Chemical engineering education ( Journal Site )

Material Information

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


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


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

Record Information

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

This item is only available as the following downloads:

Full Text

h Ad
e"IC me*



lo*uctloo Via W046ftdSW&WWuORr'SIMWROOO Tools (pg. 2-42) ...... -11, Takoudis
WWWbUV*Jw(pg, 24,0, a--HaAwg4 spiogs
$w Skfiar, Price, 1,Iojr
OmIM Pr*ctsin TwXbiogPro"s5 CoptMk",254)'
Ttwq)ki Probkm wbaroem is ao stg*-stw 2m) papavowas

ty 9 266) 4WAft, $a*!%

At l1b,"INWOW Unt v(Shippore(N. 250)'--l

Net"h Pmew Cooft-0 Ldmi-ft 314)

T" Uen& 10,T"Khb

ShmIoXosvoo"u Mass Trood" A"V80 a pwow p4embftm (pg. N6)
AfagaUOrs,10 nd,
Night SOMWWHWpped tbeTMM 27$r

c Carji' kwtutim Teo,
oqgo*) 27,0),
Aepa rd:


~..DN "D E. X ..

Graduate Education Advertisements

..Ak wn, Univei..ty..O .. .............. 325 Iowa, University of .......................... 360 Oregon Sata University ......... 39.:. i. : : .
: Alabaa University of ...................... 326 Iowa State Univrsity ........................ 361 Pesylvania University of ........... 3
;Alberta.iUnierisiy of..;................ 327 Johns Hopkins Uivety .................. 362 Pennasylvaia State University .......;, 395
Aitona :.iiver o328 Kansas, Uiversity f ............... 429 PiIabi; ty of ............. 39
; :Aian Jtate Uni .......... .329 Kiasas State U sity....... ....... Polyteclitc University .......... .....
..i raty ....t.......:....;.... 330 at vityof .......... ..364 na lUnivmity ......................
Sou Uiersy............ 42 Kig FahdUnivivesity ..................... 365 Purde Unawviity ............... :': :
,: btish Colu mbia Uiversity of ......... 425 Lamar University ................... 430 Quead U, U wersty of............ .
SBiown veraty .... ................ 439 Laval University .. ..................... 366 Rensla. yteci h ....-.
B: ii It J eif sity i ..................... 426 Lehigh Univerit ........,..i.... ..... 367 Rhode 1b a. University of.;...... ..;,.. :: :4
a : i yiU university ........ .............. 331 Loughborough Unv ersity................ 430 Rio niv sit............... :40 ;
Srai Bi erkeley,:University of .. 332 'Lousiana State niversity.................. 368 Roheste, Uiversity of................ ..... ':
aitil l.ia Nvis. University of ......... 333 Lousiana Teh Unive ity ............4 ..... 431 Raeoiatu aitute of ib :"logy 435. :
i ria ne Univsityof ......... 4 Louisville, University f.................... Rutgetsi ity ........... ... .. 403
S .. lif..d. 1AigelosUniv. of ...... 33 Maine, Utnivesity of.............., 369 Sigapo aiol University of... .44

Cf.. : uB .Sa-taeV q Uiv. of,.....:336 Manhat college ........................ 370 s south a a. Uisity of............ 40
lifo tiittetchnlogy ......337 Maryland Baltioe Cotsy Univ. of 371 Soutifhaltaii School of Mine ..s..... 440I
a: iegiMe lt university ..... ......... 338 Mary .land, College P irk. Uiv. of. 372 South Foida Univity of........ 4
e ]WestatmReserve University...... 339 Massachusetts, AlthIrst Univ. of.,.. 373 SouthnmtFlbroia, Uversity of ..... 436
nint iveity 340 Masacbse#sLowel. Univ. of ..... 439 State Uisity ofNei w York.Ruffalo 4A..
aersiy........... 34 Mssachusethsti t tofTechology37 SteYensinJitute ofTeboloy ........ 406 :
C..ein U( veti.. ity .. e .... ..... ..... M......... cpMas'Uivrs...... ........ 432- Sydey, ia itof .............. 437'
el. u d` sfte Unie.. ity....'. B.....".46 Michigan., UniErsity .......... .........375 .Syraci U .is Uritrsy........... .... 4

lo doa Sch~ lof Mines............ 344 itigat Technological Univeity,.. 377 Texas, Uniersity ...3............ ".40
lsa .: ..:. State Un.iversity ........ .....345 Minnesota, UMnivesi.tyof o;,.............. 378 Te xas A&lI;ity.. ......4........
Co: li: : ~, Uri yriy, ........ .,...t.. ......:. t Miisounri Coumbia,. U iver of,,... 79 Toledo, t aity of ..............:. 4
C-aSi : l ctib~ Univ y ........;.. 246: issi Roll Unisit of... 380 iTulanj nivesity.................... .......
SConell University .....;...... .......... 347 Monash University .............. 432 Tla. University of ................. 412
*a fut h Colle&gp. ...........;.....348 Montana StateUniversity................440 Utahl, Univityof........... 43.
DelawiuJinUveri!sity of........ 349 Nebraska, University of.............. 381 Vanderbik University.......... 413.
D: ri liveSit,.... ................... 427 Nada : Rn.T U. ity ............382 Vi U ityof............. 4

.l... 'Polytechiiqu tital 350 N uew esey l titute ofTecoiogy. 3 Vrginia; T ....... ........... 4 :.."
S:. ~ bir Um varsity ............. 42S New Mexio, Uiversity of ........ ;.; 384 ,Wahi Untoizw i..t.y............4...o .:
Sai : i. iveisityf ...... ........ 351 New Me xi t ... ......:. WiT hou lSt Univity .....

: i r: sia, E.g.,ifesa.c.,, iite.:..s. ......,. 428 New South Wales. Unvlrsity of 433 Washi. i lBU i esit, ....................... 418 :
foridAMM Univerity ........ 352 North Carolina State Univerisity. 386 Waterloo. Ul eiity of .....;....... 438
S" .. i..... ,,..,.................."h it ..........:.. 3........... 4i

l ZiToaio.....ey.....;................ 353 North Dakota. Usitvrsity of ............440 Wayne ate Univity .....................S: 4
GeA r..l.t..ofTecho y......... 354 NiOrtheaster Univfruity ........433 West Virgia Udiversity .......420 ;
Hato University of ............ 355 NorthwesternUnivrsity ...387 Wieter UniWarfimity ....................... 439
:Ht .. n vr ................ ...... 356. Not* Daml Univeaityof.. ............. 388 WIscosian l vsrity of ..................... 424.
SIdato Uniesof ....... ...429 Ohio State Uni eaii y ................. 8 : Fi ,,.,. ...... 4 : : W
Wii firs Chicag Universitf ........ 357 O:ho university ..;.. ............ s390 Wyoming;,iversiy of......... 423
i. .llinois. Urna.-e nstig, ..... 358 OklahNorn i nivtesityof........... 31 ale University.................. ....42
: lindis nsitute of Tec gy ......... 359 dlahoma State University :............ 392. ::.

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

T. J. Anderson

Phillip C. Wankat

Carole Yocum

James O. Wilkes and Mark A. Burns
University of Michigan
William J. Koros
University of Texas, Austin

E. Dendy Sloan, Jr.
Colorado School of Mines

Gary Poehlein
Georgia Institute of Technology
Klaus Timmerhaus
University of Colorado

Dianne Dorland
University of Minnesota, Duluth
Thomas F. Edgar
University of Texas at Austin
Richard M. Felder
North Carolina State University
Bruce A. Finlayson
University of Washington
H. Scott Fogler
University of Michigan
David F. Ollis
North Carolina State University
Angelo J. Perna
New Jersey Institute of Technology
Ronald W. Rousseau
Georgia Institute of Technology
Stanley I. Sander
University of Delaware
Richard C. Seagrave
Iowa State University
M. Sami Selim
Colorado School of Mines
James E. Stice
University of Texas at Austin
Donald R. Woods
McMaster University

Fall 1998

Chemical Engineering Education

Volume 32

Number 4

Fall 1998

242 Instruction via Web-Based Semiconductor Simulation Tools: The Case of
"Fundamentals and Design of Microelectronics Procesing"
Sanjit Singh Dang, Christos G. Takoudis
246 Educational Tools for Pollution Prevention Through Process Integration,
Mahmoud M. El-Halwagi, H. Dennis Spriggs
250 Postgraduate Environmental Engineering Program in the Department of
Chemical Engineering at the National University of Singapore,
K.G. Neoh, Y.P. Ting, R.R. Stanforth, C. Tien
254 Experimental Projects in Teaching Process Control,
Mikhail Skliar, Jesse W. Price, Christopher A. Tyler
260 Linear Unsteady Transport Problems When There is an Initial Steady State,
Kyriakos D. Papadopoulos
262 Graduate Education for Particle Science and Technology at the NSF Engineer-
ing Research Center: A Model for the Future, Anne E. Donnelly
266 Conducting a Multi-University Graduate Student Symposium: Goals,
Guidelines, and Experiences, Jamie A. Hestekin, Marazban Sarkari, Melissa
A. Summers, Katherine S. Ziemer, Leonard P. Zuba

270 Process Safety in the Curriculum: Explosion Prevention Technical Elective,
Charles M. Sheppard
290 Integrating Design Throughout the ChE Curriculum: Lessons Learned,
Douglas E. Hirt

278 The Night Someone Slipped the Truth Serum in the Punch Bowl,
Richard M. Felder

280 Batch Distillation Optimization Made Easy, Massimiliano Barolo

286 Single-Component Mass Transfer Across a Porous Membrane,
Ferndo D. Magalhdes, Adelio Mendes
296 Three Trends in Teaching and Learning, Donald R. Woods
302 Teaching Packed-Column Design from a Plate-Column Perspective,
Daniel J. Lacks

306 What is Inside That Black Box, And How Does It Work? Shanaya Godiwalla

308 The Importance of System Selection on Compressible Flow Analysis: Filling
Vessels, S.E. Forrester, G.M. Evans
314 Network Process Control Laboratory, Barry Lennox, Michael Brisk
318 Ultrafiltration of Dairy Products as a ChE Laboratory Experiment,
Thomas D. Conlee, Helen C. Hollein, Charles H. Gooding, C. Stewart Slater

1 269 ASEE Chemical Engineering Division Awards
> 268, 277 Book Reviews

CHEMICAL ENGINEERING EDUCATION (ISSN 0009-2479) is published quarterly by the Chemical Engineering
Division, American Society for Engineering Education, and is edited at the University of Florida. Correspondence
regarding editorial matter, circulation, and changes of address should be sent to CEE, Chemical Engineering Department,
University of Florida, Gainesville, FL 32611-6005. Copyright 1998 by the Chemical Engineering Division, American
Societyfor Engineering Education. The statements and opinions expressed in this periodical are those of the writers and
not necessarily those of the ChE Division, ASEE, which body assumes no responsibility for them. Defective copies replaced
if notified within 120 days of publication. Write for information on subscription costs and for back copy costs and
availability. POSTMASTER: Send address changes to CEE, Chemical Engineering Department., University of Florida,
Gainesville, FL 32611-6005. Periodicals Postage Paid at Gainesville, Florida.

Graduate Education



The Case of "Fundamentals and Design of

Microelectronics Processing"

University of Illinois at Chicago Chicago, IL 60607

An increasing number of chemical engineers are en-
tering the field of microelectronic materials and
processing, in part because basic knowledge of this
fast-growing field lies in chemical engineering. Oxidation of
silicon, surface reaction chemistry in microfabrication tech-
niques, diffusion of impurities through the films, and effect
of processing environments (flow rate, temperature, pres-
sure, etc.) on the nature and quality of films are some repre-
sentative example systems.-"91 Chemical, material, and elec-
trical engineering principles in the fundamental understand-
ing and design of microelectronics processing are bringing
about great changes in industrial process control, automotive
electronics, and other fields in which data acquisition, com-
putation, or controls are necessary. Thus, several chemical
engineering departments have been either offering courses
in microelectronic materials and processing or incorporating
several examples and case studies in core curriculum chemi-
cal engineering courses.[eg., 10
A year and a half ago, our department started a dual-level
class (offered to graduate and advanced undergraduate stu-
dents) titled "Fundamentals and Design of Microelectronics
Processing." The objective of this course is to provide par-
ticipants with the basic principles and practical aspects of
the most advanced state of electronics processing. The em-
phasis of the course is on basic aspects of thin film deposi-
tion, doping, substrate passivation, lithography, and etching
coupled with chemical kinetics, reactor design, thermody-
namics, optimization, and other engineering concepts as they
apply to fundamental processes that are useful for feature
sizes down to the order of about 0.1 tpm.
Last spring, the scope and effectiveness of this course was
substantially enhanced and augmented with the introduction
and implementation of web-based semiconductor simulation
tools. Two simulation tools have been implemented in the

course so far: ThermoEMP and TSuprem-4. The first (devel-
oped, in part, by our group) tackles the thermodynamics of
microelectronic materials and their processing as well as the
a priori predictions of process-properties relationships in
advanced processes and design of fabrication techniques."11
The second simulates the processing steps involved in the
manufacture of silicon integrated circuits, discrete devices,
and micro-electro-mechanical systems (MEMS)."12'
Several references and relevant journal articles as well as
case studies based on those semiconductor simulation tools
were used. In order to maintain the pace needed to cover
the course outline shown in Table 1, some topics were
covered in depth.
- ThermoEMP (Thermodynamics of Electronic
Materials Processing)
ThermoEMP is a computer program that calculates 1) the
chemical equilibrium compositions of microelectronic mate-
rials processing, and 2) the thermodynamic and transport
properties of the equilibrium mixture (formed after reac-
tion). Results are generated through a procedure that mini-
mizes the Gibbs free energy of the system via a rigorous
thermodynamic analysis."'1 This approach is useful to de-
scribe chemical equilibrium because it does not require a set
of reactions to be specified a priori. The variables that must

Sanjit Singh Dang is a chemical engineering graduate student at the
University of Illinois at Chicago. He received his Bachelor's degree in
chemical engineering from Panjab University, Chandigarh, India in 1997.
With Professor C. G. Takoudis, he is conducting research on silicon
oxynitrides for use as future ultra-thin dielectrics. He is also interested in
growth and passivation of silicon-germanium, and characterization of
thin films. His hobbies are movie-making and freelance writing.
Christos G. Takoudis is a Professor at UIC. His research interests are
in the areas of microfabrication, microelectronic materials, surface chem-
istry, and heterogenous catalysis.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Graduate Education

be specified prior to analysis
are the reacting species, the
relative amounts of the reac-
tants, and two state functions
to indicate the thermodynamic
state of the system. The two
state functions typically speci-
fied are the system tempera-
ture and pressure. The program
is capable of handling multiple
thermodynamic states and in-
puts so that a range of state
values can be simulated. This
is particularly convenient if
calculations are to be obtained
for a number of initial reactant
compositions and processing
environments. This software is
restricted to students taking the
class and users authorized by
the instructor (C.G. Takoudis).
The input file consists of two
main sections: 1) identification
of reactant species and their
relative amounts, and 2) speci-
fication of process variables
(e.g., pressure, temperature).
For each pressure and tempera-
ture, results include the spe-
cies composition obtained at
equilibrium as well as the ther-
modynamic properties of the
final system in a tabular form.

> TSuprem-4 (Two-
Dimensional Process-
Simulation Program)
TSuprem-4 (licensed from
Technology Modeling Asso-

Course Outline

Crystal Growth
Basics of Crystal Growth Processes
Heat and Mass Transfer, Thermodynamics
Doping, Design of Crystal Growth Processes
Modeling and Simulation, Examples
Chemical Vapor Deposition (CVD)
A Priori Process-Property Relationships
ThermoEMP Case Studies
Chemical Kinetics-Transport Phenomena
Reactor Design, Modeling and Simulation
Doping, Thin Film Characterization-Yield
Plasma-Assisted/Enhanced CVD
ThermoEMP and TSuprem-4 Case Studies
Physical Vapor Deposition (PVD)
Elements of Vacuum Systems-Process considerations
Molecular Beam Epitaxy
Systems Design
Three-Dimensional Integration
A Priori Prediction of Process-Property Relationships
Reactor Analysis and Design
Selective Epitaxial Growth-Other Silicon-on-Insulator Techniques
Three-Dimensional Integration and Microfabrication
ThermoEMP and TSuprem-4 Case Studies
Degradation and Characterization of Dielectric Thin Films
ThermoEMP Case Studies
Ion Implantation
Fundamentals-Kinetic Considerations
Design and Process Considerations
Analysis and Design of Masking Films for Ion Implantation
TSuprem-4 and ThermoEMP Case Studies
Advanced Lithography
Chemistry and Physics of Lithographic Materials
Fundamentals of Surface Preparation
Resists-Multi-Level Resists
Design and Control of Lithographic Processes
Advanced Lift-Off Techniques, Problem Areas
ThermoEMP and TSuprem-4 Case Studies
Dry Etching
Low-Pressure Discharges, Physical and Chemical Phenomena
Selectivity-Feature and Pattern Size Control
Fundamentals of Dry Etching
Design and Process Considerations
ThermoEMP and TSuprem-4 Case Studies

ciates, Inc.) is a computer program for simulating the pro-
cessing steps involved in the manufacture of silicon inte-
grated circuits, discrete devices, and MEMS.[12' A wide range
of processing steps can be modeled by this program. Ex-
amples are ion implantation, inert ambient drive-in, silicon
and polysilicon oxidation and silicidation, epitaxial growth,
low temperature deposition, and etching of various materi-
als. For each such process, plots can be easily obtained for
many variables (e.g., dopant concentration profiles). This
software is restricted to authorized users only.
A tandem and organized use of both semiconductor simu-

lation tools has been developed
and introduced for the first time
to advanced-level courses in mi-
croelectronic materials and their
processing for better understand-
ing of the basic principles and
practical aspects of the most ad-
vanced state of the processes and
design of microfabrication tech-

The simulation tools were used
in a manner that the scope and
instruction effectiveness would
be enhanced and augmented in
the classroom. There was one
lecture for each of these two
semiconductor simulation tools
with step-by-step explanations
and several examples done in the
classroom through a live com-
puter connection and projection
onto a screen. This was particu-
larly important for the students
to understand the basics and all
important aspects of these tools.

