer changes.
Following mesh generation, convergence criteria were selected and a
transient-results file was established for input of the transient results
once the solution was obtained. The velocity and tracer mass fraction
were chosen for the output. CFX then solved the transient problem
using the continuity and momentum equations that are already built
into the software. Following the solution, a file was generated giving
the spatial velocity and tracer mass fraction with time at all spatial
points within the reactor. Obviously, the velocity components should
not change with time (Eq. 3 can be used to calculate the axial velocity
component as a function of radius) since the laminar flow profile was
established as an initial condition and the inlet laminar velocity pro-
file should not affect the established profile. For this example, the
velocity profile from the CFX solution confirmed that the initial ve-
locity profile remained constant throughout the simulation. A
nonchanging flow profile is imperative for evaluating the RTD since
the tracer is used to characterize the species movement in a steady-
state flow regime. The spatial tracer mass fraction will change with
time, however, due to the introduction of tracer at the inlet.
Once temporal velocity and tracer mass fraction profiles were ob-
tained as a function of spatial position, the profiles at the tube exit
were used to calculate (Cmix/Co) as a function of time according to Eq.
(1). Rectangular coordinates were given in the output file so the ra-
dius was calculated according to

r~2 2"
r = x +y

The RTD function [E(t)] was then calculated from the time derivative
of (Cm/Co) according to Eq. (2).

RTD RESULTS
After solving the transient problem, CFX Visualize was used to run
a transient animation of the tracer mass fraction within the tubular
reactor. Figure 1 shows snapshots of the 10% tracer solution (black)
as it travels through the reactor initially filled with pure water (white).
The intermediate gray colors representing less than 10% tracer are
also shown and demonstrate dispersion of the tracer as it travels through
the reactor. A plug-flow reactor would show a solid black area up-
stream of a solid white area changing with time. The parabolic veloc-
ity effects on the tracer profile due to laminar flow are also evi-
dent, as demonstrated by the radial gradients in the tracer compo-
sition. The visual animation is an excellent tool for demonstrat-
ing nonideal reactor characteristics such as radial composition
gradients and dispersion.
Figure 2 shows an example calculation for the dimensionless "mix-
ing cup" tracer concentration using the CFX output data at 30 sec-
onds. Using the data at the reactor outlet, the product of the dimen-
sionless tracer concentration (C/C0), velocity (V), and radius (r) was
plotted as a function of the radius. The area under the curve repre-
sents the numerator of Eq. (1). Similarly, the product of the velocity
and the radius (r) was plotted as a function of the radius, with the area
under the curve representing the denominator of Eq. (1). Thus, the


ratio of the two areas represents the dimensionless "mix-
ing cup" tracer concentration (Cmix/C0) at 30 seconds.
Cm/C0 was similarly calculated for all other times.
Figure 3 shows the calculated (Cm, /Co) as a func-
tion of time. (Cml/Co) obtains steady state at around
40 seconds and approaches unity. Figure 4 shows E(t)
versus time as obtained from Eq. (2). The area under
the E(t) curve is approximately 1.03, which is con-
sistent with the theoretical value of one. The E(t) curve
can also be used to calculate the residence time ac-
cording to














Osec 10 sec 15 sec 20 sec 30 sec 40 sec
Figure 1. CFX flow profiles for a step tracer input in
a tubular reactor. Snapshots of the 10% tracer solu-
tion (black) as it travels through a tubular reactor
initially filled with pure water (white) are shown.
Intermediate gray colors representing less than 10%
tracer are also shown. The tubular reactor is 30 cm
long with a 2.5-cm internal diameter.

1.0
0 8--------- -- s ._. --
0.9
0.8
0.7 P %

0.6

0.4
0.3 -
0.3 (C/Co)*V*r
0.2 .
0.1
0.0


0.0 0.2 0.4 0.6 0.8
Radius, r (cm)


1.0 1.2 1.4


Figure 2. Dimensionless mixing cup (Cmx/C) tracer
concentration analysis. At the reactor outlet, the prod-
uct of the dimensionless tracer concentration (C/C),
velocity (V), and radius (r) was plotted as a function
of the radius at 30 seconds. The area under the curve
represents the numerator of Eq. (1). Similarly, the
product of V and r at the reactor outlet was plotted as
a function of the radius with the area under the curve
representing the denominator ofEq. (1). The ratio of
the two areas represents the dimensionless mixing cup
tracer concentration (Ci/Co) at 30 seconds.


Chemical Engineering Education










(4)

the
esi-
di-
or a
via-
t in
4A


tres = J tE(t)dt
0
A value of t = 29.5 seconds was obtained from
data, which is in agreement with the theoretical r
dence time of 30 seconds [reactor length (30 cm)
vided by the radial-averaged velocity (1 cm/s)]. Fc
plug-flow reactor, E(t) would be a spike at tre. De'
tion from plug flow (nonideality) would result
spreading of the spike, as demonstrated in Figun
As deviation from plug flow becomes greater (and
preaches a well-stirred reactor), the spread becoi
wider and more asymmetrical. The F(t) curve, wh
is the came as the (Cm,x/Co) curve shown in Figur
was calculated according to

F(t) = E(t)dt
0
Both E(t) and F(t) are used to calculate the convers
limits for a nonideal reactor, as described below.


1.20

1.00

0.80

0.60

0.40

0.20

0.00
0 10 20 30 40
Seconds


Figure 3. The dimensionless mixing cup tracer con-
centration (C i/C) was calculated using Eq. (1) and
is shown as a function of time. The inlet tracer con-
centration (C) is 2170 mol/m3. As defined byEq. (5),
F(t) was obtained from the E(t) curve (see Figure 4).


0 10 20 30 40 50
Seconds
Figure 4. The RTD function [E(t)] obtained from
Eq. (2) is shown as a function of time.


APPLICATION OF RTD FOR NONIDEAL REACTOR
MODELING
To demonstrate the application of using E(t) and F(t) to model the
conversion in nonideal reactors, the nonelementary liquid-phase reac-
tion A -+ D was chosen for the reaction in the tubular reactor with a
rate law of

-R kCA kCA(1- XA) (6)
(1+ KMCA) [1+KMCA0(- XA)]


For the constant volumetric flow conditions, the conversion (xA) defi-
ap-
n ition of CA = C A( XA) was included in Eq. (6). Parameters of C =
mes AO
ich 1.0 M, k = 0.05 s-', and K = 1.0 M-' were used.
e 3, The simplest method for predicting xA exiting the nonideal reactor is
using the rate law together with the E(t) and F(t) curves to estimate
bounds for xA. This method uses two zero-adjustable parameter mod-
(5) els; a segregation model and a maximum-mixedness model. The seg-
regation model assumes that molecules of the same residence time (or
ion "age" in the reactor) travel in groups and each group acts as a batch
reactor. The maximum-mixedness model assumes that molecules of
different "age" are completely mixed as they enter the reactor. For a
more detailed explanation, see reference 2 (pp. 838-851).
To obtain one bounded prediction for conversion (xAseg), the segre-
gation model uses the predicted conversion for a constant-volume batch
reactor [(xA)batch] together with the E(t) curve according to


XA,seg = (XA)batch E(t)dt (7)
0
For a constant-volume batch reactor, the batch conversion is obtained
by integrating the following equation using the reaction rate given in
Eq. (6):

d(xA batch kCA01 (XA )batch]
-CA0 -A at (8)
dt {1+KMCAO -(XA)b

Following integration of Eq. (8) to obtain an (xA)baCh curve as a func-
tion of time, a new curve is generated by plotting the product of (xA)batch
and E(t) versus time. According to Eq. (7), the area under the curve
represents xA,eg. For this example, xAsg is 0.59.
To obtain the other bounded prediction for conversion (xA.maxmi), the
maximum-mixedness model uses both the E(t) and F(t) curves, along
with RA. As noted by Fogler,1[2 the conversion is predicted by


xi+1 Xi+(AX) xi + (9)
1- F(Xi) CAO

where X is the time and AX is the time step. To apply Eq. (9), ,X is
chosen to represent the latest time of the E(t) curve in which E ap-
proaches zero (XI = 40 seconds for this example), x, is zero, and nega-
tive time step is chosen (AX = -2 seconds for this example). RA (Eq. 6)
is evaluated using x, and the values of E and F are obtained at X,.
Using Eq. (9), a value for x2 is obtained. RA is then evaluated using x2
and the values of E and F are obtained at X2 = X1 + AX = 38 seconds to
obtain x3 from Eq. (9). The process is repeated until X = 0 seconds,


Spring 2004


50
50









wherein x is the bounded conversion Amaxmix For this ex-
ample, xA maxmix is 0.60.
The above values XA,seg = 0.59 and xA,maxmix = 0.60 signify
that the conversion of A in the nonideal tubular reactor is
between 0.59 and 0.60 for the given reaction rate. Since the
E(t) curve showed only a small deviation from plug flow, the
predicted conversion should be close to the conversion pre-
dicted with a PFR model. For this example, xA is 0.60 for
plug flow, demonstrating that the nonideal tubular reactor has
a conversion only slightly less than plug flow. More com-
plex reactor geometries are likely to show greater con-
version deviations from plug flow, however, as the flow
regime becomes more nonideal.

NONIDEAL ANALYSIS
OF COMPLEX REACTOR GEOMETRIES
The power of using CFD to obtain the RTD function is that
more complex reactor geometries, rather than simple tubular
geometries, can also be evaluated. Thus, students can be given
any reactor geometry to obtain an RTD function as well as
visualize the nonideal flow patterns in the reactor. In addi-
tion, CFD can be used directly with kinetic models to design
reactors. 5l To demonstrate the nonideal flow patterns in a more
complex geometry, a tank reactor (2.25" tall, 3.5" diameter) is
shown in Figure 5 with a perpendicular inlet (0.50" diameter)
entering the topside of the reactor (0.5" from top) and an outlet
(0.25" diameter) directly at the bottom center of the reactor.
With more complex reactor geometries, the steady-state
velocity flow profile within the reactor is not known a priori
(like the tubular reactor represented by Eq. 3). As previously
mentioned, the steady-state flow profile is used as an initial
condition when using CFD to predict the RTD from a step
tracer input. A steady-state flow profile is necessary since
experimental tracer studies rely on the flow profile being es-
tablished and not changing prior to the step tracer input. There-
fore, before applying a step tracer input to predict the RTD,
the steady-state flow profile must be determined so the pro-
file can be used as an initial condition.
The steady-state velocity profile is obtained using CFX by
defining the fluid domain with a single fluid, such as water.
For this example, the inlet boundary condition was chosen as
the velocity profile given in Eq. (3). The outlet boundary con-
dition was atmospheric pressure. Upon solving, the steady-
state profile is obtained throughout the geometry and is
stored in a results file.
The transient tracer mass fraction following a step tracer
input is obtained using the same procedure described for the
tubular reactor. The initial velocity profile is read from the
above results file, however, rather than being typed in, as
was the case for the tubular reactor. The tracer mass fraction
is still specified as zero for the initial condition. The output
file will give the temporal and spatial velocity and tracer mass


fractions at the outlet that can then be used to obtain E(t) and
F(t) curves for predictions of conversion.
Although for brevity of this paper, the E(t) and F(t) pro-
files and xA bounded predictions were not obtained for the
tank reactor, CFX Visualize was used to run a transient ani-
mation of the tracer mass fraction within the tank reactor.
Figure 5 shows snapshots of the 10% tracer solution as it
travels through the reactor initially filled with pure water.
Intermediate shades of gray represent less than 10% tracer.
The nonideality of flow is strongly evident as there are wide
variations of the tracer mass fraction within the reactor. Evi-
dence of stagnant areas can also be observed. A well-mixed
reactor would show the same color throughout the reactor
and would change from black to white. Again, the visual
animation is an excellent tool for demonstrating nonideal
reactor characteristics such as stagnant regions and large
concentration gradients.

