Chemical engineering education ( Journal Site )

Material Information

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


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


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

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 01151209
lccn - 70013732
issn - 0009-2479
lcc - TP165 .C18
ddc - 660/.2/071
System ID:

Full Text

chemi c ee e ucation

acinowledges and tkanki....


w d a deow itio aj jutds.


Department of Chemical Engineering
University of Florida
Gainesville, Florida 32611
Editor: Ray Fahien (904) 392-0857
Consulting Editor: Mack Tyner
Managing Editor:
Carole C. Yocum (904) 392-0861
Publications Board and Regional
Advertising Representatives:
Gary Poehlein
Georgia Institute of Technology
Past Chairmen:
Klaus D. Timmerhaus
University of Colorado
Lee C. Eagleton
Pennsylvania State University
Richard Felder
North Carolina State University
Jack R. Hopper
Lamar University
Donald R. Paul
University of Texas
James Fair
University of Texas
J. S. Dranoff
Northwestern University
Frederick H. Shair
California Institute of Technology
Alexis T. Bell
University of California, Berkeley
Angelo J. Perna
New Jersey Institute of Technology
Stuart W. Churchill
University of Pennsylvania
Raymond Baddour
Charles Sleicher
University of Washington
Leslie W. Shemilt
McMaster University
Thomas W. Weber
State University of New York


Chemical Engineering Education

106 George Stephanopoulos of M.I.T.
Marc J. Chelemer

112 The John Hopkins University, Carol Hyman

118 Teaching Heat Exchanger Network Synthesis
Using Interactive Microcomputer Graphics,
Anthony G. Dixon
122 UC ONLINE: Berkeley's Multiloop Computer
Control Program, Alan S. Foss
126 The Burning of a Liquid Oil Droplet: A Simple
Mathematical Analysis for Teaching Purposes,
A. N. Hayhurst, R. M. Nedderman
130 Tips on Teaching Report Writing,
R. R. Hudgins

134 The Milliken/Georgia Tech Rising Senior
Summer Program, Pradeep K. Agrawal,
Jude T. Sommerfeld

138 A Reverse Osmosis System for an Advanced
Separation Process Laboratory,
C. S. Slater, J. D. Paccione
146 A First Chemical Engineering Lab Experience,
Vito L. Punzi

144 Estimating Relative Volatility of Close-Boiling
Species, Allen J. Barduhn

152 Letter to the Editor

152 Books Received

110, 117, 132, 133, 143, 145, 152 Book Reviews

CHEMICAL ENGINEERING EDUCATION is published quarterly by Chemical Engineering Division,
American Society for Engineering Education. The publication is edited at the Chemical Engineering Depart-
ment, University of Florida. Second-class postage is paid at Gainesville, Florida, and at DeLeon Spngs,
Florida. Correspondence regarding editorial matter, circulation and changes of address should be addressed
to the Editor at Gainesville, Florida 32611. Advertising rates and information are available from the adver-
tising representatives. Other advertising material may be sent directly to the printer: E. O. Painter Printing
Co., P. 0. Box 877, DeLeon Springs, Florida 32028. Subscription rate U.S., Canada, and Mexico is $20 per
year, $15 per year mailed to members of AIChE and of the hE Division of ASEE. Bulk subscription rates
to ChE faculty on request. Write for prices on individual back copies. Copyright 0 1987 Chemical Engineer-
ing Division of 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 of the ASEE which body
assumes no responsibility for them. Defective copies replaced if notified within 120 days.
The Internatonal Organization for Standardization has assigned the code US ISSN 0009-2479 for the
identification of this periodical.



of MIT

Massachusetts Institute of Technology
Cambridge, MA 02139

F ONE WERE TO select current chemical engineer-
ing faculty members whose innovative research,
visionary thinking, and creativity match those of the
original educators in the field, George Stephanopoulos
would stand high on the list. By his prodigious output
of research papers, his contribution to teaching and
education through textbooks and monographs, his
rapport with and motivation of students, and his ad-
vancement of knowledge through a willingness to
tackle new, ill-defined problems, George has distin-
guished himself as one of the best professors teaching
today. His insight and keen intellect have made great
strides in our understanding of his chosen fields of
process synthesis, control systems engineering, and
knowledge-based computer applications to these and
other problems. Most surprising, George is only thir-
teen years out of the student life himself and is only
forty years old. To understand his contributions, it is
appropriate to examine the beginnings of George's al-
ready important career in chemical engineering.


Stephanopoulos was born on June 1, 1947, in the
port city of Kalamata, Greece, near the ruins of an-
cient Sparta. He attended the National Technical Uni-
versity of Athens, receiving his diploma in chemical
engineering in 1970 with the highest honors among all
graduating engineering students, including the
"Thomaidion" and "Chryssovergeion" Awards. From
there, he moved to Hamilton, Ontario, where he
began his master's studies under Cameron Crowe at
McMaster University. Crowe recalls that the new
arrival was "a delightful student to work with. I wish
I had more like him." In his year of study,
Stephanopoulos worked in "a very difficult, delicate
area." The research, optimizing a tubular reactor
C Copyright ChE Division ASEE 1987

system with catalyst decay, contained "many traps for
the unwary student," but George avoided them all,
grasping concepts very quickly and integrating them
into his thesis. This became George's trademark: a
quick, analytical mind, the ability to understand new
ideas rapidly, and a questioning nature which took
him deeper into his fields of investigation.
After receiving his MS, Stephanopoulos went on
to the University of Florida at Gainesville to work
with Arthur Westerberg in process engineering re-
search. His PhD effort focused on understanding how


to analyze, synthesize, and optimize large-scale pro-
cesses. The Primal-Dual Bounding problem, a com-
puter algorithm, had just been completed, but the im-
plications required elaboration. George took the reins
of this project and forged ahead. Westerberg says that
Stephanopoulos was a "joy to work with, a very strong
student destined to be a great professor." Advisor and
advisee argued frequently, with each side winning
about half the time.
In one instance, George postulated a theorem and
then, on demand, provided a complete proof three
days later. Although the proof was invalid, the theory
proved correct. In a similar, equally memorable
event, Dr. and Mrs. Westerberg asked George if he
would prepare some Greek recipes for a small dinner
party if they supplied the ingredients. George in-
formed them that the three whole bulbs of garlic
supplied for his culinary work would be insufficient to
prepare the food for six guests. After the necessary
additions had been procured and the meal prepared,
his claim again proved correct.
George finished his PhD in three years, matriculat-
ing in 1974. During that time, he wrote or co-wrote
seven technical papers. This marked the beginning of
Stephanopoulos' major contributions to chemical en-
gineering research and education.
The now Dr. Stephanopoulos took his first teach-
ing position at the University of Minnesota in 1974,
where he followed the tenure track to full professor-
ship by 1980. That same year, he acted as a tourguide
in Greece for his colleagues. On that trip, he both met
his wife-to-be, Eleni, and accepted a post at his alma
mater, the National Technical University in Athens.
Although he maintained his position at Minnesota, his
colleagues missed his day-to-day presence on the staff.
During his years in Athens, George saw the pos-
sibilities of assuming a leadership role in state science
within Greece's technical community, but also felt a
pull to conduct exciting and innovative research in the
USA. Sometime during that same period, George first
made the acquaintance of MIT's chemical engineering
faculty at the Exxon suite of an AIChE meeting. "A
long admiration and courtship" followed, according to
MIT Department Head James Wei, which culminated
in Stephanopoulos' return to the USA and MIT in
1984, where he has held the J. R. Mares Professorship
in Chemical Engineering ever since.
In the thirteen years since his PhD was granted,
Stephanopoulos has distinguished himself as one of the
brightest and most forward-thinking faculty members

This became George's trademark:
a quick, analytical mind, the ability to
understand new ideas rapidly, and a questioning
nature which took him deeper into his fields of
investigation . [he] went on to the University of
Florida to work with Arthur Westerberg . .

in the profession. His prodigious output of papers,
monographs, and textbooks has ensured his place in
chemical engineering education for many years. Col-
leagues describe George's work as "extremely original
and inventive," "very exciting," "intellectually superb
and refreshing," and "leading the way in pioneering
fields." He is a methodical man: "Before he says any-
thing," explains a colleague, "he has thought it all
out-all the consequences too." He "sees things in an
entirely different light, turns ideas around." "He al-
ways tackles new problems, looks into new areas. He's
ahead of everybody else. He has a great vision and a
long time horizon." For his research work, he was
selected as a Camille and Henry Dreyfus Teacher and
Scholar in 1977, and received the Allan P. Colburn
Award of the AIChE for 1982. An examination of his
research contributions illustrates how his work has
enriched the basis for chemical engineering:
Process Synthesis: George's early work focused on
branch and bound search, and evolutionary synthesis
strategies for process synthesis. He developed the
first rigorous techniques to solve the former, and pro-
vided the basis of logical techniques to replace ad hoc
methods in use at the time for the latter. Later, he
wrote about developing physical insights to augment
the available mathematical techniques. These efforts
have helped engineers adopt a more integrated ap-
proach to plant and process design. Work on chemical
reaction pathways in a Gibbs' free energy versus tem-
perature space helped understanding of the effi-
ciency of alternative reaction pathways to desired
chemical products.
In studying separations, Stephanopoulos' research
in heat-integrated distillation revealed some unknown
properties which allowed sequential selection of sep-
aration splits and more ideal heat integration between
columns. This work justified some common industrial
practices which until that time had no quantitative
basis. It has also yielded better, more efficient indus-
trial procedures.
Control Engineering: Other early professional re-
search applied experimental ideas and techniques use-
ful for process design to control systems. His efforts


almost single-handedly changed the field from one
which studied individual unit control to understanding
complete processing system control. As industrial
practice follows the latter procedure, George's rigor-
ous whole-system approach became applicable im-
mediately. He also introduced the concepts of "con-
straints-based" supervisory control, and suggested
that operating improvements could result from what
he called "variable control structures." He also contri-
buted to understanding the need for data reconcilia-
tion and gross error identification to ensure reliable
and accurate data from which to base a control strat-
egy. He considered not only steady-state operation,

George's patience and his ability to motivate students
are the keystones of his teaching success.

but start-up, shutdown, changeover, and optimization

Model-based control systems: To ensure the effec-
tiveness of the plant-wide strategies developed in the
previously discussed research, Stephanopoulos com-
comittantly developed theoretical novel control
hardware for implementing them. Two significant re-
sults of this effort are the 2-port controller and the
strategy of structured control. The first concept pro-
vides a device with two control elements of different
but complimentary design goals. One "port" provides
command-following control within specified stability
robustness levels and performance. The second port
maintains current system status based on an input
model, thus providing regulation against process dis-
turbances. Depending on whether steady-state or
changeover operation is desired, one port is given
priority over the other. This idea is being investigated
extensively; one potential use currently under study
relates to in-vivo insulin control systems for diabetics.

Structured control is the application of rule- and
knowledge-based computer systems to real-time, on-
line responses to process disturbances. Using process
measurements from the plant, the computer system
responds in varying fashion depending on its instruc-
tions and "reasoning." This work, now in develop-
ment, promises a new generation of control systems
more powerful than any predecessor.
Artificial Intelligence: Since arriving at MIT,
George has pursued applications of powerful AI com-
puters using the LISP programming language. He has
established an industry-sponsored research labora-
tory to study the impact of these "intelligent systems"
on process engineering. In this new area of study,
George is leading both MIT's effort and the profession
as a whole.
Some of the ongoing projects are amplifications of
earlier work. Having studied heat exchanger net-
works, separation processes, whole chemical plants,
and control engineering, he is looking at some of the
basic science involved. For example, he is attempting
to design solvent molecules with characteristics useful
for extractive separations, heat-pump fluids with effi-
cient economics, and polymer molecules with desired
mechanical, chemical, and other properties. This ef-
fort relies on the use of factor analysis and additive
group contributions, relationships so complex and in-
tertwined that rule-based computer systems are
needed to assist engineers in devising likely molecular
candidates. He is also attempting to use AI and know-
ledge-based systems to screen alternative production
routes and develop process flowsheets, plantwide con-
trol schemes, and operating strategies for problems
ranging from individual process units to entire chem-
ical plants.
Biotechnology: George's research looks at three
areas in this new field: modeling of bacterial biochem-
ical pathways, design and development of mammalian
cell bioreactors, and synthesis of separation and re-
covery schemes for bioproducts, notably proteins. The
first effort helps scientists understand genetic modifi-
cations necessary to obtain certain desirable products
and to identify those bacteria most amenable to that
genetic engineering. The work in the second area aims
at the development of an "intelligent" system which
coordinates the knowledge from three different
areas-molecular biology and genetics, reactor en-
gineering, and operational analysis and control-to
design a bioreactor system with optimum characteris-
tics. George's brother Greg, who is collaborating on
these projects, played a pivotal role in stimulating in-
terest in this research area. As MIT has established



f,!*~j,, i
O Z.~2'0'

At Minnesota he was awarded the G. Taylor Teaching Award for his efforts in the classroom, yet was
playfully chided for his pronunciation of certain words and kidded about his habit of not wearing a tie.
The latter joke involved the presentation of a hair-covered undershirt for colder days in Minneapolis.

a new research center for biotechnology processing,
George plays a vital role in both fields-AI and cell
biology-new and perhaps unfamiliar to many other
engineers. True to form, Stephanopoulos and his stu-
dents are already producing paper after paper de-
monstrating high-quality research and understanding.

Stephanopoulos' written contributions to chemical
engineering education includes books, monographs,
and course notes. His best-known opus, Chemical
Process Control: An Introduction to Theory and
Practice, is used by more than 50% of the chemical
engineering departments in the US, only two years
after publication. His review papers on process syn-
thesis, on control systems for complete chemical
plants, and on AI in process engineering are unique
in their respective fields. In addition to his writing,
George is a leader in developing MIT's chemical en-
gineering curriculum. James Wei describes George's
role: "He thinks about what the meaning of education
and research really is. He wonders how we should be
teaching our students, and what changes should be
made." George has strengthened the MIT department
by helping to attract his brother Gregory and his sis-
ter-in-law Maria to MIT, where they have assumed
important roles in the areas of biotechnology and pol-
lution control science, both traditional MIT strengths.
Stephanopoulos has contributed to new course de-
velopment in all of his teaching positions, and many of
his PhD students have gone on to take faculty posi-
tions in other universities. These professors are them-
selves among the most important control system/pro-
cess design faculty working today. Altogether, he has
supervised thirty graduate students at the three uni-
George enjoys the same popularity with his stu-
dents at all levels as he does with his peers. He is
described as being very patient, understanding, and
slow to anger. At Minnesota he was awarded the G.
Taylor Teaching Award for his efforts in the class-
room, yet was playfully chided for his pronunciation
of certain words and kidded about his habit of not
wearing a necktie. The latter joke involved the pre-
sentation of a hair-covered undershirt for colder days
in Minneapolis. When one irate student expressed his
dissatisfaction with a particular exam by placing a

Stephanopoulos sharing a relaxed moment with his stu-

"Nuke Stephanopoulos" ad in the school paper, other
students rallied. The culprit was waiting for a lecture
soon after the ad appeared when four "gangsters"
marched into the room in zoot suits and wing-tips,
carrying violin cases. They promptly dispatched the
culprit with vanilla cream pies.
According to some of his current students, he
"does a tremendous job of getting people motivated."
In the area of AI, the original student skepticism has
been replaced by a positive, team attitude. Meetings
in the Lab for Intelligent Systems in Process En-
gineering (LISPE) often last twice as long as sched-
uled, with computer code, ideas, and camaraderie
shared within the group. George is particularly ap-
proachable, his students say. Perhaps that explains
why he is the most sought-after professor in MIT's
free-market advisor selection system. George is also
a participant in soccer, volleyball, and outdoor recrea-
tional activities with his students. "He's just as ag-
gressive on the field as off, and a fine athlete, too,"
says one.

In the midst of this hectic and intensively dedi-
cated professional life, George commits himself to
many private, philosophical, and artistic pursuits. "He
is a deep thinker and has more cultural interests than
many of his colleagues." "George is a true renaissance
man, particularly in music and in art." "He keeps a


healthy attitude about leading a balanced life." He is
a friend and confidant to many, and is a dedicated
family man. George and Eleni visit their families in
Greece annually, and are the parents of two small chil-
dren, Nikos and Elvie. George enjoys working with
his family members in the department, although his
suite-mates indicate that he shows a kindred spirit
with all of them.
Chess occupies George's technically creative time.
A computer program which might teach novices open-
ing moves, standard defenses, and some basic
strategies is one ongoing project.
Many colleagues mention his deep and profound
interests in literature and philosophy. George explains
that his Greek poetry and short stories are a very
private creative outlet for him. He always gave public
talks and poetry readings of other authors at Min-
nesota's annual AIChE spring banquet, much to the
delight of the audiences. Stephanopoulos has read ex-

1 book reviews

2nd Edition
by Foust, Wenzel, Clump, Maus, and Anderson
John Wiley & Sons, 1980, 768 pages
Reviewed by
Davis W. Hubbard
Michigan Technological University
This is a revision of the first edition published in
1960 and retains the same large-page format, arrange-
ment of subject matter, and emphasis on engineering
practice and design. The type size has been increased
and two chapters have been added, increasing the
number of pages from 578 to 768. The new edition
bears a 1980 copyright and quite a number of refer-
ences have been added to most chapters since the first
edition. The book is intended to be an undergraduate
textbook. It is quite easy to read, yet the material is
covered in such a way that the reader has the feeling
of progressing beyond the simplest basic material.
There are extensive equipment diagrams and an ex-
cellent emphasis on practical design. One of the attri-
butes of the book is the treatment of the common fea-
tures of the various topics in a unified way. This is
applied both to separations processes and to transport
phenomena topics. British engineering system units
are used mostly, but there are scattered examples in
which SI units are used. There is a good selection of

tensively from 20th Century American writers, the
Russian literature, and German philosophy. He fol-
lows Greek poetry, notably Nobel laureates George
Seferis and Odysseus Elytis. Lately, the South Amer-
ican writers, notably Garcia-Marquez (a strong influ-
ence, he says), Ernest Sabato, Jorge Luis Borges, and
Vargas Llossa are his favorites.
Commenting on his recent winning of the Curtis
McGraw Award for Research from ASEE, George
notes, jokingly, that engineering is rarely a route to
the Nobel Prize, and that he personally would try to
win it only for literature or peace. Considering that
he ably handles his commitments to MIT, his editor-
ship on two journals, and his membership in eleven
technical and professional societies, while still main-
taining strong ties to family, friends and students, the
Nobel seems an appropriate long-term goal for this
man of vision, technical excellence, and commitment
to learning. O

design data, eliminating the need for searching for
data in other sources and making the book easy to use
for beginning students.
The first section of the book deals with continuous
stagewise processes. The emphasis is on the
similarities among the different processes discussed.
This works well for distillation and absorption, but it
is a little strained for leaching and adsorption. The
Ponchon-Savarit graphical method of analyzing distil-
lation processes and other related graphical methods
are the focal point of this section. The material is writ-
ten in a straightforward manner and is easy to under-
stand and apply. The ideas of minimum solvent flow
rate and minimum reflux ratio are handled well.
People who believe that using the Ponchon-Savarit
method or graphical methods in general is outmoded
and should be discontinued in favor of using numerical
methods will not like this section. People who believe
that the Ponchon-Savarit method helps provide in-
sight related to energy balances and changes in molar
flow rates between stages will be pleased with this
At the end of the section about stagewise proces-
sing there is a new chapter on multicomponent calcu-
lation methods. This chapter replaces one about un-
steady distillation processes in the first edition. The
discussion focuses on distillation processes and em-
phasizes fundamentals. Computer codes are not em-
phasized though computer methods are apparently


used in solving the example problems. A very clear
example dealing with a five-component mixture is
solved using the Fenske-Underwood-Gilliland short-
cut method.
The second section of the book treats transport
phenomena. There are chapters about transport coef-
ficients, molecular transport, and turbulent transport.
The discussion of the various concepts seems quite
straightforward and easy to understand. The treat-
ment of unsteady conduction has been changed quite
a bit from the first edition. The discussion is somewhat
confusing, but empirical equations which could be
used for computer calculations are supplied. There is
a good discussion about the limitations of semi-empir-
ical turbulence models.
A chapter on heat transfer comes next. This chap-
ter uses British engineering units. The detailed exam-
ple calculations emphasizing assumptions are good,
and there is good coverage of practical aspects of shell-
tube heat exchanger design. Condenser design
methods are restricted to pure vapors. Mixed vapor
condensation is not discussed. The references cited
are quite old.
The next chapter treats mass transfer with em-
phasis on continuous contractors. The differences be-
tween gas-liquid and liquid-liquid operations is blur-
red sufficiently so that the similarities can be seen.
The mass transfer coefficient correlations presented
are easy to use. There is a clear development of design
equations with good explanations of the assumptions
involved and the methods of integrating the equa-
tions. Some good numerical examples are given. Stage
efficiency is mentioned in this chapter, but the treat-
ment is not extensive.
The next section is devoted to simultaneous energy
and mass transfer with chapters on humidification and
water cooling, drying, and evaporation and crystalli-
zation. There is good coverage of water cooling-a
topic sometimes treated in a superficial way in
textbooks. The discussion includes methods for deter-
mining the gas phase temperature and evaluating heat
and mass transfer coefficient data from experimental
data. The discussion of drying is quite standard. There
is an extensive discussion of spray drying, but other
types of continuous drying are not mentioned. The
chapter on evaporation is written using SI units, and
evaporator calculations are described clearly. There
is a good review of mass and energy balances and an
introductory discussion of mass transfer considera-
tions in crystallization.
The final chapters are devoted to fluid mechanics
and fluid mechanical separations processes. The dis-
cussion of the Bernoulli equation seems somewhat

ponderous, and the notation for friction loss is not the
same as used by other writers. The discussion of fric-
tion factors for straight pipe is good, but fitting losses
are discussed exclusively in terms of equivalent
length. This is a time-honored method, but for
smooth-walled pipe it introduces a small Reynolds
number dependence which is not supported by data.
The discussion of flow meters is good. It includes
examples of meters suitable for remote reading as well
as the traditional types.
The chapter about pumps contains excellent pic-
tures of pump assemblies and excellent discussions of
NPSH and specific speed. There is not much discus-
sion of pump efficiency, though the examples of pump
characteristic curves which are given include effi-
The chapter about fluid-solid separation begins
with a discussion of drag coefficients which includes
the effect of particle shape. Many topics are covered
in this chapter, and there is a certain lack of con-
tinuity. It might have been better to make separate
chapters focusing on moving solids and fixed solids.
The example of thickener design calculations is com-
plete, detailed, and easy to follow. The filtration calcu-
lations based upon using an equivalent volume of fil-
trate to represent the filter medium resistance are
fairly easy to understand, but the method has always
seemed artificial. The filter analysis calculations are
presented in terms of the filtration rate even though
this requires graphical or numerical differentiation.
The integrated equations are presented but are not
used in the analysis. Applications to continuous filtra-
tion are ignored.
The appendices are well written and useful. There
is a lucid discussion of g, and J putting these factors
into their proper perspective as unit conversion pro-
portionality constants. The dimensional analysis dis-
cussion focuses on the Buckingham Pi method, hitting
the high points without excessive mathematical de-
velopment. The unit system which includes force as a
fundamental dimension is used exclusively. The ap-
pendix in which screen analysis and the characteristics
of particles are discussed is well written and easy to
understand. There is also an appendix containing most
data needed for solving the problems. The sources are
sometimes not cited fully.
This book is a gold mine for problems designed to
be challenging but easy enough so that students begin-
ning their study of unit operations can be expected to
complete them. Between ten and sixty problems are
included with each chapter, and only a few chapters
have fewer than twenty-five. This is a major strong
point of the book. l[


n department





The John Hopkins University
Baltimore, MD 21218

versity, John's Hopkins University, The Johns
Hopkins University. Never has a university's name
produced so many variations. In brief, Johns was the
founder's mother's maiden name, and unfortunately
for him, he inherited a last name for a first name.
Despite the frequent misspellings and confusion,
The Johns Hopkins University has a long and presti-
gious history. Founded in 1876, Hopkins established
itself as the first true American university on the
European model; a graduate institution in which
knowledge would be created as well as taught. As
early as 1913 engineering became an integral part of
this university, creating the foundation for what
would later become the G.W.C. Whiting School of En-
The Schools of Engineering, Continuing Studies,
and Arts and Sciences, and the Space Telescope Insti-
tute are located on the Homewood campus in north
Baltimore on a 140-acre wooded campus in a residen-
tial area. The campus was originally the Homewood
estate, built for Charles Carroll, Jr., son of a signer
of the Declaration of Independence. The university
was given the estate in 1902.
In addition to the facilities at the Homewood cam-
pus, The Johns Hopkins University's academic divi-
sions and research institutions include the world-re-
nowned schools of medicine, public health, and nurs-
ing, all located at the East Baltimore campus; the
Copyright ChE Division ASEE 1987