Table 1 shows the use of these
semiconductor simulation codes
throughout the syllabus of the
course. After covering a particu-
lar topic, students were assigned
a homework problem set, the first
part of which included problems
to be solved without the use of
any simulation tools. This was

Seemed important so that they
would understand the basics of
the topic before investigating case studies that would make
use of semiconductor simulation codes. The second part of
the homework set consisted of case studies requiring simula-
tion tools. In this fashion, the students had to understand and
comprehend basic concepts as well as work with sophisti-
cated case studies based on advanced computational soft-
ware. Occasionally (e.g., after ion implantation), case stud-
ies were based on an interplay of TSuprem-4 and ThermoEMP
as applied to multiple microelectronic processes already cov-
ered in the course up to that time (since ThermoEMP alone,
for instance, would be of no value to ion implantation alone).

Fall 1998

Graduate Education

Case Studies (in part based on TSupreme-4)

1. A p-n junction is formed by implanting 100 keV As through an opening in a
layer of thermal SiO,. If a dose of 5 x 10'5 ions/cm2 is used and a background
concentration of I x 10o' cm exists in the substrate, calculate the depth of the
junction. Assume a Gaussian distribution for the implanted dose. Sketch the
junction. (HINT: Rp=0.06 pm, and ARp=0.021 tm for 100 keV As.)
2. Use TSuprem-4 to simulate the process run that had the implantations and
oxidations as listed below (show your "program" output).
N-type (100) wafer, phosphorous doped ND=4 x 101/cm3
(a) Boron ion implantation at 80 keV and a dose of 6 x 10l2/cm2. Plot the
LOG,, of #/cm3 vs distance into the substrate. What is the junction depth?
(b) 1100C anneal for 15 min in N,. Plot the LOG,,, of #/cm' vs. distance into
the substrate, before and after annealing, on the same plot as (a). (i) What is the
junction depth? (ii) What is the approximate peak doping concentration?
(c) After parts (a) and (b), an additional arsenic ion implantation at 60 keV
with a dose of 5 x 104/cm2 is applied and then annealed at 10500C for 8 min in
N,. Plot the LOGI0 of boron, phosphorus, and arsenic as well as the net doping

One example, based in part on TSuprem-4, is presented in Table 2.
These were parts of assignments given to the students after the topic of
'ion implantation' had been taught in the class. The first problem in-
cludes questions about the junction depth and a sketch of the junction for
a particular processing environment. The second problem includes mul-
tiple ion implantation processes coupled with an annealing step; such
realistic situations require numerical calculations that a sophisticated
simulation tool (TSuprem-4 in this case) should be able to take care of.
After this stage, students were asked to investigate and explore the
effects of varying system parameters along with possible additional
processing steps through the use of TSuprem-4. At this point, the simula-
tion tools' ease of use and the ability to give fast results were found to be
very helpful. Figures 1 and 2 show results from the Case Studies #3 and
#4 of this Table, respectively.

Table 3 includes some of the case studies assigned as homework, all of
which make use of ThermoEMP. It can be easily seen that the case
studies are based on state-of-the-art technologies and concepts in the
microelectronics industry. For example, the first problem requires the
determination of the minimum temperature above which oxide-free sili-
con growth (a very important requirement in the microelectronics indus-
try) can take place; the solution of this Case Study is presented in Table
4. Next, oxide-free silicon carbide growth is investigated in specific
reaction environments. The last problem is based on the effects of
dichlorosilane flow rate and temperature on the selective epitaxial growth
of silicon. A remarkable aspect of these case studies is that students
experience some of the key issues in real-life problems and they have the
opportunity to see that solutions may be obtained, in some cases at least,
within a very short period of time without doing any experiments (e.g.,
from fundamental knowledge-driven simulation tools)! On the other
hand, it should be noted that ThermoEMP is based on the chemical
equilibrium of a system; hence, if a system is at conditions where
kinetics becomes a dominating factor, chemical equilibrium-based re-
sults should be viewed in the context of the usefulness of such results in
situations of interest. In any case, however, thermodynamics would

#/cm vs. distance into the substrate on the same set of axes. Are there any p-n
junction(s)? If yes, how far are they from the surface?
3. A bird's peak-shaped structure can be formed by the following processing steps:
(i) deposit 0.03-pm-thick silicon dioxide; (ii) deposit 0.1 -pm-thick silicon
nitride; (iii) etch one-sixth of the nitride laterally from one side; (iv) oxidize at
10000C for 100 min in wet 0, ambience. Use TSuprem-4 to simulate these steps.
In the final structure formed, show the regions under compression and those
under tension, besides giving the numerical values of stress in each of the cases
at different points. Also, give reasons) for the formation of the bird's peak. (Use
1.5 pm as the lateral dimension for the structure.)
4. Use TSuprem-4 to plot the impurity distribution profile (LOG,, of boron,
antimony and arsenic concentration versus thickness from the top) in a bipolar
transistor formed after the following processing steps on a B-doped substrate
with a background concentration of lxl0'l/cm3: (i) Grow an oxide (for masking
of the buried layer) at 11500C for 120 min in steam ambience; (ii) Etch the
oxide; (iii) Implant antimony at 75 keV with a dose of lxl0's/cm2; (iv) Oxidize
in dry O ambience at 11500C for 30 min; (v) Anneal at 11500C and 360 min;
(vi) Etch the oxide; (vii) Grow 1.8-lm-thick As-doped epitaxial layer at 1050C
for 6 min with an As concentration of 5xl0'/cm2; (viii) Grow a pad oxide at
10500C for 30 min in dry 0,; (ix) Deposit a 0.12-lm-thick nitride layer.

Contours of Hydrostotic Pressure

Compress1 n

S5 7E9
'I | 0 -9IES I

1 J

0e.e 0.80 i .?e ,40 0 I .ae
i .tConC0 Ge mancflrn

Figure 1. Contours of hydrostatic pressure of a substrate
subjected to 0.03 pm oxide deposition, 0.10 pm nitride
deposition, nitride etching, and wet oxidation at 10000C
for 100 min (solution to Case Study #3, Table 2).
Active, Epitaxy

0 ,,

',nt Imony
16 Al-r..n c ..

i /1
Is I ----on


0.00 2.00
Oistonce (microns)


Figure 2. Impurity distribution in a bipolar structure
after steam oxidation of B-doped substrate at 11500C for
120 min., etching, antimony implantation at 75 keV, dry
oxidation at 1150"C for 360 min., etching, As-doped
epitaxy at 1050 C, dry oxidation at 10500C for 30 min,
and 0.12 pm silicon nitride deposition (solution to Case
Study #4, Table 2).
Chemical Engineering Education

Graduate Education

result in the regions within which materials or processes of
interest would be possible (or not).
In the near future, our plan is 1) to add more simulation
tools for microelectronic materials and their processing, and
2) to introduce one (or two) term projects) per student on
the design and fabrication of an integrated circuit, a discrete
device, or a MEM system. In such term projects, ThermoEMP
would be used for a priori design of the processing steps
involved. These steps could then be studied with TSuprem-4
(and/or further simulation tools to be added), after which the
final structure characteristics could be determined and com-
pared with the desired ones.

Case Studies: Homework Problems (based on ThermoEMP)
1. Determine the minimum temperature above which oxide-free
silicon growth can take place from a mixture of 45 slm hydrogen,
0.15 sim DCS, 0.1, 1, and 10 ppm of water vapor, at a total
pressure of 0.001, 0.1, 10, and 760 Torr.
2. Determine the minimum temperature above which oxide-free
silicon carbide growth can take place from a mixture of 45 slm
hydrogen, 0.15 slm DCS, 0.05 sim C3,H, 0.1, 1, and 10 ppm of
water vapor, at a total pressure of 0.001, 0.1, 10, and 760 Torr.
3. Use ThermoEMP for the following parts. Please include the plots
and a print of the output data for each question.
For Selective Epitaxial Growth (SEG) of silicon and for Epitaxial
Lateral Overgrowth (ELO), the temperature range is typically from
600C to 10000C. In the reactor we use 60 slm of H, and 0.22 slm
of SiH,C1, (DCS) with 0.66 slm of HCI added to the system. The
system pressure is 40 Torr.
(a) Plot on a semi-log scale the steady-state mole fractions of the
main gases in the system over the temperature range of 600C
to 1000C; also include Si(s). Plot, on the same axis, only
those that are greater than 107. Explain why the HC1 and SiC1,
are increasing with increasing temperature and the DCS is
(b) Repeat part (a) with the DCS flow rate being tripled, with all
other flow rates being held constant.
(c) Repeat parts (a) and (b) at 150 Torr.

Solution to Case Study #1, Table 3
ppm of HO in Pressure Minimum Temperature above which
the system SiO,-free Silicon CVD Takes Place
(Torr) (0C)
0.001 561
0.1 0.1 646
10.0 749
760.0 864
0.001 601
1.0 0.1 695
10.0 810
760.0 946
0.001 646
10.0 0.1 749
10.0 880
760.0 1037

Fall 1998

The feedback and written evaluations of the students on
the scope and instructional effectiveness of the web-based
semiconductor simulation tools were overwhelmingly posi-
tive: students strongly agreed that it was easy to figure out
how to use the simulation tools; they strongly agreed that the
overall class experience was enhanced by the use of the soft-
ware; they strongly agreed that it was convenient to have
universal access to programs via the web; and they would like/
have liked to use the simulation tools in other classes also.

Numerous helpful discussions with Professors Gerold
Neudeck and Mark Lundstrom from the School of Electrical
and Computational Engineering at Purdue University are
gratefully acknowledged. Also, the tremendous help of Nirav
H. Kapadia, Joseph P. Geisler, and Jay R. Hamilton from
that same school is greatly appreciated, as are the contribu-
tions to the development of the current state of ThermoEMP
by several past students of C.G. Takoudis. C.G. Takoudis is
especially thankful to Avant! Corporation (TMA, Inc.) for
licensing TSuprem-4 for classroom use. Partial financial
support provided by the National Science Foundation (ENG/
EEC/CRCD) is also gratefully acknowledged.

1. Lee, H.H., Fundamentals of Microelectronics Processing,
McGraw-Hill, New York, NY (1990)
2. Middleman, S., and A.K. Hochberg, Process Engineering
Analysis in Semiconductor Device Fabrication, McGraw-
Hill, New York, NY (1993)
3. Ruska, S.W., Microelectronic Processing, McGraw-Hill, New
York, NY (1987)
4. Wolf, S., and R.N. Tauber, Silicon Processing for the VLSI
Era. Vol.1, Process Technology, Lattice Press (1986)
5. Sze, S.M., VLSI Technology, McGraw-Hill, New York (1988)
6. Ghandi, S.K., VLSI Fabrication Principles, Wiley
Interscience (1990)
7. Campbell, S.A., The Science and Engineering of Microelec-
tronic Fabrication, Oxford University Press (1996)
8. Madou, M., Fundamentals of Microfabrication, CRC Press
9. Kovacs, G.T.A., Micromachined Transducers Sourcebook,
McGraw-Hill, New York, NY (1998)
10. Special section on Electronic Materials Processing, Chem.
Eng. Ed., 24, 26 (1990)
11. For example: (a) Gordon, S., and B.J. McBride, Tech. Rpt.
SP-273, NASA, NASA Lewis Research Center (1971); (b)
Gaynor, W.H., "Applicability of Thermodynamic Calcula-
tions on the Prediction of Chemical Vapor Deposition Reac-
tor Performance and Process-Property Relationships," BS
Thesis, School of Chemical Engineering, Purdue University
(1989); (c) I.M. Lee, A. Jansons, and C.G. Takoudis, "Effects
of Water Vapor and Chlorine on the Epitaxial Growth of
SiGe Films by Chemical Vapor Deposition; Thermodynamic
Analysis," J. Vac. Sci. Tech., B 15, 880 (1997) and references
12. TSuprem-4: User's Manual, Version 6.4, Technology Model-
ing Associates, Inc., Sunnyvale, CA, October (1996) 0

SGraduate Education





Auburn University Auburn, AL 36849

With today's increasingly stringent environmental
regulations, there is a growing need for cost- and
energy-efficient pollution prevention. This paper
is based on a senior-level/first-year graduate-level course
taught at Auburn University that provides a unique frame-
work for teaching systematic techniques for addressing pol-
lution prevention in the process industries while reconciling
environmental issues with other plant objectives such as cost
effectiveness, yield enhancement, resource conservation, and
energy reduction. (Similar aspects of the course have also
been taught at the University of Virginia.)
The overall philosophy, techniques, and tools presented in
the course are rooted in fundamental engineering principles
and process integration. This approach is significantly dif-
ferent from traditional teaching of pollution prevention where
the environmental problem is addressed with a narrow
mandate focusing on small segments or streams of the
process and experience and subjective opinion are then
used to tackle the problem.
This traditional teaching approach presents a number of
There is a bias toward end-of-the-pipe solutions
rather than understanding and solving the problem
The pollution problem can inadvertently be trans-
ferred from one medium to another (e.g., from air to
Proposed solutions to the environmental problems

" Matrix Process Integration, L.C., PO Box 2356, Leesburg, VA

lack the systematic nature that is the trademark of
chemical engineering successes. Instead, the
approaches taught to the students are typically
based on hit-or-miss procedures or attempts to
replicate other solutions regardless of the specific
nature of each process.
The solutions are, therefore, suboptimal, and result
in higher costs and less actual reduction in pollu-
Finally, there is limited opportunity for the students
to understand and learn the broader lessons of what
features of process designs cause pollution problems
in the first place.
This current state of affairs calls for a fundamental and
generally applicable approach for teaching and addressing
pollution prevention.
Mahmoud EI-Halwagi is an Alumni Professor
of Chemical Engineering at Auburn University.
He received his BS and MS degrees from Cairo
University and his PhD degree from the Univer-
sity of California, Los Angeles, all in chemical
engineering. He has served as a consultant to
a wide variety of chemical, petrochemical, pe-
troleum, pharmaceutical, and metal finishing

Dennis Spriggs is President of Matrix Pro-
cess Integration, a consulting firm specializ-
ing in integrated process designs, developing
operating strategies, and aligning technology
with corporate strategy for a sustainable com-
petitive advantage. He received his BSChE
from West Virginia Institute of Technology and
his PhD from the University of Virginia.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education


Graduate Education

The authors have intro-
duced and illustrated the
use of a complete method-
ology for addressing pol-
lution prevention and have
transferred it to the class-
room. It incorporates re-
cent developments in inte-
grated process design and
is based on understanding
the big picture first, deal-
ing with details only after
the important structural
decisions have been made.
The course has a number
of key defining features:



Mass- and Energy-Separating Sources
ted Agents in Sinks/ (Back to
es Generators Process)


Mass- and Energy-Separating
Agents out
(to Regeneration and Recycle)

Figure 1. Schematic representation of mass-integration strate-
gies for pollution prevention: segregation, mixing, interception,
recycle, and sink/generator manipulation.1,5,61

It is based on the fundamentals of chemical
engineering, process integration, thermodynamics,
mathematics, and principles of problem decomposi-
The students learn how to set robust performance
targets for pollution prevention, yield improvement,
energy integration, and cost effectiveness ahead of
detailed design.
Trade-offs are made at the targeting stage, allow-
ing the students to avoid unnecessary design detail.
Process synthesis methods and chemical engineer-
ing tools are then employed to systematically make
the trade-offs and to realize the targeted perfor-
Throughout the course, extensive use is made of
tools for representing complex process interactions
in ways that are both rigorous and easy to inter-
pret. These tools are illustrated with numerous
industrial applications.

A chemical process is an integrated system of intercon-
nected units and streams, and it should be treated as such.
Process integration is a holistic approach to process design,
retrofitting, and operation and emphasizes the unity of the
process. In light of the strong interaction among process
units, streams, and objectives, process integration offers a
unique framework for fundamental understanding of the glo-
bal insights into the process, methodically determining its
attainable performance targets and systematically making
decisions leading to the realization of those targets. Process
Fall 1998

mal-pinch techniques that can be used to identify minimum
heating and cooling utility requirements for a process.
On the other hand, mass integration1'5'61 is a systematic
methodology that provides fundamental understanding of
the global flow of mass within the process and employs this
understanding for identifying performance targets and opti-
mizing the generation and routing of species throughout the
process. Mass-allocation objectives such as pollution pre-
vention are at the heart of mass integration. Mass integration
is more general and more involved than energy integration,
and because of the overriding mass objectives of most pro-
cesses, mass integration can potentially provide much stron-
ger impact on the process than can energy integration.
Both integration branches are compatible. Mass integra-
tion coupled with energy integration provides a systematic
framework for understanding the big picture of the process,
identifying performance targets, and developing solutions
for improving process efficiency, including pollution pre-
vention. The core of the course is dedicated to mass-
integration techniques.