CONCLUSIONS
This work describes the incorporation of CFD in a junior-
level chemical reaction engineering course to demonstrate
nonideal reactor principles, as well as to model nonideal re-
actors. Important concepts of nonideal reactors such as con-
centration gradients, stagnation, dispersion, and conversion
can be visualized and modeled using CFD to enhance a
student's understanding of nonideal reactors. In addition,
transport concepts are also emphasized. An advantage of us-
ing CFD is that an RTD profile can be predicted for any reac-
tor geometry without performing an experiment.
Applications include changing the flow profiles and reac-
tor configurations that allow a student to understand the role
of mixing in reactor kinetics. Only two examples are shown
in this paper, but there are numerous possibilities for devel-
oping problems specific to other reactor geometries. Current
plans are to extend the simulation to include stirrers, pres-
sure drops, and nonisothermal conditions to make the prob-


Figure 5. Snapshots of the 10% tracer solution as it travels
through a tank reactor initially filled with pure water. Di-
ameter and height dimensions are shown.


Chemical Engineering Education









lems more realistic. In addition, the RTD can be applied to
other models, including a one-parameter dispersion model, to
characterize nonideal reactors. The validity of model assump-
tions must be checked before applying any model, however.
The nonideal reactor analysis using CFX was implemented
in the junior-level chemical reaction engineering course in
the spring of 2002. Student feedback led to the importance of
considering some issues when implementing CFD into the
curriculum. First, since students were only exposed to CFX
the prior semester, assigning the project for a group, rather
than an individual, enabled students to spend more time on
the analysis. Second, students were often frustrated when
drawing the geometry in CFX. Perhaps an instructor should
consider developing a geometry database so students can
spend more time analyzing reactor geometries and under-
standing nonideal concepts rather than drawing the geom-
etries. In conclusion, the use of CFD programs such as CFX
will enable students to be more prepared to enter today's
workforce and to solve the difficult problems that arise
with nonideal reactors.

REFERENCES
1. Levenspiel, O., Chemical Reaction Engineering, 3rd ed., Wiley, New
York, NY (1999)
2. Fogler, H.S., Elements of Chemical Reaction Engineering, 3rd ed.,
Prentice Hall, Upper Saddle River, NJ (1999)
3. Davis, R.A., "Nonideal Reactors in Process Simulators," Proc. AIChE
Conf., Los Angeles, CA, p. 166, (2000)
4. Dutta, S., and R. Gualy, "Build Robust Reactor Models," Chem. Eng.
Prog., p. 37, October (2000)
5. Bakker, A., A.H. Haidari, and E.M. Marshall, "Design Reactors via
CFD," Chem. Eng. Prog., p. 30, December (2001) 0




Experiments Using
Natural Dye Materials
Continued from page 135.
is a reasonable starting approximation. After experimental
determination of Ceq (either from the asymptote of the con-
centration/time results or from a separate experiment), con-
centration vs. time data at constant processing conditions
(temperature, stirring, dyestuff particle size, etc.) can be ana-
lyzed graphically or analytically to determine the mass trans-
fer coefficient from the time dependence of the concentra-
tion

Ceq -C =kCeq exp(-kt) (2)
Further experiments can be designed to readily examine the
effects of solvent, temperature, stirring, dyestuff particle size,
and soaking on Ceq and on k.

ADDITIONAL OPPORTUNITIES
Aspects of this approach that were listed at the start of this


article suggest some further opportunities in using natural
dyestuffs:
* Other ChE Experiments
This paper describes implementation for supporting the
learning of statistical design of experiments relevant to early-
stage process development. Variations of this experiment
could support other ChE topic areas and concepts, including
thermodynamics (dye solubility with temperature, solvent),
unit operations (post-extraction filtration, operation costs as
a function of dyebath concentration and production rate, coun-
tercurrent stage-wise extraction), and reaction engineering
(for dyestuffs requiring some chemical reaction as part of the
processing, such as indigo).
* Distance Experiment
A kit for the conduct of this experiment or something like
it could be generated for a true distance experiment; the safety
of the materials makes it possible to perform an aqueous ex-
traction experiment in a home kitchen-analysis could be
performed by color comparison to a printed standard sheet.
* Hands-On Recruiting Activity/Contest
Students could be challenged to most-closely match a tar-
get color using the smallest amounts of materials. They could
be supplied with samples of a variety of dyestuffs, water, vin-
egar, baking soda, and perhaps some nontoxic metal salts.
Materials could be price-labeled and students required to com-
plete their work within a budget.

SUMMARY
The main purpose of this article is to spark interest in the
use of natural raw materials in a variety of chemical engi-
neering educational contexts, especially laboratory experi-
ences. We have found that students respond positively to the
laboratory use of these materials in the form of natural dye-
stuffs. Such materials can provide safe, environmentally
friendly, aesthetically engaging hands-on experiences to sup-
port a variety of learning goals in the chemical engineering
context.

REFERENCES
1. "Facts and Figures for the Chemical Industry," Chemical and Engi-
neering News, 42 (2002)
2. Bloom, Benjamin S., Ed., Taxonomy of Educational Objectives: Book
1, Cognitive Domain, Longman, New York, NY (1984)
3. Cannon, John, and Margaret Cannon, Dye Plants and Dyeing, Herbert
Press, Kew, England (1994)
4. Liles, J.N., The Art and Craft of Natural Dyeing, University of Ten-
nessee Press, Knoxville, TN (1990)
5. "Making Natural Dyes from Plants," pioneerthinking.com at pioneerthinking.com/naturaldyes.html> (2003)
6. Green C.L., Natural Colorants and Dyestuffs: A Review of Produc-
tion, Markets, and Development Potential, Food and Agriculture Or-
ganization of the United Nations, Rome, Italy (1995)
7. Software sources: Minitab from Minitab Inc., State College, PA; JMP
from SAS Institute, Inc., Cary, NC; Design-Expert from Stat-Ease,
Inc., Minneapolis, MN; Fusion Pro from S-Matrix, Inc., Vancouver,
BC; SC-State from Addinsoft, Brooklyn NY O


Spring 2004










aS classroom


USE OF DYNAMIC SIMULATION

TO CONVERGE

COMPLEX PROCESS FLOWSHEETS



WILLIAM L. LUYBEN
Lehigh University Bethlehem, PA 18015


Commercial process simulators are widely used for de-
signing new processes and for analysis of existing
processes. Most senior design courses contain a sig-
nificant component of computer simulation of process
flowsheets using these tools. The most widely used commer-
cial process simulation software is that developed by Aspen
Technology-AspenPlus for steady-state simulation and
AspenDynamics for dynamic simulation-and these tools are
used in the examples in this paper. The standard Aspen nota-
tion is used. For example, distillation column stages are
counted from the top of the column: the condenser is Stage 1
and the reboiler is the last stage.
The simulators contain models of most common unit op-
erations, which can be connected into a process flowsheet. If
the units operate in series, with upstream units feeding down-
stream units, the simulation is usually reasonably straight-
forward. If the flowsheet contains recycle streams, however,
the simultaneous solution of the typically very large number
of simultaneous nonlinear algebraic equations that make up
the steady-state model can be quite challenging. There is no
guarantee that any algorithm will find a solution. In addition,
there are sometimes multiple solutions in these nonlinear sys-
tems. The convergence of these recycle loops (or "tear" streams)
is a major challenge in steady-state process simulation.
Energy integration can also produce complications because
of "energy recycle" between different units. These difficul-
ties can sometimes be avoided by using the plant utility sys-
tem to break the energy linkage.
Commercial steady-state simulators contain a variety of
algorithms. For example, the user of AspenPlus can try such
methods as Wegstein, Broyden, and Newton. Convergence
tolerances and the maximum number of iterations can also
be adjusted.
The experience of many users, particularly students, has
been that the convergence of recycle loops is the most diffi-


cult part of steady-state simulation. The normal procedure is
to assume some conditions of a recycle stream (flow, tem-
perature, pressure, and composition) and work down through
the flowsheet until the calculated values of the recycle stream
are available. If the assumed and calculated values are not
sufficiently close, new guesses must somehow be made. The
process is repeated until convergence between the assumed
and calculated values has been attained. Often, however, con-
vergence does not occur.
One would think that this convergence should be fairly
easily achieved if the user has adjusted the design and oper-
ating parameters so the assumed and calculated conditions of
the recycle stream are fairly close, but this all too frequently
does not occur. For example, in one of the cases discussed
later, an assumed recycle stream had a composition 98 mol%
methanol and 2 mol% water, while the calculated stream is
only slightly different (0.4 mol% dimethyl ether; 97.2 mol%
methanol; 2.4 mol% water). The assumed flowrate is 72 kmol/
hr, while the calculated flowrate is 72.9 kmol/hr. The tem-
perature and pressures are identical. After connecting the
recycle stream and defining it as a "tear" stream in the
"Convergence" section of the "Data Browser" in
AspenPlus, the recycle loop does not converge when us-
ing any of the algorithms.
This paper illustrates that recycle loops can be easily con-
verged if the steady-state AspenPlus simulation (with the re-

William L. Luyben earned degrees in chemical
engineering from Penn State (BS, 1955) and
Delaware (PhD, 1963). His industrial experience
includes four years with Exxon, four years with
DuPont, and three decades of consulting with
chemical and petroleum companies. He has
taught at Lehigh University since 1967 and has
participated in the development of several inno-
vative undergraduate courses.


Copyright ChE Division of ASEE 2004


Chemical Engineering Education









cycle loops not connected) is "exported" into the dynamic
simulator AspenDynamics and the recycle connections are
made on the dynamic model. The steps in going from a steady-
state simulation to a dynamic simulation are discussed in the
next section.
It should be noted that the issue of requiring good initial
guesses of plant conditions in order to converge is not a prob-
lem since all the units have been converged individually in
AspenPlus before going into dynamics. The "guessed" and


the "calculated" values of the tear streams have
also been adjusted to be fairly close to each other.