School of Advanced International Studies in Washing-
ton, DC, with centers for foreign studies in Bologna,
Italy, and Nanjing, China; the Peabody Institute, one
of the leading music schools in the United States, lo-
cated in downtown Baltimore; and the Applied
Physics Laboratory in Columbia, Maryland, a scien-
tific and engineering research facility.

The Hopkins community shares in the exciting, na-
tionally recognized, urban renaissance of Baltimore.
Baltimore is no longer just a place to drive around on
the way to New York from Washington. The city now
boasts the "Inner Harbor," a waterfront area that in-
cludes the National Aquarium, the Maryland Science
Center, and shops and restaurants in two glass pavil-
ions. The Baltimore Museum of Art, adjacent to the
Hopkins campus, houses excellent permanent collec-
tions and attracts important traveling exhibitions.


The Baltimore Symphony Orchestra, which has a
superb new symphony hall, the Morris Mechanic
Theatre, which presents Broadway touring companies
and pre-Broadway tryouts, Center Stage, and the
Baltimore City Opera are just a few of the many in-
stitutions providing entertainment for Baltimoreans.
And who could pass up watching the Baltimore
Orioles play at Memorial Stadium, just a short walk
from the Homewood campus. A good bet for seeing a
home team win would be watching the University's
own championship lacrosse team, the Blue Jays. Since
their first season in 1888, the Blue Jays have had 73
winning seasons and have won 41 national champion-
Festivals abound in Baltimore. Its many ethnic
communities stage weekend galas throughout the
spring and summer, and the city sponsors the annual
City Fair and the Artscape Festival. Not to be left
out, Hopkins holds its own Spring Fair, "3400 On
Stage." The fair is organized and run by students,
with revenues benefitting student organizations. The
Hopkins Fair draws Baltimoreans from every corner
of the city.

Perhaps it is appropriate that just about the time
the Baltimore urban renaissance began in 1979 the
G.W.C. Whiting School of Engineering was founded.
Today a full complement of undergraduate and
graduate-level programs exist, including the largest
part-time graduate engineering program in the coun-
try. Along with chemical engineering, the Whiting
School departments include biomedical engineering,
civil engineering, electrical engineering and computer
science, geography and environmental engineering,
and materials science and engineering.
Renaissance is also an applicable term to use when
discussing chemical engineering at The Johns Hopkins
University. After existing in some form or another
from the 1930s until 1967, the department was reestab-
lished in 1979 with the rest of the engineering school.
The full-time faculty now numbers seven, but plans
are underway to increase the size of the department.
The department also has ten part-time members, in-
cluding several who are on the staff of the Applied
Physics Laboratory. In 1988 the department will oc-
cupy part of a new engineering building, adding to its
existing facilities. Chemical engineering's facilities
now include laboratories for research in fluid mechan-
ics, heat and mass transfer, nucleation, rheology,
acoustics, phase-equilibria, electrochemical engineer-
ing, separation processes, and biochemical engineer-
The department places a great deal of emphasis on

the use of computers; both graduate and under-
graduate students have access to the department's
computers which include four Micro-VAX computers,
a PDP 11/45, a PDP 11/40, and a PDP 11/34, several
PC's and the School of Engineering's VAX 8600 and
AT&T 3B20.
Despite its apparent youth, the department has a
history of distinguished alumni, including several now
teaching at a variety of colleges and universities in
this country and abroad. These include Simon Goren
(Berkeley), Robert Anderson (McMaster University),
Robert Sparks (Washington University), George
Frazier (University of Tennessee), Gerald Esterson

Renaissance is also an applicable term
to use when discussing chemical engineering at
The Johns Hopkins University. After existing in
some form or another from the 1930s until 1967, the
department was reestablished in 1979 with
the rest of the engineering school.

(Hebrew University), Eric Bauer (Case-Western Re-
serve), Irvin Glassman (Princeton), Kenneth Keller
(President, University of Minnesota), John Falconer
(University of Colorado), Ralph Kummler (Chairman,
Wayne State) Stanley Middleman (University of
California, San Diego), James Douglas (University of
Massachusetts), Robert Edwards (Chairman, Case-
Western Reserve), and Edward Fisher (Chairman,
Michigan Technological University).
Marc Donohue, chairman of the department since
1984, came to Hopkins in 1979. He saw the potential
for a stimulating environment that would allow close
interaction with the students. "An outstanding fea-
ture of Hopkins is the intimate atmosphere," he says.
The university runs on a system much like the British
system in which students receive considerable indi-
vidual attention (almost comparable to private tutor-
ing) from the faculty. This is possible because of the
small student/faculty ratio in the department. It is
common to find Donohue in his office with one of his
advises, hashing out a problem.
Donohue teaches a popular undergraduate course,
"Ethical Questions in Engineering," and he stresses
its importance. "Engineering is the discipline that
translates scientific advances into products for soci-
ety. As such, an engineering education must include
attention to the adverse effects of that technology,"
he says. "We strive therefore to both provide the tech-
nical foundations necessary for students to function as
engineers and to instill a sense of sensitivity to social,
political, and environmental issues that the future
leaders of the engineering profession will face."


He is particularly proud of the undergraduate re-
search program, funded by a grant from the Exxon
Foundation. The program enables a dozen under-
graduates each year to participate in meaningful re-
search while earning salaries competitive with indus-
try. Last year several students published papers that
resulted from their work in the program, and all of
the participants continued studies in graduate schools.
In fact, since the program began three years ago, only
one participant has not gone on for further study.
Donohue, his wife, and his two small children like
to spend their free time hiking, camping, and bicycle
riding. Many weekends find them riding the trail that
leads from Washington, DC, to Mt. Vernon, Virginia.
William Schwarz has the longest association with
Hopkins. He received his BS, MS, and DrEngr de-
grees at Hopkins. His areas of research include non-
Newtonian fluid dynamics, rheology, physical acous-
tics of fluids, turbulence, and biotechnology; He is cur-
rently collaborating with physicians and speech
pathologists from Hopkins and Good Samaritan Hospi-
tals on the study of dysphagia, or swallowing disor-

The research of Joseph Katz involves nucleation
processes (e.g., condensation of supersaturated vap-
ors, boiling of superheated liquids, condensation in
flames, void formation in solids) and equations of
state. He has also worked at the nearby National
Bureau of Standards, studying combustion-generated
ceramic materials. Katz, like Donohue, came to Hop-
kins in 1979 from Clarkson with the expectation of
"high-quality students and shorter winters." His ex-
pectations were met. And, like Donohue, he ap-
preciates the intimacy that "comes with a class of fif-
teen instead of ninety students." At Clarkson, Katz
began his teaching career after a number of years in
industry. "I prefer teaching students how to do re-
search," he says. "You can't do that in industry." Katz
initiated the participation of undergraduates in re-
search projects at Clarkson. In fact, Donohue did un-
dergraduate research there with Katz.
Robert Kelly specializes in separation processes
(chemical absorption and stripping, in particular) and
biochemical engineering. After spending some time
with DuPont at Marshall Lab in Philadelphia, he re-
turned to school at North Carolina State University

Lower quad on a crisp winters day. photo by Carol Hyman


where he worked with Ron Rousseau and Jim Ferrell
on a project involving the removal of acid gases from
coal gasification streams. Since coming to Hopkins he
has returned to earlier research interests in biochem-
ical engineering, an area particularly appropriate
given Hopkins' strengths in the biological sciences.
Here the primary emphasis is on engineering prob-
lems related to bacteria from extreme environments,
especially extremely thermophilic archaebacteria.
"One of the great things about Hopkins is the oppor-
tunities for collaboration," he says. "Not only have we
been able to work with faculty in the biology depart-
ment, but with scientists at NIH, NBS, and other
government laboratories."
Kelly's other interests include sports; he often can
be found involved in lunchtime basketball games ("I
may be slow, but I can't jump.") or on the jogging
circuit around campus. He and his wife spend most of
their spare time trying to keep track of their two
young daughters, who are living proof of the second
law of thermodynamics.
Geoffrey Prentice traces his interest in chemical
engineering back to his early teens. "In the post-Sput-
nik era, do-it-yourself rocket construction was a popu-
lar activity among the junior high school set," he says.
His interest in rocket fuels led naturally to an investi-
gation of optimum mixtures (stoichiometry) as well as
combustion processes thermodynamicc and kinetics).
After completing his bachelor's and master's at Ohio
State, Prentice went to work as a staff engineer in
Sweden with Goodyear International. This was a won-
derful opportunity for both professional responsibility
and extensive travel throughout Europe. His second
assignment took him to the Republic of Zaire (for-
merly the Belgian Congo). Returning to the States,
Prentice completed his PhD at Berkeley, where he
worked with Charles Tobias on the modeling of cur-
rent distribution in electrochemical systems. Among
several projects he is currently working on, he and
fellow faculty member Mark McHugh are investigat-
ing the feasibility of performing electro-organic syn-
theses in supercritical fluids.
Prentice spends his spare time with his family,
waiting for the kids to be old enough to "get away to
places we haven't visited yet." He and his son just
completed a scuba diving course, and they "hope to
get under a few new places as well."
Mark McHugh came to the department in 1985
from Notre Dame. His areas of expertise include high-
pressure phase equilibria, polymer solution ther-
modynamics, and supercritical solvent extraction.
While he admits he was originally interested in coming
to Hopkins because of family ties, it was not long be-


photo by Carol Hyman
Prof. Robert Kelly with Loy Wilkinson, Chairman of the
Chemical Engineering Department Visiting Committee.

fore he discovered the benefits of being associated
with a small but well-known university. He enjoys
telling a story which illustrates the camaraderie
among faculty here. A short time after he arrived,
Bob Kelly introduced him to a colleague in the biology
department. They chatted about their work, and
McHugh was impressed with his friendliness. It
wasn't until some time later that he discovered that
the colleague was a Nobel Laureate. "This attitude
and friendliness is typical of Hopkins," says McHugh.
"You rub shoulders with some of the best people in
the world." McHugh believes this ability to interact
with faculty in other departments makes the univer-
sity much more than the sum of its parts. "Each part
is strong on its own," he says, "but working together
makes us formidable." McHugh is impressed with the
strides the department has made in the last few years.
At an AIChE meeting in Miami, all of the faculty gave
papers and three faculty chaired sessions. "Our impact
is being felt," he says.
McHugh likes to spend his lunch hour in the gym,
shooting baskets with Bob Kelly or lifting weights.
His free time away from Hopkins is spent discovering
Baltimore. "We go downtown as much as we can," he
says. "The geographical location is fabulous."
Chemical engineering's newest faculty member
came on board last November. Timothy Barbari's
areas of research include diffusion in polymers, mem-
brane science, and separation processes. He was at-
tracted to Hopkins because "there is a sense of
creativity and innovation here that is hard to find at
other universities. The opportunities for collaboration
within the department and across departmental lines


appear limitless." Still discovering much about Balti-
more, Barbari has been spending much of his free time
exploring the area and "getting lost in museums in
Two more faculty members will be joining the de-
partment this year. Michael Betenbaugh, whose spe-
cialty is biochemical engineering, and Mark Saltzman,
who works in transport phenomena and controlled re-
lease, are welcome additions to the department.

The program currently has about thirty graduate
students, most of whom are PhD candidates. At Hop-
kins, the PhD degree is an individualized research de-

photo by Carol Hyman
Grad student Galen Suppes injecting a sample into a
high-pressure equilibrium cell.

gree with few formal requirements. Each candidate
chooses courses with the help of an adviser in an effort
to obtain the depth of knowledge necessary to carry
out successful research in a specific subject while ob-
taining the breadth and flexibility of skills needed to
expand into new areas of research. The number and
type of courses depend upon the student's academic
background and areas of interest.
This individualized approach has attracted top stu-
dents to the program. Many of the grad students
agree that a drawing factor of Hopkins is the opportu-
nity for research experience coupled with classwork
related to this research. John Walsh, a PhD candidate
who came to Hopkins with a master's degree, says
that what attracted him to Hopkins was the size of the
department. He appreciates the chance to use his in-
itiative and to have a say in the direction of his re-
search. Due to the small student to faculty ratio, in-
itiative is encouraged and supported by faculty input.

The department also offers a Master of Science
degree. The course requirements are more well-de-
fined in this program, although specific programs are
chosen based on consultation with the student's re-
search adviser. In addition to coursework, the student
performs research culminating in a master's thesis.
Because of the diversity of expertise of the faculty,
students at Hopkins may choose from a wide variety
of areas of specialization.

Undergrads in chemical engineering at Hopkins
also reap benefits of individualized attention from fac-
ulty. Beside the Exxon program, many juniors and
seniors participate in research with faculty members.
They also have access to an array of computers, and
coursework involving computer applications is intro-
duced early in the program. The undergraduate cur-
riculum emphasizes chemistry as well as engineering,
mathematics and physics. Students with a degree
from Hopkins are well-prepared to continue to an ad-
vanced degree or to go right into a professional career.
And although many students pursue further studies
in chemical engineering, a number of students have
gone on to study business, law, and medicine.

And so, though it has only been eight years, "The
Hopkins" has returned. The chemical engineering de-
partment, beginning in 1979 with a few faculty mem-
bers but with a solid history to build upon, has seen
tremendous growth and change.
Since Marc Donohue has been with the department
from its inception, he has a clear picture of the growth
and changes that have occurred. He feels the depart-
ment is over those inevitable initial struggles. "The
department has finally gotten to the point where we
have an identity; we have stability; we have

This article is respectfully dedicated to the mem-
ory of Stanley Corrsin, Theophilus Halley Smoot Pro-
fessor of Fluid Mechanics at Hopkins until his death
June 2, 1986. Prof. Corrsin, at Hopkins since 1947, set
an example of style and tone for his many graduate
students and associates. His availability to workers in
all fields was an invaluable gift to Hopkins. Though
Stan Corrsin's legacy may well be his contributions to
fluid mechanics, he will be remembered by his friends
for his sense of humor, strength of convictions, untir-
ing pursuit of knowledge, and love of academic life. D


book reviews

by Raghu Raman
Elsevier Applied Science Publishers, 1985,
592 pages, $90
Reviewed by
Ihab Farag
University of New Hampshire
This is a very nice reference book which contains
a wealth of algorithms and computer program listings
(in FORTRAN) for solving a wide range of chemical
engineering problems. The author set out to show how
to develop algorithms and obtain solutions for a
number of practical problems. In addition, it has a
wealth of references and literature citations. I think
the book would be useful for those seeking to solve
modeling and simulation problems, but who do not
have access to well-developed process simulators,
e.g., ASPEN PLUS (trademark of Aspen Technology,
Cambridge, MA).
The seven chapters in the book are: Introduction,
Estimation of Gas and Liquid Properties, Mass Trans-
fer Operations, Flow of Fluids in Pipes, Heat Trans-

fer, Chemical Reaction Engineering, and Chemical
Process Simulation. The appendix has several useful
sections on: Matrix Methods, Solution of Equations,
Polynomial Approximation, Numerical Integration,
Ordinary Differential Equations, Function Extermi-
nation, and Computation Errors. Within each chapter
a number of models of varying degree of complexity
are described. A sample list of examples given include:
Application of UNIQUAC equation to obtain bubble
point of a four-component mixture, four-component
hydrocarbon mixture distillation, sizing of pipes for
non-Newtonian flow, shell-and-tube heat exchanger
calculation with phase change, process furnace
analysis using Hottel's zone method, residence time
distribution in a CSTR, three-phase fluidized bed
countercurrent backmixing model, and aniline man-
ufacture. The 100 computer programs are well-
documented and have been tested to insure correct-
The author, correctly so, assumes a basic under-
standing of unit operations and formulation of
mathematical models. I believe this book should be a
very useful reference for those interested in process
simulation. I only wish it was possible to get the pro-
grams in the book on an IBM compatible floppy dis-
kette. Fn


Making Significant Advances In Technology

The Amoco Research Center represents continued advancement in Amoco Corporation's support of
research and development. Petroleum products and processes, chemicals, additives, polymers and
plastics, synthetic fuels, and alternative sources of energy are only a few of the areas in which the
Amoco Research Center has made important contributions.
Located on 178 acres of spacious landscaped grounds in Naperville, Illinois, just 30 miles west of
Downtown Chicago, the Center employs over 1500 people. We are currently in need of enthusiastic
researchers who have received their degree in chemical, mechanical, or electrical engineering, to help us
improve the products and services we provide. You'll be part of a team that continually pushes back the
parameters of known technology.
Amoco is proud of its dedicated personnel and furnishes them an environment that encourages
creativity and is conducive to professional advancement. If you have the desire and proven ability to
work on mind-stimulating projects, we are prepared to offer a very attractive benefits package and
salary that reflects your expertise.
The research field provides a backbone for modern development-guiding industry through the future.
And you can be part of this.

Please send your resume to:
Amoco Research Center
Professional Recruiting Coordinator
Dept. CEE/12
P.O. Box 400
Naperville, Illinois 60566


\ An equal opportunity employer M/F/H/V


J classroom




Worcester Polytechnic Institute
Worcester, MA 01609

A MAJOR GOAL of the senior year process design
course at Worcester Polytechnic Institute (WPI)
is to expose chemical engineering students to the ap-
propriate application of a variety of computer tools
within the design task. These range from material and
energy balance calculations using a flowsheeting pack-
age (such as FLOWTRAN) on a mainframe machine,
to interactive uses of computer graphics and input
screens on a microcomputer. One area where interac-
tive computing can be particularly helpful is in the
teaching of process synthesis topics such as heat ex-
changer network and separation train design.
Process synthesis involves the partial or complete
invention of the flowsheet to achieve specified ends.
In the early stages many alternatives can be gener-
ated for evaluation, and computer graphics can allow
a student to screen (in both senses of the word) com-
peting flowsheet designs rapidly and easily. This
paper describes the HENS (Heat Exchanger Network
Synthesis) program used at WPI as an aid to perform-
ing the energy integration step in process design.

The problem of heat exchanger network specifica-
tion, or energy integration, arises for a partially com-
pleted flowsheet in which major equipment items and
process streams are known. Streams are identified
which have heating or cooling needs, and the task is
to meet these needs at minimum cost. Other objec-
tives may also be considered, such as flexibility in the

One area where interactive
computing can be particularly helpful is in
the teaching of process synthesis topics such as heat
exchanger network and separation train design.