Mass integration is based on fundamental principles of
chemical engineering combined with system analysis using
graphical and optimization-based tools. The first step in
conducting mass integration is development of a global mass
allocation representation of the whole process from a species
viewpoint, as shown in Figure 1. For each targeted species
(e.g., each pollutant) there are sources (streams that carry the
species) and process sinks (units that can accept the species).

integration has recently
been reviewed in various
educational literature. [e.g.,.2]
Two key branches of pro-
cess integration have been
developed: mass integration
and energy integration. En-
ergy integration is a sys-
tematic methodology that
provides fundamental un-
derstanding of energy utili-
zation within the process
and employs this under-
standing for identifying en-
ergy targets and optimizing
heat-recovery and energy-
utility systems. Numerous
articles on energy integra-
tion have been
published.'"e.3'4] Of particu-
lar importance are the ther-

Graduate Education

Process sinks include reactors, heaters/coolers,
biotreatment facilities, and discharge media. Streams leav-
ing the sinks become, in turn, sources. Therefore, sinks are
also generators of the targeted species. Each sink/generator
may be manipulated via design and/or operating changes to
affect the flowrate and composition of what each sink/
generator accepts and discharges.
In general, sources must be prepared for the sinks through
segregation and separation via a waste-interception net-
work (WIN). Effective pollution prevention can be achieved
by a combination of stream segregation, mixing, intercep-
tion, recycle from sources to sinks (with or without inter-
ception), and sink/generator manipulation. Therefore, is-
sues such as source reduction and recycle/reuse can be
addressed simultaneously. The following sections summa-
rize these concepts.

Segregation simply refers to avoiding the mixing of
streams. In many cases, segregating waste streams at the
source renders several streams environmentally acceptable
and hence reduces the pollution-prevention cost. Further-
more, segregating streams with different compositions
avoids unnecessary dilution of streams. This reduces the
cost of removing the pollutant from the more concentrated
streams. It may also provide composition levels that allow
the streams to be recycled directly to process units.

Recycle refers to the utilization of a pollutant-laden stream
(a source) in a process unit (a sink). Each sink has a
number of constraints on the characteristics (e.g., flowrate
and composition) of feed that it can process. If a source
satisfies these constraints it may be directly recycled to or
reused in the sink. But if the source violates these con-
straints, segregation, mixing, and/or interception may be
used to prepare the stream for recycle.

Interception denotes the utilization of separation unit
operations to adjust the composition of the pollutant-laden
streams to make them acceptable for sinks. These separa-
tions may be induced by the use of mass-separating agents
(MSAs) and/or energy separating agents (ESAs). The
design of these interception networks can be handled
using a process integration technique referred to as the
mass-pinch analysis.[7-10]

Sink/generator manipulation
Sink/generator manipulation involves design or operat-
ing changes that alter the flowrate or composition of pol-
lutant-laden streams entering or leaving the process units.

Course Outline

Overview of Process Integration and Pollution Prevention (2 lectures)

U Extracting global insights from a flowsheet
BE Branches of process integration
U Pollution prevention hierarchy

The Role of Simulation in Preventing Pollution (1 lecture)

U Process simulation versus process synthesis
U Basic modeling concepts

Mass Pinch Analysis (6 lectures)

E Development of waste composite representations
U Development of lean composite streams
E Simultaneous screening of mass-separating agents
Graphical mass-pinch analysis
E Optimization-based mass-pinch analysis

Graphical Techniques for Mass Integration (6 lectures)

B Segregation, mixing, recycle, interception, and unit manipulation
U Establishing pollution-prevention targets from a breadth analysis
E Waste source-sink mapping analysis
U Systematic development of solutions to meet the target
0 Integration of pollution prevention with other process objectives

Energy Effects in Preventing Pollution (5 lectures)

U Thermal-pinch analysis
U Combined heat and mass exchange networks
E Energy-induced systems

Integration of Process Synthesis and Process Simulations (3 lectures)

Mathematical programming and optimization
Path diagrams for tracking pollutants
Integration of path and pinch diagrams

Reactive Systems (5 lectures)

Establishing thermodynamic targets for reactive systems
Reactive mass-pinch analysis
Synthesis of environmentally acceptable reactions
Design of benign species

Chemical Engineering Education

Graduate Education

These measures include temperature/pressure changes, unit
replacement, catalyst alteration, feedstock substitution,
reaction-path changes, reaction system modification, and
solvent substitution.

The course was first offered at Auburn University in the
fall of 1990 as part of the senior-level design sequence and
graduate-level electives. Since then, the course has been
refined and updated. On the average, it is attended by
about seventy-five students per year. Table 1 shows the
course outline.
A related course is also offered at the University of Vir-
ginia. The syllabus for that course is available in a recent
CEE article. 21

In order to illustrate the applicability of the techniques
taught in the course, the students are required to address a
pollution prevention problem in an integrated manner. The
projects are chosen from literature or from actual operating
facilities. Recent projects include
Water reduction in a pulp and paper mill
Desulfurization in a coal-processing facility
Metal removal from a copper-etching process
Reduction of VOC emissions for a vinyl-chloride process
Sweetening of coke-oven gas facility
Dephenolization in an oil refinery
Reduction of solvent loss from a coating facility
Benzene removal in a polymer facility

The main reference for the course is the textbook Pollu-
tion Prevention through Process Integration: Systematic
Design Tools.11 Other useful books have also been recently
published.1'"12] There are also many archival papers that
provide design techniques and case studies that can be used
as term projects for the course (see References 1, 2, and 5 for
a citation of available literature).

The course provides a unique approach to pollution pre-
vention-one that is rooted in fundamentals, is generally
applicable, and is not left to subjective opinion and arbitrary
solutions. It emphasizes the unique role of chemical engi-
neering in preventing pollution. The course provides a thor-
ough and comprehensive analysis of both mass and energy
flows in processes and results not only in pollution preven-
tion but also in yield improvement, energy use reduction,
solvent recovery, and cost reduction. These are all compat-
Fall 1998

This paper is based on a
senior-level/first-year graduate-level
course... that provides a unique
framework for teaching systematic
techniques for addressing
pollution prevention in
the process industries.

ible goals when undertaken properly and when the inte-
grated nature of the process is well understood.
Due to its systematic nature, the course is well suited for
educational purposes and it also prepares the students to
tackle real problems after graduation. This has been vali-
dated by the feedback provided by hundreds of under-
graduate and graduate students who took the course and
later indicated that it had a major impact on their first
industrial job.

1. El-Halwagi, M.M., Pollution Prevention through Process In-
tegration: Systematic Design Tools, Academic Press, San
Diego, CA (1997)
2. Carta, G., M.D. LeVan, H.D. Spriggs, G.A. Cleotelis, II, and
J.E. Ryan, Jr., "Process Integration and Industrial Pollu-
tion Prevention," Chem. Eng. Ed., 31(4), 242 (1997)
3. Linnhoff, B., "Use Pinch Analysis to Knock Down Capital
Costs and Emissions," Chem. Eng. Prog., 90, 32 (1994)
4. Shenoy, U.V., Heat Exchange Network Synthesis: Process
Optimization by Energy and Resource Analysis, Gulf Pub-
lishing Company, Houston, TX (1995)
5. El-Hawagi, M.M., and H.D. Spriggs, "Solve Design Puzzles
with Mass Integration," Chem. Eng. Prog., p. 25, August
6. El-Halwagi, M.M., A.A. Hamad, and G.W. Garrison, "Syn-
thesis of Waste Interception and Allocation Networks,"
AIChE J., 42(11), 3087 (1996)
7. El-Halwagi, M.M., V. Manousiouthakis, "Synthesis of Mass
Exchange Networks," AIChE J., 35(8), 1233 (1989)
8. El-Halwagi, M.M., and B.K. Srinivas, "Synthesis of Reac-
tive Mass-Exchange Networks," Chem. Eng. Sci., 47(8), 2113
9. El-Halwagi, M.M., B.K. Srinivas, and R.F. Dunn, "Synthe-
sis of Optimal Heat Induced Separation Networks," Chem.
Eng. Sci., 50(1), 81 (1995)
10. El-Halwagi, M.M., "Optimal Design of Membrane Hybrid
Systems for Waste Reduction ," Sep. Sci. Tech., 28(1-3), 283
11. Allen, D.T., and K.S. Rosselot, Pollution Prevention for
Chemical Processes, John Wiley and Sons, New York, NY
12. El-Halwagi, M.M., and D.M. Petrides, eds., "Pollution Pre-
vention via Process and Product Modification," AIChE Symp.
Ser., 90(303), AIChE, New York, NY (1995) O

Graduate Education



in the Department of Chemical Engineering

at the National University of Singapore

National University of Singapore 10 Kent Ridge Crescent Singapore 119260

During the past three decades, there has been rapid
industrial and economic development in East Asia
and Southeast Asia. This development has now
reached the stage where it is imperative for nations in the
region to begin serious and concerted efforts to protect the
environment. The environmental issues in these regions are
deemed sufficiently important to have warranted special cov-
erage in Environmental Science and Technology."'
Efforts to address environmental concerns in Singapore
not only include new legislation and strict enforcement of
laws and regulations resulting from the legislation, but
they also cover a wide variety of activities and programs,
including education and training of skilled manpower in
environmental science and engineering, promotion of re-
search and development work in relevant disciplines, and
manufacturing and production of devices, products, and
software that are important to protect and preserve the
The interests and activities aimed at environmental protec-
tion have also had the effect of creating significant opportu-
nities for manufacturing interests. The National Science and
Technology Board (NSTB) of Singapore estimates that over
the next five years, Asian countries will be spending $10
billion annually on environmental equipment, systems, and
services. In line with this development, NSTB has desig-
nated environmental technology as one of the key technolo-
gies for Singapore's economic development of the next de-
cade and has set up an Environmental Technology Institute
to help propel Singapore's environmental technology indus-
try to a higher level of competitiveness.
In its educational functions, the National University of
Singapore (NUS) has long recognized the relevance and
importance of the environment. Teaching and research ac-

tivities dealing with various aspects of the environment can
be found in many of its academic units. In view of the
heightened interest in environmental technology, the Uni-
versity established a formal program in environmental engi-
neering in 1995 at the postgraduate level, which was fol-
lowed by an undergraduate program in 1997.
Before describing the features of this postgraduate pro-
gram, it may be useful to highlight the rationale for launch-
ing an environmental engineering program at the postgradu-
ate level with the Department of Chemical Engineering as its
academic home. The program was designed to be a
coursework-centered program that allowed candidates to build
on their prior educational background and to acquire skills
for the solution of advanced environmental engineering prob-
lems. It had to be sufficiently flexible to accommodate stu-

K.G. Neoh is Associate Professor and Head of the Department of Chemi-
cal and Environmental Engineering. She received her degrees (BS and
ScD) in Chemical Engineering from MIT Her research interests include
catalytic microparticles, electroactive polymers, surface functionalization,
and stability enhancement of polymeric materials.
Y.P. Ting is a Senior Lecturer. He received his BSc in Chemical Engi-
neering from the University of Manchester Institute of Science and Tech-
nology in the UK, and his PhD from Monash University, Australia. His
research interests are in the area of biosorption and bioremediation.
R.R. Stanforth is a Senior Lecturer and conducts research in stabilization
of heavy metals and adsorption of metals and anions. He received his BS
in chemistry from Heidelburg College and his MS and PhD in Water
Chemistry from the University of Wisconsin-Madison. He has developed
undergraduate and postgraduate courses in environmental chemistry and
set up the analytical facilities in the environmental engineering laborato-
C. Tien is Director of the Environmental Engineering Program and the
Environmental Technology Enterprise. He received his BS from National
Taiwan University and his MS and PhD from Kansas State University and
Northwestern University, respectively. His research interests are in deep
bed filtration, aerosol deposition in granular and fibrous media, adsorp-
tion, and cake formation in fluid/solid separation.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Graduate Education

dents from different back-
grounds and to provide prac-
ticing engineers with an op-
portunity to upgrade and en-
hance their technical qualifi-
cation, enabling them to par-
ticipate and contribute to
Singapore's push to de-
velop an environmentally
based industry.
Historically, environmen-
tal engineering has been
linked to civil engineering (or
its sub-specialty, sanitary en-
gineering), reflecting the
early need for water and
wastewater treatment. But
many of the tasks facing en-
vironmental engineers in the
developing Asian countries

[The program] had to be sufficiently flexible to
accommodate students from different backgrounds
and to provide practicing engineers with an opportunity to
upgrade and enhance their technical qualification ...

Course Requirements for
Postgraduate Diploma and MSc Degree

Option 1

Postgraduate Diploma

4 Foundation Subjects
4 Elective Subjects

Option 2

4 Foundation Subjects
2 Elective Subjects

Minor Project

Master of Science Degree

are extremely diverse and go
far beyond the traditional
"end-of-the-pipe" approach.
Additional programs to complement the sanitary engineer-
ing approach were needed; programs that covered areas such
as process modification and waste minimization to avoid
pollution problems, environmental chemistry and microbiol-
ogy, and transport processes.'2' These topics fit more eas-
ily into a chemical engineering program, with its empha-
sis on industrial process design and on chemistry, than
into a civil engineering program; therefore the environ-
mental engineering program is based in the Chemical
Engineering Department at NUS.

In establishing the environmental engineering program,
we were cognizant of the fact that environmental engineer-
ing is a multidisciplinary subject of study and that the pro-
gram must be accessible to candidates with varied educa-
tional and professional backgrounds. We established a two-
tiered program that offers either a Postgraduate Diploma or a
Master of Science (MSc) degree, and which can be pursued
on either a full-time or a part-time basis.
In order to be considered for admission into the program,
applicants must at least possess a Bachelor's degree in sci-
ence or engineering. Generally, the selection is based on the
applicant's undergraduate academic record and the relevancy
of his/her work experience (if any). Students may also be
required to undergo an interview. The selection process for
the MSc applicants is more stringent than for the Diploma
applicants. Postgraduate Diploma students who have per-
formed well may upgrade to the MSc program.
Fall 1998

4 Foundation Subjects
6 Elective Subjects

4 Foundation Subjects
2 Elective Subjects

Major Project

The course require-
ments for the program
are given in Table 1.
The foundation sub-
jects are the more fun-
damental subjects and
serve as a mechanism
for filling in gaps in
the background of the
students from a vari-
ety of disciplines who
are attracted to this
program. Up to two of
these foundation sub-
jects may be waived
for students who have
demonstrated profi-
ciency in these areas,
with elective subjects
replacing them. The
elective subjects are
the more specialized
subjects and are fur-

their divided into two groups: Group 1 includes courses on
advanced engineering science relevant to environmental en-
gineering, and processes and devices used in pollution con-
trol and abatement; Group 2 covers more diverse topics such
as risk, impact assessment, and other courses related to envi-
ronmental engineering that are offered in the postgraduate
chemical and civil engineering programs.
For the MSc environmental engineering students, under
Option 1 at least three of the elective subjects must be from
Group 1, while under Option 2, at least one elective subject
must be from Group 1. For the Postgraduate Diploma stu-
dents, the corresponding number of subjects from Group 1
are two and one, respectively. The project option (in lieu of
four subjects for MSc students and two subjects for Post-
graduate Diploma students) gives the students an opportu-
nity to conduct research related to their interests.
The classes are conducted from 5:30 pm to 8:30 pm,
giving practicing engineers and technologists an opportunity
to upgrade and enhance their technical qualifications. Each
semester lasts thirteen weeks. Up to 40% of the final grade
may be based on continuous assessment, which comprises
homework, quizzes, term papers, etc., with the final exami-
nation accounting for the balance.