TRANSITION FROM STEADY STATE
TO DYNAMIC SIMULATION
There are several items that must be taken care
of to convert a steady-state simulation into a
dynamic simulation: all equipment must be sized
and a control structure must be developed.
Luyben'I presents many details of these neces-
sary steps, which are summarized below. Not
all of the units that are available in steady-state
AspenPlus are supported in AspenDynamics, so
this limitation must be kept in mind. For ex-
ample, neither the "separator" (a fictitious com-
ponent splitter) nor a liquid-liquid extractor is
supported in the current version (Version 11.1)
of AspenDynamics.
When the steady-state simulation in AspenPlus
is exported into AspenDynamics, a "pressure-
driven" dynamic simulation should be used. This
requires that all the "plumbing" must be speci-
fied in the flowsheet. Pumps and compressors
must be inserted where needed to provide the


the easiest heuristic approach is to fix the distillate and bot-
toms specifications (using the Design Spec and Vary tools in
AspenPlus) and keep increasing the number of stages until
the required reflux ratio stops decreasing-this gives the mini-
mum reflux ratio. Then the actual reflux ratio is set at 1.2
times this minimum. Finally, the optimum feed stage can be
determined by varying the feed stage until the minimum
reboiler energy consumption is found.
The Tray Sizing section of a distillation column block in


The
convergence
of steady-state
simuklions of
flowsheets
with recycle
streams is
frequently
very difficult.
An alternative
is suggested in
this paper and
an. example
illustrates the
proposed
method.


required pressure drop for material flow. Control valves must
be installed where needed, and their pressure drops selected.
This is one of the more important educational aspects of
the procedure since most students have a poor grasp of plumb-
ing. Common errors include inserting two valves into a liq-
uid-filled line, inserting a valve in the suction of pumps, or
inserting a valve at the discharge of compressors (com-
pressor speed or its equivalent compressor work should
be manipulated).
Equipment Sizing For steady-state simulation, the size
of the equipment is not needed, except for reactors. For dy-
namic simulation, the inventories of material contained in all
the pieces of equipment affect the dynamic response, so the
physical dimensions of all units must be known.
In distillation columns, the diameter of the column, the weir
height, and the sizes of the reflux drum and the column base
must be specified. Of course, before these can be calculated,
the number of stages and the feed stage location must be set
by some heuristic or rigorous optimization method. Perhaps


AspenPlus can be easily used to provide the
column diameter. The default weir height of
0.05 m can be used. The volumetric flowrates
of liquid into the reflux drum (Stage 1, the
"condenser" in Aspen terminology) and the
liquid into the base of the column (the last
stage, or "sump" in Aspen terminology) can
be used to size the two vessels by using the
heuristic of a 10-minute holdup time. These
volumetric flowrates are given in the Hydrau-
lics page tab of the Profiles section of the col-
umn block. To have these results made avail-
able, you must go to the Report section of the
column block, select the Property Options
page tab and click the Include Hydraulic Pa-
rameters box before running the program.
For example, the liquid holdup in the reflux
drum of a column with a total condenser is
calculated from the volumetric flowrate of liq-
uid leaving the drum distillatee plus reflux).
[Drum volume (m3)]
= [Liquid volumetric flowrate (m3/
min)][10 minutes]
If an aspect ratio (length to diameter, L/D) of


2 is used, the diameter of the drum is
D = [2(Volume)/t]J"
The same procedure can be used for flash tanks and vaporiz-
ers. Flash tank vapor velocity should also be checked.'21
Heat exchanger tube-and-shell volumes can be calculated
from the heat-transfer area, which is known from the steady-
state design, if a tube diameter D (typically 0.0245 m) is se-
lected.
Area = [number of tubes] [nD] [tube length L] = N.(TtDL)
[Volume of tubes] = [iD'/4][L][Nu]
= [iD2 /4][L][Area]/[tDL] = D[Area]/4
Shell volume is approximately equal to tube volume in most
tube-in-shell heat exchangers. If the process streams in the
heat exchanger are gases, the dynamics are very fast and can
usually be ignored (specify Instantaneous in the Dynamic
section of the heat-exchanger block.
Plantwide Control When the file containing the flowsheet
is opened in AspenDynamic, a default control scheme is al-


Spring 2004









ready installed on some loops. For example, level and pressure con- eters. The chemistry is the exothermic reversible decom-
trollers are inserted on all distillation columns and reactors in the position of methanol to form DME and water in an adia-
flowsheet. This default control must be modified and supplemented batic, tubular, gas-phase reactor.
with other control loops to incorporate a stable basic regulatory 2 MeOH DME + Water
control structure.
The reaction is exothermic, and the adiabatic temperature
Plantwide Control Structure A simple heuristic method for de-
velopment of an effective plantwide control structure is presented
by Luyben, et al.13] General principles and many examples are given -
in great detail. The proposed nine-step procedure has been suc- -
cessfully applied to a large number of realistically complex indus-
trial processes. Some of the key concepts are: (1) the control struc- S
ture should guarantee that all chemical components fed into the
system are either reacted or can leave in some exiting stream; (2) a
flow controller should be installed somewhere in all liquid re-
cycle loops; and (3) all liquid levels must be controlled and
pressures in gas systems, which can be multiple units connected
together, must be controlled. j__
Length (meters)
Controller Tuning Most of the controllers are easily tuned by L
simply using heuristics. All liquid levels should use proportional- DR=1-i m LR=2 .5/5/10
only controllers with a gain of 2. All flow controllers should use a
gain of 0.5 and an integral time of 0.3 minutes (also enable filtering 00 ----- -
with a filter time of 0.1 minutes). L=10 =5 L2.
The default values in AspenDynamics for most pressure control-
lers seem to work reasonably well. Temperature controllers often 50 o00
need some adjustments. The default transmitter ranges are usually T
too large, and spans should be set at about 10% of the absolute 7
temperature level (typically a span of 100 K for moderate-tem- sso -
perature processes). --- ---
550
Distillation columns are typically controlled by manipulating S,
reboiler heat input to control the temperature on some selected tray. 450-50 5005 600
The heuristic procedure of finding a tray where the temperature Tin (K)
changes from tray to tray are large is easy to use and provides ef- Figure 1.
fective composition control in most cases. Direct composition mea- (A) Adiabatic reactor temperature profile (10 meters
surements can be used if temperature changes are too small. If very length).
(B) Effect of inlet temperature for different reactor
large temperature changes occur in the column (over 100 K), an lengths.
average temperature can be used (measur-
ing the temperatures at three or four trays,
calculating the average, and using this for m
control).
It should be kept in mind that the objec-
tive at this point is not to come up with the ,,v
"best" control structure or the optimum T c M *,
controller tuning. We only need a struc- "
ture and tunings that drive the simulation 1 ,2
to a steady state.
DMIE Product P12

EXAMPLE ("DME" PROCESS) n 0
To illustrate the use of a dynamic simula-
tor to converge a flowsheet, we select the Me .ol '
dimethyl ether (DME) process discussed by Watr Produ
Turton, et al.[4] This design text gives a flow-
sheet and some preliminary design param- Figure 2. DME flowsheet.


Chemical Engineering Education









rise is about 120 K. The temperature profile is shown in Fig-
ure 1A for a reactor that has a diameter of 0.72 m and is 10 m
long. Figure 1B shows how reactor inlet temperature Tin af-
fects the production of DME and the reactor exit temperature
Tout for three different reactor lengths. The inlet temperature
required to achieve maximum conversion decreases as the
reactor is made bigger. Since the reaction is exothermic,
the maximum conversion decreases slowly as tempera-
tures increase.
The equilibrium constant is about 6 at a reactor tempera-
ture of about 600 K, so the per-pass conversion is about 80%.
This requires a recycle of methanol back to the reactor from
the separation section. Figure 2 gives the AspenPlus flow-
sheet, and Figure 3 gives stream conditions. The NRTL physi-
cal property package is used in the simulation.
Fresh methanol and recycle methanol are vaporized and
heated to 555 K. The vapor-phase reaction occurs at about 15
atm. Reactor effluent is cooled and fed into a two-column
separation section. The low-boiling DME is the distillate prod-
uct in the first column C1, which operates at a pressure of 10
atm so that cooling water can be used in the condenser (re-
flux-drum temperature is 318 K with 99.9 mol% DME pu-
rity). The column has 32 stages, is fed on Stage 15, and is
operated with a reflux ratio of 0.5.
The bottoms of Column C is fed to the second column C2
in which water is removed from the bottom and recycled
methanol is removed from the top. This column operates at
1.1 atm and has a reflux-drum temperature of 340 K. It has
22 stages and a reflux ratio of 1.5.

Reaction Kinetics One of the most difficult parts of de-
veloping a flowsheet is getting the reaction kinetic param-
eters correctly specified. Any errors in unit conversions are


amplified by the exponential expressions.
Since the reaction is reversible, both the forward and the
reverse reaction rates must be specified. Thurton, et al., [2 pro-
vide a kinetic expression for the forward reaction rate

RF(kmol/hr/m) = .21x10 e-0.480/RTPMEOH (1)

where PMEOH has units of kPa and the activation energy has
units of kJ/kmol. Aspen insists on expressing reaction rates
in kmol/sec/m3, so we must divide by 3600. More seriously,
Aspen also insists that partial pressure be in Pa, not in kPa. A
common error is to multiply by 1000 to convert the pressure
in Eq. (1) from kPa to Pa. This is incorrect. The pressure in
Eq. (1) is in kPa, and Aspen uses Pa, so the pre-exponential
factor must be divided by 1000.

RF(kmol/sec/m3)= .21x 106 e-80480/RT pkPa ]/3600
F Pa]
= 1.21x106 e-80'480/RT 1-p /3600
1000

= 0.3361e-80,480/RTpPa (2)

Turton, et al., [4 also give information about the equilibrium
constant, but state that the published data does not seem to
match the calculations using free energies. This was confirmed
by running an R-Gibbs reactor in AspenPlus, which gave a
much lower reactor conversion than would be predicted by
the published data.
The heat of reaction is X = -11,770 kJ/kmol. so we can
estimate the activation energy of the reverse reaction from
that of the forward reaction.
EpF ER = X ER = 80,480- (-11,770)= 92,250 kJ/kmol


Figure 3. DME stream data.


Spring 2004









To find a reasonable pre-exponential factor for
the reverse reaction, the conversion data (re-
actor size; inlet flowrate, temperature, and
composition; and exit composition) given
Turton, et al., were used to back-calculate this
parameter by trial and error. The resulting ki-
netic expression is

RR(kmol/sec/m3)
= 2.7 x 10 e-92,250/RT (PDME)(Pwater) (4)

where PDME and Pater have units of Pa.

Thermal Recycle Loop A feed-effluent
heat exchanger (HX1) is used to preheat reac-
tor feed, using a portion (201 kmol/hr) of the
hot reactor effluent to heat the vapor
from the vaporizer from 427 K to 555
K. Some of the reactor effluent (134
kmol/hr) is bypassed through valve V2
to control reactor inlet temperature.
Thus there is a thermal recycle loop
in the process. Unless the reactor ef-
fluent stream is known, the HX1 equa-
tions cannot be solved. The conven-
tional procedure to get started is to use
a tear stream (see Figure 4). We specify
a stream "HIN" and provide estimates
of its flow (193 kmol/hr), temperature
(672 K), pressure (14.4 atm), and com-
position (40 mol% DME, 20 mol%
methanol, and 40 mol% water).
When this flowsheet is converged,
the source of "HIN" is changed to the
splitter T2 (deleting stream "5" in Fig-
ure 4). In the Convergence section of
the Data Browser, we select Tear and
specify HIN to be a tear stream. Then
we rerun the program to get the con-
verged flowsheet around the reactor
and heat exchanger HX1.