Copyright ChE Division ASEE 1987

Anthony G. Dixon is an associate professor in the Department of
Chemical Engineering at Worcester Polytechnic Institute. He received
his BSc in mathematics and his PhD in chemical engineering from the
University of Edinburgh, and was a post-doctoral fellow at the Univer-
sity of Wisconsin from 1978-1980, when he joined WPI. He teaches
process design, mathematical modelling and transport phenomena,
and his research interests include fixed bed reactor modelling, zeolite
technology and semiconductor technology.

face of process disturbances, and constraints of plant
layout and process safety.
The extent to which the heating and cooling needs
may be met by heat transfer between process streams
is limited by both the enthalpy contents of the streams
and the minimum approach temperature (ATmin) of
two streams in a heat exchanger. This minimum tem-
perature driving force is usually fixed by rule of
thumb before the synthesis task is begun and is taken
as constant for all exchangers. Utility heating and
cooling are used to bring streams to temperature
when further process heat exchange is not possible.
This problem is a relatively mature one, and one
method, the pinch design method, has been exten-
sively developed [1, 2, 3] and applied to real-life prob-
lems [4]. It is particularly suitable for undergraduate
instruction, as it is simple to apply, has a high graphi-
cal component, and requires active participation and
decision-making by the designer at all stages. The key
elements of the method are 1) preanalysis of the prob-


lem data to give minimum possible utility require-
ments, minimum number of units (both process and
utility), and identification of the location of a temper-
ature "pinch point" (if one exists); 2) specification of
process stream matches beginning at the pinch point,
as the design is most constrained there, to give a
"minimum utility" network; 3) evolution of the
minimum utility network to examine trade-offs of
energy recovery versus units, i.e., operating versus
capital costs. The preanalysis step involves repetitive
calculations and can easily be relegated to a computer.
The two design steps (2 and 3) involve drawing and
re-drawing the exchanger network for graphical ma-
nipulations. The HENS program allows this to be done
interactively, saving time and allowing students to in-
vestigate a wider range of design alternatives.
Several other computer packages for heat ex-
changer network synthesis exist, although none serve
the function of the HENS package. The HEXTRAN
[5] simulation package allows for design of the net-
work by repetitive calculation of heat and mass bal-
ances and economic evaluation of modified networks.
It provides an alternative approach to the use of pinch
technology. Both Chemcalc 5 [6] and the recently-de-
scribed program by Govind et al [7] design the net-
work for the user, which is not the objective of in-
structional use.
SuperTargetTM presents a comprehensive frame-
work for preanalysis and investigating the effects of
process modifications for both grassroots and retrofit
design [8]. Its graphical displays include composite
curves on the temperature-enthalpy diagram, and an
annotated design grid. SuperTargetTM does not pre-
sent the entire network design on the grid, but as a
design aid gives the user a full analysis of the conse-
quences of any process stream match.
RESHEX [9, 10] allows preanalysis of data, in-
teractive synthesis of a new network or modification
of an existing one, or automatic synthesis of a network
by the program. In addition, considerable emphasis is
given to post-analysis of a network for optimization,
feasibility and resilience testing. RESHEX is a com-
prehensive program of great utility, but appears to be
limited in its use of graphics and user-friendliness. As
an instructional aid, it may be more suitable for ad-
vanced classes and as a research tool.
A more mathematically rigorous approach than the
pinch method to the synthesis of optimal heat ex-
changer networks has been developed [11] and em-
bodied in the MAGNETS program [12]. This is
another very powerful package, able to handle stream
splits, multiple pinch points and restricted matches,
and to automatically synthesize the network struc-

This problem is a relatively mature one,
and one method, the pinch design method, has been
extensively developed and applied to real-life
problems. It is particularly suitable
for undergraduate instruction . .

ture. Again, this may be a more advanced tool than is
suitable for beginning students.

The fundamentals of the pinch design method are
covered in class before the students are given the
HENS program. The minimum utilities are derived
for an example problem both on the temperature-
enthalpy diagram using composite curves, and alge-
braically by the "problem table" [1, 2]. This introduc-
tory material is limited to constant heat capacity-flow
rate (CpW) streams, although linearization of the tem-
perature-enthalpy relationship is mentioned. The sim-
ple minimum units formulas [1, 2, 3] are also pre-
sented during this target-setting stage.
Particular attention is paid to the physical inter-
pretation of the pinch point and the rules regarding
decomposition of the problem at the pinch [3]. The
grid representation of the network is introduced and
an example worked by hand in class. Stream splitting
and feasibility criteria at the pinch are discussed.
The final part of the lecture material explores the
incompatibility of minimum units and minimum
utilities, and the generation of a small number of near-
minimum-utility networks by a utilities/units trade-
off. Depending on time, unpinched problems, resi-
liency and multiple utilities pinches may also be in-
WPI has recently endorsed the AT&T 6300"
microcomputer for undergraduate instruction, and al-
most all seniors have experience on this machine. As
the computer is integrated into the curriculum most
students will own one. There is thus no need for spe-
cial instruction on the microcomputer use itself.

The HENS program is written in Turbo PascalTM
for the AT&T 6300 microcomputer. The main program
and associated files take up 70K of memory. The in-
teractive input procedures are modified versions of
the general-purpose code described by Wood [13]. The
graphical part of the program was developed with the
aid of the Turbo Graphix ToolboxTM.
The program is distributed to the students in a
compiled form; the source code is not made available
nor can they abstract parts of the package for future
use. This protects the interests of the commercial


software vendors. Students may either use the distri-
bution disks in any one of several computer labs on
campus, or copy a disk for home use on their own
The HENS program operates in two parts: data
entry and analysis, and graphical network design. The
design part is nested within the data entry part, so
that students can abandon ( Exit) a design to
change the problem data without returning to system


The students enter data to the program by filling
in blanks on an input screen. The initial screen with
some default values is shown in Figure la. Prompts
appear at the bottom of the screen, and menu options
are selected by keying in the initial letter of the op-
tion. All input is user-friendly and controlled, so that
if inappropriate characters are entered they are trap-
ped, an error message is issued and the prompt for
the required input is repeated.
The Compute and Design options are not active

Number of Streams...: 0
Minimum approach temperature...:


Pinch point location...: None
Minimum hot utility...: 0.00 Minimum cold utility...:

IMP- Press a CMD: key to enter selection -=>
CMD: Compute / Modify / Design / Fl HELP / Exit


Number of Streams...: 4
Minimum approach temperature...:
1 150.0
2 90.0
3 20 0
4 25.0


Pinch point location...: 90/70
Minimum hot utility...: 107.50 Minimum cold utility.

4J : ,20.0
MSG: Enter minimum delta-T. Up to 6 digits inc. decimal.
CEMD: t Prey Pld / Clear Pld / Exit




until the Modify option has been selected at least once.
Pressing "M" causes a cursor to appear at the first
entry, the number of streams (between two and ten).
The cursor is moved to the next entry by pressing
Enter, or can be made to go back to the previous
entry by the Up arrow. In Figure lb the input screen
is shown for a four-stream design [3] in which the
cursor has been placed at the second entry, the ap-
proach temperature. The current value of the entry
appears at the bottom of the screen, along with a suit-
able prompt and cursor control options. If Enter is
pressed the current value is accepted, otherwise a
new value should be typed in. Again all input is con-
trolled, and only a positive real number will be ac-
cepted in this case. Pressing at this point
causes a return to the data entry and analysis menu.
The Compute option causes a calculation of
minimum utilities and the pinch point location to be
performed. This just follows the Problem Table al-
gorithm [1, 2] and is restricted, like the HENS pro-
gram, to streams of constant CpW with no phase
change. Figure lb shows the results of this option for
the four-stream problem. If a pinch point does not
exist, the algorithm recognizes this and reports
"Help" screens are available by pressing the Fl
special key when this appears as a menu option. A
new Help Menu appears, from which the students can
select a Help screen on each data entry menu option.
These screens consist of explanations of program op-
eration, warnings and occasionally a reminder from
the lecture material. An exit from the Help menu re-
stores the data input screen.


Selecting the Design option on the data entry and
analysis menu passes the user to the design stage.
The input screen is replaced by an initial stream grid
and design menu (Figure 2). The student can return
to the input screen by pressing during the
DO design stage. On the initial grid the problem data are
o0 shown associated with hot (upper) or cold (lower)
streams. The minimum utilities targets are displayed
as a convenient reminder.
A design according to the pinch decomposition
principle is not forced upon the students, therefore
40.00 the pinch is not shown at first. It can be turned on by
selecting the Pinch option, upon which a cursor is pro-
vided at the top of the screen that can be moved hori-
zontally, to position the pinch lines (see Figures 3 and
5). The pinch can be removed by selecting the Pinch
option again, and can be turned on and off and re-
positioned as often as desired.




100 25


IHP: Press a CHD: kex to enter selection ==>

CHD: Add / Delete / Split Strea / Pinch / Fl HELP / (Esc) Exit


CMD: Add / Delete / Split StreaI / Pinch / Fl HELP / (Esc> Exit


For the Add, Delete and Split Stream options, po-
sitions must be indicated on the stream lines. When
one of these options is selected, a cursor appears
which can jump vertically from one stream to another,
or move along a stream in discrete jumps. The stu-
dents use the arrow keys to move the cursor. For a
process heat exchanger the cursor movement is re-
stricted to vertical leaps once the first stream position
has been marked (by pressing Enter); similarly the
cursor can only move horizontally once the first split
stream position has been given. These features pre-
vent creative drawing by the more artistic of our en-
gineers. The cursor can follow along split streams, the
branch taken depending on whether it approaches
from the left or the right.
The Add option inserts both utilities and process
exchangers. For a utility, the student must give inlet
and outlet temperature, and the program calculates
load, displayed beneath the unit. For an exchanger,
either the hot stream or cold stream can be the first
stream specified. Both inlet and outlet temperatures
must be given, as well as the inlet temperature on the
second stream. The program calculates load and sec-
ond stream outlet temperature. To save space, utility
units can be placed with the outlet temperature over-
writing the stream target temperature (see Figure 3).
The Delete option removes both types of units.
The cursor is simply placed over the unit-either end
for an exchanger-and Enter is hit. The unit and as-
sociated numbers are removed. Both Add and Delete
are very rapid, allowing students to recover quickly
from mistakes.
A stream can be Split, but the split cannot be
erased short of beginning the design again ().
The student has to specify CpW for the top branch,
and the branch CpWs are displayed beside the split.
Continued on page 156.

1" ""0 .
3t7 32t 3t3 2 30 --

003 'I
u220 220s i2 1tt lot0
-------- --- Q --------- ..
^" ---- -- -- -^ -------" ...

INP: Press a CHD: keg to enter selection ==>
CHD: Add / Delete / Split Strea / Pinch / Fl HELP / (Ese> Exit



j classroom


Berkeley's Multiloop Computer Control Program

University of California
Berkeley, CA 94720

M ULTILOOPS ABOUND. They are all around us.
We invent these intricate control systems and
apply them to distillation columns, fired heaters, reac-
tors, steam systems. They are the "brains" and "nerv-
ous system" of chemical processes. Processes need
them to operate and to operate safely.
That's news? Not at all. Everyone has known it for
generations. Everyone, that is, except the students in
our process control courses here in the States.**
That's got to change! And the change has to be
made in the first course in process control-the first


Author Alan Foss writes that "after a quarter of a century of search-
ing for ways to tell Californians about process control, I am still search-
ing. The article published here reports one of the 'finds' along the
way. Richer veins assuredly lie somewhere farther along the tunnel."
Professor Foss came to the academic world after five years of industrial
practice with the DuPont Company and studies at Worcester Polytechnic
Institute and the University of Delaware.

*Copyright 1986 by P. H. Gusciora, C-H Mak, L. Poslavsky, and
A. S. Foss
**There is probably a handful of departments to which this state-
ment does not apply, and I know colleagues there will forgive this
slight overstatement in the recognition that it is very close to the

course, my colleagues, because there is seldom a sec-
ond. Not everyone will agree with that of course, and
those that do will ask, "How?"


Imagine that you have a computer program that
permits the user to configure any multiloop control
system he desires for a particular process, say the
system shown in Figure 1 for a distillation column.
And suppose such a control system accepts process
"measurements" from a dynamic simulation of the col-
umn and delivers its "commands" to that same simula-
tion. With the keyboard command

YC, KP = 10, KI = 20, FREQ = 5
the user sets the proportional- and integral-gain pa-

Distillate D
Y Mole Frac.

Reflux, L

Bottoms B
X Mole Frar

FIGURE 1. Control system for regulation of top and bot-
tom product concentrations by manipulation of distillate
and boilup flow rates. The relative gain for this control
configuration is 0.68.

0 Copyright ChE Division ASEE 1987


Imagine that you have a computer program that permits the user to configure any multiloop control system
he desires for a particular process, say the system shown in Figure 1 for a distillation column. And suppose
such a control system accepts process "measurements" from a dynamic simulation of the column . .

C. ONLINE Plot of variables vs, time


Date 87/ 2/ 2 Tie 8: 6: 6, 8


' ^ -.. .. .... . . .. ... ...



Title 'D-V CONFIG,,XP:5,KI:18 TOP; KP:15,XI:30 BOTTOM

FIGURE 2. A page of the screen display of trends in top
and bottom product concentrations. Response to a step
increase in setpoint of bottom product concentration
with control system of Figure 1.

rameters of the top-product concentration controller
YC and the sample time to 5 seconds.
The command

turns the controller to ON status and the user now
has an operating process running in a real-time mode
under the action of a simple control loop. He may com-
mission the bottom-product concentration controller
XC in the same way with the command

XC, KP = 10, KI = 20, FREQ = 5, CD = ON

at which time he will find himself in the land of mul-
tiloop control, control loop interactions, active input
constraints, reset windup, and multiple alarms. Get-
ting such a system to work when the column feed rate
is varying will likely require a little tuning, all of
which can be done "online" while the process and con-
trol system is running by simply typing commands for
setting KP and KI similar to those above. The user
knows how his system is shaping up by viewing
periodically updated tabular data about the measured
and manipulated, variables and controller states or
graphical displays of trends in any set of selected sys-
tem variables, such as the group in Figure 2. The
graphs shown there come from a full tray-by-tray cal-
culation of a 39-tray column.
Now suppose the user is dissatisfied with the best
performance he can squeeze out of this particular con-
trol system configuration and is curious about the



Bottoms B
X Mole Froc.

FIGURE 3. Control system using the ratio D/V and V as
manipulated inputs. The relative gain is 1.92.

claims found in Shinskey's book [4] for the superiority
of the configuration shown in Figure 3. He is curious
because the argument about the reduction in loop in-
teraction given by Shinskey seems to be just what he
is looking for, but he will have to see the performance
improvement to believe it; Shinskey does not show
performance. Some reconfiguration obviously needs to
be done to convert the control system of Figure 1 to
Figure 3.
No problem, as we say. Turn off both controllers,
define a new variable as the product of D/V and V in
the overhead system, and redirect the output of con-
troller YC to the multiplier. The sequence of com-
mands for these changes, which can be made in a few
minutes at the keyboard once one has decided what to
try, are shown in Table 1. Turn both controllers to

Keyboard Input to Convert the
(DN) configuration to (D/V,V)
ASUM,AL=+,V1=DM,V2=LM,AO= .0,A1=1.0,ML=0.0,MH=1200,UN=MPH,VA=600
D*,AFL=*,IO=24,V1=ASUM,V2=DVR,AO= 1, A=0,MH=300,ML=0,UN=MPH,VA=100.0


Now, things are a little more involved than
I have made out. I am sure that that is no news.
There are a lot of details about maximum and minimum
values of variables everywhere in the control
system that need specification . .


Plot of variables us, time

Title :DM- CONFIC,,X1P,BQ85,XI:,816? TOP; KP:15,KI:39 BOTTOM

FIGURE 4. Top and bottom concentration responses to a
step increase in setpoint of bottom product concentra-
tion. Control system of Figure 3.

ON and retune. Sure enough; performance is indeed
better. The top product concentration is barely influ-
enced by a change in the set point of the bottoms
controller (see Figure 2 and 4).
"Professor, just a moment," you interject. "What
about the need to recompile the control system pro-
gram and to relink it to the process simulation?"
That's not necessary these days. Everything about
the control system is stored in tables. The program
simply moves data from here to there when asked. All
of that was worked out years ago by the computer
scientists; we simply adopt the technique.


Now, things are a little more involved than I have
made out. I am sure that that is no news either. There
are a lot of details about maximum and minimum val-
ues of variables everywhere in the control system that
need specification, else alarms announcing over-rang-
ing would never reach the operator's eyes or ears.
The declaration of such maximum and minimum val-
ues for all variables is an added chore for the user, to
be sure, but is not the consideration of process limits
important to safe process operation? It is, and we
should expose students at least once to this component
of the process control task.
The input of such information is easy. The screen
display of the set of process variables for Shinskey's

control configuration is shown in Figure 5. The vari-
ables named in the list are identified by the labels
shown on the process diagram in Figure 3. All vari-
ables are free to be named by any 4-character symbol
the user desires. So also may controllers be named.
The command structure of ONLINE is designed so
that the user need only type the name of the variable
or controller followed by the attributes to effect what-
ever setting or change he desires to make in those
attributes. Important information for the top product
concentration y, for example, is that its input channel
(IO) is 1, that the type of algorithm (AL) is an analog-
to-digital conversion (AD), and that the linear conver-
sion of the concentration transducer signal to mole
fraction has a slope of 1.0 (Al) and an intercept of zero
To set this information, the user simply types

y, IO = 1, AL = AD, Al = 1, AO = 0

This information appears immediately on the screen
upon completion of typing this line. Error messages
appear should there be any miskeying. Other vari-
ables in this example are seen to be identified as mul-
tipliers, summers, and digital-to-analog (DA) conver-
sions. The use of input-output A/D and D/A data chan-
nels is a carry-over from a version of this program
used with experimental apparatus. Their retention
here serves two purposes: a decoupling of the simula-
tion and control program is achieved, and students
are made aware early that input and output channels
must be specified when communicating with physical
processes. Variables such as summers and multipliers
have their inputs named under the VI and V2 col-
umns. With such declarations, the user specifies that
part of the control system configuration. The
maximum and mininum values mentioned earlier are

Ln NAme AL 10 VI V2
I Y AD 1
2 X AD 2
3 LA AD 3
4 LB AD 4
5.DM AD 6
7.BM AD 5
8 VM AD 8
9 DP AD 9
ll.L DA 3
12:B DA 1
13*V DA V
14,Z DA 5
15.F DA 6
16b0 DA 7
170YSP 0
19.DVR 0
20,ASUM + 0 DM LM
210D* 24 ASUM DVR

24.0 DA 2


Date 2/ 2/87
1.0 .000
1.0 .000
1.0 0.
1.0 0.
1.0 0.
1.0 0.
1.0 0.
1.0 0.
1.0 .000 1
1.0 0.
1.0 0.
1.0 0.
1.0 .000
1.0 0.
1.0 -.2
1.0 .600
1.0 .001
1.0 .00
1.0 .0 1
.00 .0

Time 0: 6:47. 0

79.763 .00 1.0 .00 300.00 79.76 MPH

FIGURE 5. Screen display of process variables for control
system of Figure 3.


listed under the columns labeled MH (measurement
high) and ML (measurement low). These limits repre-
sent the operable range of the measurement trans-
The remaining part of the system configuration is
established by naming the inputs and outputs of all
the controllers. Figure 6 displays the video screen
"page" that provides that information for each control-
ler. The measurement source (MS), setpoint source
(SS), and output destination (OD) for each controller
is declared by variable name. These "connections" can
be altered easily by the user through a few key
strokes like those just mentioned. Such ease of recon-
figuration is a feature indispensable to the efficient
use of ONLINE in coursework. Students need to im-
plement their conceptions in a matter of minutes, not
days. Figure 6 also displays maximum and minimum
declarations for the setpoint, the measured variable,
and the controller output. The significance of these
limits differs from those of the process variables just
described. The limits on the measured variable, for
example, are considered alarm limits, which when
transgressed trigger an H or L message to the
operator. The setpoint limits, normally set "inside"
the limits on the measured variable, constrain the de-
sired range of the controlled variable. Output limits

Page 1 LOOPS
Ln CmmD CScd NAme

MeaS UNit OUtput
.950 MLFR .13
.035 MLFR 594.
514. MPH 514.
120. MPH 120.

Date 2/ 2/87

Time O: 8: 3. 0
1.67E-02 .00
30. .00
.00 .00
.00 .00

FIGURE 6. Screen display of controller page for system
of Figure 3.

reflect the rangeability of the process manipulatable
variable driven by the controller. The ON-OFF-
COND-FAIL status of the controller and the PID pa-
rameters are also displayed. All of the information for
the process variables and controllers just described is
"dynamic" and is updated in the data base and on the
video display at intervals selectable by the user.
Suppose the exercise for the day concerns the
tuning of a multiloop system already configured. The
complete slate of information just described about sys-
tem variables and controllers (excluding the PID pa-
rameters) can be prepared by the instructor ahead of
time in a disk file. The user merely types


to load the entire configuration. Keyboard work is
then necessary only for setting controller parameters.
Portions of such a SETUP file are shown in Table 2.
These files also serve as a permanent record and
documentation of each control system.

Key Portions of the File Used to Set Up the
(D,V) Configuration for the Example of this paper


Y ,IO=1,AO=0.0,A1=1.0,MH= 1.0,ML=O.O,SF=3,AL=AD,UN=MLFR
X ,IO-=2,AO=O.,A1=I.O,MH= I.0,ML=O.OSF=3,AL
LA,IO=3,AO=O.O,A1=1.O,MH= 600.,ML-O.
LB, IO=4,AO=O.O,AI=. O,MH-

D ,IO=2,AO=O.O,A=1.0,MH= 239.,ML=0.0,SFO=,AL=DA,UN=MPH ,VA= 80.
L ,O=3,AO=0.O,Al=1.O,MH=1534.,ML=O.O,SF=O,AL=DA,UN=MPH ,VA= 511.
B ,IO=1,AO=O.O,A1=1.0,MH= 361.,ML=0.O,SF=O,AL=DA,UNMP "
V IO=4,AO=O.O,Al=1.0,MH-H=1183.,ML=O
Z IO=5,0=0. 0 _A. ...............