The list of subjects offered in the postgraduate environ-
mental engineering program are given in Table 2. The foun-
dation subjects are formulated to meet the challenge of teach-
ing advanced topics in environmental engineering to stu-

E d u c a ti~eo n -e -- -~

dents with different backgrounds. Many of the students
are practicing engineers/technologists who may have
some relevant practical experience, but who are lack-
ing in the fundamentals. The foundation courses are
briefly described in Table 3.
The students are strongly encouraged to take the
foundation subjects in the early part of their candida-
ture before embarking on the more specialized elec-
tives. Some Group 1 electives (Aquatic Chemistry, At-
mospheric Chemistry, and Biological Waste Treatment)
require background knowledge covered in the Basic
Environmental Science foundation subject. In the three
remaining Group 1 electives (Water Pollution Control
Technology, Air Pollution Control Technology, and
Solid Waste Management), the students are assumed to
possess sufficient knowledge of mathematical and en-
gineering principles to understand the basic theory and
unit operations of the treatment processes covered in
these electives. For example, in Air Pollution Control
Technology, the basic designs of the various control
equipment for particulates (cyclone, venturi scrubber,
etc.) and gaseous pollutants (scrubbers, packed bed
adsorbers, catalytic reactors, etc.) are studied. In these
electives, the philosophy that the best method of pollu-
tion control is the minimization of pollutant formation
is also emphasized. For example, on the topic of the
control of NO, emission, the various mechanisms of
NOx formation during combustion and the factors
affecting the formation processes are highlighted.
The use of combustion technologies such as flue gas
recirculation, off-stoichiometric or stage combus-
tion to reduce NO, formation are discussed in addi-
tion to the post-combustion control technologies that
rely on the introduction of reactants to destroy the
NO, in the flue gas. 3-51
The Group 2 electives (Table 2) provide the students
with an opportunity to acquire more in-depth knowl-
edge on chemical engineering topics such as reaction
engineering, separation technology, process control, and
modeling, as well as topics more closely related to civil
and sanitary engineering. In addition, two electives,
Quantified Risk Analysis and Environmental Impact
Assessment and Auditing, highlight issues of current
interest and importance. In Quantified Risk Analysis,
the major hazard control legislation, practices and poli-
cies, hazard identification techniques and evaluations,
risk assessment, and various dispersion and explosion
models are covered. The Environmental Impact As-
sessment and Auditing elective covers impact analysis
of land use and erosion, noise, air and water quality,
vegetation and wildlife, as well as socio-economic
and geological impact analysis and environmental

Subjects in the Postgraduate
Environmental Engineering Program

Foundation Subjects
* Basic Environmental Sciences
* Mathematical Methods for Environmental Engineering
* Physical Principles of Environmental Engineering
* Process Engineering Design Principles

Elective Subjects
Group I
* Aquatic Chemistry
* Water Pollution Control Technology
* Atmospheric Chemistry
* Air Pollution Control Technology
* Biological Waste Treatment
* Solid Waste Management

Group 2
* Chemical and Biochemical Reaction Engineering
Instrumentation and Process Control
SQuantified Risk Analysis
* Environmental Impact Assessment and Auditing
* Membrane Separation Technology
SAdvanced Separation Processes Offere
SPrinciples of Adsorption and Adsorption Processes Chemi
* Advanced Reaction Engineering Postgr
SProcess Modeling and Optimization

d in the
cal Engineering
aduate Program

ed in the
graduate Program

* Water Quality Management
* Industrial Wastewater Control
* Toxic and Hazardous Waste Management
* Advanced Hydraulics
* Water Resource Systems Analysis
* Environmental Health Engineering
*Urban Environmental Management

- Civil

Brief Descriptions of Foundation Subjects

Basic Environmental Science
Basic environmental chemistry, including concepts from general and physical
chemistry, organic chemistry, and biochemistry Fundamentals of environmental
microbiology, including microbiological principles and biogeochemical cycles *
Illustrations from water and wastewater treatment.
Mathematical Methods for Environmental Engineering
Linear algebra Solution of systems of linear and non-linear equations Numeri-
cal solution of initial value and boundary value ordinary differential equations *
Parameter estimation and regression analysis Introduction to optimization tech-
niques Statistical methods for process analysis Examples and applications.
Physical Principles of Environmental Engineering
Fundamentals and applications of mass, momentum, and heat transport in envi-
ronmental engineering Advection, diffusion, dispersion, settling, and surface
transfer in air and water Applications in natural environment and treatment
systems Quantitative applications.
Process Engineering Design Principles
Introduction and classification of processes Steps in process design Material
balances involving phase equilibria, chemical reaction, and recycle Design of
simple stagewise processes Energy balances Introduction to and use of
flowsheeting packages.

Chemical Engineering Education

Graduate Education

auditing. In addition, ISO 14000 issues are also covered
in this elective.
Currently, only a minority of the students take up the
project option, since most of the students who are working
full-time have difficulty conducting a project (especially an
experimental one) and meeting with the staff members dur-
ing office hours. Some examples of the MSc projects cur-

Examples of MSc Projects

El Measurement of Chemical Species in Rain Water
El Synthesis of Spinel Oxides for N,O Decomposition
E Long-term Stability of Treated Lead-Contaminated Soil
El Nutrient Removal from Domestic Sewage Using a Modified
Sequential Batch Reactor
El Enzyme Mimic Catalysis : Decomposition of Pollutants in
Wastewater Treatment
E Modeling for the Prediction of Flue Gas H,S/SO,

Students' Choice of Candidature

Postgraduate Diploma Master of Science Degree
Part-Time Full-Time Part-Time Full-Time
26 7 69 8
-33- 77

Students' Educational Background
and Working Experience

First Degree
Science 45
Engineering 65

Working Experience
(a) Number of Years
<1 13
1-3 28
4-6 16
>6 53

(b) Employment Category
Petroleum/Petrochemicals 11
Specialty Chemicals/Pharmaceuticals 12
Project Engineering/Construction 18
Environmental Services 9
Electronics 8
Commercial Laboratories 6
Government/Statutory Board 21
Foreign Companies 6
Not Employed/Fresh Graduates 8
Others 11

Fall 1998

rently in progress are shown in Table 4.
As illustrated by the projects in Table 4, there is a strong
chemistry/chemical engineering component in these projects.
The projects in the postgraduate program are supervised by
staff from the Chemical Engineering Department (which
include both chemical engineering and environmental engi-
neering staff), Civil Engineering Department, and the De-
partment of Microbiology.

The majority of the students in the environmental engi-
neering program are practicing professionals who enrolled
in the program on a part-time basis (see Table 5). About 60%
of the students possess a Bachelor's degree in engineering,
with the remaining having a degree in science (see Table 6).
Few of the students embark on the postgraduate program
directly after their first degree. About half of the students
have more than six years of working experience and had
worked in a wide range of sectors reflecting the industries in
Singapore. There are a few foreign students (under the cat-
egories "Foreign Companies" and "Not Employed/Fresh
Graduates"), mainly from Indonesia, Myanmar, India, and
China. So far, thirteen students have graduated from this
program; five of them were full-time students while the
remainder were enrolled on a part-time basis. These full-
time students found employment upon graduation in envi-
ronmental consulting and engineering companies. The part-
time students continued with their jobs upon graduation.

At the National University of Singapore, the choice of the
Department of Chemical Engineering to host the environ-
mental engineering program reflects our philosophy that
environmental engineers need knowledge of process engi-
neering principles and applications as well as chemistry and
microbiology. Since the establishment of the postgraduate
environmental engineering program in 1995, thirteen stu-
dents have graduated and 110 students are currently en-
rolled. The program is sufficiently flexible to accommodate
students with varied backgrounds due to its unique features:
a two-tiered (Postgraduate Diploma, MSc) program that can
be pursued on either a full-time or a part-time basis and the
provision of foundation courses to make up for any deficien-
cies in the students' background.
The rapid growth of this program and its popularity among
working professionals demonstrate that there is a growing
and previously unmet need for environmental engineering
training in Singapore. In view of the success of the post-
graduate program, the department has launched an under-
graduate degree program in environmental engineering and
the first students will graduate in the year 2000. The under-
Continued on page 259.

Graduate Education




University of Utah Salt Lake City, UT 84112-9203

S ince the early 1950s, the University of Utah has been
a place where an industrially oriented chemical engi-
neering education has been emphasized through its
pilot-scale unit operations laboratory. At that time, experi-
ments were performed using pneumatic controllers with strip
chart recorders attached. Between the 1980s and the mid
1990s, our laboratory was finally instrumented with several
roll-around carts with a computer and data acquisition equip-
ment that could be attached to various unit operations for
data logging purposes. Process control experiments at this
time were instrumented with electronic PID controllers and
strip chart recorders.
Today, the roll-around workstations remain an important
fallback solution in our current laboratory operation. At
some schools (the University of Colorado Integrated Teach-
ing and Learning Laboratory is probably the best example'"),
the flexibility of the well-designed and instrumented roll-
around workstation remain the primary method of experi-
ments instrumentation. We, however, decided that our
large process units are better served by developing a
permanent networked distributed control system (DCS),
the solution found in industry.

A number of chemical engineering departments have se-
lected a proprietary DCS for their unit operations. For in-
stance, laboratories at the Arizona State and Michigan Tech-
nological Universities have implemented the Honeywell TDC
3000 distribute control systems.[2,'3 In our case, we found
that the industrial-strength open-architecture PC-based
DCS is well suited for the needs and environment of the
university laboratory.
Our unit operations laboratory is instrumented with an
industrial Opto 22 DCS networked to both PCs running

Windows 95 and Unix workstations. The combination of the
Opto 22 open architecture, multithread digital controllers,
modular I/O racks (bricks), and support of all major net-
working and communication protocols provides us with
needed functionality, flexibility and scalability. The costs of
implementing such a system proved to be reasonable com-
pared to the proprietary DCS technology.
Opto 22 provides both hardware and software for control
and data acquisition purposes. The hardware consists of I/O
units and controllers. At present, our DCS has 450 input-
output points around a four-floor 10,000 ft2 laboratory that
are concentrated in three smart bricks that are nodes on the
university network. I/O units are available in both high- and
low-density bricks. The high-density bricks accept multiple
I/O boards, each with 16 inputs or outputs. The high-density
bricks do not allow input of more than two types of different
analog signals and have limited scaling capabilities, using
the brick's built-in processor.
The low-density bricks accept a total of 16 inputs/outputs,
but allow a mix-and-match of input-output analog and digi-
tal signals of all types. Each brick has a co-processor that
allows scaling at the brick, plus the low-level implementa-

Mikhail Skliar is Assistant Professor of Chemical and Fuels Engineer-
ing at the University of Utah. He received his MSEE degree from
Odessa Technical University, his Candidate of Science degree in Con-
trol of Technical Systems from Kiev Polytechnic Institute, and his PhD
from the University of Colorado, Boulder. His research interests include
process control and identification, and inverse problems.
Jesse W. Price is a graduate student in chemical engineering at the
University of Utah. He is currently pursuing a MSc degree, having
received his BSc in Chemical Engineering from the University of Utah in
1996. His research is centered on crude fractionation processes, includ-
ing simulation and inferential control.
Christopher Tyler (not pictured) is a graduate of the University of Utah
in chemical and fuels engineering, with an emphasis in mathematics
and applied science. He helped build, debug, and design the department's
Opto 22-based DCS. He is now attending graduate school at the Univer-
sity of Minnesota, seeking his PhD in chemical engineering.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Graduate Education

tion of the PID control algo-
rithm at the brick. The on- The combination
brick calculation saves the architecture, multith
controller from complex cal- modular I/O racks (bi
culations and decreases the major networking and
network data traffic, liberat- provides us with i
ing system resources for al- flexibility a
ternative tasks and at the
same time providing an ad-
ditional robustness and se- I/O Brick I/O Brick
No. 1 No.2
curity to the DCS operation. No 1 No. 2
If some I/Os and control-
ler communication fail, the
brick co-processor will RS 48
maintain the functionality of
all PID control loops with
unaffected inputs and out-
puts. The digital controller
is in essence a hardened in-
dustrial microcomputer con-
trolled by a specialized real- Remote Use
time operating system and
networked to PC hosts. The
controller's processor RemoteUse
handles all calculations and
interactions between differ-
ent I/O bricks. A controller
can support 255 I/O units per
communication port.
New I/O units can be Figure 1. D
added to the controller at any
time. In our case, I/O units are distributed around the labora-
tory and are specific to the particular piece of equipment. All
I/O are connected to a controller, which simultaneously runs
multiple control strategies for several different experiments.
The control server routes the acquired data to an appropriate
computer used as a man-machine interface (MMI) for a
particular unit operation.
The Opto 22 controller interacts with PCs throughout the
laboratory and faculty's offices used for MMI. The design
and testing of the new or modified control strategy is accom-
plished on the host computer, referred to as a control server.
The network connection between the server and unit-opera-
tion-specific MMI computers is used to provide online infor-
mation about the process to the student operators. Figure 1 is
used to illustrate hardware architecture used in our labora-
tory. It shows multiple I/Os that send and receive data from
the controller. The control server is used to develop, test, and
download control strategies and to direct the flow of infor-
mation to and from the controller. Multiple PCs are used as
an interface with the student operators. The operator is pre-
sented with the real time and trended process information.
Fall 1998


CS a

Based on this information
he Opto 22 open and the goals of the experi-
d digital controllers, ment, the operator changes
s), and support of all operating and control param-
imunication protocols eters of the process.
ed functionality, Three software packages
scalability. (OptoControl, OptoDisplay,

and OptoServer) provide the
graphical interface for de-
/O Bnck Third Party velopment of the control
No. 3 Devices
S. strategies and MMIs and for
network communication.
OptoControl combines a
RS 232 Third Party graphical flowchart pro-
i gramming language with in-
,to 22 tegrated debugging and de-
trollers velopment tools. It uses
command and logic blocks
to implement a control
strategy. The flowchart
Control Server code of the control strat-
egy is fairly easy to write
and understand.
TCP/IP Ethernet OptoControl allows sub-
routine programming using
both user-defined and li-
MMI .. I brary subroutines. The li-
brary subroutines include
many frequently used algo-
rchitecture. rithms, including PID con-
trol, cascade control, and
gain scheduling. Once developed and tested by executing
the algorithm one step at a time, the control strategy is
compiled and downloaded to the digital controller where
it is executed independently from the host computer.
Advanced control algorithms can be implemented on the
PC host or networked Unix workstation by obtaining
real-time process variables from the controller to calcu-
late the desired control moves and downloading the cal-
culated setpoints back to the controller.
OptoDisplay is a graphical interfacing tool. It allows the
creation of MMIs for interactive control and on-line data
acquisition using live graphical displays. It is simple in
structure and similar to many commercial MMI packages. It
allows students to have exposure to industrial MMI tools.
OptoServer provides for a networking environment
between control, PCs, and the outside world. With
OptoServer, many PCs can access data from one control-
ler, as shown in Figure 1.
The implementation of the laboratory DCS required sig-
nificant initial commitment of time and effort from several

Graduate Education

faculty members and technical personnel. But once imple-
mented, the DCS allows us to have almost infinite
expandability with only incremental cost and effort. Many
of the enhancements and
extensions to the existing
system are made by dedi-
cated graduate and under-
graduate students of this
department, done for aca-
demic credit.

An important feature of
the implemented system is
its open architecture: Any
third party hardware that
uses RS232 or RS485
communication can be
used with Opto 22 DCS.
For example, we were able
to integrate Valtec smart
valve and stand-alone

Figure 2. Inline pH co

Honeywell PID controllers
with relative ease. The typical way to integrate an existing
process unit with the DCS is to simply network the existing
control and instrumentation equipment with Opto 22 and
to develop the new MMI. New I/O can be added for any
equipment as needed.
The process control curriculum has benefited the most
from the introduction of the DCS in the unit operations
laboratory. In addition to the dedicated control experiments,
every process unit integrated with the DCS with a little
imagination becomes a control experiment. The students are
no longer confined to the boundaries of rigidly packaged
control experiments, but are free to implement any advanced
control method they see fit.

The first graduate process control course (also open to
seniors) in our department covers a broad range of time and
frequency domain analysis and synthesis methods for linear
single-input-single-output (SISO) and multivariable control
systems. The course emphasis is on the design of robust
controllers and performance limitations and design tradeoffs
imposed by difficult dynamics, constraints, and process in-
teractions. The course textbook141 gives an excellent cover-
age of many topics traditionally considered difficult for
chemical engineering students and is supplemented by
lecture notes to give an extended coverage of time do-
main methods and practical implementations of the model
predictive control.
As a part of the course, students are required to complete a

substantial project that until recently involved confirmation
of the theoretical and simulation results from a recent paper
published in one of the archival control journals. To encour-
age independent work
by students, we usually
DCS require that the chosen
Controller --. project does not en-
pH signal tirely rely on the mate-
Position rial covered in lectures
and the textbook.
In-line pH probe Last year, for the first
time, we added an ex-
perimental component
Rotameters to the graduate control
course. Now students
have a choice of select-
ing a theoretical/compu-
tational project or de-
ntrol experiment. signing and implement-
ing in a laboratory set-
ting one of the advanced
control methods. The ease of implementation of the control
strategies using the DCS makes it possible to take a project
from proposal to implementation in a short period of time.
The students are strongly encouraged to select an experi-
mental project. If the SISO system is selected, the students
are asked to give generalization on multivariable systems.

Our initial experience shows that experimental projects,
though more time consuming, give our students valuable
experience in bridging abstract theoretical concepts and prac-
tical process control.

As an illustration, consider three student projects carried
out on the same piece of equipment-the inline pH control
apparatus, a unit operation used on many effluent plant
streams. (Additional examples of the students' projects can
be found in Reference 5.)

Inline pH control using nonlinear gain scheduling The
objective of the first project was to design and implement a
controller for the inline neutralization process, which gives
an adequate performance for the range of the pH setpoints
from pH 5 to pH 9. The experimental setup is shown in
Figure 2. A weak solution of sodium hydroxide (pH 10) is
used to neutralize a weak solution of hydrogen chloride (pH
4). The controlled variable is the pH of the effluent stream,
controlled by closing or opening the control valve on the
base stream. The flow rate of the acid is a disturbance to the
After initial experiments, students quickly discovered that
Chemical Engineering Education


Graduate Education

due to the process nonlinearity, an adequate tuning of a
single PID controller over five orders of magnitude of the
proton concentration is not possible; there always will be a
range of pH where the response will be either too oscillatory
or too sluggish, depending on the pH value for which the
PID controller was tuned. In order to get a better understand-
ing of the process, students performed a number of experi-
ments. As an initial estimate of process nonlinearity, the
students obtained titration curves off-line using bench-top
experiments. But on-line investigations suggested that the
process peculiarities, such as nonlinear (equal-percentage)
characteristics of the control valve and slight flow interac-
tions, made the off-line titration ineffective in characterizing
the process. The students then modified the Opto 22 code in
order to obtain the online titration curve by continuously
ramping the control valve over a set interval of time. As a
result of online titration experiments, they discovered that
the control valve nonlinearity has a significant effect on the
process gain as a function of pH.
After the process was understood, the students selected
three values of pH (pH 5, 7.5, and 9) for which they obtained
three sets of the tuning parameters for the PI algorithm. The
students' linear gain-schedul-
ing algorithm interpolated the
values of the controller gain
Negative Set Z
and the integral reset time lin-
early between the previously 1 ----
determined sets of tuning pa-
rameters. The results of the
closed-loop step tests showed
good controller performance
over the required range of ef- 0
fluent stream pH.
The students decided to take
the project one step further. Figure 3. Fuzzy sets used
Since they have information
about the process gain as a
function of the pH, they thought that it should be fairly easy
to implement a controller with nonlinear gain that changes
with pH like an inverse of the online titration. The students
have correctly concluded that once implemented, such non-
linear gain scheduling will in effect linearize the process,
allowing for better closed-loop controller performance. The
students have implemented an exponential gain scheduling
for the PI controller. Closed-loop experiments have shown
that the exponentially gain-scheduling controller exhibits
superior performance.
This simple experimental project carried out within sev-
eral short weeks exposed students to all essential steps of
the control-system life cycle: from problem understand-
ing, to process characterization, to algorithm develop-
ment, all the way to the implementation and testing of
Fall 1998

ero S



in t

the developed control system.