Material Recycle Loop Figure 5
shows the flowsheet with a tear stream
RECYCLE, whose stream conditions
have been estimated to be 72 kmol/hr,
345 K, 17 atm, 99 mol% methanol,
and 1 mol% water. The calculated con-
ditions for stream RCALC are only
slightly different from those assumed
for RECYCLE (73.2 kmol/hr, 342 K,
17 atm, 99.3 mol% methanol, and 0.1
mol% water).
The default convergence method


Figure 4. Flowsheet with energy tear stream.


Figure 5. Flowsheet with material tear stream RECYCLE.


Figure 6. Default control scheme; DME process.


Energy Tear Steam


Chemical Engineering Education









Wegstein is used (with a limit of 30 iterations), the source of the
RECYCLE stream is specified to be the discharge of pump P22
and it is specified as a Tear under the Convergence Section of the
Data Browser The program is run.
The convergence loop fails to converge.
This is a typical result in many cases. It should be noted that
Design Specs and Vary features are used on both distillation col-
umns. In Column Cl, the distillate DME product purity is speci-
fied to be 99.9 mol% DME, and the distillate flowrate is varied.
In Column C2, the bottoms water product purity is specified to
be 99.9 mol% water, and the bottoms flowrate is varied. The re-
flux ratio is fixed in both columns.
Increasing the maximum iterations does not achieve conver-
gence. Switching to the Broyden algorithm is equally unsuccess-
ful.
This failure to converge certainly does not occur in all cases
for all flowsheets, but it does occur in many cases. Senior students
spend many frustrating hours trying to get recycle loops to con-
verge.
Converting to Dynamic Simulation The diameters of the two
columns are calculated in the Tray Sizing section: 0.61 m for Cl
and 0.94 m for C2. The liquid flowrates into the reflux drums of
Cl and C2 are 0.242 and 0.131 m3/min, respectively, as found in
the Hydraulics page tab. Reflux drum sizes (D x L) in the two
columns are set at 1.2 x 2.4 m and 1 x 2 m, respectively, using an
aspect ratio of 2. The liquid flowrates into the base (to the sump
from the next-to-last stage in the column) of C and C2 are 0.213
and 0.0853 m3 min, respectively. Reboiler sizes in the two col-
umns are set at 1.1 x 2.2 m and 0.82 x 1.64 m, respectively.
The heat exchangers are assumed to be "instantaneous" since
they are gas phase. Reactor size is already specified for the steady-
state simulation. The only other dynamic unit in the flowsheet is
the vaporizer. Its liquid feed is 10,700 kg/hr with a density of


Suggested Action Message


Figure 7. Error message and suggested action message.


785 kg/m-, which gives a diameter of 1.1 m and a length
of 2.2 m to provide ten minutes of holdup.
The file is pressure checked and exported into
AspenDynamics. The default control scheme is shown in
Figure 6. Note that each column has pressure and level con-
trollers, some of which are fully connected and others with-
out the controller output signal (OP) connected to a valve.
This default control scheme must be modified to provide
an effective regulatory control scheme.
Before doing anything, an Initialization run and a short
Dynamic run should be made to confirm that all the plumb-
ing is okay and the process is correctly configured.
Closing the Recycle Loop The procedure for changing
the process structure in AspenDynamics is the same as in
AspenPlus. The stream RECYCLE is deleted. The destina-
tion of the stream RCALC is made the inlet to valve V22.
Another set of Initialization and Dynamic runs should be
made. Now the default control structure can be modified. It
is a good idea to perform Initialization and Dynamic runs
after each new change in the control structure so that any
error in controller installation can be detected individually.
The most common error is to have the wrong action in the
controller (for example, specifying reverse action when it
should be direct action).
An obvious alternative to deleting RECYCLE and reat-
taching RCALC is to delete RCALC and reattach RECYCLE.
If this is attempted, the little green box at the bottom of the
AspenDynamics screen turns red, indicating that something
is wrong. Double-clicking the red box opens the window
shown at the top of Figure 7, which states that the problem
is overspecified. Clicking the Analyze button opens the win-
dow shown at the bottom of Figure 7, which says that the
temperature and pressure of the RECYCLE stream must be
changed from fixed to variable. Clicking the ACCEPT but-
ton turns the box green again. Then Initiation and Dy-
namic are again run to make sure all is okay. Now the
default control scheme can be modified.
Plantwide Control Structure Using the method pro-
posed in Luyben, et al., 31 a plantwide control scheme is de-
veloped that features the following loops:
Thefeedflow to thefirst column is flow controlled. This
puts a flow controller in the liquid recycle loop. It also has
the advantage of keeping a fixed steady flow to the column.
Since the final DME product is produced in this column,
product quality variability is minimized by not permitting
disturbances to enter this column.
The vaporizer level is controlled by manipulating the fresh
methanol feed. This guarantees that only the amount of
methanol that is being consumed in the reactor will be fed
into the process. If more DME production is required, the
setpoint of the flow controller on Column Cl feed can be
increased. Using reactor inlet temperature to change
production is not effective because of the reversibility of the
reaction. For example, sometimes increasing reactor inlet


Spring 2004









temperature produces a decrease in the production of
DME.
Vaporizer pressure is controlled by manipulating steam
(heat input Q) to the vaporizer
Reactor inlet temperature is controlled by the HX1 heat
exchanger bypass flow (valve V2).
Condenser HX2 exit temperature is controlled by manipu-
lating cooling water (heat removal Q).
The pressure in each column is controlled by condenser
heat removal.
The reflux-drum level in each column is controlled by
manipulating distillate flow.
The base level in each column is controlled by manipulat-
ing bottoms product flow.
A temperature in each
column is controlled by
manipulating reboiler
heat input. The steep part VAPw
of the temperature profile LO
in Cl is at Stage 17 and !-0 -------i
in C2 at Stage 19, so MEOHT.
these tray locations are T
selected. The temperature
setpoint in Cl is 380 K
and in C2 is 370 K.
The reflux ratio in each
column is controlled.
The distillate flowrate is
measured, multiplied by
the desired reflux ratio
(0.5 in Cl and 1.5 in m
C2) and this signal sets
the flowrate of reflux.
Figure 8 gives the



TABLE 1
DME Controller Parameters

Transmitter Valve Integral
Range Range Time
Controller min/max min/max T,
(Action) (units) (Units) K (min)


TCin (Dir) 500/600 (K) 0/100 (%) 2 5


TCcond (Rev) 300/400 (K) -2030/0 (kcal.sec) 2 20


TC1 (Rev) 300/400 (K) 0/690 (kcal/sec) 1 20


LC12 (Dir) 0/3.4 (m) 0/100 (%) 2


TC2 (Rev) 300/400 (K) 0/688 (kcal/sec) 1 20


LC22 (Dir) 0/2.52 (m) 0/100 (%) 2
_If1J;.E- Qlc2.f sae) 20 --5


AspenDynamics flowsheet with this control structure installed. All
level controllers are proportional with KC = 2. Table 1 gives con-
troller tuning parameters and transmitter spans. The dynamic simu-
lation is run until it reaches a steady-state condition. This may take
several "process hours" (several minutes of computer time), de-
pending on the complexity of the flowsheet. When running the dy-
namic simulation out to a steady state, valve V22 in the recycle line
went wide open and the reflux-drum level in Column C2 began to
rise. The power to the pump P22 was increased to solve this valve
saturation problem. The details of how this is done are given in
Luyben.~'~
The same valve saturation problem occurred in V12 when the


Figure 8. Plantwide control structure: DME process.


Figure 9. Dynamic responses to 20% increase in feed to
column C1.


Chemical Engineering Education










setpoint of the Cl column feed flow controller was increased 20%
(from 236 to 402 kmol/hr). The power to pump P12 was increased.
Notice that the steady-state signals to valves V22 and V12 are 20%
and 29%, respectively, instead of the normal 50%, because of the
higher valve pressure drops.
Figure 9 shows the dynamic response of the system to a 20%
increase in column Cl feed. The initial condition is the steady-state
condition. The fresh feed of methanol increases from 262 to 308
kmol/hr and DME product increased from 131 to 154 kmol/hr
(17.5% increase in production rate). The purities of both the
DME and the water products are maintained for this large dis-
turbance. The system takes about two process hours to come to
the new steady state.
The one-recycle process illustrates how easily the flowsheet can
be converged by using dynamic simulation. Figures 10 and 11 give
an example of a process with two recycles. The methyl acetate and
butanol reactants not consumed in the reactor are separated in a three-
column separation section and recycled in two different streams. More
details of this example are available from the author.

CONCLUSION
The convergence of steady-state simulations of flowsheets with


Figure 10. Flowsheet for two-recycle butyl acetate process.


recycle streams is frequently very difficult. An alterna-
tive is suggested in this paper and an example illustrates
the proposed method. The steady-state simulation is con-
verted into a dynamic simulation, and the recycle loops
are converged by letting the dynamic simulation run to
steady-state conditions.
The method depends on the development of a base-
level regulatory plantwide control structure, which can
be obtained by following a simple heuristic design pro-
cedure. Simple controller tuning rules can be applied to
eliminate detailed and lengthy controller tuning efforts.
There is, of course, an additional benefit for this ap-
proach. The dynamic simulation can also be used to look
at the dynamic effects of alternative design conditions
(flowsheet structure, operating conditions, equipment
sizes, etc.). This approach, which is called "simultaneous
design," is a design philosophy in which both the steady
state and the dynamic performances of a process are
considered at all stages of the development of a pro-
cess. The book by Seider, et al.,151 discusses this ap-
proach in more detail.

NOMENCLATURE


D diameter
DME dimethyl ether
E activation energy
FF fresh feed
K controller gain
L length
P partial pressure of component j
TCn tray temperature controller in column n
Tin reactor inlet temperature
Tout reactor exit temperature
T, integral time constant in PI controller
(minutes)
X heat of reaction

REFERENCES
1. Luyben, W.L., Plantwide Dynamic Simulators for
Chemical Processing and Control,
Marcel Dekker (2002)
2. O'Brien, N.G., and R.G. Franks, "De-
velopment and Application of a General
Purpose Analog Computer Circuit to
Steady-State Multicomponent Distilla-
tion," Chem. Eng., 55, 21 (1958)
3. Luyben, W.L., B.D. Tyreus, and M.L.
Luyben, Plantwide Process Control,
V1 McGraw-Hill (1999)
4. Turton, R., R.C. Baille, W.B. Whit-
ing, and J.A. Shaeiwitz, Analysis, Syn-
VJ v thesis and Design of Chemical Pro-
cesses, 2nd ed., Prentice Hall, pg. 939
(2003)
5. Seider, W.D., J.D. Seader, and D.R.
Lewin, Product and Process Design
Principles, John Wiley and Sons, New
York, NY(2004) 0


Spring 2004


PFi
Figure 11. Plantwide control structure for two-recycle process.









[e] f classroom


USING SMALL BLOCKS OF TIME

FOR ACTIVE LEARNING AND

CRITICAL THINKING


STEPHEN J. LOMBARDO
University of Missouri Columbia, MO 65211


During lecture periods, five- or ten-minute time slots
sometimes arise that do not warrant introducing new
topics in a course but do afford the opportunity for
learning. I use these short periods of time with two types of
learning exercises: "Estimation Problems" and the "Top 10
Chemicals." Both are active learning activities that involve
critical thinking.11-7 In addition, they help maintain student
interest and motivation, and they also can serve to provide a
break in the class period.
Because both of these activities are used in an introductory
thermodynamics course, the estimation problems are slanted
to reinforce principles of thermodynamics. They also include,
however, problems that touch on subject matter found in other
engineering and physical science courses. The Top-10-Chemi-
cals exercise reinforces both principles of chemical engineer-
ing and process chemistry more typically covered in chemis-
try courses, or (possibly) not at all in the curriculum.