YC ,ST=I.OMH= l.,ML=O,SH=O.999,SL=0.001,OH=300.,OL=O
XC ,ST=1.0,MH=I.,ML=O,SH= .999,SL=0.O0l

If the exercise asks the user to invent his own
control system, then the instructor merely lops off
that segment of the SETUP file defining the control
links, leaving only the process variables for loading.
In Table 2, everything from the entry LD (loop defini-
tion) and below would be omitted in such a SETUP
file. Creating that slate of information would consti-
tute the exercise for the day. Alternatively, two
SETUP files could have been prepared, one defining
process variables only, the other the control links; the
instructor supplies whichever file combination is ap-
Or "preanalysis" programs can be developed that
prepare a complete SETUP file defining process vari-
ables. We have such a "front-end" program for binary
distillation columns that calculates and displays the
relative gains and steady-state operating conditions.
A full set of information for the distillation simulation
and the SETUP file for process variables, like that in
Table 2, is written to disk upon user command and is
read into ONLINE's data structure during the initiali-
zation phase. Such "front-end" programs, particularly
for distillation columns, make it practical for students
Continued on page 154.


l classroom



University of Cambridge
Cambridge, CB2 3RA England

THE QUASI-STEADY STATE combustion of a single
isolated liquid oil droplet has been considered by
several authors [1-5]. We have found it necessary in
our teaching to seek a simpler treatment than those
usually presented. Below we consider such a single
drop, burning in stagnant air without any effects from
natural convection or radiative heat transfer. Of
course, once this, the simplest situation, has been de-
scribed mathematically, it is possible to treat more
realistic and complicated cases. The experimental ob-
servations are clear, and excellent reviews are avail-
able [1]. The fundamental empirical fact is that the
radius, a, of a liquid oil droplet decreases with time,
t, according to

Allan Hayhurst is a lecturer in chemical engineering at Cambridge
University and a Fellow of Queens' College. He received his PhD de-
gree from Cambridge where he has spent most of his career, apart
from seven years at Sheffield University. His research interests include
combustion and reactions in fluidised beds. (L)
Ronald Nedderman is a lecturer in chemical engineering at Cam-
bridge University and a Fellow of Trinity College. His research interests
include stress and velocity predictions in flowing granular materials
with particular emphasis on silo design. He has never done any research
on combustion but enjoys interfering in other people's algebra. (R)

flame front

mo /s

mol /s )

(nM mol/s)

FIGURE 1. Sketch of burning oil droplet surrounded by
spherical flame front and sphere of general radius r. The
total radial flow rates of fuel, 02 and products are as

-ddt = constant, (1)
during its burning. This paper gives a rigorous deriva-
tion of kb. Unlike the standard textbooks [2, 3], the
approach below uses entirely molar, rather than
mass, units because of the much simpler description
of diffusion which results. Like Long's analysis [4],
but in contrast with Spalding's [2], this treatment
takes the products of combustion into account. How-
ever, unlike Long, we calculate, from first principles,
the temperature of the flame front surrounding the
burning droplet. All these approaches rely on God-
save's early experimental and theoretical work [5].

To formulate the equations for diffusion we first
consider a binary mixture of two gases A and B mov-
ing together in a one-dimensional situation. The molar
flux of A in the x-direction (for a stationary observer)
is given by

Copyright ChE Division ASEE 1987


The oil droplet, depicted in Figure 1, is assumed to be at a constant temperature which is just less than,
but close to, the boiling-point of the liquid fuel. While the droplet evaporates, vapour of the fuel
diffuses outward; oxygen from the surrounding air diffuses inward.

c dc dc
NA = NZ'--D = NyAD (2
where CA and CB are the molar concentrations of A and
B, c (= CA + CB) is the total molar concentration, N
is the total molar flux, i.e., NA + NB, and YA is the
mole fraction of A. Eq. (2) derives from the fact that
the flux of A consists of a convective term, NyA, and
a diffusive term, -D dcA/dx. If A is in a multicompo-
nent mixture of gases (temperature T and total pres-
sure P) and confined to a one-dimensional situation,
then Eq. (2) becomes
DP dyA
NA = A A (3)

This assumes that multicomponent diffusion can be
simplified to a pseudobinary description with a con-
stant effective diffusivity, DA. Eq. (2) is identical with
the possibly more familiar [6] form:

By analogy with Eq. (2) the equation for the enthalpy
flux, Q, is
Q Nh k (4)

where k is the thermal conductivity and h is the molar
enthalpy of the particular gas mixture.

The oil droplet, depicted in Figure 1, is assumed
to be at a constant temperature, To, which is just less
than, but close to, the boiling-point of the liquid fuel.
While the droplet evaporates, vapour of the fuel dif-
fuses outward; oxygen from the surrounding air dif-
fuses inward. Fuel and oxygen meet in stoichiometric
amounts and react very rapidly in a thin reaction zone,
or flame front, at a distance rf from the centre of the
dropet. Here the oxidation reaction is assumed, to be

1 fuel + s 02 + n products

and is taken to occur almost instantaneously. Con-
sequently, the ratio of the molar fluxes of 02 and fuel
into the flame front is s. No distinction will be made
at this stage between the products CO2 and H20, al-
though this is a straightforward thing to do, if it were
considered necessary. Thus a total of n moles of prod-

distance from centre of droplet/arbitrary units

FIGURE 2. Approximate sketch of mole fractions of fuel
vapour, 02, N2 and combustion products around liquid
droplet, together with gas temperature.

ucts diffuses radially outwards for each mole of fuel
consumed. The situation in Figure 1 has spherical
Figure 2 sketches the expected concentration pro-
files of fuel, 02, N2 and products. It will be noticed
that for a radius r < rf there is no oxygen, and for r
> rf there is no fuel. The concentrations of both oxy-
gen and fuel in the reaction zone are taken to be in
effect zero, i.e., reaction is rapid. As the droplet con-
tracts, both N2 and products diffuse inward to fill part
of the space formerly occupied by the liquid fuel. The
fluxes of N2 and products toward the shrinking drop-
let's surface must be small, since the upper limit of
their sum is (Pgas/Pliq) times the flux of fuel outwards
away from the drop. Here (pgas/Puq) is the ratio of the
molar density of the gas around the drop to that of the
liquid fuel; its value will be roughly 1.5 x 103. These
considerations of the limiting case, when all the
evaporating liquid is replaced by N2 and products, es-
tablish that the flux of, e.g., N2 is everywhere negli-
gible, as also is the flux of products in the region a <
r < rf.
Of the heat liberated in the reaction zone, some is
conducted back to the droplet and provides the latent
heat of vaporisation. The rest is convected or con-
ducted outward from the flame front. Radiative ef-
fects are ignored here. A rough sketch of the temper-
ature of the gases around the droplet is also given in


Figure 2. The temperature is assumed to have a
maximum value of Tf in the reaction zone, which is
well above T., the temperature of the surrounding
air. A calculation given below indicates that Tf can be
as high as 900 "C.
The above model is one controlled by diffusion and
heat transfer, with chemical kinetics being relatively
fast. Suppose that M mol/s of fuel evaporate from the
drop. The steady state rate of heat production in the
flame front, where the temperature is a maximum, is
accordingly MAH (N.B. the molar heat of combus-
tion of the fuel, AH" < 0). Of this heat an amount MX
is conducted inward to the droplet surface, X being
the molar latent heat of evaporation. Thus for r > rf,
the total net outflow of heat past any spherical surface

(AH + X) M

In addition, the total inflow of oxygen toward the
reaction zone is sM mol/s and the total outflow of prod-
ucts is nM mol/s (see Figure 1).
In principle, there are five basic unknowns in this
problem: the temperature of the drop, To, and the
corresponding vapour pressure, Pv, of the fuel; the
evaporation rate, M; the radius, rf; and temperature,
Tf, of the flame front. These can be evaluated by con-

1. Heat transfer from the flame front to the drop
2. Heat transfer outward from the flame front
3. Mass transfer of fuel from the droplet to the flame front
4. Mass transfer of oxygen inward to the flame front
5. The dependence of vapour pressure of the fuel on temper-

In fact, it turns out that the drop is close to its
boiling point, so that Pv is approximately atmospheric
pressure. Consideration of 1, 2 and 4 is now sufficient
to predict M, rf and Tf and the remaining two relation-
ships can be used to confirm the starting assumption
that the drop is very close to its boiling point.

Here we consider oxygen diffusing from the sur-
rounding air to the reaction zone. The flux of 02 past
a sphere of radius r is


with the negative sign denoting an inward flux. Sub-
stitution into Eq. (3) gives

sM (n-s)My02 D02P dy02
4rr2 4nr2 RT dr

This equation assumes the total flux of all species to be


i.e., the difference between the fluxes of products out-
ward and 02 inward. That the flux of N2 in this region
of space is negligible was demonstrated above. The
above equation gives

S dr D2 fY dy02
4i r2 RT (n-s)y + s
r, 0

for a mole fraction, y-, of 02 far away from the drop-
let. Integration yields

M D P (n-s)ye
47rr- RTln- n 1 S+ J -S

In deriving Eq. (5) it has been assumed that Do/T
is independent of temperature, so that D02 is a mean
value for the range T, < T < Tf. In fact, Do2 a T53
would have been a better approximation [7], but the
need for simplicity is paramount and a careful averag-
ing is not attempted at this stage.

There is only fuel (M mol/s) diffusing outward in
the region a < r < rf, with heat being conducted in-
ward to the droplet at a rate MX. In the steady state
the total enthalpy flow rate past an arbitrary surface
of radius r is given by Eq. (4) as

4rr2 Q = 4rr2 (- kf ) + M(H' + cfT)

Here kf is the thermal conductivity of the fuel-rich
gases for r < rf, cf is the molar specific heat of the fuel
and h = Hf + cfT is the molar enthalpy of the fuel at
temperature T. If T is in "C, then Hf is the fuel's
standard molar enthalpy referred to 0C as the datum
for enthalpy. It is convenient to use o0C as the refer-
ence temperature for enthalpy, as it avoids an explicit
statement of where the enthalpy datum lies. The
above expression for the net energy flow rate past
any sphere of general radius r can be calculated at the
droplet's surface. Here the rate of heat conduction
inward to the droplet from the surrounding gases is
MX (see above) and the molar enthalpy of the vapour
leaving the droplet is (Hf + cfTo), where To can be


taken to be the boiling point of the fuel in C. Equating
the enthalpy flow rates past spheres at r = r and r =
a gives

4rr2 (- kf A) + M(H; + cfT) = MA + M(Hf + cfT,)

This simplifies to

47r2 kf = M {t+ cf(T-To))

M rf dr

Tf k dT
0 X + c,(T-T)

M = kf f(Tf-T)
^ ft-y c (1 -)

Eq. (6) assumes mean values of kf and cf for the inter-
val To to Tf.

If combustion in the flame front is adiabatic, so
that, e.g., there are no gains or losses of energy by
radiation, then the rate of energy flow past any sur-
face r = r equals the rate of flow across the droplet's
surface at r = a. In the steady state this can be ex-
pressed as

MX + M(H* + cfTo) = 4r2k dr
+ M{n(H + c T) s(H 0+ c T)} (7)
P P 02 02

In Eq. (7) positive signs represent outward radial
flows. Again, H2 and H are, respectively, the stand-
ard molar enthalpies of oxygen and products (lumped
together, with no distinction between the triatomic
species CO2 and H20), but referred to 0C. The ther-
mal conductivity of this oxygen-rich mixture is ko. By
definition AH, the heat of combustion at 0C, equals

nH sH H*
p 02 f

so that Eq. (7) simplifies to

M(AH + X) = M{T(nc scO2) cT } 4r2k d
p 2 fo f dr

The left-hand side of this equation corresponds to
there being an outward flow rate of heat equal to M
(AH + X) past a surface of given r (> rf), as discussed
above. Rearrangement yields

0 T
dr kr k dT
r2 (nCp SC2)T + AH + X cfT
rf f

M k (ncp sc2)T + AHo + X- cfTo
47rf (nc Sc 2) ( nc sc )Tf + AH' + X To (8)

This assumes that specific heats and ko do not vary
with temperature. Addition of Eqs.(6) and (8) gives

M kf n c (T T )

k -(nc-pC sc)T +AHn++-cT
k i(ncP c 02)T + AH + c T
+ (nc -sc) (nc SC2 )Tf + AH + X c T

(6) The subsequent calculation can be simplified by as-
suming that

AHI >> (X c T )

as is borne out in reality. Also, if a mean specific heat,
Co, for the fuel-lean gases at r > rfis defined, so that

(n-s)c = ncp sc0

then Eqs. (8) and (9), respectively, become

= k k(n-s)c T + AH
4irf in [(n-s)coTf + AHO (10)

M k i cf(Tf- To)] k (n-s)coT + AHO
4T-a "c in 1n X -s+ (n-s)co+ AH

Eqs. (5) and (10) can now be equated to eliminate (M/
4'nrf) and give

PD (n-s)yj k (n-s)coT + AHO
RT_ n I + in n-sc T + AHO

which gives the value of Tf. The last, but strictly-
speaking unnecessary, assumption can now be made
to simplify the algebra. This is to recognize the dimen-
sionless group
(k RT/PD c )

as the Lewis number, or the ratio of the Schmidt and
Prandtl numbers for the oxygen-rich gases at r > rf.
Simple versions of the kinetic theory of gases give a
value of unity for the Lewis number. This is often a
Continued on page 149.




University of Waterloo
Waterloo, Ontario, Canada

MANY PEOPLE IGNORE their bodies until they
start to pain. Similarly, many engineers ignore
report writing skills until their bosses complain.
Normally, such feedback produces a strong re-
sponse. Indeed, among recent alumni, I have come
across many born-again report writers. Most of them
favour greater emphasis on report writing in the un-
dergraduate curriculum. Listening to them, I marvel
that they look so much like the people who only a
short time ago condemned the very thing they now
praise. St. Paul would certainly understand.
And so we address that perennial problem: how to
make instruction in report writing efficient and palat-
able for both students and instructors. In this article,
I would like to offer a few tips that I have found help-
ful. Among them will not be found the suggestion to
simply write more reports. Instead, the focus will be
on a more considered approach to writing and evaluat-
ing the reports already assigned.

Whenever I open a style manual, I feel smothered
by detail. As for those earnest, homemade style
guides that professors sometimes grind out . no
comment; by reflex, students cheerfully ignore the lot.
As an alternative, I have opted for brevity. Table 1 is
a one-page cribsheet on which I have organized the
key points of style in an attention-getting but boiled
down form [1]. Evidence suggests it does get read.
Those who want more detail are referred to a recent
research journal. This confers the added benefit of
introducing the student to some current literature.

Whenever I open a style manual,
I feel smothered by detail. As for those
earnest, homemade style guides that professors
sometimes grind out ... no comment; by reflex,
students cheerfully ignore the lot. As an
alternative, I have opted for brevity.

Bob Hudgins is a professor of chemical engineering at University
of Waterloo, Canada, and holds degrees from University of Toronto
and Princeton University. He teaches reaction engineering, staged oper-
ations, and laboratories that go with them. His research interests lie in
periodic operation of catalytic reactors and in the improvement of grav-
ity clarifiers.

One trick that amuses as it teaches is the "Fog
Index," a term coined by Robert Gunning [2]. The
Fog Index for a particular sample of prose is given by
the relation
F = 0.4 (W/S + 100P/W)
where F = Fog Index, a number representing ap-
proximately the years of schooling needed
for a reader to readily understand the
written passage
W = number of words in a passage composed of
several consecutive sentences. W must be
at least 100 words.
S = number of sentences in the sample.
P = number of polysyllabic words (three syl-
lables or more) in the word sample. Capi-
talized words, simple fusions such as
"manpower," and verbs containing three
syllables by virtue of a suffix "es" or "ed,"
must be omitted from P.
To survive in the marketplace, popular magazines

Copyright ChE Division ASEE 1987


have evolved Fog Indices between 8 and 12. Further-
more, according to Gunning, passages in which F >
17 cannot be fathomed by a general reader.
Clearly, the Fog Index can nudge a writer into
using short sentences and small words. However, it
does not encourage the higher elements of style such
as clarity, imagery, or diction. Even so, chopping the
ratios W/S and P/W may represent an improvement
for many.
Someone calculating an F for the first time may be
shocked to find it well above the 8 to 12 range. A little
experience will show it is easier to keep F small in
descriptive passages (e.g., experimental details) than
in more convoluted discussions. On the other hand,
students often observe that an unavoidable repetition

of technical terms quickly inflates the P/W ratio. In
fairness to them, Gunning's suggested range (8 < F
< 12) may be a few points too low for technical writ-
ing. (Incidentally, F is about 10 for this article.)


Marking reports is a highly individualistic task.
Even so, a legitimate concern of the student author is
how consistently the reports have been graded. For
example, did the marker use a set of objective stand-
ards for judging each report? In an effort to achieve
consistency in grading, I have worked up a checklist
of the type shown in Table 2. Its use assures that the
same minimum number of questions is asked of each

Style Cribsheet for Formal Reports

S R. R. "U -'T i t4S \99g
- i>tlp \n opi. cen+rad y-ti'tut'ho cLu.'or~, Co-aWLKr
group nuL'm6er, S3ignat're, ID Wo., aLoafe pLeaf'ir! Sfeei6y
"An r&>fc Y r~i -s p axfL(&,ld of0 Al TIPS
-tb S+nl ..,, b M n S o.-,,, / k Vl |s L-

ked e. ,, r -LCOms + s,-. r& "+ , iu t
P',yi,4,'Vd (Va as -1m _,am r Itr.
cLA- pr-L m ni aam rA5Si
C--- -ro r-r41-c..4C s. .~A**f~4^ _
_aJcs' .vt r,-& CLEA EN SLISH
4 .s -e $ S +; +-- ioS .'t."
*'\... 0o fpr cm( of +Kt
es +-o loo w-0 qd-+; As sb*omid *b
- pj kPaoihge ) PO-y c~eIeUW- So un'teULLji'rd ltC\o

- euSai, aJs 54lyf+_j rB-revcfa
- se App").5 xic r jCS5 ^re- 5


re s-firtpa c8flai-U1 ,t Y mr -to (ea-- | A
fm Seb2AL 'CM*, rseur 4 5 vv. io-, w.

npciLw. y steo> or Lbn.
-e J_; 4-0144 .a o ruW
9tcJar- (lec wblb(A- kb toPr Lafx.
St wsd9s pM-F4red3 4.A oJr;ipcanCS.
-$Ulloul $SLLsL6N Cfa FryLr^S *4 t TAAo!l5
p.4ec- raw cL S, hf tidub d a(4 ^^y've)
.^* .

DIscussiw4 oF RESULTs
-~ .r rTs~ ,S c -Zot',A l l? rdzsuA A. . i 7.
- cl ,.lvu rouAi 4u H 04 ,'U ,.O (Cp~- ~ ~-" rplP6.a .

,g mmMes c^- s ( _T :.
o- r $ a c^ ='^ c .S JL ;m ses.i),
-r_ k l -,-d-- "--, ,^+ -.,
L 60-5 L-f 9", its PbsIC'S !
,C$~. O r~-.eMN ofiGCoPs-,'U (;*1cA( Cu,~.e~r
_r61 w 6 'pr; l rie)
N OTATLrOh ecre, .er .
- C4LU Sijc* l A \Nf6Ee> OUT 3ARQON

"t cR1- I-I "-
C:. -dpQ i, ia.a-,-

,- ,- es A. r o- L rt .-k- -.

fz. '5o _f c a.-)_- 7r)S-
~~t.Fjo URES TABLES 6--


report and that none is overlooked. It also serves as
a quick, written record of the report's strong and
weak points.
Once a checklist has been completed for a given
report, how is a grade to be assigned? My preference
is to give separate grades for form and for content.
This assures that both of these vital aspects of good

Checklist for Report Writing

AUTHOR .........................................................................
EXPERIMENT ........................................ ...................
GENERAL APPEARANCE .................................................
(Y N) Describes essential results, conclusions, recomenda-
(Y N) Informative rather than descriptive?
INTRODUCTION: I suggest developing this section from the
general to the particular.
(Y N) Objective stated rather than just implied? Objectives
are guides to content.
(Y N) Essential assumptions of model presented?
(Y N) Relevant equations identified? (No need to develop.)
(Y N) References cited?
(Y N) Complete criterion given for determining the onset of
steady state?
(Y N n/a) Justification needed that data were taken at steady
(Y N) Adequate discussion of relative importance of errors?
(Y N n/a) Mass/Energy balances presented?
(N Y) Needs reorganizing into smaller segments using sub-
(N Y) Tortured English / logic?
Comment on style (optional)
imperatives should be avoided.)
(Y N) Are they appropriate, i.e., do they complement the ob-
(N Y) Is new material introduced? (Should only summarize.)
(Y N) Reference cited at least once in the text?
(Y N) Referencing given in proper form?
APPENDICES: Appendix should be used to hold ...
(Y N) Spot-checked sample calculation OK?
TABLES & FIGURES (Schematic rather than pictorial draw-
ings of apparatus are preferred.)
( ) Tables . that duplicate data in Figures should be
placed in Appendix.
(Y N) Appearance OK?
(Y N) Informative independently of the text?
(N Y) Information obscured in binding?
(Y N n/a) Theoretical lines properly referenced?
INACCURACIES of statement or analysis?

report writing are evaluated. The over-all grade is
simply the mean.


Finally, if the teaching of report writing is to be
taken seriously, one-on-one discussions between the
student and the instructor should be encouraged.
Marking a report in the style of a final examination
paper produces a grade, but no feedback, so an inter-
view with the student is useful for reviewing the re-
port's content and structure. The checklist can serve
as a helpful focus for this purpose. In my experience,
most students welcome such meetings in proportion
to their need to improve.