Adaptive model-based controller for inline pH control *
A student who participated in the just-described laboratory
project decided, upon consultation with faculty, to follow up
on the obtained results and to design an adaptive controller
to control the pH of a weak-acid-strong-base system. This
system is not easily modeled and is the traditional test-bed
for adaptive control algorithms. The student modeled the
process as a dynamic system with nonlinear static pH-
dependent process gain, the functional form of which is
determined from theoretical considerations.'61 The devel-
oped adaptive algorithm has an internal model-control
structure with the model gain updates based on on-line
least-squares regression.
The developed adaptive controller became the topic of the
student's senior honors project. The student himself became
interested in process control and is likely to choose control
theory and applications as a subject of his PhD research.

Fuzzy-logic controller for inline pH mixing The next
example involved implementation of a fuzzy-logic control-
ler (FLC) on the pH neutraliza-
tion experiment. A fuzzy-logic
controller uses imprecise
et Positive Set knowledge about the process
/-- to determine the appropriate
/ control actions. Instead of a
/ mathematical model of the pro-
,/ cess, the control algorithm re-
/ lies on a set of fuzzy IF-THEN
2 rules that describe the physics
of the process in simple terms
or and attempts to capture the
knowledge and experience of
he fuzzy-logic controller skilled operators. Overlapping,
triangular fuzzy-membership
functions are typically used to classify the input variables
based on their membership value in each fuzzy set. The
membership value of each input variable is then used to
activate multiple rules, and a de-fuzzification process deter-
mines a crisp value for the output variable.
The graduate student implemented a simplified FLC algo-
rithm, largely based on Reference 7. The error between the
pH setpoint and the measured pH value is classified in three
fuzzy sets: negative, zero, and positive, as shown in Figure
3. A rules-base included three simple rules, and a centroid
weighting method was used to determine the desired control
move sent to the valve. Implementation of the FLC algo-
rithm within the Opto 22 system involved translating the
fuzzy algorithm into a sequence of condition and action
blocks using the flowsheet programming language. The main

Graduate Education

flowsheet developed by the student is shown in Figure 4,
where pt is the membership function.
As expected for this highly nonlinear process, the de-
signed simple fuzzy-logic controller resulted in a poor closed-
loop performance. Figure 5 shows sluggish performance of
the FLC at higher pH and oscillations in the process re-
sponse at lower pH. But the project was considered to be a
success since its real value was an exceptional educational
experience for the student.

The new level of availability of the laboratory experiments
brought forth by the DCS technology begins to change the
paradigm of process control teaching by emphasizing project-

Receive signal from pH electrode

Convert Signal to pH Value

Calculate Error
Set point Process Value

Calculate the Set Membership
of the Error: RNEG' ZlERo' POS

Apply Centroid Method to Calculate
Output using Rule Base (Valve Action)

Convert Valve Action to Signal

Send signal to valve

Figure 4. Flowsheet implementation of FLC code.

based learning. Graduate and undergraduate students often
view process control courses as an abstract exercise in math-
ematics and usually have difficulty applying what they have
learned to actual processes.
The laboratory control experiments allow us to bridge
theory and practice of process control. After gaining experi-
ence with control topics using Matlab computer simulations
for idealized processes, students are typically surprised to
discover during experiments that such phenomena as pro-
cess nonlinearity, large time delays, measurement noise, and
nonlinearity of the final elements (such as valve hysteresis)

Setpoint Change from 7 to 5

jP 6

Z., Z -. N p. N '_. N, 1,3. b, r6- A. A".

Time, sec

Sequence of Setpoint Changes


4- IZP NSI IP +'NbS-434P0

U' *' *"_ ..0 0
o,. ,. ,- S o
Time, sec

Figure 5. Response of the FLC to set point change.

DCS-Integrated Experiments

Control Experiments Data Acquisition Only Ongoing Integration
Bubble-cap distillation Double pipe heat exchanger Glass-lined CSTR
pH control Shell and tube heat exchanger Vacuum oven (data acquisition)
Fermentation reactor Liquid flow bench
Batch distillation Fluidized beds
Liquid level control CSTR with multiple steady states
Multivariable control bench Different catalytic reactors
Water-heater control

Chemical Engineering Education

Graduate Education

substantially complicate the design and tuning of the
actual control loops. In general, the laboratory experi-
ments provide students with valuable experience with
physical systems and give them much needed practice in
applying theory to the real process.
The distributed control system in our laboratory creates a
dynamic environment in which control experiments are con-
stantly evolving. The incremental improvements and the
addition of new equipment (see Table 1) are accompanied
by long-term projects aimed at increasing the functionality
and availability of the experiments. The most important
enhancement that we are currently working on will allow the
instructors to run experiments online from the classroom
using the standard Ethernet connection to the controller or
the control server. Ideally, we would like to see not only
professors running experiments remotely, but also students
accessing the system from campus computer laboratories
and their homes to collect the necessary data for their home-
work assignments or the design projects.
In the future we plan to make our laboratory available for
the distant-learning students and for those using the Internet.
Our initial experience (during the first remote demonstration
of one of our control experiments,"s] the unattended control
laboratory was flooded), however, shows that before the
system can be made available to the students and the outside
world, the security of the systems must be substantially
improved and the equipment must be modified to make it
fool- and foul-play-proof before making it available for un-
supervised remote experimentation. Despite potential prob-
lems, we see enhanced networking as a way to maximize the
positive effect that the control experiments have on the edu-
cational experience of our students.
In addition to the effect on control courses, we are also
witnessing a positive impact on research in process control
theory since practical implementation of the theoretical re-
sults can be quickly and easily achieved. With the creation
of the Teaching Center for Advanced Control Technology
(TCACT), we are now in a position to enhance cooperation
with industry by providing flexible and customized training
of engineering and technical personnel in practical methods
of modern process control using our laboratory.

We want to acknowledge the importance of the close
cooperation with industry in designing and upgrading the
engineering laboratory. Our industrial partners helped us
design the relevant laboratory experiments and maintain the
cutting technological edge in teaching engineering students.
We are particularly thankful to Robert G. Engman (BSEE
University of Utah '53), President of Opto 22, for the
generous donation of Opto 22 hardware and software
used in our laboratory.
Fall 1998

1. Clough, D.E., "Bringing Active Learning into the Tradi-
tional Classroom: Teaching Process Control the Right Way,"
1998 ASEE Annual Conference and Exposition, Seattle, WA
2. Rivera, D.E., K.S. Jun, V.E. Sater, and M.K. Shetty, "Teach-
ing Process Dynamics and Control Using an Industrial-
Scale Real-Time Computing Environment," Comp. Appls.
in Eng. Ed., 4, 191 (1996)
3. Pintar, A.J., D.W. Caspary, T.B. Co., E.R. Fisher, and N.K.
Kim, "Process Simulation and Control Center: An Auto-
mated Pilot Plant Laboratory," ASEE Summer School of
Chemical Engineering Faculty, Snowbird, UT (1997)
4. Skogestad, S., and I. Postlethwaite, Multivariable Feedback
Control: Analysis and Design, Wiley, Chichester, England
5. Skliar, M., J.W. Price, C.A. Tyler, T.A. Ring, and G.A. Silcox.,
"Integration of Laboratory Experiments in the Chemical
Engineering Curriculum Using a Distributed Control Sys-
tem," accepted for publication in Comp. Appls. in Eng. Ed.
6. Gustafsson, T.K., B.O. Skrifvars, K.V. Sandstrom, and K.V.
Waller, "Modeling of pH for Control," Ind. Eng. Chem. Res.,
34,820 (1995)
7. Rhinehart, R.R., and P. Murugan, "Improve Process Control
Using Fuzzy Logic," Chem. Eng. Progress, 92(11), 60 (1996)
8. Skliar, M., "Undergraduate Process Control Experiments
Using Distributed Control System," ASEE Summer School
for Chemical Engineering Faculty, Snowbird, UT (1997) O

Postgraduate Environmental
Engineering Program
Continued from page 253
graduate program is motivated by the need to provide quali-
fied engineers to address the environmental challenges asso-
ciated with the large petroleum refining and chemical indus-
trial base in Singapore. With the establishment of the envi-
ronmental engineering program, the Department is the sole
institution responsible for the training of both professional
chemical engineers and environmental engineers in
Singapore. To better reflect our academic and research pro-
grams in environmental engineering, the department has
been renamed the Department of Chemical and Environ-
mental Engineering, effective July 1998.

1. Special Issue, "Southeast Asia Facing Development Chal-
lenges," Env. Sci. and Tech., 27(12) (1993)
2. Cortese, A.D., "Education for an Environmentally Sustain-
able Future," Env. Sci. and Tech., 26(6), 1108 (1992)
3. Wood, S.C., "Select the Right NO, Control Technology,"
Chem. Eng. Prog., 90, 31 (1994)
4. Bosch, H., and F. Janssen, "Catalytic Reduction of Nitrogen
Oxides: A Review on the Fundamentals and Technology,"
Catalysis Today, 2, 369 (1998)
5. De Nevers, N., Air Pollution Control Engineering, McGraw-
Hill, New York, NY, Ch. 12 (1995) J

Graduate Education




Tulane University New Orleans, LA 70118

his article will attempt to share a way of setting up
linear unsteady-state transport models in beginning
graduate and upper-level undergraduate transport phe-
nomena courses, a method that is more general than the one
presented in texts and published articles. When setting up a
transient problem in momentum, heat, or mass transport, the
usual formulation states that before a change in the bound-
aries or in the driving force takes place, the time- and space-
dependent variable is at some homogeneous value in the
space domain of interest.'1-4 Thus, in fluid mechanics we say
that the velocity is zero throughout when suddenly either a
pressure gradient is applied or one of the boundaries is set in
motion. Similarly, in heat and mass transfer, the space do-
main in question is allowed to originally have a certain
space-independent temperature (molar concentration) before
a new value is imposed on its boundaries, or alternatively the
heat (moles) generation experiences some change.
In the following two representative examples, it will be
shown how, when the initial condition is not constant with
respect to the spatial variable(s) but is instead some steady-
state profile, the resulting linear differential equations and
initial and boundary conditions can be made to be identical
to those of the constant-initial-condition case through the
use of a "deviation" dependent variable. The first example is
a classic driving-force-change problem, while in the second
example the change is imposed on the boundary.

Kyriakos Papadopoulos is Professor and Chair
of the Department of Chemical Engineering at
Tulane University, having joined its faculty in
1981. He received his BS (1978), MS (1980),
and DEngSc (1982) in chemical engineering from
Columbia University. His research is in the sta-
bility of dispersions and their transport through
porous media. He has taught a variety of under-
graduate and graduate courses and has been
honored by three departmental and one school-
wide teaching awards at Tulane.

Copyright ChE Division of ASEE 1998

For the flow of a liquid of constant density p and viscosity
tp in a tube of length L and radius R, when the fluid is at rest
and suddenly a constant pressure gradient, (p0-pi)/L, is ap-
plied, the differential equation in dimensionless form is"'

=4+ (1)


*= v/[(po- pl)R2/ 4L]
T = tg/pR2
and v=v(r,t) is the velocity of the liquid in the axial direction
of the tube. For this partial differential equation (PDE), the
dimensionless initial and boundary conditions are
O(T = 0)= 0; (4 = 0)= finite; ( = 1)= 0
Now, instead of having a zero initial condition, let us state
that the fluid is originally at a steady state, which we will
henceforth refer to as the "original steady-state velocity pro-
file," ,,ss (4). Let the driving force at the original steady state
be a pressure gradient (po -Pl)/ L and at t=0 the pressure
gradient is step-changed to (po P2) / L. The differential equa-
tion is[11

pv Po -P2 +1 a ( )v
at L r r r )
with initial and boundary conditions of

(1t = 0)= O"s, = 1- 2; (( = 0)= finite; (4( = )= 0
Equation 2 can also be written as

aV Po-P PI-P2 +1 (av3)
at L L r r or) (
If we define a "deviation" velocity w, as the v's departure
from the original steady-state velocity, w=v-v,,,, and then
Chemical Engineering Education

Graduate Education

substitute v=w+vssi in Eq. (3), the following PDE is ob-
aw p -p, 1 8 w poP 1 d ( dvs,)
at L r rr ar L r dr- dr ) (4)
The last two terms in Eq. (4), however, add up to zero
because the ordinary differential equation for the original
steady state is

d r dssl ) o -Pl (5)
rr ddr r ) L
If we now define the dimensionless deviation velocity, 9(O, t),
(Pi -P2)R2 /4pL
then Eq. (4) takes the form
ae a 8aeo .
a 4+ 1 (7)

which is identical to Eq. (1) and also has the same initial and
boundary conditions
0(T = 0) = 0; 0(4 = 0) = finite; 90( = 1) = 0
and has as solution for 9(4,z), Eq. 4.1-40 in Bird, Stewart,
and Lightfoot.l'i
The heat transfer analogue of Hagen-Poiseuille flow is the
problem of heat conduction in an electrically heated wire or
coil.J' For the unsteady case, the problem statement lets the
wire or coil operate at some original steady state under a
constant heat source Se,. Then at t=0, the heat source is
step-changed to S,2. The resulting treatment and differen-
tial equations are identical to the ones for the unsteady
Poiseuille flow.

The steady-state differential equation for diffusion and
first-order chemical reaction in a sphere, with constant
diffusivity D and reaction rate constant k, is

1 d 2 _C 2C =0 (8)
42 dSi d)
where C is the concentration of the reactant, 4 = r /R, and
S= R kk/ If the surface concentration of the reactant is
kept constant at Co, the two boundary conditions are
C(4=0)=finite; C(=I1)=C,

If we are interested in the transient response c(4, r) when
the catalyst is originally free of reactant, and suddenly a
surface concentration Co is imposed, the PDE is

1 8 2)2 -C 2C (9
42 4 a' a- (9)

with initial and boundary conditions

C(t = 0) = 0; C(4 = 0)= finite; C( = 1) = Co

where t = t / R2 is the dimensionless time.
Now, let us suppose that the catalyst is operating at some
original steady state and has a concentration profile C,,,
under a constant surface concentration C, when suddenly
the surface concentration is raised to C2. The PDE will still
be given by Eq. (9), but the conditions will be
C(T=0)=Cs,,,()=C, sinhop/[4sinhp]; C(4=0)=finite; C(= 1)=C2
Defining a deviation concentration r(;, r) = c(4, T) Csi (4),
we may substitute C in Eq. (9) by Cssi + F. The equation thus

1 d 2 dCssi j 1 a 2 B (10)
pC+ r F-- (10)
_2 d4 ( d4 s 2sc a )1 2
with the first two terms adding up to zero because of Eq. (8).
If we define Fo = C2 -C1, the PDE for r(, ) and its condi-
tions are

1 8a2 a8 _2[ = ar.

r(t = 0) = 0; r( = 0)= finite; Fr( = 1) = o
which are identical to Eq. (9) and its conditions. The solution
for r(4,T) can then easily be obtained following Example
19.1-2 of Bird, Stewart, and Lightfoot.


It has been shown how models of transient transport, when
a boundary or driving-force change is introduced following
an original steady-state operation, can be made to be math-
ematically identical in terms of the governing partial differ-
ential equation and initial and boundary conditions, to the
respective constant-initial-condition models. This method
should give transport teachers an alternative approach for
not only setting up new unsteady-state classroom problems,
but also for casting existing literature examples in a more
general context.

1. Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport
Phenomena, Wiley, New York, NY (1960)
2. Middleman, S., An Introduction to Fluid Mechanics, Wiley,
New York, NY (1998)
3. Middleman, S., An Introduction to Mass and Heat Transfer,
Wiley, New York, NY (1998)
4. Chan Man Fong, C.F., D. De Kee, and P.N. Kaloni, Ad-
vanced Mathematics for Applied and Pure Sciences, Gordon
and Breach, Amsterdam, The Netherlands (1997) O

Fall 1998

Graduate Education
\^ ^




At The NSF Engineering Research Center

A Model for the Future

University of Florida Gainesville, FL 32611

he need for change in the traditional graduate engi-
neering curriculum was described in 1995 by the
Committee on Science, Engineering, and Public
Policy'1 (COSEPUP), which documented that less than half
of all new PhDs enter academia. In recognition of this shift,
the Committee called for graduate education to move away
from a narrowly defined, single-department focus to a
broader student experience. There was acknowledgment
by the Committee that students were not being prepared
to function in the multidisciplinary environment that char-
acterizes industry today.
This call for a "new PhD" coincided with publication of a
paper[2' by three industrial researchers, Ralph Nelson and
Reg Davies of DuPont, and Karl Jacob of Dow. This group
also called for a shift in engineering education to allow the
chemical industry in the United States to remain competitive
in the global market. The following summarizes the key
concerns described in this report:
0 Particle processing is of major importance to many
C The need for researchers trained in Particle Science

Copyright ChEDivision of ASEE 1998

and Technology (PS&T) is increasing
E Foreign companies have an advanced mastery of the
application of particle science to chemical processing
U The U.S. engineering curriculum has not adequately
prepared students in the field of particle technology
B To meet industrial needs, U.S. companies must
recruit graduates from abroad
A growing recognition of the need for an educational
program that focuses on PS&T, coupled with a call for a
move away from traditional single-discipline degrees, coin-
cided with the development of a program at the NSF Engi-
neering Research Center for Particle Science and Technol-
ogy at the University of Florida. The NSF concept behind
the development of engineering research centers (ERC) was
to foster collaboration between government, universities,
and industry to enhance the competitiveness of American
industries and to promote new educational models to meet
the changing needs of research and industry. The establish-
ment of a national center for PS&T with a focus on
multidisciplinary research provided a venue for answering
industry's need for students adequately trained in this here-
tofore neglected, but critical, field.