ESTIMATION PROBLEMS
The short estimation problems are used to improve the abil-
ity of students to use limited information to quickly find ap-
proximate order-of-magnitude solutions. These problems
are typically presented to the students as short statements on
the chalkboard; the students then spend 2-5 minutes working
together in small groups to develop a problem-solving strat-
egy and an order-of-magnitude solution estimate. I ask the
students to volunteer at the blackboard to present their an-
swer to the other students, and very often students are will-
ing to do this. I take this as an important indicator that they
feel comfortable and confident in demonstrating their knowl-
edge in front of their peers and me. This type of problem
solving, where students teach each other, also serves to rein-
force their own intellectual capabilities and combats the per-


ception that only the professor can solve problems.
Table 1 is a partial list of the estimation problem statements
used so far, and the solutions to two of them (#13 and #2) are
presented in more detail below. The first estimation problem,
"Estimate the power usage of a home microwave oven," re-
inforces principles of thermodynamics. For this problem, the
students invariably choose to heat a cup of water from room
temperature to the boiling point. This leads to an energy, which
can be computed from the enthalphy change of water,
AH=mC AT, which is then converted to power by dividing
AH by a heating time of 3-5 minutes. With this "seat of the
pants" approach, the students determine a power usage of
several hundred watts, which is certainly a valid estimate for
small microwave ovens and is within an order of magnitude
of the power for large units.
At this point, the student who has presented the answer sits
down to a well-deserved round of applause from the instruc-
tor and the other students; the applause is certainly well de-
served, not only for presenting in front of the class but also
for presenting extemporaneously.
Because we are in a thermodynamics class, I then lead a
discussion of how the result can be formally obtained from
the First Law of Thermodynamics for a closed system ap-

Stephen J. Lombardo received his BS de-
gree from Worcester Polytechnic Institute and
his PhD degree from the University of Califor-
nia, Berkeley, both in chemical engineering.
He worked for seven years in industry in the
areas of ceramic materials and ceramic pro-
cessing before joining the Department of
Chemical Engineering at the University of Mis-
souri-Columbia in 1997.


Copyright ChE Division of ASEE 2004


Chemical Engineering Education









plied to the initial and final states of water in the cup
TABLE 1(see Figure 1). For this problem, the terms for changes
in kinetic and potential energies are taken as zero, and
Problem Statements for Estimation Problems the heat transfer, Q, and the work, W, are also taken as
Estimate... zero. The type of microwave (mw) energy that crosses
the system boundary, however, is added to the First Law
1. ...the number of BS chemical engineers graduated each year in the as Qmw This aspect of the problem reinforces the prin-
USA ciple that when we use the First Law as a bookkeeping
2. ...the annual residential energy use in the United States tool to conserve energy, we may need to add terms to
3. ...the number of cars sold per year in the United States account for the process under consideration. Although
4. ...the concentration of atoms in the surface of a solid the First Law of Thermodynamics for a closed system
5. ...the annual gross domestic product of the United States is most often formulated in terms of the internal en-
6. ...the motor size of a garage door opener ergy, for liquids the constant volume and constant pres-
7. ...the energy required to climb two flights of stairs sure heat capacities are approximately equal, and thus
8. ...the power output of the human heart the internal energy and enthalpy changes are approxi-
9. ...the water pressure in town mately equal as well.
10. ...the distance an auto skids during braking from high speed At this point, after the students realize how their "es-
11. ...the tons of cooling an air conditioner must supply for a timated" solution is, in fact, well grounded in funda-
residential home mental aspects of thermodynamics, it is instructive to
12. ...the carbon dioxide production from automobile exhaust per day define the system as the magnetron or, alternatively, as
in town the microwave oven itself. Figure 1 demonstrates the
13. ...the power usage of a home microwave oven application of the First Law of Thermodynamics for
14. ...the number of toothbrushes sold per year in the United States these cases as well. If the magnetron is taken as the
15. ...the entropy production of a student while sitting in class system, the First Law reduces to a form where values
16. ...the energy obtained from a windmill to calculate the electrical work into the magnetron, Welec,
17. ...the cost to produce a gallon of gasoline in a refinery or the microwave energy emanating from the magne-
tron, Qmw, are not readily available. Thus, it is not pos-
sible to obtain a numerical value for the energy use with
this approach from the given problem statement.
Estimate the power usage of a home microwave oven. If the entire microwave oven is taken as the system,

Q 1) First Law with cup contents as system: then again we cannot immediately estimate the power
input; equating the internal energy change of the mi-
A liqud kA p = Q. a,. crowave with the internal energy change of the con-
/ / tents of the cup, however, leads to the same solution
,AUii,d AH mJ Cp dt = 0.3x4.18(100-20)= 100 kJ
AUli AH m dt 0.3x4.18100-20) 100 kJ presented earlier. Thus, the students can observe how
Power Required: judicious selection of the system as the contents of the
-- Q,,, 100,000 J cup facilitates quickly achieving a solution estimate.
P. =- 500 Watts These other forms of the First Law also naturally lead
Time 3 minx60 s/min to a discussion of where thermal heat losses or storage

W.1. 2) First Law with magnetron as system (AU,.,. 0): may occur (the cup, the magnetron, the power cord)
---.-.- and how these influence the energy efficiency of the
. We = Q,. device. At the end of this exercise, which examines a
Magnetron ) common process in simplified terms, the utility of esti-
-------. Unless we have electrical values,
"Q -- we cannot proceed. mation and the ramifications of process simplification
are reinforced for the students.
3) First Law with microwave as system: A second type of estimation problem, determining the
....... W... W.I. = AU,, AUI,q,d m C, dt annual residential energy use in the United States, is
Microwave not directly related to course content. Here, students
.. same as approach 1) above typically elect to use their monthly electrical energy bill
and an estimate of the cost of electricity of 5-10 cents
per kW-h to arrive at an order-of-magnitude answer (see
Figure 1. Representation of the chalkboard for estimating Figure 2).
the power usage of a home microwave oven. Often, for this problem some students can conceptu-


Spring 2004









alize the solution strategy but lack specific values for
the U.S. population, for the number of U.S. house-
holds, or for the energy cost per kW-h. Collectively,
however, the class typically knows one or more of
these values, and thus an estimate can be computed
and presented to the other students. To conclude this
problem, we often use a simple error analysis to as-
sess how certain we are of specific numbers and thus
the overall estimate. We may know the population or
number of households to within 10%, the cost of en-
ergy to within a factor of two, the number of months
per year exactly, and the monthly dollar usage to within
a factor of five. Taken together, these uncertainties
allow us to bound the error in the overall estimate.
For this particular problem, the estimate of annual resi-
dential energy consumption of 1012 kW-h is within
25% of the reported value.Es8 I usually conclude this
particular problem by mentioning that the total U.S.
electricity usage is approximately 1/3 residential, 1/3
industrial, and 1/3 commercial.
For some of the estimation problems, no student is
willing to present a solution strategy at the chalkboard.
In these cases, I usually draw the solution from the
students by asking a sequence of questions that steps
them through the thought process necessary to ar-
rive at an estimate. Often, the students actually do
know the solution strategy, but are not 100% cer-
tain. After observing the "estimation" methodol-
ogy a few times, additional students then develop
the confidence to share their solution with the rest
of the class.

THE TOP 10 CHEMICALS

The discussion of each top-10 chemical is initiated
by having the students
guess what is the next
chemical on the running
list. Usually, the students Estimate the ann
know about 3-5 of the
chemicals, but generally # of US househo
# of US househo
not in order. The top 10
chemicals on a mass basis Average househo
produced per year in the Cost of energy pi
United States are listed in
Table 2.E[9 A discussion of
each chemical is conducted Total electricity
first by querying the stu-
dents for the source of raw
materials for the chemical.
For sulfuric acid, the source
of sulfur is either mined
sulfur, spent sulfur, or sul- Figure 2.
fur obtained as hydrogen


sulfide gas from hydrodesulfurization units in refineries.
The students are next queried as to the sequence of likely reactions
to convert the raw material to the desired product. The reactions to
convert sulfur or hydrogen sulfide to sulfur dioxide, then to sulfur
trioxide, and then to sulfuric acid are thus obtained from the students,
who use their knowledge of general chemistry to ascertain what might
be likely reactions based solely on chemical formulas. A representa-
tion of the appearance of the chalkboard for the reactions leading to
sulfuric acid is shown in Figure 3.
The uses of sulfuric acid are next discussed. The main use is to
digest phosphate-containing rock to produce phosphate fertilizer. I
tell the students that it is comforting to know that the main use for the
number-one chemical produced is ultimately so that we can feed our-
selves.
Each time we cover one of the top 10 chemicals, the list is aug-
mented, and as the semester progresses, knowledge gained from pre-
vious chemicals can be applied to subsequent ones; the parallels be-
tween ethylene (#3) and propylene (#7), for example, are obviously
quite pronounced. The fact that sulfuric acid is the number-one chemi-
cal, primarily used to digest phosphate-containing rock, should make
it no surprise that phosphoric acid, used in phosphate-based fertilizer
production, also appears in the top ten.
In a similar analogy, polyethylene, the sixth most-produced chemi-
cal, is derived from ethylene, the third most-produced chemical and
the number-one organic chemical. Other commonalities among the

TABLE 2
Top 10 Chemicals Produced in the
United States in 2001[9',1a

1. Sulfuric acid 6. Polyethylene
2. Nitrogen 7. Propylene
3. Ethylene 8. Ammonia
4. Oxygen 9. Chlorine
5. Lime 10. Phosphoric acid


Representation of the chalkboard for estimating the annual
residential energy use in the United States.


Chemical Engineering Education


ual residential energy usage in the United States.

lds 100 million
Id monthly electrical bill $100
er KW-h z $0.07

108 household x $100/month-household x 12 montl
usage -
$0.07/KW-h

S1012 KW-h










top ten chemicals also become apparent during the course of
the semester. For example, three of the top ten (sulfuric acid,
ammonia, and phosphoric acid) are used in the manufacture of
fertilizers for growing food.
At the end of the semester, the students are given a copy of
the Top 100 Chemicals.E91 We briefly examine it to glean some
long-term trends. For example, thirty years ago, nitrogen was


Top 10 Chemicals
1. Sulfuric Acid (HzSO4) ~ 80 billion Ibs produced per year
Source of Raw Materials:
S from sulfur mines
HS2 from petroleum refineries and natural gas
Spent sulfuric acid (contaminated and diluted)

Reactions:
S
H2S V catalyst
+ 02 -+ SO + Oz -+ 503 +H20 -+ H2504
Spent
Acid

Major Uses:
~ 2/3 to digest phosphate-containing rock to produce
phosphate fertilizers for agricultural uses
~ 1/3 miscellaneous uses
metal/ore processing
Petroleum refining
-pulp and paper manufacturing

Figure 3. Representation of the chalkboard for
discussing the number-one Top-10-Chemical,
sulfuric acid.