1. ChE 410 Undergraduate Laboratory Manual, Dept. Chemical
Engineering, University of Waterloo.
2. R. Gunning, How to Take the Fog Out of Writing, Dartnell
Corp., Chicago, Ill., 1964. (My thanks to Dr. B.M.E. van der
Hoff for this reference). O

book reviews

by Welty, Wicks and Wilson
Wiley & Sons, Somerset, NJ 08873 (1984) $36.95
Reviewed by
Hugo S. Caram
Lehigh University

This is the third edition of a popular undergraduate
textbook and comes some twenty-three years after
the publication of the book that started it all; Bird,
Stewart and Lighfoot's Transport Phenomena. A new
edition of that book is expected to reflect what we have
learned since then about transport phenomena and the
teaching of it.
As done in BSL, the three transfers are treated in
series with the option of covering them in parallel for
the integrated approach where similarities between
the processes are emphasized. This seems to be the
current trend with the exception of the unique book
of R. Fahien where the three processes are initially
treated simultaneously. Where is the book different
from the original BSL?
First, this is a finite undergraduate textbook and
this means that one can essentially cover all of the
book in two semesters without leaving out a para-
graph. This is in opposition to "open" survey books
where all the material in the book cannot possibly be
covered in any reasonable amount of time. Our stand-


ard BSL is closer to an open text of which a leading
example would be Batchelor's Introduction to Fluid
Mechanics. This closed-end approach also, unfortu-
nately, means that a number of things are left out.
They include, for example, non-Newtonian fluids,
multicomponent diffusion and simultaneous heat and
mass transfer when discussing the traditional material
and diffusion through membranes and in ionic solu-
tions, turbulent diffusion and mixing in jets and
plumes and multiphase flows when thinking of newer
material of technical relevance.
Second, the authors begin with macroscopic bal-
ances that are the natural extension of the first chem-
ical engineering course on material and energy bal-
ances and are the everyday tool used by the engineer.
They are then used to derive the microscopic equa-
tions (although I find it unfortunate that after the con-
ceptual effort involved in the derivation of the macros-
copic equations, the authors were not willing to intro-
duce the divergence theorem). Traditional books, like
BSL, will derive them independently and lack some
internal unity. The book also treats some of the ap-
proximations like boundary layer theory and ideal
flow with great clarity while, again, leaving out com-
pletely the low Reynolds number hydrodynamics as
applied, for example, to flow about a sphere. It should
also be pointed out that in recent years a well-de-
veloped body of theory has appeared that allows the
engineer to recognize the order of magnitude of the
terms in an equation and make the approximations
described above in a more or less scientific way. These
techniques are, however, barely mentioned in associa-
tion with the boundary layer discussion.
Finally, the book covers a number of applications
to the design of equipment of industrial interest like
flow in pipes, heat exchangers,and packed absorption
columns. Comparison with a common book in unit op-
erations would show, with the exception of the discus-
sion of staged operations and distillation, a wide over-
lap with those texts. The missing parts correspond
mostly to equipment description but not to fundamen-
tal concepts since the book provides enough tools to
solve a large fraction of the problems found in unit
operations textbooks.
End of the chapter problems are one of the strong
points of the book. There are many short, numerically
simple, attractive exercises that, while lacking over-
whelming industrial flavor, will be of great help in the
teaching of the subject. Missing, however, are prob-
lems discussing non-traditional chemical engineering
applications of the methods to biochemical-biomedical,
product engineering, or environmental situations.
They would be desirable to broaden the outlook of the
chemical engineering student. In summary, this is a

very good junior-level textbook that adds to the teach-
ing of the subject as is traditionally known, but does
not bring in any of either the new problems that are
starting to fascinate chemical engineers or new tech-
niques developed to deal with the old problems. It
must, in that area, be supplemented from other
sources to cover more advanced topics or to find the
description of specific industrial equipment. O

by L. Ehrlich and David S. Holmes
John Wiley and Sons
Reviewed by
George T. Tsao
Purdue University
Biotechnology is an old field which has taken on a
great deal of new excitement since the 1970's due to
the advances made in molecular genetics. There are
those who consider biotechnology involves nothing
else but genetic engineering and production of pro-
teins for pharmaceutical uses. There are also those
who prefer a broad definition of biotechnology to mean
technology based upon biological activities of one type
or another. Biotechnology and Bioengineering Sym-
posium No. 16 is the proceedings volume of the work-
shop on "Biotechnology for the Mining, Metal-Refin-
ing and Fossil Fuel Processing Industries," held in
May 1985 on the campus of Rensselaer Polytechnic
Institute in Troy, New York. The workshop brought
together many top experts in this field from different
parts of the world to review biotechnological applica-
tion in the metal-mining industry, the current state of
the technology, the industry's and the government's
view on the subject and the latest advances in molecu-
lar biology and the application of genetics and genetic
engineering in the metal-mining industry. The volume
should be a useful reference to those who have been
working in this field; it should also serve as an infor-
mative introductory volume for technical managers,
policy makers, life scientists, process engineers, and
others who wish to quickly become somewhat know-
ledgeable on the subject.
While the volume may be an excellent review of
biotechnology in the metal-mining industry, it does
not address specifically what work may be important
for the future advancement of this subject. This re-
viewer believes that a logical follow-up event could be a
workshop on the identification of general and specific
research needs in biotechnology for the metal-mining
industry. O


Sp program



Georgia Institute of Technology
Atlanta, GA 30332-0100

M ILLIKEN AND COMPANY is one of the leading
textile manufacturers in the country. The com-
pany, which produces a wide range of fabric materials,
is headquartered in Spartanburg, South Carolina, and
has approximately sixty manufacturing locations in
the Carolinas and Georgia. Its Interior Furnishings
Division, for example, produces carpeting, drapery
and upholstery materials. Apparel products include
athletic uniform and tennis shoe materials in addition
to conventional clothing fabrics (polyester, nylon,
polycotton). It also manufactures a wide variety of
automotive upholstery materials. Specialty industrial
fabrics include materials for tires and conveyor belts.
The company also has a small chemicals manufactur-
ing facility located in Inman, South Carolina (near
Spartanburg); products of this facility include spe-
cialty chemicals for a variety of applications in addi-
tion to textile chemicals for both internal consumption
and external marketing.
With its extensive manufacturing capabilities and
its exceptionally strong emphasis on research and de-
velopment, Milliken is a large employer of engineers
and scientists. The company has been responsible for
a number of revolutionary developments within the
domestic textile industry. Chief among these is its
computer-controlled system for continuous dyeing (of
carpeting and upholstery), developed in the 1970's and
known as the Millitron system, which completely
changed this industry. The company also developed

Georgia Tech was instrumental in helping
to set up the company's cooperative work-study program
in the 1970's. Shortly thereafter, the School of
Chemical Engineering at Georgia Tech inaugurated
its graduate residency program.

0 Copyright ChE Division ASEE 1987

Pradeep K. Agmwal, currently associate professor of chemical en-
gineering, has been a Georgia Tech faculty member since 1979. He
received his BChE degree from the University of Roorkee (India) in
1975 and his MS and PhD degrees in chemical engineering from the
University of Delaware in 1977 and 1979. His research interests in-
clude heterogeneous catalysis, reaction engineering, and modeling of
chemical vapor deposition processes. (L)
Jude T. Sommerfeld is professor and associate director of the School
of Chemical Engineering at Georgia Tech. He received his BChE degree
from the University of Detroit and his MSE and PhD degrees from the
University of Michigan. His 25 years of industrial and academic experi-
ence have been primarily in the area of computer-aided design, and
he has published over 70 articles in this and other areas. (R)

the well-known VISA fabric treating process to en-
hance washability, and the CAPTURE dry-process
carpet cleaning technology.
Georgia Tech has a long history of synergistic
cooperation with Milliken and Company. Georgia Tech
was instrumental in helping to set up the company's
cooperative work-study program in the 1970's.
Shortly thereafter, the School of Chemical Engineer-
ing at Georgia Tech inaugurated its graduate resi-
dency program (a form of cooperative program at the
master's level), with Milliken as the prototype par-
ticipating company and which has since been extended
to a number of other companies (including Tennessee-
Eastman, 3M, Phillips Petroleum, Philip-Morris, and
the Institute of Paper Chemistry). The president of
Milliken and Company-Thomas J. Malone, a Georgia
Tech chemical engineering graduate-served as the


This program is specifically oriented to rising senior engineering students, i.e., students
who have completed their junior year. Occasional exceptions have been made for outstanding underclassmen
or recent BS graduates bound for graduate school . Applicants are required to be U.S. citizens, in the
top 10% of their class or with a grade-point average of 3.5 and with demonstrated interpersonal skills.

first chairman of the School's Industrial Advisory
Board. A number of the institute's faculty members
have also been engaged in summer employment and/or
served as consultants to the company in the past fif-
teen years.

As indicated above, Milliken and Company has had
an extensive cooperative education program for some
time, as well as a summer employment program for
both students and faculty. The unique feature of the
program described in this article, however, relates to
direct faculty involvement in the summer projects of
undergraduate student interns. The idea for this spe-
cific program originated in discussions between
Malone and Gary W. Poehlein-then Director of Geor-
gia Tech's School of Chemical Engineering-in late
1984. Known as the Rising Senior Program, it is de-
signed to provide student interns with an opportunity
to apply engineering principles to real problems re-
lated to the broad scope of Milliken business interests.
Students in this program benefit from the guidance of
both Milliken professionals and Georgia Tech en-
gineering faculty.

This summer work program is specifically oriented
to rising senior engineering students, i.e., students
who have completed their junior year. Occasional ex-
ceptions have been made for outstanding under-
classmen or recent BS graduates bound for graduate
school. Student applicants to this program are re-
quired to be U.S. citizens, in the top 10% of their class
or with a grade-point average of 3.5 (out of 4.0), and
with demonstrated interpersonal skills.
The chemical engineering component of this pro-
gram operates with the equivalent of one full-time fac-
ulty member supported by the company for the sum-
mer. During the first two years (1985 and 1986) of
operation of this program, the actual Georgia Tech
chemical engineering faculty contribution has been
two faculty members-the authors of this present ar-
ticle-each on a 50% time basis. Included among the
students they have advised have been a few textile
engineering students. A program of comparable mag-

nitude also operates in the mechanical engineering
sector, again with a similar contribution from Georgia
Tech faculty members. This latter program has also
included a small number of industrial engineering stu-
dents. Milliken supports summer programs of smaller
scope in the electrical engineering (including computer
engineering) and textile technology areas. Students
who participate in this program are recruited from
the major engineering schools in the Southeast, plus
a few other targeted schools. Thus, in comparison
with conventional cooperative and summer work pro-
grams, the only additional costs for this Rising Senior
Summer Program are faculty salary, travel expenses,
and associated overhead. These additional costs con-
stitute a small fraction (approximately 20%) of the
total program cost.

As presently implemented, planning for the sum-
mer program begins in the preceding fall quarter or
semester. At this time (around Thanksgiving),
brochures and letters are sent to the department
heads at the selected schools, asking them to invite
qualified students to apply. Depending upon recruiter
schedules, some interested students are interviewed
on campus by Milliken representatives while others
are invited directly to company facilities for inter-
views. All prospective summer employees are eventu-
ally interviewed at their proposed work location. The
candidate selection process is finalized near the end of
the winter quarter, at which time offers of summer
positions are made. The summer work terms are typ-
ically of ten to twelve weeks duration. Students from
semester schools usually begin their work in May and
finish in August, while those from quarter schools
start in June and conclude in September.
Concurrently with the candidate selection process,
suggestions for appropriate summer projects are
being solicited and received from the various company
manufacturing and technical locations. These projects
are also reviewed for suitability by the contributing
Georgia Tech faculty members. Summer interns are
matched with projects at a specific location and are
assigned to a Milliken professional who acts as his/her
supervisor (or sponsor). This entire activity is coordi-
nated by Milliken's Director of College Relations.


ChE Students in
Milliken's Rising Senior Summer Program
(Q = quarter; S = semester)
of Year
School Term 1985 1986
Auburn Q 3 2
Clemson S 2 1
Florida S 2 3
Georgia Tech Q 1 1
Mass. Inst. Tech. S 0 2
N.C. State S 2 1
South Carolina S 1 0
Tennessee Q 2 2
Yale S 1 1
TOTALS 14 13

The schools and number of students participating
in the chemical engineering component of this pro-
gram during the summers of 1985 and 1986 are listed
in Table 1. The total number of student interns (all
disciplines) engaged in the summer 1986 program was
approximately forty. This was in addition to a half
dozen or so faculty members (not advising students
per se) from different schools, including Georgia Tech,
who participated as faculty summer interns on various
independent projects.
In operation, a faculty advisor meets with each
student intern on the average of once a week. Given
the widespread locations of Milliken's various opera-
tions, a considerable amount of travel by the con-
tributing faculty members is required. This travel
time is considered in the determination of a faculty
member's contribution to the program.
Some of the chemical engineering projects from
this summer program are summarized in Table 2.
Chemical engineering students are typically placed in
textile finishing plants, wherein the operations of
washing, dyeing, drying, etc. are more akin to chemi-
cal engineering unit operations. Mechanical engineer-
ing students, on the other hand, are usually engaged
in projects in the upstream areas of cloth knitting and
Milliken Research Corporation (also located in
Spartanburg), in addition to its research and develop-
ment functions, is also strongly oriented to customer
technical service and manufacturing support ac-
tivities. Thus, many of the summer interns have direct
(sometimes hands-on) experience with advanced
analytical methods, e.g., differential scanning

calorimetry (DSC), infrared (IR) analysis, and sophis-
ticated colorimetric techniques, not to mention a
myriad of computer applications.
Milliken, like the automotive and other industries,
is a state-of-the-art practitioner of statistical quality
control (SQC) procedures; thus, most of the summer
interns experience, during their term, considerable
exposure to statistics-an important topic virtually
absent in most undergraduate chemical engineering
curricula. These various work experiences are clearly
quite valuable to engineering students returning to
their senior year who are about to embark on profes-
sional careers.
In their weekly meetings, of typical duration of an
hour or so, the student intern and faculty advisor (and
sometimes the student's sponsor) review progress on
the project and discuss future plans, experiments, any
problems, etc. The faculty advisor is quite available to
all of the student interns, and may be contacted at his
Georgia Tech office in Atlanta at any time by the stu-
dents. In addition to providing professional guidance
and technical support, the faculty advisor functions as
a resource person. The student interns are often di-
rected to certain textbooks and articles relevant to
their projectss. Copies of the latter are often directly
supplied to the student interns from the faculty
member's personal files or from the Georgia Tech li-
brary although, depending upon the timing and the
student's specific location, his/her needs may be
served more readily by Milliken's excellent technical
library in Spartanburg. In some cases, the faculty ad-
visor has served as a communications link between
two or more different plants wherein students were
working with a newly introduced technique. The fac-
ulty advisor is also in a unique position to suggest
advanced technical solutions which the student intern
(and company sponsor) may be unaware of. Specific
examples from our two years of experience have in-
cluded applications of linear programming, cascade
control, queuing theory, and sophisticated numerical
integration techniques.
Certainly, a highlight of the past two years has
been the two-day sharing meeting. This meeting, at
company headquarters in Spartanburg, is held in
July-the only month in which the summer interns
from both semester and quarter schools are all em-
ployed. The student interns, faculty advisors and fac-
ulty interns are all assembled for this meeting. On the
first day, they are given a general overview of the
company's business and operations by representatives
of senior corporate management, followed by tours of
the company's central facilities (research center, cus-


tomer center, etc.). On that evening, social events in-
clude a picnic and a volleyball tournament (dominated
to date, by the way, by the chemical engineering
team). On the second day, the student and faculty
interns all make brief (five-minute) presentations on
their summer projects. Again, senior corporate mana-
gers attend these presentations. The planning and
scheduling for this event is quite complex, but all par-
ticipants have felt it to be a very valuable experience.
Thus, the unique aspects of this Rising Senior

Sample of ChE Projects in Milliken's
Rising Senior Summer Program

* Energy conservation in batch textile dyeing operations
* Design of a heat transfer system for a batch chemical reactor
* Interfacing of laboratory colorimeters with personal com-
Determination of kinetic parameters in batch textile dyeing
Extraction of oils from knitted or woven greige fabrics
Computer control of a batch chemical reactor
Development of chemical process for manufacture of a de-
foaming agent
Determination of the time-temperature histories of finished
Viscosity studies of various gum blends as admixtures with
Analysis of oxidation/reduction processes in the post-dyeing
Effect of oxygen as an impurity in steam on the dye-fixing
Study of process variables in the wick-proofing (air dif-
fusion resistance) of chafer fabrics

Summer Program, in comparison with conventional
Summer or cooperative work programs, can be sum-
marized as follows:

The weekly meetings between the student and faculty
member can be best described as brainstorming sessions.
The thought process developed during these sessions can
often be quite different from that employed by someone
who has worked closely in the area for a long time. It thus
permits a student to critically examine the various ap-
proaches available to define and solve a problem.
This interaction provides an opportunity and encourage-
ment for the student to relate the project to science and
engineering fundamentals. Many young engineers be-
come disillusioned with the academic learning process
when they first encounter an industrial problem. The
presence of a faculty member facilitates a smoother tran-
sition from an academic to an industrial environment.


Startup problems associated with the Rising
Senior Summer Program have been minimal. Cer-
tainly, not all of the various projects have yielded
quantifiable results. On the other hand, savings of
several hundred thousand dollars per year can be
clearly identified with a good number of these pro-
jects. In any event, the informational value resulting
from most of these projects has been substantial.
Company and student response to this program
has been enthusiastic. Critiques are solicited from
both the company sponsors and student summer in-
terns at the end of their terms. Following are some
selected comments from these critiques:

It is good to be able to discuss your project with someone
knowledgeable from outside the plant [the faculty advisor]
in order to get different viewpoints.
I applaud Milliken and Company for designing and or-
ganizing both the summer intern and summer professor
I feel the strengths of the intern program are the challeng-
ing projects that the interns face and the help of the pro-
fessor advisors. ... The two days at the Milliken Research
Center were the highlights of the internship.
My preferential pickup project required me to learn about
heterogeneous surfaces and mass transport.
I wish other companies would follow your example so
more students could benefit from this type of experience.
. Having someone from outside the plant to explain my
project to was very helpful in organizing my thoughts.
I have friends [at my school] that would love to have a
chance to do the kind of work I did this summer. ... It
was extremely helpful to have a professor from Georgia
Tech come to visit and give technical advice.
I feel that the summer professors added to the program,
supplying additional technical insight.
I have worked internships with two other large companies
and I feel the Milliken program is a step above the rest.

Cooperative and summer work programs have
long been recognized by companies as important ad-
juncts to their overall corporate recruiting programs.
Thus, for example, at Georgia Tech approximately
55% of our co-op program graduates have traditionally
accepted their initial employment after graduation
with their co-op employer. In 1986, Milliken and Com-
pany hired about half of the chemical engineering stu-
dent participants in the 1985 Summer program after
their graduation (recall the qualifications for participa-
tion in this program). Clearly, this Rising Senior Sum-
mer Program has redounded favorably to both the
company and the student participants. O





Manhattan College
Riverdale, NY 10471

M EMBRANE PROCESSES are one of the new tech-
nologies being introduced in today's engineering
curriculum. The membrane process reverse osmosis
(RO) is being utilized by a broad spectrum of indus-
tries for a variety of uses. Applications are found in
the agrichemical, biochemical, chemical, electrochem-
ical, food and beverage, metal finishing, petrochemi-
cal, pharmaceutical, pulp and paper, and textile indus-
tries [1]. Reverse osmosis membrane processes are
competing with the more traditional separation tech-
niques such as distillation, evaporation, and filtration.
Reverse osmosis is considered a mass transfer unit
operation and as such, theory relating to membrane
transport should be presented in a mass transfer
oriented course. System design and operation can be
discussed briefly in a process or plant design course.
Most engineering curricula present membrane
technology in graduate courses, but some knowledge
of the theory and operation should also be presented

C. Stewart Slater is an assistant
professor of chemical engineering
at Manhattan College. He has held
the position since 1983 and had
prior industrial experience with
Procter & Gamble Co. and Arthur
W. Ponzio Co. He received his PhD,
MPh, MS and BS degrees in chem-
ical engineering from Rutgers Uni-
versity. His research and teaching
interests are in separation and re-
covery technology, membrane pro-
cesses and biotechnology. (R)
John D. Paccione is currently a
graduate student in chemical en-
gineering at Rensselaer Polytechnic Institute. His involvement in this
project was part of a Senior Honors Project at Manhattan College. He
was active in the department's laboratory development activities for
two years. He has industrial experience as a research technician with
the Union Carbide Corporation. (Not pictured)

*This paper is based on a paper previously published in the ASEE
1986 Annual Conference Proceedings.

to undergraduates. The demonstration of reverse os-
mosis should occur in a senior level chemical engineer-
ing or "unit operations" laboratory. In such an experi-
mental setting the student can understand the theory,
operation and design of these systems and see the
applications to industry. This could be in the form of
a half-day experiment or a full semester project.
This paper focuses on the development of a small
pilot unit for use in an advanced separations process
laboratory. The end goal is to develop experiments
with advanced separation processes such as reverse
osmosis, ultrafiltration, adsorption, chromatography,
etc. This paper presents one step in that direction.