Particle science and technology is by its very nature a
multidisciplinary field. Fluidization operations, particulate
separation, and colloidal and interfacial phenomena are just
several of the areas of PS&T relevant to chemical engineer-
ing. Particle science issues can also be found in materials
science and engineering, chemistry, physics, microbiology
and cell science, computer and information sciences, and in
Chemical Engineering Education

Anne E. Donnelly is the Associate Director for
Education and Outreach at the ERC. She man-
ages the ERC educational programs, including
Graduate Student activities, the Undergradu-
ate Research and Scholarship Awards, module
and textbook production, and consults with the
Industrial Partners Program on particle science
and technology issues of mutual concern.

Graduate Education )

environmental, biomedical, aerospace, mechanical, coastal,
and agricultural engineering. This makes an interdiscipli-
nary approach necessary.
The fit between engineering research cen-
ters across the country and the recommenda- The 1
tions made by the COSEPUP report were behind th
discussed at the 1996 ERC annual meeting of engine
and were summarized by Costerton.31 The centers
ERC for PS&T has established a graduate foster c
education program that simultaneously meets
the mandate for a new PhD described by
COSEPUP and the request from industry for unive
U.S.-trained students in PS&T. The Center industry
currently funds fifty graduate students. They compe
are assigned to work with one of the ERC Americ
faculty members and are registered with the and to
home engineering or science department of e
their major professor. They receive their de-
gree from the home department while partici- meet t
pating in research projects funded by the ERC. needs oj
This increases the breadth of the student ex- in
perience while at the same time giving the
student a firm foundation in a traditional dis-

ERC graduate students are assigned to one of five research
thrust groups:
Advanced Measurements and Characterization
Advanced Separation Processes
0 Dispersion, Agglomeration and Consolidation
> Engineered Particulates
Transport and Handling
There is also a recently added Targeted Technology Team
Project initiative in Chemical and Mechanical Polishing
Chemical engineering students work in these research
groups with students from many other departments. For
example, chemical engineering students in the Advanced
Separation Processes thrust group work with graduate stu-
dents from materials science and engineering, environmen-
tal engineering, and microbiology and cell science. These
students meet weekly with the thrust group faculty members
to make research presentations and to discuss current projects.
Faculty representation in the thrust groups also reflects the
multidisciplinary nature of the field, so in addition to being
exposed to students from other departments, these graduate
students also have regular contact with faculty members
from other disciplines. The ERC currently has 34 faculty
members representing eleven different engineering and sci-
ence departments. This "cross-talk" across departments al-
lows students to become familiar with other departments and
prepares them to function on the kinds of interdisciplinary

Fall 1998

teams that typify industry today.
Interdisciplinary team training is also a
critical part of the training of those students
entering academia. In order to train the stu-
:oncept dents of the future using this new model,
velopment faculty must be able to function and demon-
g research state the skills necessary to work in inter-
C) was to disciplinary teams and the ability to col-
boration elaborate across departments and with indus-
ement, try. The ERC for PS&T has brought to-
gether faculty from across campus, who
es, and themselves have had to learn to function in
chance the collaborative teams in order to foster these
eness of same skills in their students.
industries In addition to daily research and weekly
ote new group meetings, the ERC provides students
models to with other opportunities to meet with and
hanging discuss particle science with students from
rch and all of the research thrusts. A monthly Stu-
earch and
dent Seminar series provides a forum for
ty. students to practice presentation skills and
allows students from other groups to learn
about research projects in other labs. This

series is also open to undergraduates, to encourage them
to learn about research across the ERC.

The engineering curriculum has been enhanced by the
addition of several courses in particle science available to
both ERC and non-ERC students who desire to broaden their
academic exposure to PS&T. Particle Science and Technol-
ogy: Theory and Practice, Particulate Interfacial Systems:
Science and Engineering, Physics of Colloids, and Optimi-
zation, Scale-Up, and Statistical Experimental Design have
been previously described. 41 In addition, two other graduate-
level courses dealing with particle science have been of-
fered: Biomimetics and Biomineralization, which explores
the mechanisms used by organisms to control mineralization
in order to identify new biomimetic strategies for the engi-
neering of particles and composites, and Numerical Simula-
tion Techniques for Particle-Dynamics (see Table 1). These
courses allow students to supplement their coursework with
material covering a variety of particle science topics.

An equally important feature of ERC graduate training is
that the research conducted in the ERC is industrially rel-
evant. The ERC has a network of industries that support its
purpose of investigating particulate materials to enable the
invention and development of innovative particulate pro-
cessing technologies for new products and devices. These

1SF c
e de
to en
an i
he cl

Graduate Education

industrial partners recognize the value of the type of training
ERC students receive. In 1997, NSF published a survey of
ERC industrial partners from across the country51' that con-
cluded that industrial partners who have hired ERC gradu-
ates ranked this connection to top-quality students as the
most valuable benefit of ERC participation. Furthermore,
industries that hired ERC graduates reported that they were
superior to non-ERC-trained employees in items including
breadth of technical understanding, ability to work in inter-
disciplinary teams, depth of technical understanding, and
overall preparedness. These observations have been echoed
by the industrial partners of the ERC for PS&T, who have
made it clear that a key feature of the Center is the interdisci-
plinary model and that the traditional degree path is no
longer adequate to meet industrial needs.
Graduate students benefit from this partnership with in-
dustry in many ways. They are invited to present their re-
search work at the semi-annual meetings of the Industrial
Advisory Board. All students are also given the opportunity
to participate in poster sessions held at each meeting, during
which industrial representatives have the opportunity to meet
and discuss research progress with the students involved.
These posters typically reflect the team approach by listing
faculty, graduate students, and undergraduates who have
contributed to the work.
The ERC culture provides additional opportunities for
graduate students to enhance their knowledge of PS&T. The
ERC sponsors several short courses each year on a variety of
topics. These courses are offered to industrial representa-
tives and attract approximately 150 attendees each year. A
listing of courses held in the past two years is found in Table
2. ERC graduate students attend these session at no cost,
exposing them to additional training in selected PS&T top-
ics. Participation in these meetings facilitates communica-
tion between students and industrial representatives, thereby
improving student understanding of issues confronting in-
dustry today. ERC students are also encouraged to attend
and make technical presentations at professional society meet-
ings and provides financial assistance when possible to fa-
cilitate these experiences.

ERC graduate students have a unique opportunity to ex-
pand their awareness of global issues in PS&T as a result of
the Visiting Eminent Scholar Program where world-renown
researchers in PS&T are invited for extended visits to the
ERC. While in residence at the Center, these researchers
present seminar series or workshops that graduate students
are invited to attend. Students may also arrange to meet one-
on-one with the visitors to discuss areas of mutual interest.
This exposure to internationally known researchers adds to
the students' appreciation of the global perspective of PS&T.

Topics Covered in the Graduate Course
"Numerical Simulation Techniques for Particle-Dynamics"

Section 1A Overview of multiple-particle simulation techniques
Section 1B Instantaneous collision models (primarily for
spheres in 1, 2, or 3D)
Section 2 Hard-sphere MD, collision detection, searching and
computer-memory considerations
Section 3 Soft-sphere simulations with engineering-mechanics
contact models
Section 4 Soft-particle code structure
Section 5 Details of realistic-contact models (why quasi-static
force-displacement relations work for dynamics)
Section 6 Elastic frictional contacts (Mindlin's analysis,
implications, and limits of validity)
Section 7 Inclusion of other forces (Van der Waals, JKR,
liquid meniscus, lubrication, and viscous effects)
Section 8 Diagnostics
Section 9 Simulation results
Section 10-15 Detailed discussion on selected subtopics, including
non-spherical particles, suspension models, dilute
gas-solid models, continuum models, and others

Selected Short Courses Offered by the ERC for PS&T

* Dispersion Technology: Fundamentals, Processing, and Applications
* Improving Standards in Particle-Size Measurement
* Surfactants: Principles and Applications
* Stability of Dispersions: Theory and Practice
* Solids Handling: Flow and Conveying of Bulk Solids
* Special Topics in Particle Science and Technology

Topics Covered in the Seminar Course on
Professional Development

Section 1: Basic Tools for Conducting Successful Research
Introduction to resources available for doing research
Lab safety
Organizational skills
Technical writing and presentations
Time and stress management
Section 2: Resource Development and Professional Ethics
Academic research
Funding for research in science and engineering
Basic tenets of professional ethics
Section 3: Post-Graduation Issues
Requirements of a typical research position in industry
Academic and governmental positions

Chemical Engineering Education

Graduate Education

ERC graduate student Aaron Clapp dis-
cusses his research work with Industrial
Advisory Board member Ralph Nelson and
ERC faculty member Richard Dickinson at
a recent poster session. V


A Jim Fitz-Gerald. ERC gradu-
ate student, describes his re-
search work through a formal A ERC students Byron Palla, Steve
presentation to the Industrial Truesdail, and Kathryn Shaw par-
Advisory Board. ticipate in an ERC short course.

Recent exit surveys completed by ERC graduates asked
them to describe how the ERC experience was different
from that of their colleagues who did not participate. They
reported that the ERC provided them with a solid link to
industry, provided experience in interdisciplinary team work,
and improved their technical writing and presentation skills.
One student commented that "I will go into industry with an
increased confidence in my ability to take on and success-
fully complete a project."

In an update of their earlier remarks, Nelson and Davies
state that there has been some progress in addressing the
concerns raised by their article.1"6 They note that in addition
to the ERC, new courses are being added at City College of
New York, New Jersey Institute of Technology, University
of Pittsburgh, Yale University, and the University of Cincin-
nati. They then challenge universities to take the next step by
not only educating students in PS&T but also encouraging
them to enter management.
Leadership skills that are important in the development of
managerial abilities are fostered by the ERC in several ways.
The ERC funds undergraduate research work each semester.
These students are assigned to a graduate student, who as-
sumes primary responsibility for supervising them. This ex-
perience helps train students to be the managers of the fu-
ture. A leadership skills class has also been developed in
which students discuss topics outlined in Table 3. The ERC
has established a Student Leadership Committee composed
of graduate students from a variety of thrust groups to advise
ERC administrators on issues of concern to students and to
serve as a liaison between the staff and the students.
The ERC focus on interdisciplinary teamwork is the new
paradigm for graduate education called for by COSEPUP. It
is interesting to note that the type of interdisciplinary train-
Fall 1998

ing that ERC students are being exposed to has anticipated
the new ABET Engineering Criteria 2000. Under these guide-
lines, universities must demonstrate that graduates have ac-
quired the skills necessary to function on multidisciplinary
teams. It is clear that the ERC experience will help engineer-
ing colleges meet this criteria.
The establishment of a national Center for Particle Science
and Technology by NSF was a major step toward meeting
the challenge of improving the state of PS&T education in
this country. The Engineering Research Center for Particle
Science and Technology offers graduate students an
unparalleled opportunity to become the industrial and aca-
demic leaders of PS&T for the future.

The author would like to acknowledge the financial sup-
port of the Engineering Research Center for Particle Science
and Technology at the University of Florida, the National
Science Foundation NSF Grant #EEC-94-02989, and the
industrial partners of the ERC.

1. Committee on Science, Engineering, and Public Policy, "Re-
shaping the Graduate Education of Scientists and Engi-
neers," Report Brief, National Academy Press (1995)
2. Nelson, Jr., R.D., R. Davies, and K. Jacob, "Teach 'Em
Particle Technology," (summary of 1992 talk), Chem. Eng.
Ed., 29, 12 (1995)
3. Costerton, J.W., "The ERC Model for a 'New PhD,"' report
from the Center for Biofilm Engineering, Bozeman, MT
4. Donnelly, A.E., and R. Rajagopalan, "Particle Science and
Technology: Educational Initiatives at the University of
Florida," Chem. Eng. Ed., 32, 122 (1998)
5. Parker, L., "The Engineering Research Centers (ERC) Pro-
gram: An Assessment of Benefits and Outcomes," National
Science Foundation (1997)
6. Nelson, Jr., R.D., and R. Davies, "Industrial Perspective on
Teaching Particle Technology," Chem. Eng. Ed., 32, 98 (1998)

Graduate Education



Goals, Guidelines, and Experiences

University of Kentucky Lexington, KY40506-0046

Effective communication is an integral part of high-
quality research. While in graduate school, students
need opportunities to practice their presentation skills,
to meet and interact with industrial and academic research-
ers, and to receive feedback from a variety of sources in
order to enhance their professional development. While pro-
fessional meetings may foster this type of growth, these
opportunities are often limited to senior students, and the
size of the meetings may not allow extensive interaction
between students and industry.
In September of 1997, West Virginia University (WVU)
hosted a regional symposium based on the goals and guide-
lines developed by the University of Kentucky (UK) in
symposiums held at UK the previous four years. The re-
gional symposiums gave graduate students an opportunity to
present their research and to interact with industry.
The original idea and format of hosting a symposium
came from Carnegie Mellon University (CMU). CMU's in-
ternal symposium gave graduate students a formal opportu-

Jamie A. Hestekin is a fourth-year PhD student in chemical engineering
at the University of Kentucky. He received his BS in chemical engineer-
ing from the University of Minnesota, Duluth, in 1995. His current area of
research is in functionalized microfiltration membranes for heavy metal
Marazban Sarkari is a fourth-year PhD student in chemical engineering
at the University of Kentucky. He received his BS in chemical engineer-
ing from the University of Bombay in 1993 and his MS in bioprocess
technology in 1995. His current area of research is in the application of
supercritical fluid media to enzyme-based processes.
Melissa A. Summers is a fourth-year PhD student in chemical engineer-
ing at the University of Kentucky. She received her BS in chemical
engineering from the University of Kentucky in 1995. Her current area of
research examines the effect of stimuli on the adhesion of cancer cells to
endothelial monolayers.
Katherine S. Ziemer is a third-year PhD student in chemical engineering
at West Virginia University conducting research in the processing of
electronic materials. She received her BS in chemical engineering from
Virginia Tech in 1989 and worked as a chemical engineer for DuPont for
seven years.
Leonard P. Zuba received his MS in chemical engineering from West
Virginia University in 1998. He received dual BS degrees in 1992; in
chemical engineering from the University of Pittsburgh and in chemistry
from Gannon University. He is currently an engineer for Rehau, Inc.

* Address: West Virginia University; Morgantown, WV 26506-

nity to share their research with their peers. Many schools
hold graduate student symposiums with similar formats. Simi-
larly, the UK event began as an internal one; but over the
years it grew to become a regional event and to incorporate
several novel concepts. Currently, the symposium draws
upon various schools and industries within the region. This
allows graduate students from the host school to gain experi-
ence in organizing a regional event and allows industrial
representatives to sample graduate research throughout the
region. In addition, moving the symposium location through-
out the region allows graduate students to make several new
industrial contacts each year.
This article demonstrates how a regional symposium might
be hosted. Included are a brief history of the event, goals and
guidelines for hosting such a symposium, and an illustration
of a successful symposium. We hope that this article will
encourage other universities to host a similar symposium in
their region.

The UK Chemical Engineering Graduate Student Associa-
tion (ChEGSA) held its first symposium in September of
1993. The idea to host the event was initially proposed by
then-ChEGSA faculty advisor, Dr. Kimberly Anderson. Hav-
ing graduated from CMU, she suggested modeling a meet-
ing on the annual ChEGSA Symposium that CMU had been
holding since 1979.11] The primary intention of the event was to
provide a forum that allowed graduate students to improve
their presentation skills and to give them industrial exposure.
From an internal event involving representatives from a
few local companies, UK expanded the symposium to in-
clude regional universities and industries. Due to this growth
and success, UK felt it was time to encourage other schools
in the area to jointly host the symposium. As a result, the
Fifth Annual Chemical Engineering Graduate Student Sym-
posium was hosted by the WVU Graduate Student Organi-
zation (GSO) in September of 1997. It attracted graduate
student participants from seven major chemical engineering

Copyright ChEDivision ofASEE 1998

Chemical Engineering Education


Graduate Education

programs and involved eight industrial corporations.