TABLE 3
Selected Student Comments from End-of-Year Evaluation

I liked the Top 10 Chemicals, estimation problems, etc. They are
a nice way of "taking a break" while still learning.
Instructor was very enthusiastic, encouraged involvement in
problems and estimation.
Changing the topic several times a class helps us to remain
focused and to continue to learn. Has done a wonderful job.
I liked the Estimation Problems and Top 10 Chemicals. I felt this
course related well to real-life problems.
Other non-lecture based activities are especially good.
Top 10 Chemicals and Estimations are good teaching parts of
this course.
I like the fact that the Top 10 Chemicals and the Estimation
Problems...get the class communicating and thinking.
Top 10 Chemicals and Estimation Problems keep us interested
and force us to think like engineers (problem solvers).
The Estimation Problems show us that we know and can
accomplish a lot more than we think we can.


not in the top ten.
Methyl-t-butyl ether (MTBE) is another interesting ex-
ample to consider. Its emergence as a top 10 chemical can
be linked to its role as an octane enhancer as lead-a known
carcinogen in gasoline-was phased out. Now, MTBE is a
suspected carcinogen, and its production can be expected
to decline as new octane enhancers are introduced into fu-
els. Two pairs of chemicals-nitrogen and oxygen, and caus-
tic and chlorine-which are often produced in tandem, af-
ford an opportunity to bring in some economic consider-
ations related to pricing and the imbalance of product sup-
ply and demand.

SUMMARY
Two activities, "Estimation Problems" and the "Top 10
Chemicals," are used as active learning and critical think-
ing exercises to use small blocks of class time. Although
each exercise addresses different aspects of chemical engi-
neering, they have a common basis in delivery in that both
involve extensive class participation, which helps to main-
tain student attention. In addition, these two learning op-
portunities can serve an important educational role by ex-
panding the knowledge base of the students. Probably even
more important-the students enjoy both interactive exer-
cises and more than half of them comment favorably on
one or both of these exercises in their end-of-year course
evaluations; some selected comments are listed in Table 3.
Finally, because of the participatory nature of these exer-
cises, they help build a strong rapport between the students
and the instructor.

REFERENCES
1. McKeachie, W.J., Teaching Tips, 8th ed., D.C. Heath & Co., Lexing-
ton, MA (1986)
2. Kurfiss, J.G., Critical Thinking: Theory, Research, Practice, and Pos-
sibilities, ASHE-ERIC Higher Education Reports, Washington, DC
(1988)
3. Bonwell, C.C., and J.E. Eison, Active Learning: Creating Excite-
ment in the Classroom, ASHE-ERIC Higher Education Reports,
Washington, DC (1991)
4. Felder, R.M., "It Goes Without Saying," Chem. Eng. Ed., 25, 132
(1991)
5. Felder, R.M., "How About a Quick One?" Chem. Eng. Ed., 26,18
(1992)
6. Johnson, D.W., R.T. Johnson, and K.A. Smith, Active Learning: Co-
operation in the College Classroom, Interaction Book Company,
Edina, MN (1998)
7. Souza, D.A., How the Brain Learns, National Association of Sec-
ondary School Principals, Reston, VA (1995)
8. Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed., (elec-
tronic resource), Wiley, New York, NY (2000)
9. List of top 10 chemicals produced in the United States obtained from
Guide to the Business of Chemistry, American Chemistry Council,
Arlington, VA (2002)
10. A description of the production methods, uses, and sources of raw
materials for each of the top ten chemicals can be obtained from
Reference 8. 0


Spring 2004










r classroom


TEACHING NONIDEAL REACTORS

WITH CFD TOOLS


Luis M. MADEIRA, MANUEL A. ALVES, ALiRIO E. RODRIGUES
Universidade do Porto 4200-465 Porto, Portugal


Behavior of nonideal reactors and their flow pattern
characterization are issues taught in a majority of
chemical reaction engineering (CRE) courses[I'41 (de-
pending on the curriculum structure, usually at the under-
graduate level) and therefore they are addressed in most CRE
textbooks.[g.,1-58] In this respect, the classical stimulus-response
tracer experimentsE91 are essential to obtain theoretical func-
tions such as the residence time distribution (RTD), which
are crucial for diagnosis of equipment operation, reactor
modeling, and prediction of conversion. Comprehension of
the involved concepts, which are not easily grasped by stu-
dents, can be improved with laboratory sessions-but this is
not always feasible. This disadvantage can be overcome, how-
ever, by implementing computer simulation of tracer experi-
ments,[110 particularly with a commercial computational fluid
dynamics (CFD) tool.'"31]
CFD codes have been widely used for simulating flow pat-
tern in real systems, and their use in a chemical engineering
undergraduate course has several advantages, as pointed out
by Sinclair.[41 Among those advantages are: first, the use of
color in CFD plots allows students to visualize flow behav-
ior, and this visualization of the flow phenomenon can sig-
nificantly facilitate and enhance the learning process. Sec-
ond, students can explore the effects of change in the system
geometry, system properties, or operation conditions. All these
advantages, apart from the fact that students like to experi-
ence software tools used in industry (which are usually user-
friendly), justify implementation of CFD codes in CRE
courses, particularly when teaching RTD theory.
Such tools have some disadvantages and/or limitations,
however. Commercially available codes may be extremely
powerful, but their operation requires a high level of skill and
understanding to obtain meaningful results. In addition, CFD
cannot be adequately used without continued reference to ex-
perimental and/or analytical validation of numerical results.113
The case study described here can be implemented as home-
work for students taking a CRE course dealing with nonideal
reactors. Simulations can be performed using a commercial


package such as Fluent, but it is advisable that a brief tutorial
be provided so students can quickly familiarize themselves
with the program. In this tutorial, the essential steps that must
be followed for any simulation should be outlined. Different
reservoir/reactor geometries can be provided for different
groups of two-to-three students, but each group should per-
form its own parametric study (e.g., evaluate the effect of
space-time, or Reynolds number, in flow pattern character-
ization, or Damkbhler number in reaction simulations). Fi-
nally, the students should compile the results obtained by other
colleagues and discuss them in a final written report.
Using 2-D reservoirs/reactors with laminar flow (described
in detail in the next section), students should perform the fol-
lowing tasks, which are also the objectives of this paper:
1. Characterize the hydrodynamics in the vesselss.
2. Determine the RTDfrom tracer experiments, which
includes diagnosis of reservoir/reactor operation.
3. Predict the conversion in a continuous-flow system.

CASE STUDY FORMULATION
AND SIMULATION WITH FLUENT
It is well known that the mean residence time (fr) has an
important effect on the performance of some large natural
Luis M. Madeira is Assistant Professor of Chemical Engineering at the
University of Port (Portugal). He graduated in Chemical Engineering
(1993) and received his PhD (1998) from the Technical University of
Lisbon. He teaches chemical engineering laboratories and chemical re-
action engineering. His main research interests are in catalytic mem-
brane reactors, heterogeneous catalysis (including environmental
catalysis), and fuel cells.
Manuel A. Alves is a Teaching Assistant in the Chemical Engineering
Department at the University of Porto (Portugal). He graduated in Chemi-
cal Engineering from the University of Porto in 1995. His main research
interests are in the simulation of viscoelastic flows and the development
of numerical algorithms for computational fluid dynamics.
Alirio E. Rodrlgues is Professor of Chemical Engineering at the Univer-
sity of Porto (Portugal). He graduated in Chemical Engineering from the
University of Porto in 1968 and received his Docteur-lngenieur degree at
the University of Nancy (France) in 1973. He is Director of the Laboratory
of Separation and Reaction Engineering-LSRE (www.fe.up.pt/lsra).His
main research interests are in cyclic separation/reaction processes (simu-
lated moving bed, pressure swing adsorption, and parametric pumping
technologies) and chemical reaction engineering.
Copyright ChE Division of ASEE 2004


Chemical Engineering Education









conversion systems, such as biological lagoons, since it af-
fects the biological conversion of biodegradable matter. More-
over, the geometry of the reservoirs seems to affect the RTD,
and thus ir. It is therefore very important to perform tracer
experiments in these systems (or try to estimate the RTD func-
tion) in order to better design such wastewater treatment
plants. The mean residence times in these large reservoirs or
lagoons are extremely high (ranging from one day up to several
months'l5'6]), however, so tracer experiments are impractical.
Possible strategies to overcome this problem involve ob-
taining the RTD on pilot-scale setups with various geometries,
where tracer experiments are easily conducted, and perform-
ing a scale-up analysis, or deriving the RTD by solving the
Navier-Stokes and diffusion-convection equations that stu-
dents learned in the fluid mechanics curricula.[eg"161 Commer-
cially available CFD packages (e.g., Fluent, CFX, Fidap,
Phoenics, STAR-CD, FLOW3D, etc.) can readily solve the
balance equations for reactor operation coupled with the Navier-
Stoke equations. Thus, we adopted the second approach.
The case study considers a 2-D reservoir with dimensions
L (length) and H (height) in laminar flow and isothermal con-
ditions (see Figure 1). Reservoirs with different aspect ratios
(L/H) were considered, varying between 0.5 and 20 (with H
= 0.1 m). Both the inlet and the outlet boundaries of the res-
ervoirs have a height of 0.01 m, with distances from the bot-
tom of the reservoir of 0 and 0.02 m, respectively. A fully
developed parabolic velocity profile is imposed at the inlet
boundary

ux =Umx y-H/20] (1)

H /20

where Uma is the fluid velocity at the center of the inlet bound-
ary, U, = 1.5 Umean. A constant species concentration profile
is assumed at the inlet boundary. On the walls (see Figure 1),
no-slip conditions are assumed (ux = u = 0) and a null flux
(zero gradient) of species concentration is imposed. At the
outflow boundary condition, the CFD code extrapolates the
required information from the interior cells (a zero diffusion
flux is assumed for all flow variables in the direction normal
to the exit plane). In the simulations, the Reynolds number,


Stationary walls
0.7H
H
Inlet boundary Outflow boundary
condition condition

) 0.2 H
u (Y) $0.1H 0.2H
x
L
Figure 1. Sketch of reservoir geometry.


here defined based on inlet conditions, ranged from 1 to 100.
Changing Re for a given fluid and reservoir is equivalent to
changing the fluid velocity (and thus the residence time).
The governing equations to be solved are (for an incom-
pressible fluid)

Continuity
au+ = 0 (2)
ax ay
Momentum


u a(u2 +) a(uxuy) >p (a2u a2 '|
Pux =- -+ 2Uax +2Ux 3)
at ax ay ax x2 (3)

auy a(uxuy) (u) ap (u a2U
p- +P y+ -- --gp + -+- (4)
at ax ay ay x2 x 2 )y J


Species Transport
ac a(uxC) a(uyC) (a2C 2C"
+ x =DIx2 + -y + S(C) (5)
at ax ay Dax2