In simplest terms, RO uses a thin semipermeable
membrane that allows the transport of certain species
while retaining others. A feed stream is introduced,
and the membrane separates it into a purer stream,
the permeate, and a more concentrated stream, the
retentate or concentrate. The permeate is the stream
that permeates the membrane barrier. Mass transfer
through an RO membrane can occur by several mech-
anisms for which many models have been proposed [2,
3]. The solution-diffusion model describing water and
solute transport through the membrane is utilized
here [4]. In this model each species in solution dis-
solves and diffuses through the membrane at a rate
corresponding to the applied transmembrane pres-
sure, AP, and the concentration gradient ACs, across
the membrane.
The permeate flux, J,, which in most applications
is water, is directly related to the transmembrane or
hydraulic pressure driving force, AP, minus the differ-
ence in osmotic pressure, Aur, on both sides of the

J = A (AP Ad)
w w

The flux and pressure gradients are related by the
water permeability coefficient, Aw.
The solute flux, Js, is related to the concentration
Copyright ChE Division ASEE 1987


gradient on both sides of the membrane, ACs, by a
solute permeability coefficient, Bs

J, = (Bs)(AC) (2)

Recovery is the ratio of permeate production, Qp,
to feed rate, Qf.

Recovery, Y = (3)

Solute rejection, R, can be measured in several
ways. It can be denoted as a relationship between
feed and permeate solute concentrations
R = 1 -- (4)

Other ways to express rejection compare permeate to
retentate concentrations or to an average of the feed
and retentate concentrations. The recovery can also
be incorporated into the expression [5]. It can be
demonstrated from the solute and permeate flux equa-
tions that rejection is a function of pressure and con-
centration gradients. Water flux is dependent on pres-
sure; therefore, an increase in pressure will increase
water flux at constant solute flux, i.e., decrease per-
meate solute concentration and increase solute rejec-

Since the membrane is the critical element in the
RO system, proper understanding of structure and
configuration is necessary before system design can
commence. Reverse osmosis membranes are charac-
terized by a high degree of semipermeability, high
water flux, mechanical strength, chemical stability
and economically acceptable cost. The early RO mem-
branes were made out of cellulose acetate, but restric-
tions on process stream pH and temperature, includ-
ing low rejection of some organic, spurred the de-
velopment of non-cellulosic and composite materials.
Polysulfones, polyamides, among others, and compos-
ite structures are popular alternatives because they
do not have the draw-back of cellulose acetate [6]. The
conventional composite membranes, normally called
thin film composite membranes, do have a drawback
in their ability to tolerate chlorine.
The membranes are configured into certain
geometries for system operation. The four basic con-
figurations are plate and frame, tubular, spiral wound,
and hollow fiber. Since the studies employed in this
paper use the spiral wound module, emphasis will be
given to its characteristics with the reader referred to
Leeper et al [7] for more design details. The spiral

This paper focuses on the development of a small
pilot unit for use in an advanced separations process
laboratory. The end goal is to develop experiments with
advanced separation processes such as reverse osmosis,
ultrafiltration, adsorption, chromatography, etc.

FIGURE 1. Spiral wound membrane configuration (from
McNulty, K. J. and P. R. Hoover, EPA-600/2-80-084, U.S.
Environmental Protection Agency, Cincinnati, OH, 1980,
page 5).

wound configuration consists of two membrane sheets
forming an envelope for a permeate channel with feed
channels on the outside of the envelope wrapped in a
"jelly-roll" pattern around a perforated tube (Figure
1). The feed flows axially through the unit, exiting as
retentate. The permeate enters the membrane en-
velope and travels spirally to the perforated center
tube where it exits the membrane at the retentate
end. The plate and frame and tubular membrane mod-
ules employ far more simple geometries, as their
names imply. The hollow fiber membrane configura-
tion utilizes thousands of "hair-like" hollow fibers in a
fiber bundle.
The aforementioned membrane modules can be ar-
ranged in series and parallel to design a commercial
scale unit. In most large-scale commercial systems,
this is needed to accommodate the higher feed rates
and produce a high recovery. The functioning of one
module is the basis for commercial system design and
scale-up. Therefore, experiments on one membrane
element can be utilized by a student to develop a sys-
tem for a process or plant design course.
A simple laboratory RO system was designed to
be as versatile as possible. It can be operated in such
a way that different process parameters can be
evaluated. A single (or small multiple) membrane sys-
tem was chosen. It has the capability to independently
vary the following process parameters: membrane
feed rate, operating pressure and temperature. It is
designed so that modifications can be easily done and
membrane replacement can be quickly accomplished.
Use of other membrane configurations is also a design
consideration. Since spiral wound RO membrane sys-


teams are the most prevalent in the industry the sys-
tem was designed on this basis. The system can be
easily retrofitted for hollow fiber and tubular config-
System design rational and a detailed discussion of
the system's components are presented in Slater and
Paccione [8]. A list of manufacturers of spiral wound
membrane modules can be acquired by writing the
author. Table 1 and the accompanying Figure 2 pres-
ent the components and layout of the system. Some
key features of the system are

Stainless steel construction to allow for durability and for
the processing of a broad range of fluids
Sizing to accommodate all 2.5 inch diameter spiral wound
membranes and most of the 4 inch diameter modules
Independent setting of feed rate (0-10 gpm) and pressure
(0-1000 psi) to obtain desired recoveries
Construction with tubing and standard compression fit-
tings to allow for easy alterations
Layout which permits easy control of system and direct
observation of flow patterns

System cost is always a consideration in a labora-
tory development project. It is encouraging to note
that many membrane vendors and equipment
suppliers give educational institutions discounts rang-
ing up to 20% and some will even supply membranes.
The system described in this paper, including a vari-
ety of membrane types, costs approximately $12,000.
Obviously this figure does not include the labor in-
volved in design and fabrication. It is important to
note that this cost could be reduced by not using stain-
less steel construction, by using a modest pump and
drive, and by using house water for temperature reg-
ulation. A simple system on the order of $5000 can be
constructed for basic experiments if the future use
and utility of the system is not important. Regardless
of the complexity of the system involved, it is always
cheaper to build your own system, and the involve-
ment of students in the design, fabrication, and start-
up activities broadens the scope of the project.

II -13

10 14

Several modes of operation are possible with the
RO system and are shown in Figure 3. The steady-
state recycle pattern of operation was used in the de-
tailed experiments that follow. The critical process
stream characteristics are flow rate and solute concen-
tration. Permeate flux is obtained by dividing the per-
meate flow rate by the respective membrane surface
area. Typical units are m3/m2d, cm3/cm2s and gal/

Reverse Osmosis System Component Listing

Component Model Distributor

1. Temperature
2. Agitator

3. Feed Tank

4. Prefilter

5. Pulsation
6. Pump

7. Drive

8. Permeate Tank

9. Pressure Relief
10. Feed Rota-
11. Feed Pressure
12a. Membrane
12b. Membrane
13. Retentate
14. Back Pressure
15. Permeate

T-Line Lab Model

Model SSD-55

Big Blue Model

Hydra-Cell D 10S

Reeves Vari-Speed
5 Hp-Size 331

Model CC-55


10A 2227A

Master Series
E 9672B

SW 30-2521

see component 11



Forma Scientific
Marietta, OH

Talboys Engineer-
ing Corp.
Emerson, NJ
Utensco, Inc.
Port Washing-
ton, NY
Cole-Parmer In-
strument Co.,
Chicago, IL
Greer Products,
Los Angels, CA
Wanner Engineer-
ing, Mahwah,
Reliance Electric
Co., Columbus,
Utensco, Inc.
Port Washing-
ton, NY
Nupro Co., Wil-
lingboro, OH
Fischer & Porter
Warminster, PA
Marsh Instrument
Co., Skokie, IL
Advanced Struc-
tures, San
Marcos, CA
FilmTec Corp.
Minneapolis, MN

Tescom Corp.
Elk River, MN
Gilmont Instru-
ments, Great
Neck, NY

FIGURE 2. Reverse osmosis system (numbering refers to
Table 1).

*Used in the specific experiments described in the paper. Other
suppliers and specifications listed in ref. [8]


to a separate retentate
tank or discard
FIGURE 3. System operational schematic. Flow options:
a) single pass, b) recycle steady state, c) recycle un-
steady state.

dayft2, also written as gfd. Solute concentration for a
simple solute system is determined by measuring pro-
cess stream conductivity and correlating the value to
solute concentration.


Several simple experimental studies can be con-
ducted to examine process parameters and membrane
mass transfer characteristics. Data can be analyzed
quickly to determine if the experiment was carried
out correctly. In most of these experiments certain
process parameters are kept constant and others
varied. Process feed rate, concentration, tempera-
ture, pH, recovery, and pressure can be adjusted
within certain operating limits.
The two major characteristics to be observed in a
reverse osmosis experiment are separation efficiency
and permeate production. The effect of process vari-
ables on these, and the verification of membrane mass
transfer models, can be accomplished with simple ex-
periments. A summary of some typical experiments is
presented below. Specific details are omitted so that
instructors can construct the proper experiment for
the appropriate student audience and time frame.

Pressure Study

The effect of pressure on permeate production and
solute rejection can be easily demonstrated. A par-
ticular membrane module should be chosen for this
study-particularly one that will allow for a wide
range of operating pressures. A simple solute, e.g.,
NaC1, can be utilized as the feed at a certain concen-
tration, e.g., 5000 mg/L. Process parameters such as
temperature and pH are maintained at certain values.
The recovery for each run can be held constant by
regulating the feed flow rate. It is important to re-
member to operate within the appropriate feed flow
and/or recovery limits established by the manufac-
turer to lessen the problems associated with concen-
tration polarization. Runs are conducted at various
pressures, at 25 or 50 psi increments, depending on

the amount of data to be collected. Membrane man-
ufacturers give a maximum operating pressure for
each membrane along with a normal or recommended
operating range. This range is usually 10 to 25% less
than the maximum. The system's safety relief valve
can be set at the upper end of the pressure range to
protect the membrane from excessive pressures re-
sulting from operator error or mechanical malfunc-
In the experiment the student first observes the
minimum pressure needed to produce permeate. As
the pressure increases, the permeate flow will in-
crease and the permeate concentration will decrease.
The effect of increased pressure on permeate produc-
tion can be explained by the permeate flux model (Eq.
1). By plotting permeate flux vs pressure the relation-
ship is observed. At each of the pressures a sample of
the permeate is analyzed using the conductivity
meter, and its salt concentration is determined. The
permeate concentration will decrease exponentially

5 40
3 J0
2 20




99.5 m
99.0 Z

I I 98.0
0 200 400 600 800 1000


FIGURE 4. Results of a typical experimental study ex-
amining the effect of pressure on permeate flux and
solute rejection.

with increasing pressure. This phenomena is a result
of solute flux being independent of applied pressure
(Eq. 2). The membrane's solute rejection can be calcu-
lated by one of the expressions given earlier (Eq. 4).
The rejection will increase and appear to approach a
maximum as pressure is increased.
An example of typical experimental output is
shown in Figure 4. This study utilized a FilmTec
SW30-2521 (2.5 in. diameter x 21 in. length) thin film
composite spiral wound membrane. A 10,000 mg/L
solution of NaC1 was used as the feed at a temperature
of 200C and pH of 6. The study was run at a recovery
of 5.0% at pressures ranging from 300 to 900 psi.

Feed Concentration Study

The effect of feed solute concentration on permeate


The two major characteristics to be
observed . are separation efficiency and permeate
production. The effect of process variables on
these, and the verification of membrane mass
transfer models, can be accomplished . .

flux and concentration can be readily observed. In this
experiment the only parameter varied is the solute
concentration. The feed and standard processing con-
ditions utilized in the previous study can be employed
here. A particular operating pressure is selected for
all runs. The recovery can also be maintained.
Runs are conducted at various concentrations from
a low value to a high value e.g., 1000 mg/L to 35,000
mg/L. Increments are chosen depending on time; five
to ten different concentrations are usually sufficient.
At each different concentration the permeate produc-
tion and its concentration are determined. As concen-
tration increases, permeate flux decreases because of
the increased osmotic pressure of the feed (Eq. 1).
The solute flux is dependent on the concentration (Eq.
2) so it will increase as the feed concentration is in-
creased. Therefore an increase in the permeate con-
centration and a decrease in solute rejection with in-
creased feed concentration will be demonstrated.
Experimental output from a series of runs with a
FilmTec SW30-2521 membrane at increasing NaCl
feed concentrations can be seen in Figure 5. The plot
illustrates the decrease in permeate flux and the in-
crease in permeate concentration as the concentration
of the NaCl feed solution is increased. The runs were
performed at an applied pressure of 600 psi at 20C
with a feed rate of 3.0 gpm.
An extension of this and the initial study would be
to examine different membrane models from each
manufacturer. This comparison would show the differ-

50 250
P = 600 psi
40 200 m

o -Iso
_J 0
S20 100
o 0
ic 10 50 3
C0O SW30-2521
0 .n- 0
0 5 10 15 20 25 30 35

FIGURE 5. Results of a typical experimental study ex-
amining the effect of feed concentration on permeate
flux and concentration.

ent membrane characteristics within a particular
membrane product line.

Additional Studies

Many other studies can be performed utilizing the
RO system. Publication of a thorough description of
each study is planned for some later date and will

The effect of temperature on permeate production and sep-
aration efficiency
Determination of mass transfer coefficients
Analysis of concentration polarization and related models
The effects of fouling on permeate production and separa-
tion efficiency
Operational characteristics of other membrane configura-
tions, e.g., tubular and hollow fiber
The effect of different membrane materials on organic sol-
ute rejection
The effect of recovery and feed rate on permeate produc-
tion and separation efficiency
The effect of pH on separation efficiency and membrane


Reverse osmosis theory and system operation can
be demonstrated with a series of experiments on a
small pilot-scale unit. The system design is simple and
yet quite versatile. The system is made with an "open-
end" design and can be easily changed to incorporate
future design modifications and be used with different
types of membranes. The system that was developed
for an advanced separations laboratory was based on
using small spiral wound membranes and can operate
at feed rates to 10 gpm and pressures to 1000 psi.
Feed flow, solute concentration, temperature, pres-
sure, and recovery can all be independently varied.
The system can operate in various flow schemes. Ex-
periments investigate simple operational parameters
and mass transfer characteristics. Permeate produc-
tion and solute rejection are studied. More detailed
experiments involving organic separation, concentra-
tion polarization and fouling and other membrane con-
figurations can be performed.


Partial support for this work was provided by the
National Science Foundation's College Science In-
strumentation Program through grant #CSI-8551851.
The authors would like to thank Paul Carney for his
outstanding technical assistance in system design and
fabrication. The authors would also like to thank
Richard Ide of Desal Desalination Systems and David


McGovern of FilmTec for their assistance with this

A, Water permeability coefficient [L-it]
Bs Solute permeability coefficient [Lt-1]
Cf Feed solute concentration [ML3]
C, Permeate solute concentration [ML3]
Cr Retentate solute concentration [ML4]
C, Solute concentration [ML-]
J, Solute flux [LL-2t-'], [ML-2t-1]
J, Water or permeate flux [L3L-2t-'], [ML-2t-']
AP Applied pressure gradient [ML-'t-2]
Qf Volumetric flow rate of feed [L3t-']
Q, Volumetric flow rate of permeate [L3t-']
Y Recovery, single-pass operation [unitless]
Arr Osmotic pressure gradient [ML-'t-2]

1. Slater, C. S., R. C. Ahlert, and C. G. Uchrin, Desalination,

P book reviews

Edited by L. Liberti and J. R. Millar
Martinus Nijhoff Publishers, Dordrecht,
The Netherlands, 1985. 484 pgs, $65.50
Reviewed by
Friedrich G. Helfferich
Pennsylvania State University
Fundamentals and Applications of Ion Exchange
is a collection of thirty contributions to a NATO Ad-
vanced Study Institute held in 1982 in Maratea, Italy.
A wide range of topics is covered, from a crystal-ball
perspective "Ion Exchange Towards the Twenty-
First Century" by the unforgettable Calvin Calmon,
to highly specialized industrial problems such as re-
duction of regenerant acid use in water desalination
(Hendry), start-up of a RIM-NUT plant for recovery
of nutrients from eutrophic aqueous discharges
(Liberti et al), and copper and nickel recovery from
plating plant effluents (Stortini), and to complex
theoretical studies e.g. of the microscopic basis and
limits of the Nernst-Planck-Poisson system (Buck)
and use of the Stefan-Maxwell flux equations in mul-
ticomponent ion exchange (Graham).
The attractively produced volume constitutes the
third crop harvested from the Maratea NATO Ad-

48, 171 (1983).
2. Sourirajan, S., and T. Matsuura, Reverse Osmosis/UltraFil-
tration Process Principles, National Research Council of
Canada, Ottawa, Canada, 1985.
3. Belfort, G. (ed.), Synthetic Membrane Processes: Fundamen-
tals and Water Applications, Academic Press, New York, NY,
4. Weber, W. J., Jr., "Membrane Processes" in Physiochemical
Processes for Water Quality Control, Wiley, New York, NY,
5. Burns and Roe Industrial Services Corp., Reverse Osmosis
Technical Manual, Office of Water Research Technology, De-
partment of the Interior, Washington, DC, 1979.
6. Turbak, A. F. (ed)., Synthetic Membranes: Vol I., American
Chemical Society, New York, NY, 1981.
7. Leeper, S. A., D. H. Stevenson, P. Y. Chiu, S. J. Priebe, H.
F. Sanchez, and P. M. Wikoff, Membrane Technoloqy and Ap-
plications: An Assessment, EG & G Idaho, Inc., Idaho Falls,
ID, 1984.
8. Slater, C. S., and J. D. Paccione, "Design of a Pilot-Scale Re-
verse Osmosis System for an Advanced Separations Labora-
tory," 1986 Annual Conference Proceedings of the American
Society for Engineering Education, Washington, DC, pp. 812-
820 (1986). [

vanced Study Institute. The thirteen main lectures
were published in a previous volume of the NATO
ASI Series (Mass Transfer and Kinetics of Ion Ex-
change, No. 71, Nijhoff, The Hague, 1983) and a selec-
tion of fifteen other contributions of general interest
appeared in Reactive Polymers (Vol. 2, Nos. 1, 2,
January 1984). The residue collected here is of mostly
high, if uneven, quality. Perhaps the harshest criti-
cism that can be voiced on this score is that the volume
has no cohesion, no common denominator other than
a loose relation to ion exchange. The seemingly unor-
ganized side-by-side of specialized pragmatic, abstract
complex theoretical, and review-style papers is a little
Perhaps more disturbing is that at least five of the
thirty papers of the book have been published previ-
ously [Hogfeldt et al on a method of summarizing
equilibria data, and Meagher et al on Mossbauer and
electron microprobe studies of precipitation in Nafion
membranes, in Reactive Polymers, 2 (1984) 19 and 51,
respectively; Bolto et al on recycling of waste water
constituents, in Effluent Water Treat. J., (1983) 23;
Buck on the Nernst-Planck-Poisson system, in J.
Membrane Sci., 17 (1984) 1; Drummond et al on kine-
tics in Zeolite A in J. Phys. Chem., 87 (1983) 1967].
No references to such prior publications are given in
the book.
The ion exchange expert will welcome this volume
as a reasonably priced collection of specialized infor-
mation. No other reader is likely to be interested, and
the didactic value for course work is nil. D


Siclass and home problems

The object of this column is to enhance our readers' collection of interesting and novel problems in chemical
engineering. Problems of the type that can be used to motivate the student by presenting a particular principle
in class, or in a new light, or that can be assigned as a novel home problem, are requested as well as those that
are more traditional in nature, which elucidate difficult concepts. Please submit them to Professor H. Scott
Fogler, ChE Department, University of Michigan, Ann Arbor, MI 48109.



Syrcuse University
Syracuse, NY 13244


N ORDER TO separate two close-boiling species by
distillation, it is most useful to know their relative
volatility (a) since one can calculate from only this the
approximate number of stages to get any given degree
of separation, e.g. by the methods of McCabe and
Thiele or Fenske's equation for total reflux. On the
other hand, the a may not be readily available and the
data in handbooks are usually limited to only the boil-
ing points of the two species.
You are to derive an equation for the relative vol-
atility of two species knowing only the two normal
boiling points. It may be assumed that:
They boil only a few degrees apart, say no more than 10C
e.g., ortho vs para-xylene.
They are similar chemically and thus conform to Raoult's
They thus do not form an azeotrope.
For simplicity, the ideal gas law may be assumed if
You should know the exact definition of relative
volatility. Also, state any further assumptions you
need to make.

The relative volatility (a) of a pair of substances is
defined as the ratio of their volatilities (vi), i.e.