The goal of the symposium is to provide a learning envi-
ronment that promotes professional development of gradu-
ate students and an interactive exchange of ideas among
regional industries and universities. This goal is facilitated
by student posters and oral presentations that are judged by a
panel of industrial representatives on the quality of the re-
search and the effectiveness of the presentation. Additionally,
this interactive structure provides an opportunity for research-
ers in industry and academia to increase their awareness of the
ongoing research activities in their geographical region.
To promote interaction between industry and academia,
opportunities for attendees to network are designed into the
day's activities. Industrial involvement is encouraged through
the use of judges and a keynote speaker from recognized,
engineering-related industrial organizations. This network-
ing benefits all attendees by increasing contacts, giving new
perspectives, and accommodating different viewpoints. The
use of industrial judges not only ensures unbiased evalua-
tion, but also gives the graduate students an occasion to
tailor their talks for an industrial audience as opposed to the
primarily academic audience usually encountered. Simulta-
neously, the graduate students get feedback from an indus-
trial perspective on their research and presentation skills.
Industry also plays a role in the symposium through mon-
etary support. Though some money is provided by various
sources within the host school, industry provides much of
the funding. Student organizers gain experience marketing
their ideas to industrial representatives and encounter the
challenges of fundraising.
Regional school participation brings depth to the sympo-
sium. Not only does this give an advantage of peer-to-peer

Approximate Timetable for Hosting a Symposium


* Find and contact keynote speaker
* Set date and venue for symposium
* Set date for abstracts

January Send out invitations and guidelines for abstracts
Contact industry for contributions/judges
February-April Follow up on invitations and contacts
May-June Begin accepting abstracts and monetary support
July Abstract deadline
Organize lunch, dinner, and hotel accommodations



* Print program and send to judges with judging guidelines

* Host symposium

networking, but it also allows for a larger industrial draw
than if each school held their own internal symposium. To
attract student participation, monetary awards for the best
three oral presentations and posters are presented. The use of
awards creates an opportunity for unbiased industrial judges
to give valuable feedback to the students.
An inherent benefit of the symposium lies in the fact that it
is entirely managed by graduate students. They gain first-
hand experience in communication with industry, communi-
cation with academic peers, and event organization. Faculty
support is appreciated, but should be limited to advice and
attendance. To facilitate efficient organization of the sympo-
sium, the host school should have an established organization
of graduate students such as UK's ChEGSA or WVU's GSO.
Hosting the symposium involves a considerable input of
time, effort, and money. To retain a sense of continuity and
to best meet the symposium goals discussed above, a set of
guidelines was jointly outlined by UK and WVU. They
require that 1) the event be organized and run by students, 2)
the keynote speaker and a majority of the judges come from
industrial corporations, 3) the host school extend invitations
to several schools in the region, 4) no more than 50% of
participation be from the host school, and 5) cash prizes be
provided to winners of both the oral and poster presenta-
tions. An approximate timetable (Table 1) and budget (Table
2) have been prepared for use in planning activities. As
shown, hosting a symposium requires a full year's prepara-
tion and a significant budget.

The Fifth Annual Chemical Engineering Graduate Student
Symposium is an example of a regional symposium. It was
hosted by WVU and included participation by seven re-
gional schools: Carnegie Mellon University, Lehigh Univer-
sity, Pennsylvania State University, University of
Kentucky, University of Maryland College Park,
University of Pittsburgh, and West Virginia Uni-
versity. Through contacts and mailings, the host
school was able to attract approximately 100 people

Fall 1998

Budget for Hosting a Symposium
(Based on 100 Attendees)

Food and venue* $25 per person
Keynote speaker NA**
Awards $1200
Proceedings $500
Miscellaneous $800
TOTAL $5000
* Lunch, banquet, snacks
** Many companies will sponsor their employee

Graduate Education

to the day's events.
The symposium activities were planned around graduate
student research, which included nineteen oral presentations
and ten posters covering a variety of areas in chemical engi-
neering. Invited judges represented DuPont, Union Carbide
Corporation, Bayer Corporation, Air Products and Chemi-
cals, Inc., PPG Industries, Witco Corporation, and the De-
partment of Energy. Symposium organizers provided the
judges and participants with detailed judging criteria one
month before the presentations. The industry judges gener-
ously spent the extra time and effort required to give excel-
lent written comments to the participants in addition to the
general scoring. This gave participants valuable feedback
not normally received during a professional meeting. In
addition, score-based prizes in the form of monetary awards,
plaques, and certificates were awarded to first, second, and
third places in both oral and poster presentation categories.
Interaction among students and industrial judges was fur-
ther promoted during both a formal luncheon and an evening
awards banquet. In the tradition of previous symposiums, a
highlight of the day's activities was the keynote address
given by a prominent figure in the field of chemical engi-
neering. Past keynote speakers have included Dr. Rakesh
Gupta (WVU), Dr. David Ollis (North Carolina State Uni-
versity), Dr. Michael Jaffe (Hoechst Celanese Corporation),
Dr. George Keller (Union Carbide Corporation), and Dr.
Frank Derbyshire (Center for Applied Energy Research, UK).
For the Fifth Symposium, Dr. Stan Speed, Chief Scientist for
Exxon Corporation in Baytown, Texas, gave the keynote
address, "R&D at the Academic/Industrial Interface." Sym-
posium organizers developed their skills in fundraising as
the host school successfully obtained financial support
through DuPont, Union Carbide Corporation, Witco Corpo-
ration, Hoescht Celanese Corporation, Exxon Corporation,
and WVU to fund the event.

The Chemical Engineering Graduate Student Symposium
provides a learning environment that promotes professional
development of graduate students and the active exchange of
ideas among regional industries and universities. It offers
benefits beyond professional meetings for graduate students
through detailed industrial feedback and firsthand experi-
ence in leadership and organization. The success and useful-
ness of the symposium is evidenced by its growth over the
past five years. The goals and guidelines described here
will hopefully encourage other schools to host regional
symposia in their area. For further information on how to
organize this type of event, please contact symposium
organizers at or


We would like to thank the faculty at both UK and WVU
for their assistance in making the symposium possible. Fur-
ther, we would like to thank the participants from the past
symposia, especially those who have traveled from other
schools. Finally, we would like to thank both internal and
industrial sponsors for their support, both financial and
through judging and speaking.

1. Modi, A.K., and P.T. Bowman, "The ChEGSA Symposium:
A Continuing Tradition at Carnegie Mellon University,"
Chem. Eng. Ed., 23(2), 100 (1989) 3

book review

Analysis, Synthesis, and Design
of Chemical Processes
by Richard Turton, Richard C. Bailie, Wallace B. Whiting,
and Joseph A. Shaeiwitz

Reviewed by
Mark A. Stadtherr
University ofNotre Dame

This is a new textbook for the process design course (one-
or two-term) taught in almost all chemical engineering cur-
ricula. I have now used the book twice (once in a "beta"
version) in the one-term, senior-level design course taught at
Notre Dame. In general, I have found this text to be very
well written, both clear and concise, with feedback from
students positive in this regard as well. Much of the material
covered is well supported by examples.
The book begins with a relatively detailed discussion of
the various types of process diagrams, ranging from simple
input-output diagrams to detailed piping and instrumenta-
tion diagrams. This serves not only to introduce a "lan-
guage" for process design, but also, through the hierarchy of
increasingly detailed diagrams, to emphasize to the student
the evolutionary nature of design.
The first major section of the text covers engineering
economics, including cost estimation. This material is con-
cise and to the point, covering the necessary material with-
out introducing unnecessary complexities. For capital cost
estimation, an easy-to-use computer program (CAPCOST)
running under Microsoft WindowsTM is provided, a unique
feature among currently available texts. The fairly extensive
cost correlations and data incorporated in CAPCOST are also
given in an appendix (in both tabular and graphical form).
Continued on page 276.
Chemical Engineering Education

ASEE Chemical Engineering Division Awards

* CACHE Award for Excellence in Computing in Chemical Engineering Education
This award, sponsored by the CACHE Corporation, is presented for significant contributions in the development of
computer aids for chemical engineering education. The award consists of a plaque and a $1,000 honorarium and is
presented at the Chemical Engineering Division awards banquet held at the ASEE Annual Conference.

* Ray W. Fahien Award
This award is given in honor of Ray Fahien, who was editor of Chemical Engineering Education from 1967-1995 and
who was effectively the founding father of the journal, establishing it as a premier publication vehicle in the field of
chemical engineering education. Professor Fahien selflessly gave his time and talents to advance pedagogical scholar-
ship, particularly in the careers of young educators, through his dedication to the journal and the profession. The award,
which is given annually to an educator who has shown evidence of vision and contribution to chemical engineering
education, consists of a $1,000 stipend and a plaque to be presented at the Chemical Engineering Division Banquet of the
ASEE Annual Conference. Educators who have been faculty members for not more than ten years as of July 1 in the year
of the award are eligible.
Nominees will be evaluated based upon two equally weighted criteria.
1. Outstanding Teaching Effectiveness. Nominees must demonstrate evidence of excellence in the training of under-
graduate and/or graduate students in the classroom or through supervision of independent study projects. This can be
demonstrated by such evidence as standardized student evaluations over a three year period along with letters from
students or former students (not more than four).
2. Educational Scholarship. The award recipient should have made significant contributions to chemical engineering
education that go beyond his/her own institution. Examples of such evidence include development of instructional
methods or materials (textbooks, instructional software, assessment instruments), along with the associated publications in
refereed scholarly journals and conference proceedings. This evidence may be supported by the applicant's teaching
dossier, reprints of papers and up to four supporting letters, three of which should be outside the candidate's institution.
Nominators must accompany the submission with a one page citation, which will appear in Chemical Engineering

* Union Carbide Lectureship Award
This award, sponsored by Union Carbide, is presented to a distinguished engineering educator to recognize and encourage
outstanding achievement in an important field of fundamental chemical engineering theory or practice. The individual
shall demonstrate achievement through the formulation of fundamental theory or principles, improvements of lasting
influence to chemical engineering education with books and/or articles and the demonstration of success as a teacher. In
addition, evidence of the ability to conduct original, sound and productive research, and an interest in the progression of
chemical engineering through participation in professional and educational societies shall be demonstrated. The individual
receives a $3,000 honorarium and a certificate of achievement.

* William H. Corcoran Award
This award, sponsored by Bayer Corp., is presented each year to the author of the most outstanding article published in
Chemical Engineering Education. The award consists of a small honorarium or travel expenses and a commemorative

* Joseph .. Martin Award
The J. J. Martin Award is presented for the most outstanding Chemical Engineering Division paper presented at the ASEE
Annual Conference. The award consists of a commemorative plaque.

Nominations are due at ASEE by January 15, 1999. For more information, see

This information will also be contained in the ASEE awards package mailed to all members in the fall.

Fall 1998





Explosion Prevention Technical Elective

Louisiana Tech University Ruston, LA 71272

here are a number of good reasons to teach process
safety to our students: we care about them and the
community; to meet a need of industry; to reinforce
basic concepts such as thermodynamics, reaction chemistry,
and phase behavior; to illustrate the application of basic
concepts to non-traditional situations; to add a practical as-
pect to the students' education, which is often lost in the
rigor of derivations; to train students to understand the pro-
cess chemistry and equipment, including the constraints of
both; to fulfill ABET requirements; and, not unimportantly,
because it is interesting!
This paper will explore what to teach, how to teach it, who
should teach it, and when to teach it.

Crowl and Louvarl' cover the fundamentals of chemical
process safety in their text. It includes the basic topics of
toxicity, fire and explosions, ignition sources (e.g., electro-
statics), and more. Of course, each instructor will have his or
her favorite topic, but in my mind the unifying concept is the
reactive chemistry: a substance is toxic because it is reactive
in our bodies; it burns because it reacts with an oxidizer; it is
explosive because it burns rapidly and/or because it experi-
ences a rapid chemical decomposition reaction.
Three measures to handle these reactions can be sug-

Charles M. Sheppard is Associate Professor of
Chemical Engineering at Louisiana Tech Uni-
versity. He seeks to teach the upcoming gen-
eration of chemical engineers a design approach
that includes hazard or safety awareness and
environmental responsibility. His research in-
cludes two-phase flow implications to pressure
relief device design, knowledge-based design
systems, and fluidized catalytic cracker and other
f reactor modeling.

* Now at the University of Tulsa, Tulsa OK 74104-3189; e-mail

gested. First, before a hazard can be addressed, it must be
identified. Second, once a potential hazard is identified, the
best approach is to prevent the reaction from occurring.
Third, in those cases where complete prevention is impos-
sible or not economically practical, the consequence of an
incident should be minimized, that is, mitigated. Each of
these three steps is discussed below.
Identification: Violent Reactions-Oxidation. Reduction.
and Decomposition The reactivity depends on the chemi-
cals present and the energy in the system. Regarding the
chemicals present, one can examine the type of chemicals
(Bodurtha[21 lists twenty-one hazardous types of compounds,
e.g., azo compounds) or similarly, the structure of the
compound (e.g., reactive double and triple bonds). The
chemical's structure can be found via chemistry books or
through the web site
Thermodynamics can be used to calculate the energy associ-
ated with a reaction (e.g., decomposition or oxidation). Some
of the methods to perform these calculations include using
available data and calculational tools (see Table 1), simula-
tion packages, and the CHETAH program.31 The CHETAH
program allows the user to access data for many known
compounds and to build other compounds. But even without
the depth of the preliminary investigation indicated above,
one can examine the reactivity of a system. This is illustrated
in Table 2 for acrylonitrile production via the partial oxida-
tion of propylene in the presence of ammonia. This table was
constructed knowing the main reaction and considering all
possible combinations of reactants. Possible chemical inter-
action based on the above chemistry reasoning and the ex-
perimental methods mentioned next, can be recorded for
binary pairs on a chemical compatibility chart (see Table 3).
Regarding the effect of system energy on reactivity, one
can begin by looking at the chemical state (see Table 4).

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Gases are generally more reactive than liquids or solids, and the higher the
temperature the more reactive the system.

These two basic chemistry concepts of identifying possible reactions and
quantifying the energy of the system based on the chemical state can go a long
way to reduce the hazards associated with chemical processes. It behooves us
to get the next generation of chemical engineers thinking in these terms.

Beyond these initial literature and analytical screening tools, any possible
reactivity should be investigated via experiments. Table 5 (next page) lists
some common methods going from the small sample size screening tools to
the more detailed analysis tools requiring larger samples; more details are
available in Reference 5.

Prevention: Attacking the Three Sides of the Triangle The fire triangle
(see Figure 1) teaches the requisites for combustion. By seeking to eliminate

Figure 1. Fire Triangle

Welker & Springer's Safety, Health, and Loss
Prevention in Chemical Processes Problem 39:
Thermodynamics: Constant Volume Gas Phase

Chemical Engineering Topic: Thermodynamics: Constant Vol-
ume Gas Phase Reaction
Safety and Health Concept: Explosions: Pressure rise for en-
closed combustion reaction
Background: Refineries and chemical plants use a variety of
low pressure vessels as knockout drums and seal drums. Most
of these vessels are operated at very low pressures, but they
may contain flammable mixtures of vapor and air. It is quite
unlikely that ignition will occur in such a vessel because there
is usually no source of ignition. However, there is always a
chance that ignition might occur, so the American Petroleum
Institute's Recommended Practice 521 (API RP 521), 1st
edition, states; "Most knockout drums and seal drums will be
operated at relative low pressures. To ensure safe conditions
and sound construction, a minimum design pressure of 50 psig
is suggested. A vessel with 50 psig design pressure should not
rupture if an explosion occurs. Stiochiometric hydrocarbon-
air mixtures can produce peak explosion pressures in the order
of 7 to 8 times operating pressure, most flare seal drums
operate in the range of 0 to 5 psig, and ASME code-allowable
stresses are based on a safety factor of 4 to 1." Section 8,
Division 1 of the ASME (American Society of Mechanical
Engineers) pressure vessel code specifies a safety factor of
four to one (applicable at low pressures only). That safety factor
implies that a vessel with a stated mechanical design of 50 psig
should not rupture at pressures up to 200 psig.
Problem: Show that the 50 psig design pressure suggested by
API RP 521 will contain the explosive combustion of a mix-
ture of air and n-hexane with initial conditions of 77F (250C)
and 5 psig and stoichiometric concentration of n-hexane in air.
Compare your result to the estimated pressure rise of "7 to 8
times operating pressure" referred to in the API standard. You
may assume that the reaction proceeds to completion and that
the products of combustion are carbon dioxide and water.
(This problem was suggested by Mr. J.R. Phillips, a graduate
student at the University of Arkansas.)
Extension to flammability limits
Look up the LFL and UFL. If you were to repeat the pressure
calculations, under these conditions, how do you expect the
results to compare? Which limit is more like reality? Perform
the pressure calculation for this case.

Fall 1998

Chemical Reactivity Chart for Acrylonitrile



propylene, oxygen & ammonia
propylene & oxygen
oxygen & ammonia
propylene & ammonia
propylene, oxygen & ammonia

Expected Reaction
(if any)

Partial combustion

Endo- or Exo-




Chemical Compatibility Chartt14
Health* Fire* Reactivity* Component propylene ammonia oxygen
1 4 1 propylene P
3 1 0 ammonia 0
0 0 0 oxygen 4 1
* NFPA (National Fire Protection Association) ratings: 0-hazard; 1-
minimal; 2-caution; 3-danger; 4-extreme danger: P-polymerization or
precipitation; X-unlikely mixture, S-special-minimal if dry, danger if

Example of Safety Implications of
Process Stream State and ContentstsS'6

a b c d e
Temperature Medium Medium Low Low High
Pressure Low Low High High High
State at Process Conditions Gas Liquid Liquid Liquid Vapor
State at Ambient Conditions Gas Liquid Vapor Vapor Vapor
Contents (e.g., fuel, oxidizer,...) Inerts Fuel Fuel Fuel; oxidizer Fuel; oxidizer
Hazardous? Minimal Moderate Significant* Large Extreme

*According to Welkeri7 "low temperature liquid mixtures of fuel and oxidizers may be
extremely hazardous. For example, liquid oxygen conform shock-sensitive mixtures with
liquid fuels at very low temperatures. If the fuel and oxidizer are completely soluble, the
mixtures can result in detonations.

or shrink each side, we can reduce the probability of the
occurrence of a fire or an explosion and perhaps the size of a
fire if one does occur. Industry seeks to prevent the exist-
ence of flammable mixtures. Other examples are elimi-
nating ignition sources, inert storage tanks, and reducing
the amount of fuel present.
Inert gas can be added to dilute the flammable gas compo-
sition below the lower flammability limit. Calculating the
amount or flowrate of inert gas required for safety reinforces
the principles of partial pressure and revisits the concepts of
a perfectly mixed tank versus plug flow.
Mitigating: Placement of Pressure-Relief Devices and of

Equipment "Mitigation" is reducing the severity of the
consequences of an incident. Two common mitigation meth-
ods are including pressure relief
devices and placing equipment
strategically. The first seeks to
eliminate or reduce the size of Re
an explosion, and the second
seeks to reduce the number of Tool Sampli
people and the amount of equip- Mixing Calorimeter 2 drops-
ment affected by a fire or explo-
sion. Pressure-relief devices are DSC (Differential 10 m
Scanning Calorimetry)
needed on the equipment that
may be likely to be over pres-
DTA (Differential 10m
sure due to reaction, fire, com- Theral Analysis)
pression, etc. Placement of the
equipment should be such that RSST (Reactive System 10
Screening Tool)
fires and explosions that occur
in one process area are not propa- ARC (Accelerating 10
Rate Calorimetry)
gated to other process areas.