In Eq. (5), S(C) is a source term. For the transport of an inert
tracer, by convection-diffusion, S(C) = 0, while for the trans-
port of a reagent species, S(C) = -kC (assuming a first-order
irreversible reaction).
Simulations were run with the commercial package Fluent
6.0. The fluid considered is water (v = L/p = 10-6 m2s-), and
a tracer solution was created in Fluent's database with identi-
cal properties of water, so that it does not affect reactor hy-
drodynamics. A molecular diffusivity of 5x10-'1 m2s-' was
considered, which is a typical value for liquids.tl71 A tracer
step input was used, with uniform concentration across the
entrance section, together with the parabolic velocity profile
defined in Eq. (1).
Fluent can simulate both the hydrodynamics and chemical
reaction processes, and therefore the reservoirs previously
considered can be used for modeling continuous-flow reac-
tors (e.g., biological lagoons). In this case, simulations were
run by defining a reactant and a product, both with properties
identical to water. The reactor is initially full of water (inert),
the laminar flow is established; after that (time t = 0) the
reactant is fed to the reactor, similarly to the tracer step input.
Conversion of reactant is calculated based on the time evolu-
tion of species concentration at the reactor exit, obtained from
a mass-weighted average formulation.
We must point out that to achieve a high level of accuracy,
all the simulation results presented in this paper involved a
detailed analysis of the numerical algorithms, the mesh em-
ployed, and the time step adopted (in transient simulations).
For instance, the QUICK scheme of Leonard[l]l was selected
for discretization of the convective terms, a second-order


Spring 2004









implicit formulation was used during unsteady stimulation,
and the computational grid contained typically 100x100 ele-
ments (note that for reservoirs with L/H >> 1, it is conve-
nient to use a larger number of cells in the x-direction). It is
always a good practice to perform the calculations on several
meshes with different levels of refinement in order to obtain
mesh-independent results. A similar procedure should be
adopted regarding the time step used in transient calculations.
A control-volume approach is used by Fluent to numerically
solve the governing equations. "

RESULTS AND DISCUSSION
> Hydrodynamic Characterization
Figure 2 shows contour plots of the stream function within
the reservoir, with L/H = 1, which illustrates the trajectories
of the fluid elements (streamlines). For a given inlet velocity
(or more broadly speaking, a given Reynolds number), the
formation of a recirculation zone above the entrance of the
reservoir, where velocity is small, is evident, thus suggesting
formation of a stagnant region. It is noteworthy that the im-
portance of such a region increases with Re, becoming par-
ticularly large for Re values around 100 where fluid trajecto-
ries are almost linear. Lower Re numbers lead to a smaller
stagnant region and more curved streamlines. For Re < 1,
inertia is negligible and the trajectories obtained are equiva-
lent to creeping-flow conditions.
For longer reservoirs, the conclusions are similar, but now
the importance of the recirculation zone decreases (for the
same fluid inlet velocity). Indeed, the steady state stream-
lines shown in Figure 3, obtained for a reservoir with L/H =
5, show formation of a stagnant zone above the entrance,
where the size increases with the Reynolds number. Com-
parison with the stream function contours of Figure 2, how-
ever, shows that for the same Re, the fraction of dead volume
decreases when the geometric ratio L/H increases.
> RTD Determination from Tracer Experiments
After performing steady state simulations, students can
proceed to transient runs, but they must first define a tracer
step input at the inlet boundary condition. They must also be
aware that for t = 0, no tracer exists within the reservoir and
that the laminar flow is already established. They must first
initialize the entire domain with a null tracer concentration
and wait until the laminar regime is established before intro-
ducing the tracer step change at the inlet. After that, the CFD
code solves the convection-diffusion equation that describes
the tracer transport in the reservoir and the concentration field
of tracer under transient regime is obtained. Particularly in-
teresting is its concentration at the outflow boundary, Co(t).
The contours of tracer concentration throughout the reser-
voir along time are also very interesting because they pro-
vide a good perspective on the evolution of concentration
fronts. For a reservoir with L/H = 1, some frames were re-
corded at different times and are shown in Figure 4. They


show that only for 0 = t/T around 0.22 can one start to 'see'
tracer at the reservoir exit. In addition, even for a very long
time of operation (about five times the residence time), the
reservoir is not completely full of tracer, due to the stagnant
zone previously identified. To better illustrate this transient
behavior, it is possible to create an animation sequence with
Fluent, using several frames obtained from the previous simu-
lation. This was done and is available at ~mmalves/cfd/reactor/index.htm>.
With the data of transient tracer concentration at the outlet
of the reservoir, which can be exported to an ASCII file, stu-
dents can then compute the so-called Danckwerts' F curve,
the normalized response of the reactor to a step input,

F(t)= C (6)
Cin


Figure 2. Steady state contours of the stream function
for the reservoir with L/H = 1 as a function of the
Reynolds number.
Re=1
^^^_ ---------------- :::^



Re=5



-R-10




Re20




Figure 3. Steady state contours of the stream function
for the reservoir with L/H = 5 as a function of the
Reynolds number.


Chemical Engineering Education









Data can then be manipulated with a spreadsheet program
such as Microsoft Excel. Results shown in Figure 5 show the
expected F curve, which only reaches the asymptotic value
of 1 for very long times. It is worth mentioning the use of a
logarithmic time scale, showing that tracer starts to exit the
reservoir at around 0 = t/T = 0.22 due to its transport by convec-
tion, while fluid elements that enter the stagnant region only
come out (by diffusion) much later (see detail of Figure 5).
With the response to a step input, the RTD function can
now be computed by

E(t)dF(t) (7)
dt


9=0.5 9=1









9=2 9=5









Figure 4. Transient tracer concentration contours for the
reservoir with L/H = 1 (Re = 10).

1.0 T-c- -------------------
F(S)
0.8
1.000
0.6 F(9)
0 995
0.4

0.2 0.990
10 100 1000
0.0
0.1 1 10 100 1000

Figure 5. Danckwerts' F curve for the reservoir,
with L/H = 1 (Re = 10).


or, in terms of reduced time 6,
dF(0)
E()= E(t) (8)

Figure 6 shows the RTD curves as a function of the
Reynolds number, which gives an idea (for a given fluid)
how space-time affects the RTD function. In all cases, the
RTD curves evidence a very long tail and that a large fraction
of fluid elements exits the reservoir with ages younger than
the space-time, T. Both features are indicative of the exist-
ence of large stagnant regions, together with slow recirculat-
ing flows near the inlet, as can be seen in Figure 2. The fact
that Re affects the RTD curve is also visible in Figure 6. Higher
Reynolds numbers imply that the fluid elements start to leave
the reactor sooner and a higher fraction of fluid elements has
a smaller residence time.
The effect of the reservoir geometry on the RTD is shown
in Figure 7. It can be seen that when L/H increases, the curves
are shifted to the right, i.e., the mean residence time seems to
increase because the importance of the recirculating zone
decreases. This was previously found in the steady state
streamlines shown in Figures 2 and 3. It is particularly note-
worthy that for very long reservoirs, one tends toward an as-
ymptotic E(O) curve, also shown in Figure 7. This curve cor-
responds to the case of laminar flow between parallel plates,
given by"19


1 2

2
E(9)= 30
S 0< 2-
3


It must be stressed that for this situation, i.e., laminar flow
between parallel plates, the parabolic profile is characterized
by a maximum velocity at the center that is 1.5 times the
average velocity, while for flow in pipes this ratio is 2.
The RTD functions can then be used to calculate the mean
residence time, defined as
r


0 0.1 0.2 0.3 0.4 0.5
0

Figure 6. Effect of the Reynolds number on the residence
time distribution (reservoir with L/H = 1).


Spring 2004


9= 0.22






. -


O=0.1






J











t= tE(t)dt (10)
0
oro
or

Or = -= 9 E(9)d (11)
0

For a closed-closed system (i.e., with no dispersion) and if
no bypass or stagnant regions exist, it is well known that the
mean residence time and space-time are equal.[7'20] Due to the
very long tail of the RTD function, however, this is only veri-
fied if simulations are run up to very high times, typically 0
of 0(103). Otherwise, the normalization condition


f E(t)dt = f E(9)d9 = 1
0 0
is not satisfied and the computed mean residence time is
smaller than the real space-time.
Calculation of the mean residence time is very important
for evaluation of malfunctions during the reactor's operation.
Indeed, a straightforward way to diagnose the reactor's flow
performance consists of comparing the computed value of
the mean residence time from Eq. (10) with the space-time
(7 = V/Q, where V is the total volume of the reactor and Q is
the volumetric flow rate), which is equivalent to comparing
9r with 1. Disagreement between these two values may indi-
cate the existence of bypasses, dead volumes, etc. For in-
stance, as indicated by Froment,'5s if a region of the vessel
retains a portion of the fluid for an order of magnitude greater
than the mean residence time of the total fluid, then for all
practical purposes, that portion is essentially at rest and the
region is wasted space in the vessel.
The values obtained for the mean residence time shown in
Tables 1 and 2 were calculated by integration of the RTD
curves up to 0 = 10. It is evident that they strongly depend on
both the Reynolds number and the geometry of the reservoir.
As Re decreases, the mean residence time
increases (Table 1). As expected, when L/H T
increases, the mean residence time approaches Influence of
the space-time value (Table 2). on the Mean
Fraction
Because in all cases, tr < T or 5r <1, we (Reserv
can conclude that a stagnant region exists, Values calculated
which in practice would be a dead volume,
leading to a lower reactor performance. The Mean Residenc
fraction of the reactor volume occupied by Re Or = tr
the dead region is given by[6,20' 1 0.778


Vd=l (1d
V T
The dead volume fractions obtained, shown
in Tables 1 and 2, indicate that for high Re


values (or high fluid velocities) and for geometries where L/
H approaches 1, or even smaller, a large fraction of the reser-
voir will not be efficiently used for reaction purposes.
- Prediction of Conversion in the Continuous-Flow Reactor
In a real reactor, the RTD function can be used to predict
the limiting values of conversion under the two extremes of
micromixing, using the well-known total segregation or maxi-
mum mixedness models.121 For first-order reactions (linear
systems), however, the state of mixing does not affect con-
version,1201 and therefore the easy-to-use segregation approach
can be applied to predict reactor performance. The total seg-
regation model assumes that all fluid elements having the
same age (residence time) "travel together" in the reactor and
do not mix with elements of different ages until they exit the
reactor.17 Because there is no interchange of matter between
fluid elements, each one acts as a batch reactor, so the mean
steady-state conversion (X) in the real reactor is given by

f Xbatch(t)= E(t)dt= Xbatch (0) E()d (13)
0 0
where Xbach(t), for a first-order reaction, is given by

Xbatch = e(kt) = 1 e(-Da) (14)

10
E (0) L/H= 2
S2 Flow between
8 20 parallel pltes
6

4

2

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

Figure 7. Effect of the geometric ratio L/H on the resi-
dence time distribution (Re = 10).


Chemical Engineering Education


ABLE 1
the Reynolds Number
SResidence Time and
Sof Dead Volume
oir with L/H = 1)
from RTD curves up to 6 = 10.

e Time Fraction of Dead Volume
/ T V/ V
0.222


5 0.724 0.276
10 0.412 0.588
20 0.318 0.682
100 0.261 0.739


TABLE 2
Influence of Reservoir Geometry on
the Mean Residence Time and
Fraction of Dead Volume
(for Re = 10)
Values calculated from RTD curves up to 0 = 10.