12 V2

and the volatility, vi is defined as the partial pressure
of i, pi, above a liquid solution divided by its mole
Copyright ChE Division ASEE 1987

Allen J. Barduhn received his MS from the University of Washing-
ton in 1941 and worked for several years at a California oil refinery
before re-entering college at the University of Texas, Austin, where he
received his PhD in 1955. He was a professor of chemical engineering
at Syracuse University from that date until last year when he retired.
He is presently involved in writing a book and articles.

fraction in the liquid, i.e.
v. = (2)
For solutions which follow Raoult's Law, the vol-
atility is thus equal to its vapor pressure.
In this case then, the relative volatility is just the
ratio of vapor pressures, or

12 =-P

Also for this case we may estimate the relative
volatility knowing only the normal boiling points of
the two substances, especially when they boil within
a few degrees of one another, say no more than 10C
apart. The derivation of this relation follows.
Accepting Eq. (3) above with all its special require-
ments, i.e., (a) ideal solution in the liquid, (b) Raoult's
Law followed, (c) n.b.p. not far apart but known, we


P P1 P2
12 1 + (4)
2 2
We may estimate the change in vapor pressure
with temperature from the Clapeyron Equation
dP AH AH (5)
dT' TA-V T(V Vi )
gas liq
and making the usual assumption that Vga >> Vliq
and that the ideal gas law applies to the vapor,
AV V RT (6)
gas P
applying Eq. (6) to Eq. (5) we get
dP AH P (7)
dT = RT
and if we may approximate dP/dT with AP/AT over
the small temperature range involved
dP AP AH P (8)

and further that
AT PAH (9)

Now the AP is the change in vapor pressure with
a change in temperature of AT and may be applied to
either component in Eq. (4) above in which the Pi* -
P2* may be identified with the AP in Eq. (9). Thus
combining Eq. (4) with Eq. (9)
S=+ AP AT AH P AT AH (10)
12 p T RT P 2 T RTR
since P1 and P2 are the same as P from normal boiling
points, or for boiling points known at any other fixed
The ratio AH/T is Trouton's constant which is often
taken as 21 cal/(gmole)(K) at 1 atmosphere. Also since
R = 1.987 cal/gmole(K) we arrive at

= 1 T 21 = 1+ 10.6 AT (11)
T12 = 1 1.987 T 12

In Eq. (11) the temperatures must all be on the
same absolute scale, Kelvin or Rankine.
The constant 10.6 is subject to experimental confir-
mation since Trouton's constant is not a fixed number
but depends on the substances and the temperature
level at which they boil.
The constant appears to lie between 10 and 13 for
many of the close boiling pairs I have checked it on.
For the isomeric xylenes it is 12.
Also it is 17 to 20 for the system (02 + N2) at 1
atm. The temperature range is probably too large (AT

= 13C), or the temperature is too low (77 and 90K)
for Trouton's constant to hold. The o range is 3.7 to
4.2 at 1 atm and at 5 atm the mean a = 2.6.

rij = relative volatility of i with respect to j
pi = partial pressure of i
P* = vapor pressure of i
P = total pressure
R = gas constant
T = absolute temperature R or K
AT = difference in boiling points at a fixed P
vi = volatility of i (Eq. 2)
xi = mole fraction of i in the liquid E

book reviews

by Xue-jun Chen, and T. Nehat Veziroglu
Hemisphere Press, 785 pgs, $175 (1985)
Reviewed by
A. E. Dukler
University of Houston
In 1984, a small conference in Xian China on two
phase flow and heat transfer was attended by about
forty Chinese researchers, ten US researchers, and
four researchers from other countries. The conference
was organized to stimulate interaction between the
Chinese and US community working on this subject,
as has been the purpose of many other such confer-
ences. This volume contains the 49 papers presented
at that meeting.
Of the thirteen papers presented by US partici-
pants, ten were essentially review papers (pressure
drop, burnout, critical flow, heat exchanger design,
etc). These vary greatly in quality and completeness.
In many cases more complete reviews can be found
elsewhere, in some cases by the same authors.
The papers presented by Chinese investigators
were, in the most part, reports on work in progress
which was largely experimental and which were con-
cerned with global characteristics of the flow (pres-
sure gradient, Nusselt number, flow patterns, burnout
conditions, etc.). These papers contribute little to a
knowledge of the basic mechanisms underlying these
processes, although in some cases the new data are of
The individual papers appear to be unedited and
the book consists of direct reproduction of each paper
as prepared by the author. Thus the type and graphics
differ in each paper. There is a very complete index. O





Villanova University
Villanova, PA 19085

ONE OF THE major challenges for faculty involved
in a first chemical engineering lab course which
emphasizes technical report writing is the develop-
ment of appropriate experiments. Because the stu-
dents have limited chemical engineering experience
at the start of such a course, the first few experiments
must be simple and should be based on chemical en-
gineering courses completed in the sophomore year.
This paper describes a simple thermodynamics ex-
periment currently in use at Villanova University.
The specific heat of a liquid is determined using an
easily and inexpensively constructed, operated, and
maintained apparatus. The students use the results of
the experiments to formulate recommendations on
how the performance of the apparatus can be im-
proved so that it could be used in a wider range of


Vito L. Punzi joined the chemical engineering faculty at Villanova
University in 1980 after previous experience in industry. He received
his BS degree from Polytechnic Institute of Brooklyn in 1972 and his
MS and PhD from Polytechnic Institute of New York in 1974 and 1979,
respectively. His teaching interests are in chemical process calculations,
chemical engineering thermodynamics, and fluid dynamics, and he is
currently performing research on the solute rejection mechanism in
reverse osmosis and on the removal of heavy metals from industrial
waste water via adsorption.

This paper describes a simple thermodynamics
experiment currently in use at Villanova University.
The specific heat of a liquid is determined using
an easily and inexpensively constructed,
operated, and maintained apparatus.

applications than those specifically tested through ex-
Since the experiment produces accurate results
and stimulates the development of "engineering judg-
ment," a satisfying technical report writing experi-
ence usually results.

The principal components used in this experiment
are shown in Figure 1. A wide-mouthed 700 ml Dewar
flask vacuum bottle is fitted with a styrofoam insu-
lated lid which supports a 300 watt immersion heater
and a mercury-in-glass thermometer. The flask is
mounted on a magnetic stirrer to ensure proper mix-
ing and a uniform liquid temperature. The immersion
heater is connected to a variable power transformer
and wattmeter (accurate to +5%) so that the energy
input rate to the liquid can be controlled and deter-
mined. A 500 ml graduated cylinder and hydrometers
are used for the volume and specific gravity measure-
ments needed to determine the mass of the liquid sam-
ples. A digital stopwatch is used along with the ther-
mometer to collect elapsed time and liquid tempera-
ture data.

The unsteady state energy balance for a nonflow
process with negligible kinetic and potential energy
effects can be expressed as

Qnet Wnet = dU/dt (1)

where Qnet is the net rate of heat transfer into the
system from the surroundings, Wnet is the net rate of
work performed by the system on the surroundings,
Copyright ChE Division ASEE 1987


and U is the internal energy of the system, which
varies with time, t. If the system undergoes changes
at constant pressure and if shaft work effects are neg-
ligible, the thermodynamic definitions of work and en-
thalpy (H) can be used to obtain

Qnet = LdH/dt] (2)
For the calorimeter used in the experiment (a
Dewar flask vacuum bottle), the thermodynamic sys-
tem is the liquid contents of the calorimeter, the solid
components of the immersion heater and the inner
wall of the Dewar flask. In this system, Qnet is the
sum of the heat added to the system through the im-
mersion heater (Q), and the heat added to the system
through the inner wall of the Dewar flask due to the
temperature difference between the surrounding en-
vironment and the system. Since the heat added to
the system through the inner wall of the Dewar flask
is negligible when compared to the heat added to the
system through the immersion heater, Eq. 2 simplifies
Q = cp [dT/dt] (3)

where c, is the total heat capacity of the system con-
sidered, and T is the temperature of the system, as-
sumed uniform throughout the system.
The total heat capacity of the system consists of
the heat capacity of the liquid contents, cp, and the
heat capacity of the solid components of the immersion
heater and the inner wall of the Dewar flask which are
lumped into a single parameter, the calorimeter con-
stant Cpc. If the heat capacity of the liquid is expressed
in terms of the specific heat of the liquid (ci') and the
mass of the liquid (ml), then Eq. 3 can be rewritten as
Q = (mlc + cpc)[dT/dt] (4)
Thus, a well-insulated batch calorimeter can be
used to determine the specific heat of a liquid from the
rate of temperature change which results when
energy is added at a known rate to a known liquid

First, the students perform three repetitions of a
calibration experiment to determine the calorimeter
constant Cpc. A constant energy input rate between
100 and 200 watts is used to heat 500 ml of water.
Since the specific heat of water is known and is rela-
tively constant, rate of temperature change data can
be used to determine Cpc using the following rear-
rangement of Eq. 4
Cpc = (Q/[dT/dt]) mlcp (5)

FIGURE 1. Experimental Apparatus

Next, a series of experimental runs is performed
to determine the specific heat of the test solution, Cpl',
at several concentrations. The same procedure is used
as in the calibration experiments, except that a test
solution is used instead of water. Using the value of
Cpc determined above, the value of cpi' can be calcu-
lated from
S' = {((/[dT/dt]) Cpc}/mI (6)

Ethanol-water solutions and glycerol-water solutions
have been used successfully as test solutions. These
solutions have been used primarily because minimal
evaporative losses occur, and because extensive spe-
cific heat data are available in the literature.
The test solutions are heated over a narrow tem-
perature range chosen so that the middle of the range
coincides with a temperature at which literature val-
ues of the specific heat are available. For example,
glycerol-water solutions are usually heated from 5C
to 25C, since literature data are available at 15C (1).
If calibration runs are performed over the same tem-
perature range, the final results of the experiment
are usually quite accurate.
The experiment can be performed either using
standard solutions of known composition or using sol-
ution compositions chosen at random by the students.
The advantage of the former approach is that it allows
a direct comparison between experimental and litera-
ture values.

The heat capacity of the solid components of the
calorimeter (Cpc) is approximately 10% of the total
heat capacity of the system. Thus, neglecting this


parameter in an energy balance would result in a sig-
nificant deterioration in the quality of the final results.
However, the magnitude of this parameter is small
enough so that more accurate estimates would not sig-
nificantly improve the final results.
During the 1984-85 academic year, ethanol-water
solutions were used, with solution compositions cho-
sen at random by the students. All experimentally
determined values of specific heat were within 10% of
an interpolated literature value, at all concentrations.
During the 1985-86 academic year, standard
glycerol-water solutions containing 10 wt% glycerol
and 40 wt% glycerol were used; a limited number of
additional experiments were performed using a 20
wt% glycerol solution. As shown in Table 1, the ex-
perimentally determined values of the specific heat of
the 10 wt% solution were, on the average, within 2%
of the literature value at 15C. All but one of the 38
individual values determined in the experiments using
the 10 wt% solution were within 7.5% of the literature
value at 15C, with 32 of the 38 individual values
within 5% of the literature value at 15C.

Results of Experiments Using
Glycerol-Water Solutions at 15C
Specific Heat, cal/g-C
Composition Range of Average Number
(weight % Individual Experimental Literature of
glycerol) Values Value Value"' Samples
10 0.866-1.002 0.941 0.961 38
20 0.859-0.893 0.881 0.929 3
40 0.752-0.849 0.791 0.851 37

The experimentally determined values of the spe-
cific heat of the 40 wt% solution were, on the average,
within 7% of the literature value at 15C. All of the
individual values determined in the experiments using
the 40 wt% solution were within 11.5% of the litera-
ture value at 15C, with 33 of the 37 individual values
within 10% of the literature value at 15C.

The exercise described above is structured so that
the student is required to do more than collect data
and tabulate results. For many students, this experi-
ment provides them with their first opportunity to
develop "engineering judgment" in the chemical en-
gineering laboratory environment. The students must
use knowledge obtained in lecture courses, and the
results of experimentation, to make logical conclusions
and meaningful recommendations. Further, the stu-

This... experiment involves basic thermodynamics,
uses easily operated experimental equipment,
produces accurate final results ... and encourages
students to use engineering judgment ...

dents are required to organize and present their find-
ings in a technical report written in accordance with
the classroom instruction they are given.
The experiment is not intended to produce hand-
book-quality physical property data, nor to duplicate
industrial research methods. Indeed, more accurate
results could easily be obtained, for example, by using
a more accurate wattmeter or by using a direct mea-
surement of sample mass. However, despite these and
a few other built-in flaws in the equipment and proce-
dures used, reproducible results which are within
10% of handbook data are consistently achieved. The
flaws do not significantly deteriorate the quality of
the results produced by the apparatus, but are obvi-
ous enough so that the students can usually make sev-
eral worthwhile recommendations, rather than the
contrived recommendations which are quite prevalent
in undergraduate reports.

This paper describes an experiment which involves
basic thermodynamics, uses easily operated experi-
mental equipment, produces accurate final results
from a rather straightforward analysis and encour-
ages students to use engineering judgment to formu-
late worthwhile conclusions and recommendations on
how to improve the apparatus. All of these features
combine to produce a favorable first experience in
technical report writing.
In addition, the experiment requires only a small
expenditure of funds: since breakable parts are rela-
tively inexpensive, annual operating costs are usually
minimal and most of the equipment involved in the
experiment need not be committed solely to this ex-
periment since it can be used in other teaching, labo-
ratory, and research applications.

The author thanks the members of the Villanova
University Department of Chemical Engineering clas-
ses of 1986 and 1987 for providing the experimental
results discussed above.

1. Chemical Engineer's Handbook, R. H. Perry, et. al. (eds.), 6th
ed., McGraw-Hill, NY, 1984. O


Continued from page 129.

good approximation, as discussed below, in which case
the above equation becomes

(n-s)ym (n-s)coT + AH
1+ s (n-s)c T + AH (12)

s T AH y,
T = [s + (n-s)yj cs + (n-s)yj (13)

This is the simplest expression for Tf, the temperature
of the flame front and, in fact, on detailed examination
shows that Tf is not equal to the adiabatic flame tem-
perature, as assumed by Long [4]. Eqs. (12) and (13)
can now be substituted into Eq. (11), giving

M = kf f[coT AHy/s coTo(1 + (n-s)y/s)]
4i-a Cf C A 1 + (n-s) J

+o T n + (n-s) (14)

This now enables a simple expression to be calculated
for the burning constant kb (see Eq. (1)) using

-~ 4 3 2 da
M = d a pliq = 4a2pliq

d M
dt (a2) = 2b apliq

kb C fli

In 1

+ cf[CT AHoyjs CoTo(+(n-s)yjs)]J
CoA 1 + (n-s) -

2k (n-s)y
+ i-n + s (15)
Co n- )Pliq

Of course, a somewhat more cumbersome form of
Eq. (15) can be derived by not using the above approx-
imations of an averaged specific heat, co, for 02 and
the products of combustion or of unit Lewis number,
etc. In fact, Eqs. (5), (8), and (9) without these as-
sumptions lead to the general expression

Here the dimensionless group a is given by
a = PD c /RTk
02 .
and is the reciprocal of a Lewis number. Of course,
Eq. (15) is preferred over Eq. (16) for teaching pur-


Eqs. (15) and (16) each contain two terms, and in
that sense are similar to Long's result [4]. However,
our second term is different from Long's, which we
conclude to be in error. Substitution of numerical val-
ues into either Eq. (15) or (16) indicates that the sec-
ond term is 1-2% of the first term. In this case kb can
be taken to be the first term of either Eq. (15) or (16),
depending on the precision required. Fixing attention
on Eq. (15) for a fairly heavy paraffin being burnt
according to

i 2 + 0 iCO2 + (1+1) H20
CiH2i+2 + 2 2 2
we have
n/s = 2(2i+1)/(3i+1) z 4/3

provided i > 5. As a result

(1 + (n-s)y/s) = 1 + yj/3 = 1.07

which can be taken as unity. Consequently, the first
term in Eq. (15) can be simplified to

S 2k n f[(T T- AHO o (] )

which is similar to Spalding's result [2]. In fact, it is
identical to Spalding's result, when cf = co. Spalding's
derivation [2] is interestingly different from that given
above; in particular, it requires the possibly unpalat-
able assumption that diffusion for both r < rf and r >
rf can be modelled by one composite substance of fuel
and "negative" oxygen, with complete neglect of the
products of combustion. As discussed below, Spald-
ing's assumption that cf = Co is often not a good one.

C (AH + c T ) (n-s)y,1
cT + 11- 1+
2k (ncp SC02) S ]iI
k in 1 +
b CfPliq (n-s)y a
A 1+ -

- CfT 1 + -(n-s)y
-T 1+-4

2k a
+ 0
pliq(nc SC02

In 1 + -


Some insight can be gained into the above conclu-
sion that the ratio of the second term to the first in
both Eqs. (15) and (16) is small. Comparison of Eqs.
(6), (11), and (14) reveals that the ratio of the first and
second terms in both Eqs. (15) and (16) is rf/a. Sub-
stitution of the values of physical properties (see
below) shows that rf/a is roughly in the range 50-100.
Thus neglect of the second term in Eqs. (15) or (16)
generates an uncertainty of less than 2%. Perhaps the
best expression for kb from Eq. (16) neglects the sec-
ond term and assumes
1 > (n-s)yjs

but does not assume unit Lewis number. The result is

2k cf(T T- cAHay/sco
kb n 1+ (18)
k CfPliq

1. First we estimate Tf by substituting values into
Eq. (13). Consider toluene burning in C7H8 + 902 -
7CO2 + 4H20, for which AH = -3.9 x 106 J/mol, X =
3.2 x 104 J/mol, s = 9, y0 = 0.21 and n = 11. At the
quite arbitrary temperature of 700 K, the molar heat
capacities of 02, H20 and CO2 are [8], respectively,
33.0, 37.5 and 49.6 J/(mol K). This gives 2c, = 7 x 49.6
+ 4 x 37.5 9 x 33.0 or co = 100 J/(mol K). Also, these
data give Tf = 1149 K for an ambient air temperature
To = 20C. This value of Tf is less than the adiabatic
temperature of around 2300 K, in contrast with Long's
assumption [4] that the two are identical. This discre-
pancy arises partly because Co is much greater than
the heat capacities of 02, HO2 or CO2. Even so, Co is
not as large as the specific heat of the fuel, with [8]
e.g., cf = 173, 208 and 358 J/(mol K), respectively, for
benzene, toluene and n-octane. Spalding's simplifica-
tion [2] that cf = Co accordingly should be treated with
great care.
2. The above data also give rf/a = 49 (again for
toluene burning in air at 200C), from a comparison of
the two terms in Eq. (15). However, if the more pre-
cise Eq. (16) is used, then r/a = 72 is obtained. For
these calculations values of kf and ko were taken [8]
to be 0.065 and 0.050 J/(K s m), respectively, and To
= 383 K (boiling point of toluene = 110.6C) was as-
sumed. Such a magnitude for rf/a justifies the earlier
neglect of the second term in Eqs. (15) and (16).
3. The previous assumption of unit Lewis number
is worthy of comment. In fact, the value [9] for single
gases at atmospheric pressure is close to unity at 0C,
but can rise to around 1.3 at 700 K. The magnitude of

a (the reciprocal of a Lewis number) used in Eqs. (16)
and (18) turns out to be 2.6 at 700 K, but this is due
to the unexpectedly large value of co.
4. We are now in a position to calculate the burn-
ing constant, kb. Eq. (16) is expected to yield the most
precise value of it, and Eq. (17) the crudest. Eq. (18)
is relatively simple and unlike Eq. (17), does not take
the Lewis number to be unity. The resulting values
of kb are 1.7 x 10-7, 1.2 x 10- and 1.8 x 10-7 m2/s, as
calculated from Eqs. (16)-(18), respectively. The sec-
ond term in Eq. (16) was ignored in this calculation,
which assumed Pliq = 9.6 kmol/mn. The experimental
value [3, 4, 5] of kb is around 1.8 x 10-7 m2/s, which
suggests that Eqs. (16) and (18) both provide satisfac-
tory estimates. However, it will be noted that in each
case the calculated kb is proportional to the ratio k/cf,
which for pure toluene increases by a factor of 3.1
from its boiling point (110.6C) to 1000C. Hence, by
a judicious choice of mean temperature for estimating
kf/cf, a match between theory and experiment can al-
ways be obtained. Long [4] used 1300 K (i.e., greater
than Tf as calculated here for toluene), whereas 700 K
was arbitrarily used above. Clearly, for any greater
precision, the above theory has to be re-worked to
cope with kf and cf varying with temperature and also
composition, because traces of H2 from pyrolysis of
the fuel will have a'significant effect on kf. Otherwise,
it is worth noting that Eq. (18) is a useful and more
precise simplification of Eq. (16) than Eq. (17).
5. It was assumed above that To, the droplet's
temperature, was in effect the normal boiling point of
the liquid concerned. This assumption can be checked
by first considering mass transfer of fuel from the
droplet (at r = a) to the flame front at r = rf. Here
the total outward flow of fuel vapour is M mol/s and
Eq. (3) becomes
M ODP dy
47r2 47Y2 f RT dr

where yf and Df are, respectively, the mole fraction
and diffusivity of the fuel. Hence

U o
M dr DfP dyf
4 r 2 o- RT
a Yfo

where yf = yfo at the droplet's surface (r = a). The
left hand side of this equation is given by Eq. (6), so
that after integration, assuming D/T to be constant

k f(T To) DP f
cf fn 1 + xn
cRT fo


As a good approximation the Lewis number for the
fuel (kfRT/PDfcf) can be taken to be unity, leading to

cf(Tf To)
Yfo + cf(T To)
f f 0

Also the Clausius-Clapeyron equation gives the vari-
ation of yfo with To as

Yfo R To (21)

where Tbp is the normal boiling point of the liquid
fuel. Eqs. (20) and (21) are two simultaneous equa-
tions in To and yfo. Using the above data for a toluene
droplet burning in atmospheric pressure air, the solu-
tion is that yfo = 0.898 and To = 380 K for Tbp = 384
K. This confirms the previous assumption that To =
6. Finally, as noted by Spalding [2] and Long [4],
the above treatment can be used to calculate the diffu-
sion-controlled burning rate and burn-out time for an
involatile solid, such as carbon. In this case 02 diffuses
right up to the surface of the solid, and Eq. (5) has to
be modified to make a = rf, giving

Do2P (n-s)y.
M = 02P n 1 + (22)
4-- RT-n-T 1s +

which simplifies, for 1 >> (n-s) yJs, to

M 02P
4ira sRT

Eq. (23) is exactly true (see the derivation of Eq. (5)
for n = s) if the combustion reaction is Cs + 02 ->
CO2, when in fact, n = s. Alternatively, if the surface
reaction is Cs + /202 CO, then (n-s) yJs = 0.21, in
which case Eq. (23) is true to within 10%. Eq. (23) is
the Nusselt relationship [10] for the diffusion-control-
led burning of a solid particle, such as coal char.