Mean Residence Time Fraction of Dead Volume
L/H 0, = / Vd / V
0.5 0.300 0.700
1 0.412 0.588
2 0.752 0.248
5 0.935 0.065
10 0.992 0.008
20 0.997 0.003










and Da =kT is the Damkohler number.
Steady-state conversion is then computed by using Eq. (13),
with the RTD function previously determined. It is important
to remark that the segregation model can also be used for
prediction of the reactor transient behavior. In this case, the
upper integration limit in Eq. (13) must be set to t (or 0). This
was done for our case study and the results, shown in Figure
8, illustrate the reactant conversion in transient conditions,
up to steady state, for different Damk6hler values. One must
take care that the RTD used for prediction of reactor perfor-
mance depends on its geometry and on the Reynolds num-
ber. In addition, because of the very long tail of the RTD
function (as shown in Figure 5 for the F(O) curve), prediction
of steady-state conversion requires RTD data up to very large
times (note the logarithmic time scale). This interesting fea-
ture is also evident in Figure 8-a non-negligible contribu-
tion to the overall reactor performance (in terms of fractional
conversion), which is noticed at very long times. Such be-
havior can be attributed to the stagnant region and to the dif-
ferent time scales for the involved phenomena: reaction, con-
vection, and molecular diffusion.
As mentioned above, Fluent can also be used to simulate
the system in the presence of a reaction, and so the data shown
in Figure 8 can be obtained either through the total segrega-

0.90
X(O) D=5
0.80
0.70
0.60
0.650
0.40
Da=l
0.30
0.20 Da =0.5
0.10
0000
0.1 1 10 100 1000
0

Figure 8. Unsteady-state conversion obtained for the
reactor with L/H = 1, predicted from the segregation
model (Re = 10).

1.0
S1 Lamnar ow between
X
parallel plates .
0.8 PFR X a

0.6
^ *- CSTR
0.4
t o L/H= 20
SD L/H= 5
0.2 --e L/H 2
.-- LW=
0.0
0 1 2 3 4 5
Da
Figure 9. Steady-state conversion versus Damkohler
number for reactors with different geometries (Re = 10).

Spring 2004


tion model or directly from CFD simulations (obtained curves
coincide). Asking students to compare results from both ap-
proaches is important because they feel more confident about
the simulation results and calculations.
Simulation of the continuous-flow reactor via CFD can also
be used to evidence the contours of species concentration
throughout the reactor, for instance at steady state. Some color
pictures can be seen on our web site at ~mmalves/cfd/reactor/index.htm>.The color gradient inside
the reactor is particularly interesting to observe.
Finally, it is convenient to ask students to compare the
steady-state conversion attained in the real reactor with those
achieved with the ideal reactors that they learned in previous
CRE courses: continuous stirred tank and plug flow reactors.
For a first-order reaction, performance achieved by these re-
actors is given by'e g..7,8]

XCSTR Da (15)
1 + Da

XpFR = e(-Da) (16)

Data shown in Figure 9 indicate that when dead regions or
stagnant zones are negligible, i.e., for geometries where L/H
is higher than about 5 (for Re = 10), the performance of the
real reactor lies between that of the CSTR and PFR. When L/H
is close to 1, or even smaller, such anomaly (dead volume)
leads to a much lower performance of the nonideal reactor-
even lower than that achieved with a perfectly mixed reactor.
It is also noteworthy that for very long reactors (i.e., high L/H
ratios), one approaches the theoretical behavior of a laminar
flow reactor (flow between parallel plates) computed using the
RTD function given in Eq. (9) and the segregation model.

CONCLUDING REMARKS

Concepts dealing with the RTD theory and nonideal reac-
tors are not very familiar to undergraduate students, and thus
it is not an easy task to teach these matters in CRE courses.
Explanation of how stimulus-response tracer experiments
provide the theoretical functions that are crucial for a reactor's
diagnosis, prediction of conversion, etc., becomes clearer
through experimentation, which is not always feasible. For
such purposes, the use of CFD packages has been particu-
larly advantageous. Besides, CFD tools provide a graphical
portrait of flow throughout the reservoir/reactor, thus allow-
ing going beyond simply predicting what will come out of
the reactor to predicting all of the properties of interest within
the reactor. Some of the concepts involved are more easily
understood, such as the progression of the tracer concentra-
tion front, the formation of dead volumes, or the existence of a
concentration profile along the reactor. Therefore, implemen-
tation of a CFD code has a considerable pedagogical content.
The case study we have described, a 2-D lagoon with lami-
nar flow, was solved with the commercial package Fluent.


159










Using reservoirs/reactors with different geometries and/or
different operating conditions, students may be asked to
Characterize the hydrodynamics
Determine the residence time distribution from tracer
experiments, which provides diagnosis of reactor operation
Predict conversion in a continuous-flow reactor (both
steady state and transient behavior)
Our experience shows that use of the CFD code allows stu-
dents to more easily understand some of the basic concepts
taught in CRE curricula. Finally, comparison of numerical
with analytical solutions known for laminar flow between
parallel plates (i.e., for geometries with high L/H ratios) im-
proves their self-reliance regarding CFD results.
In a survey sent a few years ago to chemical engineering
departments spread all over the world, two of the main points
addressed by the departments to a question relating to the
future of CRE courses were:t21 the increasing importance of
computer applications and software packages, and putting
more emphasis on nonideal reactors. With the case study
herein proposed, both issues are dealt with. In addition, stu-
dents learn the potential of CFD codes, which have been suc-
cessfully used in practice to design commercial-size reac-
tors, usually with complex flow processes.[221

NOMENCLATURE
C concentration of tracer, reactant, or product
(mol.m3 or kg.m-3)
D diffusivity (m2s)
Da Damkohler number, dimensionless
E(t) resident-time distribution function (s ')
E(O) normalized RTD function, dimensionless
F(t) Danckwerts' F curve, dimensionless
H height of the reservoir/reactor (m)
k reaction rate constant of the first-order reaction (s-')
L length of the reservoir/reactor (m)
P pressure (Pa)
S source term (mol mn s-' or kg m-ns ')
Q volumetric flow rate (m3s-')
Re Reynolds number, dimensionless
t time (s)
tr mean residence time (s)
U fluid velocity at the center of the inlet boundary
(ms-')
Umen mean fluid velocity at the inlet boundary (ms-')
ux x-velocity (ms-1)
u y-velocity (ms-')
V volume of reservoir/reactor (m3)
Vd dead volume in the reservoir/reactor (m3)
X conversion, dimensionless
x horizontal coordinate (m)
y vertical coordinate (m)
Subscripts


batch
CSTR
in
PFR


refers to batch reactor
continuous stirred tank reactor
inlet conditions
plug flow reactor


out outflow conditions
Greek Symbols
O=t/T reduced time, dimensionless

Or = tr / T reduced mean residence time, dimensionless
lI viscosity of the fluid (kg m-'s-')
v kinematic viscosity of the fluid (m's-1)
p density of the fluid (kg m"3)
7 space-time (s)

REFERENCES

1. Fogler, H.S., "An Appetizing Structure of Chemical Reaction Engi-
neering for Undergraduates," Chem. Eng. Ed., 27, 110 (1993)
2. Shalabi, M., M. Al-Saleh, J. Beltramini, and D. Al-Harbi, "Current
Trends in Chemical Reaction Engineering Education," Chem. Eng.
Ed., 30, 146 (1996)
3. Falconer, J.L., and G.S. Huvard, "Important Concepts in Under-
graduate Kinetics and Reactor Design Courses," Chem. Eng. Ed.,
38, 138 (1999)
4. Conesa, J.A., and I. Martin-Gull6n, "Courses in Fluid Mechanics
and Chemical Reaction Engineering in Europe," Chem. Eng. Ed.,
34, 284 (2000)
5. Froment, G.F., and K.B. Bischoff, Chemical Reactor Analysis and
Design, 2nd ed., John Wiley & Sons, New York, NY (1990)
6. Villermaux, J., Genie de la Rdaction Chimique: Conception et
Fonctionnement des Rdacteurs, Tec & Doc, Lavoisier, Paris, France
(1993)
7. Fogler, H.S., Elements of Chemical Reaction Engineering, 3rd ed.,
Prentice Hall, New Jersey (1999)
8. Levenspiel, O., Chemical Reaction Engineering, 3rd ed., John Wiley
& Sons, New York, NY (1999)
9. Danckwerts, P.V., "Continuous Flow Systems: Distribution of Resi-
dence Times," Chem. Eng. Sci., 2, 1 (1953)
10. Conesa, J.A., J. Gonzilez-Garcia, and J. Iniesta, "Computer Simula-
tions of Tracer Input Experiments," Chem. Eng. Ed., 33, 300 (1999)
11. Versteeg, H.K., and W. Malalasekera, An Introduction to Computa-
tional Fluid Dynamics: The Finite Volume Method, Prentice Hall, New
Jersey (1995)
12. Roache, P.J., Fundamentals of Computational Fluid Dynamics,
Hermosa Publishers, Albuquerque, NM (1998)
13. Roache, P.J., Verification and Validation in Computational Science
and Engineering, Hermosa Publishers, Albuquerque, NM (1998)
14. Sinclair, J.L., "CFD Case Studies in Fluid-Particle Flow," Chem. Eng.
Ed., 32, 108 (1998)
15. Metcalf & Eddy Inc., Wastewater Engineering: Treatment, Disposal,
and Reuse, revised by G. Tchobanglous and F.L. Burton, 3rd ed.,
McGraw-Hill, New York, NY (1991)
16. Brunier, E., A. Zoulalian, G. Antonini, and A. Rodrigues, "Residence
Time Distributions in Laminar Flow through Reservoirs from Mo-
mentum and Mass Transport Equations," in 8th Int. Symp.CRE, IChemE
Symposium Series, No. 87, 439 (1984)
17. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport Phenomena,
1st ed., John Wiley & Sons, New York, NY (1960)
18. Leonard, B.P., "A Stable and Accurate Convective Modeling Proce-
dure Based on Quadratic Upstream Interpolation," Comput. Meth. Appl.
Mech. Eng., 19, 59 (1979)
19. Levenspiel, O., The Chemical Reactor Omnibook, OSU Bookstores,
Corvallis, OR 97330 (1979)
20. Rodrigues, A.E., "Theory of Residence Time Distributions," in
Multiphase Chemical Reactors, A.E. Rodrigues, J.M. Calo, and N.H.
Sweed, eds, NATO ASI Series, Sijthoff Noordhoff, No. 51, Vol. I,
225 (1981)
21. Zwietering, T.N., "The Degree of Mixing in Continuous Flow Sys-
tems," Chem. Eng. Sci., 11, 1 (1959)
22. Bakker, A., A.H. Haidari, and E.M. Marshall, "Design Reactors via
CFD," Chem. Eng. Prog., 97, 30 (2001) 0


Chemical Engineering Education


160














AUTHOR GUIDELINES


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


Specific suggestions on preparing papers *

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

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

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

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

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

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

ACKNOWLEDGMENT Include in acknowledgment only such credits as are essential.

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

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


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







ASE
Anna Confrenc and Expsiton

I-II


Wokhp


Ssion




University of Florida Home Page
© 2004 - 2011 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated May 24, 2011 - - mvs