1. Hedley, A. B., and G. F. Martin, J. Inst. Fuel, 1971, 43, 38;
Williams, A. Combustion and Flame, 1973, 21, 1; Prog.
Energy Combust. Sci., 1976, 2, 167; The Combustion of
Sprays of Liquid Fuels, Paul Elek, 1976.
2. Spalding, D. B., Combustion and Mass Transfer: A Textbook
with Multiple-Choice Exercises for Engineering Students,
Ch. 7, Pergamon Press, 1979, and references therein.
3. Kanury, A. M., Introduction to Combustion Phenomena,
Gordon and Breach, 1977; Glassman, I., Combustion,
Academic Press, 1977.
4. Long, V. D., J. Inst. Fuel, 1964, 37, 522.
5. Godsave, G. A. E., Nature 1949, 164, 708; 1953, 171, 86;

Fourth Symposium (Int.) on Combustion, p. 818, Williams
and Wilkins, Baltimore, 1953.
6. Kay, J. M., and R. M. Nedderman, Fluid Mechanics and
Transfer Processes, p. 374, Cambridge University Press,
7. Fristrom, R. M., and A. A. Westenberg, Flame Structure,
Ch. 12, McGraw-Hill, 1965.
8. JANAF Thermochemical Tables, 2nd ed, DOW Chemical Co.,
Midland, Michigan, 1970; Yaws, C.L., Physical Properties,
McGraw-Hill, 1977; Reid, R. C., J. M. Prausnitz, and T. K.
Sherwood, The Properties of Gases and Liquids, 3rd ed,
McGraw-Hill, 1977.
9. Kanury, A. M., Introduction to Combustion Phenomena,
Gordon and Breach, p. 253, 1977.
10. Nusselt, W., Z. des Vereins Deutscher Ing., 1924, 68. 124.

a radius of burning droplet
CA,CB concentration of gaseous species A and B
c CA + CB
cf,Co2,Cp molar heat capacity of fuel, oxygen and
Co mean molar heat capacity defined by
(n-s)Co = (ncp sco,)
D binary diffusion coefficient
DA,Do2 diffusion coefficient of a species in a mul-
ticomponent mixture
H',H ,H' standard molar reference enthalpies for
fuel, oxygen and combustion products
kb burning constant (see Eq. (1))
kf,ko thermal conductivity for fuel-rich (r < rf)
and fuel-lean gases (r > rf), respec-
k thermal conductivity
M number of moles evaporating from fuel
droplet per sec.
NA,NB molar fluxes of species A and B (mol/
(m2 s))
N total of all molar fluxes
n number of moles of products formed by
complete combustion of one mole of
P total pressure
Pv saturated vapour pressure of fuel at
temperature To
Q enthalpy flux (per unit area)
R universal gas constant
r distance from centre of burning oil drop-
rf distance of flame front from centre of oil
s number of moles of 02 required to burn
completely one mole of fuel








temperature of gases around droplet at
distance r
normal boiling point of liquid fuel
temperature of oil droplet's surface
temperature of surrounding air well
away from droplet
temperature of gases in flame front
distance in one-dimensional situation
mole fractions of A and 02
mole fraction of 02 in air = 0.21
dimensionless group = Do,P(ncp-sCo0)/
heat of combustion per mole fuel at 0C
latent heat of vaporisation per mole fuel
density of gas adjacent to droplet
density of liquid fuel (mol/m3) O

n 91 book reviews


by Rita K. Hessley, John W. Reasoner,
and John T. Riley
John Wiley & Sons, New York, 1986. $35.00
Reviewed by
T. D. Wheelock
Iowa State University
This concise and easily read book provides a useful
and basic introduction to the field of coal science and
technology. The introductory chapter provides an
overview of the coal mining/utilization industry and
coal resources of the United States. Subsequent chap-
ters deal with a series of varied and important topics.
A description of the complex processes which form
peat and convert peat into coal provides a basis for
the physical and petrographic characterization of coal.
A picture of the organic structure and chemical reac-
tions of coal is built up through a review of coal treat-
ments involving pyrolysis, solvent extraction, hydro-
genation, and oxidation, and through a review of coal
characterization by instrumental analysis. The book
treats the chemistry and technology of a number of
methods which have been proposed and sometimes
used for converting coal into liquid and gaseous fuels.
The final chapter is devoted to a review of standard
methods (mainly ASTM) for determining the chemical
and physical properties of coal. By seeing how such

properties are measured, the reader is left with a
greater appreciation for a number of empirical proper-
ties such as the proximate analysis, free-swelling
index, and grindability. This appreciation justifies the
greater coverage given to methods of analysis than to
any other topic.
The book should serve as a useful reference for
those seeking a broad rather than a penetrating intro-
duction to coal science and technology. However, it
could also serve those with more specialized interests
by providing an entry into the technical literature.
Frequent references are made to the literature
throughout the text, and each chapter is furnished
with a lengthy list of references.
The book has been used as a college text for coal
chemistry courses taught by the authors. Students
and others using this book would benefit from a prior
knowledge of general chemistry and organic chemis-
try. O


I want to express my appreciation and that of my
colleagues for the very generous editorial entitled, "A
Department That Serves," that appeared in the
Spring 1987 issue of Chemical Engineering Educa-
tion. It is most gratifying to us that you have so
eloquently expressed our actual motives for develop-
ing the study on "Chemical Engineering for the Fu-
ture" and disseminating its findings.
Howard F. Rase
The University of Texas, Austin

books received

Adsorption Technology: A Step-By-Step Approach to Process
Evaluation and Application, edited by Frank L. Slejko; Marcel
Dekker, 270 Madison Ave., New York 10016; 240 pages, $55 (1985).
Heat Transfer and Fluid Flow in Rotating Machinery, Wen-Jei
Yang. Hemisphere Publishing, New York, NY 10016 (1987). 553
pages, $95.00.
Computational Heat Transfer, Yogesh Jaluria and Kenneth E.
Torrance. Hemisphere Publishing, New York, NY 10016 (1986).
366 pages, $49.00.
Aerothermodynamics of Low Pressure Steam Turbines and Con-
densers, M. J. Moore and C. H. Sieverding. Hemisphere Publish-
ing, New York, NY 10016 (1987). 290 pages, $59.95.
Basic Cost Engineering, Second Edition, Revised and Expanded,
Kenneth K. Humphreys and Paul Wellman. Marcel Dekker, New
York, NY 10016 (1987). 368 pages, $34.75


University Microfilms International

Please send additional information for

State Zip

300 North Zeeb Road
Dept. PR.
Ann Arbor, Mi. 48106

(name of publication)

30-32 Mortimer Street
Dept. P.R.
London WIN 7RA

Continued from page 125.
to investigate control systems for a spectrum of situa-

The controllers implemented in ONLINE are
sophisticated even though they excecute only a PID
algorithm. Here are examples of what they can do.
Suppose in the cascade system shown in Figure 7,
the operator decides to switch the flow controller to
OFF. That action leaves the level controller ineffec-
tive because the link between the measured and ma-
nipulated variable has been "broken" It is essential
that the level controller (and the process operator) be
informed of any such break in the loop linkage what-
ever its genesis and that the level controller properly
accommodate to the new situation. In ONLINE, in-
formation about the status of controllers and control
system variables is communicated through the signal
network in the direction opposite to that of the control
signal propagation. In this particular circumstance,
the receipt of information by the level controller about
the OFF status of the flow controller will trigger an
automatic switch in the level controller status from
mode the level controller discontinues excecution of
its algorithm until the flow controller status returns
to ON. When the "downstream" link of a master con-
troller is fully reestablished all the way to a valve or
other process actuator, the controller status is
switched automatically from CONDITIONAL to ON.
This process of information transmission and status
changes occurs no mater how deep the cascade or how
branched the signal network [3]. We view these fea-
tures as essential and are surprised to find them lack-
ing in some commercially marketed packages.
A closely analogous situation arises when the valve
driven by the flow controller is forced full open yet
the level in the accumulator is still rising. In the event
that integral action is used in the level controller, the
integration should be suspended as soon as the valve
saturation condition is known. That is accomplished
by a digital counterpart of the feedback signal used in
electronic controllers to achieve integral action. The
controllers in ONLINE are equipped with a
"TRACK" input variable (shown in Figure 7), which
can be named as any control system variable previ-
ously defined. In the cascade system of Figure 7, the
track variable is the output signal of the flow control-
ler. As the controller module is processed, the value
of the tracked variable is compared with high and
low limits (such as the maximum and minimum flow

FIGURE 7. Cascade level control system with Track vari-
able communicated to the master controller.

through the valve in this example). If a transgression
is detected, the integrating action is suspended as long
as the saturated condition exists. Actually, the
tracked variable is used to change the gains of all
three modes to any values the user desires. Such a
feature can also be used to fashion a 3-segment non-
linear controller. We use it in this way for gain
scheduling in the temperature control of a laboratory
heat exchanger over a range of flow rates. It adds a
bit of spice to the life of students as they discover the
interesting and useful embellishments that can be
made to simple loops.
The use of a track variable is essential when high-
or low-select operators are placed on the output of PI
controllers because the integral calculation must be
suspended for that controller that has been excluded
from the active control link. Shinskey [6] gives several
examples of this type. The track variable proves to be
an extremely useful auxiliary input.
The user also has a choice of three different PID-
controller configurations. The "classical" configuration
employs the error between setpoint and measured sig-
nal in all three modes. Another uses the error in the
integral mode only; the measured signal is used di-
rectly in the P and D modes. Such a configuration
avoids "setpoint kick." A third uses error on P and I
and measurement on D. All versions are implemented
by the incremental algorithm and all employ a first-
order filter on the derivative mode. The calculated
output is checked against the maximum and minimum
declared by the user and is not allowed to exceed those
Other "signal transformations" available in ON-
LINE are: a lead-lag, addition, multiplication, and di-
vision of two signals, high and low selection of two
signals, and square root evaluation. These operations
permit the implementation of just about any multiloop


The availability of ONLINE has "shaken the earth" out here in Berkeley.
Systems with similar capability are also causing a stir elsewhere. The possibilities for
the treatment of process control problems beyond a stirred tank or a dead time are now without bound.

control system imaginable. Indeed, the limiting link
in the control system configuration process seems to
be one's imagination.
The number of controllers placed by designers on
even simple and modest-sized processes can escalate
to the point where tuning becomes a noisome chore,
particularly when loops interact strongly. The incor-
poration of an automatic tuning procedure as one of
the controller status states is an obvious way to re-
lieve the user of such a task. We are working on it.
The program structure of ONLINE permits the addi-
tion of such new features, new control algorithms, and
new commands.
It is also obvious that "umbrella" procedures such
as process optimization procedures and knowledge-
based systems that monitor process operability could
be interfaced with ONLINE. Those matters are also
of pressing interest to chemical engineers, and our
students could benefit from some practice on these

Why use this "real-time" coupling of control sys-
tem and process simulation to practice multiloop sys-
tem operation? Why not instead use the "once-
through" type of simulation starting from prescribed
initial conditions and disturbances? The principal
merit of a "real-time" system such as ONLINE is its
interactive capabilities. The immediate observation of
the cause-effect relation in complex processes as
changes are made is a very effective teacher. Further,
the interactive capability gives the student experience
in running an operating process, which contrasts with
the experience in calculating a "once-through" type of
simulation. Making judgements about what should be
done to halt a reactor runaway, for example, with in-
formation up to the moment, is distinctly different
from making those judgements after the complete
"once-through" response is available. "Once-through"
calculations historically have been the unvaried fare
of this subject. We now have a challenging new capa-
Why not use commercially available software pack-
ages for the teaching of multiloop control? Several are
now available that run on microcomputer systems.
Some of these systems, however, require special
hardware. And some require "special" money. These
commercial programs have been developed to meet

the demands of the industrial workplace and under-
standably are much more sophisticated and complex
than UC ONLINE. Those two attributes can be a
liability for a university instructional environment,
however, because there is more to master (by both
student and instructor) and more time has to be in-
vested in gaining that mastery. User manuals for com-
mercial programs are very thick. That can easily de-
flect the focus of a course from one of learning princi-
ples and their application to one of searching for a
route through the labyrinth of multiple screens, data
files, and system conventions. UC ONLINE is not
without these, but they were built keeping in mind
the special needs of the chemical engineering student
and the constraints imposed by the university instruc-
tional environment. Ten pages of description are all
that's needed to inform students about UC ONLINE.
They pick it up very fast.
The availability of ONLINE has "shaken the
earth" out here in Berkeley. Systems with similar
capability are also causing a stir elsewhere [1, 2]. The
possibilities for the treatment of process control prob-
lems beyond a stirred tank or a dead time are now
without bound. One can now ask students to invent
those control systems that we have heretofore only
talked about, to implement them, and to make them
work. Attention can be focused on the synthesis of
control systems addressing production rate control,
control under constraints, local optimization, and vari-
able structures. Interesting examples of problems in
these areas are found in Shinskey's books [4, 5, 6],
and UC ONLINE is up to handling all of them. With
such capability we can now engage our students in an
activity that, in the author's opinion, is the activity in
which chemical engineers make their most significant
contribution to process control systems: the invention
and development of the control system configuration.

UC ONLINE was conceived and developed in 1984
by Paul H. Gusciora and Chi-Ho Mak, graduate and
undergraduate students respectively. Their im-
plementation was made in Fortran in the multitasking
environment of the Data General RDOS and RTOS
operating systems. Leonid Poslavsky, an under-
graduate student, converted their version to the
single-task version also in Fortran, described here; it
runs on the IBM PC/XT and AT. The work of these


young men holds my respect and admiration. The dis-
tillation simulation used as the example in this paper
was developed by Professor Babu Joseph of Washing-
ton University while on sabbatical leave at Berkeley.
Financial support for a portion of the conversion was
provided by the chemical engineering department at
Berkeley. The computing equipment used was do-
nated by the IBM Corporation through the UC Ber-
keley/IBM joint project on distributed academic com-


1. Koppel, L. B., and G. R. Sullivan, "Use of IBM's Advanced
Control System in Undergraduate Process Control Education,"
CEE, 20 (2), 70 (1986).
2. Lavie, R., Technion, Haifa, Personal communication, 1985.
3. Mak, C-H, and P. H. Gusciora, "ONLINE Control Program for
Multiloop Control Systems." Internal report. Dept. Chem.
Engr., Univ. of Calif., Berkeley, CA (1984).
4. Shinskey, F. G., Distillation Control, 2nd Ed. McGraw-Hill,
New York, NY, 1984.
5. Shinskey, F. G., Process Control Systems, 2nd Ed. McGraw-
Hill, New York, NY, 1979.
6. Shinskey, F. G., Controlling Multivariable Processes, Instru-
ment Soc. Amer., Research Triangle Park, NC, 1981. [

Continued from page 121.

Subcooling and overheating of streams is allowed, but
the user has the responsibility of ensuring the correct
overall heat load.
A Help menu is brought up by special key Fl, and
gives information on the design menu options. It oper-
ates similarly to the data entry Help screens.
A minimum utility network for the four-stream
problem of Figure lb is shown in Figure 3. Seven
units are used, which is two more than the absolute
minimum. In Figure 4 the network has been evolved
into a six-unit design with slightly increased utilities.
The pinch lines have been removed, as the design vi-
olates the pinch decomposition.
In a design with more process streams and hence
more units, the screen can become very busy. It may
be necessary to do only part of the design at a time,
as shown in Figure 5. For this nine-stream problem
[14] only the hot end is shown, and the pinch lines
moved far to the right.
A print-out of the network can be obtained at any
time by pressing the key.


The HENS program relieves the student of a tedi-
ous target-setting hand calculation and allows rapid

generation and change of networks in the graphical
design stage. A major guideline during program de-
velopment was that the program should allow the stu-
dents to make the same mistakes as they could with
pencil and paper. The objective of the program is to
help students think about heat exchanger networks,
not to think for them.
Student reception of the program has been good
on limited exposure to fairly simple designs. Despite
the on-line help material, several students had diffi-
culty using the program efficiently. An in-class dem-
onstration is planned for future classes that should
alleviate this problem. Some improvements for the fu-
ture might include allowing a split stream to be erased
and allowing a design to be saved to disk, along with
the problem data.
The program is available on disk at nominal cost
from the author. Please indicate whether the AT&T
version (640x400 screen) or the IBM version (640x200
screen) is wanted.


1. Linnhoff, B. et al., "User Guide on Process Integration for
the Efficient Use of Energy," Inst. Chem. Engrs., Rugby,
UK (1982).
2. Linnhoff, B., and J. A. Turner, "Heat Recovery Networks:
New Insights Yield Big Savings," Chem. Eng., (Nov. 2, 1981)
p. 56.
3. Linnhoff, B. and E. Hindmarsh, "The Pinch Design Method
for Heat Exchanger Networks," Chem. Eng. Sci. 38, 745
4. Linnhoff, B., and D. R. Vredeveld, "Pinch Technology Has
Come of Age," Chem. Eng. Prog., 80 (7), 31 (1984).
5. Challand, T. B., R. W. Colbert, and C. K. Venkatesh, "Com-
puterized Heat Exchanger Networks," Chem. Eng. Prog. p.
65 (July, 1981).
6. Lipowicz, M., "Heat-Exchanger Software," Chem. Eng.
(Aug. 4, 1986), p. 73.
7. Govind, R., D. Mocsny, P. Cosson, and J. Klei, "Exchanger
Network Synthesis on a Microcomputer," Hydrocarbon Pro-
cessing, p. 53 July, 1986.
8. SuperTargetTM, Linhoff March, PO Box 2306, Leesburg, VA
9. Saboo, A. K., M. Morari, and R. D. Colberg, "RESHEX: An
Interactive Software Package for the Synthesis and Analysis
of Resilient Heat-Exchanger Networks: I. Program Descrip-
tion and Application," Comp. Chem. Eng. 10, 577 (1986).
10. ibid; II. Discussion of Area Targeting and Network Synthesis
Algorithms," Comp. Chem. Eng. 10, 591 (1986).
11. Papoulias, S. A., and I. E. Grossmann, "A Structural Optimi-
zation Approach in Process Synthesis: II Heat Recovery Net-
works," Comp. Chem. Eng. 7, 707 (1983).
12. Floudas, C. A., A. R. Ciric, and I. E. Grossmann, "Automatic
Synthesis of Optimum Heat Exchanger Network Configura-
tions,", AIChE. J. 32, 276 (1986).
13. Wood, S., "Using Turbo PascalTM," Osborne McGraw-Hill,
Berkeley, California (1986).
14. Tjoe, T. N., and B. Linnhoff, "Using Pinch Technology for
Process Retrofit," Chem. Eng. (April 28, 1986) p. 47. O



Departmental Sponsors

The following 152 departments contributed to the support of CEE in 1987 with bulk subscriptions.

University of Akron
University of Alabama
University of Alberta
University of Arizona
University of Arkansas
Auburn University
Brigham Young University
University of British Columbia
Brown University
Bucknell University
University of Calgary
California State University, Long Beach
California Institute of Technology
University of California (Berkeley)
University of California (Davis)
University of California (Los Angeles)
University of California (Santa Barbara)
University of California at San Diego
Carnegie-Mellon University
Case-Western Reserve University
University of Cincinnati
Clarkson University
Clemson University
Cleveland State University
University of Colorado
Colorado School of Mines
Colorado State University
University of Connecticut
Cornell University
Dartmouth College
University of Delaware
Drexel University
University of Florida
Florida State University
Florida Institute of Technology
Georgia Institute of Technology
University of Houston
Howard University
University of Idaho
University of Illinois (Chicago)
University of Illinois (Urbana)
Illinois Institute of Technology
Institute of Paper Chemistry
University of Iowa
Iowa State University
Johns Hopkins University
University of Kansas
Kansas State University
University of Kentucky
Lafayette College
Lakehead University

Lamar University
Laval University
Lehigh University
Loughborough University of Technology
Louisiana State University
Louisiana Tech. University
University of Louisville
University of Lowell
University of Maine
Manhattan College
University of Maryland
University of Massachusetts
Massachusetts Institute of Technology
McMaster University
McNeese State University
University of Michigan
Michigan State University
Michigan Tech. University
University of Minnesota, Duluth
University of Minnesota, Minneapolis
University of Missouri (Columbia)
University of Missouri (Rolla)
Monash University
Montana State University
University of Nebraska
University of New Hampshire
University of New Haven
University of New South Wales
New Jersey Institute of Tech.
University of New Mexico
New Mexico State University
City University of New York
Polytechnic Institute of New York
State University of N.Y. at Buffalo
North Carolina A&T State University
North Carolina State University
University of North Dakota
Northeastern University
Northwestern University
University of Notre Dame
Nova Scotia Technical College
Ohio State University
Ohio University
University of Oklahoma
Oklahoma State University
Oregon State University
University of Ottawa
University of Pennsylvania
Pennsylvania State University
University of Pittsburgh
Princeton University

Purdue University
Queen's University
Rensselaer Polytechnic Institute
University of Rhode Island
Rice University
University of Rochester
Rose-Hulman Institute of Technology
Rutgers University
University of South Alabama
University of South Carolina
University of Saskatchewan
South Dakota School of Mines
University of South Florida
University of Southern California
Stanford University
Stevens Institute of Technology
Syracuse University
Teesside Polytechnic Institute
Tennessee Technological University
University of Tennessee
Texas A&M University
University of Texas at Austin
Texas Technological University
University of Toledo
Tri-State University
Tufts University
Tulane University
Tuskegee Institute
University of Tulsa
University of Utah
Vanderbilt University
Villanova University
University of Virginia
Virginia Polytechnic Institute
Washington State University
University of Washington
Washington University
University of Waterloo
Wayne State University
West Virginia Col. of Grad Studies
West Virginia Inst. Technology
West Virginia University
University of Western Ontario
Widener University
University of Windsor
University of Wisconsin (Madison)
Worcester Polytechnic Institute
University of Wyoming
Yale University
Youngstown State University

If your department is not a contributor, write to CHEMICAL ENGINEERING EDUCATION,
c/o Chemical Engineering Dept., University of Florida, Gainesville FL 32611, for information on bulk subscriptions.

Full Text