HOW: TEACHING VSP (Vent Sizing 100

When approaching this topic,
it is helpful to have a mix of
experiments, demonstrations,
and analytical theory. This mix Process Safety 1
both appeals to different learn-
ing styles and reinforces these Experimental*
valid approaches to problem flammability limits
solving. Table 6 gives some ex- toxicity
amples along with case studies
illustrating that unfortunately stoichiometry
these concepts are relevant to
one's life expectancy as a pro- kinetics ARC, VSP
cess engineer, dust explosions
The Louisiana Tech Univer- burning speed
sity chemistry department is very electrostatics experimei
supportive in providing the dem- Toluene water boiling point d
onstrations listed in Table 7. This El i s
Experimental indicates that
demonstration time is always a
or that experimental data are
favorite of the students. Of

course, proper safety precautions should be followed, in-
cluding but not limited to, safety glasses, hearing protection,
and in some cases blast shields. One may consult the chem-
istry department at your school or Dr. Bill Deese at Louisi-
ana Tech University for more details.

Text I have found four potential texts for this course[1,2,5,8]
They are all good, but have different strengths. Bodurtha's
textt21 is specifically on explosion prevention and covers the
basics well, but does not get into the chemistry in detail.
Frank Bodurtha gave me permission to photocopy his text,
now out of print. I believe if contacted in writing, he would
grant others this same permission. I have used this book as a
supplement both years I have taught the course. The first
year I taught the course, AIChE had a special on Stull's

active Chemical Testing Equipment Choices

eSize Data Obtained Conditions

2 ml




AH of mixing or reaction
exotherm or endotherm;
A Hrxn, Cp, some rate data

exotherm onset

dT/dt, dP/dt for a specific
heat rate 4 vent size

il dT/dt/ dP/dt, time to max rate
A Hrxn, kinetic parameters

nl dT/dt, dP/dt with minimal
energy loss, A Hrxn, flow
regime, kinetic parameters

room temperature
temperature programming, no P data, no
mixing, small sample (hard to obtain repre-
sentative sample for heterogeneous systems)
temperature programming, enthalphy change
not quantified
temperature programming, no mixing,
limited qualitative data
temperature programming, limited mixing
(stir bar) mass of bomb absorbs energy, thus
damping reaction rate
temperature programming, direct agitation,
vessel venting and injection possible

ropic Covered via Experimental or Analytical Approach


Case Studies/Problem Sets

Flixborough, England flammability limits (vapor pressure)
Seveso, Italy bond strength (-,=,=)
Bhopal, India HAZOP
Nitroaniline Sauget, Illinois energy transfer rate
Pasadena, Texas polymerization
Willey Kinetics energy of reaction A Hrxn
Welker & Springer chemical compatibility chart CHETAH

Kletz What went wrong


charge density and relaxation
liquid activity coefficient models

either one can perform the experiments under carefully controlled conditions
available for analysis.

Chemical Engineering Education

monograph"1 for only $8. At that price the students could
easily buy this text and the photocopy of Bodurtha's text.
That year, the course had a very strong chemistry-of-reac-
tion flavor. The sale was off by the second year I taught the
course, so I opted for Crowl and Louvar's more general
text.11l It also is a good text on flammability, inerting, the
DOW Fire & Explosion Index, etc., but lacks a strong reac-
tive chemistry content. (We use this text in our first capstone
design course that covers safety and economics.) I retained
much of the strong chemistry foundation via my lectures and
handouts. Finally, Barton & Rogers' textt15 is new to me. It is
strong on the chemistry and the experimental method, but
lacking the flammability limits and inerting. My current
plans are to use this text next year, supplemented by either or
both Bodurth, or Crowl and Louvar. The course text is
greatly supplemented by the SACHE slides set, and the
homework problems, discussed below.

Flammability Demonstrations Provided by the Louisiana
Tech University Chemistry Department

Burning candle; illustrates the chemical reaction (burning) occurring
in the gas phase with free radical reaction and radiant energy (from
soot that will deposit on chalk put into the flame).
Flame speed illustrated with an angle-iron channel and pentane
flammable vapors.
Flammable methane soap bubbles; gas density and energy and speed
of rection-deflagration.
Flammable hydrogen/oxygen soap bubbles; gas density and energy
and speed of reaction-detonation. One ignites a handful of the suds
in a student's hands. (Safety goggles and ear protectors required of
course.) It makes a loud bang but is safe.
Flammable hydrogen balloon; gas density and energy and speed of
Flammable hydrogen/air balloon; gas density and energy and speed
of reaction-detonation. (Safety goggles and ear protectors required,
of course.)
Carbon dioxide extinguishing a candle, illustrating gas density and
oxygen requirement for burning.
Dust explosion; ignition of dispersed Lycopodium powder in a closed
paint can; lid is dislodged (to the ceiling) by the explosion.
General chemistry concepts, including stoichiometry of reaction,
pressure as a function of temperature, and excess fuel. One ignites
rubbing alcohol (70% isopropyl) vapors in a 5-gallon glass water
jug. Just after the flame, one puts his hand on the mouth and talks
while the jug cools. The hand gets stuck to the jug as the pressure
drops; then peel off the hand and air rushes in. Immediately thrust a
burning splint into the mouth and ignite the vapors and remove the
splint. A "ring of fire" forms and burs slowly from top to bottom of
the jug.*
Deese, W.C., "The Ring of Fire Demonstration," Chem. 13 News
(November 1996). Currently, hazards associated with this demonstra-
tion, in addition to those discussed in the article, are being reviewed.
Some have tried this with other alcohols, at higher temperatures and in
oxygen-rich atmospheres with violent results. John Forman at Wright
State University is currently performing a study using various alcohols
with various amounts of water that bears on this demonstration.

Fall 1998

Lectures and Homework Table 8 gives an outline of
course content for the three-semester-credit hour course taught
in a "quarter" system. This course applies knowledge the
students have in chemistry, thermodynamics, and strengths
of materials. We review actual incident case studies because
"those who do not learn from history are bound to repeat it."
One of the areas I desire to develop is in the experimental
screening of potential reactions. At SACHE's 1996 Detroit
meeting, Tom Hoefflich[19 gave some background on chemi-
cal screening via experimental measures. Braton and Rogers151
present similar material and inchlde information on both
predictive (e.g., CHETAH) and experimental (e.g., Differ-
ential Scanning Calorimeter; DSC) approaches. The stu-
dents use a DSC in the physical chemistry laboratory to
find the heat of fusion for diphenylamine. Also, Ron
Willey has several problems related to this concept in his
kinetics problem set.
Table 9 (next page) gives a short list of potential home-
work problems. They are from the Crowl and Louvar text,['

Course Outline for Explosion Prevention Technical

Introduction (1) Crowl & Louvar, 1
1. Definitions (3) Crowl and Louvar 6; Bodurtha 3
a. Fire triangle
b. Oxidizing agent
c. Explosion type
d. Flammability limits flash point,LFL, UFL, MOC, etc.
e. Flammability diagram, reading, and construction
2. Chemistry of Combustion (4 plus exam) Crowl and Louvar 6;
Bodurtha 4,7; Stull
a. Stability: possible reactions
b. Thermodynamic measures
c. Classification based on
Classes (Bodurtha, 7)
Chemical element present
Enthalpy of formation
Bond strength
d. Measured kinetic reaction rate and activation energy (Seveso
3. Prevention and Mitigation of Fires and Explosions (3) Crowl and
Louvar 7 and 8; Bodurtha 5
a. Inerting; below LFL or above UFL
b. Ignition source (Crowl and Louvar 7; SACHE 1996 Workshop
case study)
c. Pressure relief (Crowl and Louvar 8 and 9)
d. Inventory control and plant layout
4. Case Studies (3 plus presentation) Crowl and Louvar 1 and 13
a. Flixborough (slides)
b. Phillips (video and slides)
c. ANGUS; Sterlington, LA
d. Channelview, TX
e. Students' presentations
5. Identification and Quantification of the Risk (2 plus final) Crowl
and Louvar 10
a. HAZOP studies (video)
b. Using Dow's Fires and Explosions Index (Dow case study"'61)

Welker and Springero101 problem sets, Willey's 1-15 slide and
problem sets, and original problems formulated based on
thermodynamics concepts. One can choose from the avail-
able problems to emphasize the desired concepts.
Figure 2 shows the flammability diagram similar to one
developed in the homework. This and other illustrations give
representations from which the student can reason about the
safety implication of leaks, loss of inert gas, and other inci-
dents. The balanced combustion reaction in air (21% 02 and
79% N2) is
1 C3H8+ 7 02+ 26.3 N2 3 CO2 + 4 H20 + 26.3 N2 (1)
The fuel composition is shown on the abscissa, the oxygen
concentration on the ordinate, and the nitrogen composition
is calculated as the one less these compositions (based on
this ternary mixture and the mole fractions summing to one).
The light grey in the figure shows the experimentally deter-
mined flammability region for propane and air. Also shown,
in dark gray, is a calculated estimated flammability region.
The lower flammability limit (LFL) is when just enough fuel
is present so that it all reacts and combustion is sustained-
fuel is the limiting reagent; the prediction is close. The upper
flammability limit (UFL) is when enough fuel is present so
that all the oxygen reacts and combustion is sustained-
oxygen is the limiting reagent; the prediction is high but
conservative. Below the minimum oxygen concentration
(MOC), combustion cannot be sustained. This occurs with
the minimum fuel composition and the stoichiometric amount
of oxygen. Therefore, the MOC should be at the intersection
of the LFL and the stoichiometric composition line (having a
slope of 7 02 to 1 fuel). Point F indicates a fuel concentration
below which any combination with air (e.g., from a leak)
will produce a non-flammable mixture. That is, if a nitrogen
purge keeps the vapor space in a tank to the left of the line
segment between air and F, even if there is an air leak the
resulting mixture will not be flammable.
Thermodynamics concepts like energy of compression,
auto ignition temperature (AIT), and adiabatic flame tem-
perature (AFT) tie directly into this course. The compression
adds energy to the gas being
compressed. The auto ignition

temperature is the temperature
at which a fuel in air can get
enough energy from the envi-
ronment to ignite. So if a gas is
flammable and its auto igni-
tion temperature (AIT) is ex-
ceeded, there is an explosion.
This is what is believed to
have happened in
Channelview, Texas, in 1990,
due to a faulty oxygen gauge
during a compressor
startup."71 Shanley and

Melhem[ 81 suggest that the AFT's do a reasonable job of
predicting reactivity of compounds.
An example problem from Welker and Springer"o01 is shown
in Table 1. It is based on applying thermodynamic concepts
to the design of a knock-out drum. Part of the problem has
been completed using a spreadsheet (see Tables 10, 11,12)
and part is left for the reader. The problem can also be
extended, as illustrated, to include other topics such as
flammability limits.

The students help teach the class. There are at least nine
(see Table 13) multimedia presentations based on profes-
sionally prepared case studies. Each student presents a pro-
fessionally prepared case study (see Table 14) and a pre-
sentation on the chemical process industry incident of
their choice. The latter is prepared, along with a written
report, toward the end of the course. One student had
photographs from his grandfather of the Texas City fer-
tilizer incident, which occurred after World War II, and




1 C3H
I C3H8

S air fuel mixture
Stoichiometric Composition (Cst)
----Safety limit for air leak
7 Experimental Flammable Region
E Predicted Flammable Region

0%o I I I .


Figure 2. Propane Flammability Diagram Constructed
Using Excel (after Bodurtha2').


Potential Homework Assignments for
Explosion Prevention Technical Elective (both years included)
Crowl and Louvar (CL); Welker and Springer (WS); Bodurtha (B); Stull (S)

1. Flammability
2. Flammability, AIT

Explosions: Mechanical, Chemical
Prediction via AH.omb
Chemical Classification
Inerting and Pressure Relief
Dow F&EI

S; CL 1,6 CL 6-1,3; WS 41
CL 6, B 1-2 Constructing a flammability diagram
CL 6-4-6 (LFL, UFL); Cl 6-10 (AIT)
S; CL Ideal gas; WS 14, 12, 34-39 etc.
S 1-2 CL 6-13; butane AFT; WS 37
S Assigned chemical MSDS
S Willey's Nitroaniline and Kinetics 18, 12
CL 7, 8 Willey's Seveso; CL 7-6,28; 8-2,9; 9-12
CL 10 WS 46; CL 10-7,10
CL; Dow F&EI1171 In-class completion of Dow's workbook problems

Chemical Engineering Education

% fuel

- L


which he researched and reported on. Another student
researched the reactor explosion that occurred in her
hometown of Lake Charles, Louisiana.

This course is offered to senior engineering and science
students as a technical elective each fall. By this time the
students have had thermodynamics, mass transfer, the first
capstone design course, and are in reactor design. Thus, they
have the needed background and their appetite has been
whetted via the capstone design course. The class enroll-
ment has grown with time, and representatives from local
industry have provided positive feedback on the course.

Additional materials can be found on the following web
sites: (MSDS's)

Thermodynamic, Reaction, and Composition data for Welker and Springer 39

Compound Formula Al form Tmax

Carbon dioxide
Water vapor
Exhaust gas

C6H14 -166,902
02 0
N2 0
CO2 -393,509
H20 -241,818

B C D Coeff. Cone.
in Air

1,500 3.025 53.722 -16.791 0.000


3.280 0.593 0.000
5.457 1.045 0.000
3.470 1.450 0.000
125.515 37.613 0.000

-4.665 *
(Hazardous properties of materials: physical hazards,
such as flammability and corrosivity; toxic effects such
as carcinogenicity, toxicity, and target organ informa-
tion; and regulatory requirements) (MSDS's)
dx2home.html (DX-2 high explosives science and tech-
nology) (George Glob Purdue LOX
to ignite charcoal) (structures and
physical properties of chemicals)
The following slide/video sets are also available:
* Bethea, R.M., Phillips 66 Company Explosion and Fire
at Pasadena, TX, AIChE SACHE (1996)
Bethea, R.M., Process Safety Management with Case

Student Presentations of
SACHE Modules and Their Incident

Sept 3 Explosions slides
Sept 19 Miscellaneous case slides

Sept 26
Oct. 1
Oct. 3
Oct 8
Oct 10
Oct 17

Nitroaniline reactor slides
Seveso Dioxin release slide
Dust, vapor explosions apparatus video
Explosion control video
Flixborough slides
HAZOP slides, HAZOP video

Oct 22,24 Student report presentations

Student Responsibilities for
Slide/Video Presentations

Get materials ahead of time and preview

For slide, read script and edit to make
them clearer and more concise

For both slides and video, prepare an
introduction to orient the class and follow
questions to reinforce key points

Be prepared to answer questions


1 0%
9.5 21%
35.74 79%
6 0%
7 0%

Final Temperature Calculation for Welker & Springer 39

AH rxn (J/mol) 3,886,878 =Hexane AH_form-C02_coeff*CO2_AH-form-H20 coeff
*H20 AHform
Tin (K) 298.2 Given
Tout (K) 2,839.4 Calculated < Tmax?
AH process 0 AHrxn AH sensible set = 0 by changing To,
AH sensible 3,886,878 =(xg A*(Tout-Tin) + xg B/2*(Tout^2-Tin^2)/10^3-xg D*
(1/Tout- 1/Tin)*10^5)*R

Pressure Calculations for Welker & Springer 39

P initial 5.0 psig
P final 183.1 psig =(P_initial+14.7)*Tout/Tin*SUM((N2_coeff):(H20 coeff))/
SUM((Hexane coeff):(N2_coeff))-14.7
P ratio 36.6 P final / Pinitial
P design 50.0 psig Given
safety factor 4.0
P safe 200.0 psig P design (safety factor)
P final

Fall 1998

Studies: Flixborough (England), Pasadena (Texas), and
Other Incidents, AIChE SACHE (1996)
Chevron Process Hazards Management Video
HAZOP A Team in Action, Chevron Video
HAZOP A Practical Element of Process Hazards Man-
agements, JBF and Amoco Video
The following publications are available:
* SACHE Faculty Workshop: Characterization and Con-
trol of Chemical Process Hazards, AIChE SACHE
(1996) (Electrostatics case study and experimental dem-
Dow's Fires and Explosions Index Hazard Clas-
sification Guide, AIChE Publishers, New York,
NY (1994)
Goldfarb, A.S., G.R. Goldgraben, E.C. Herrick,
R.P. Ouellette, and P.N. Cheremisinoff, Organic
Chemicals Manufacturing Hazards, Ann Arbor
Science (1981)
Guidelines for Evaluating the Characteristics of
Vapor Clouds, Explosions, Flash Fires, and
BLEVES, CCPS/AIChE Publishers, New York,
NY (1994)
Kletz, T.A., What Went Wrong? Case Histories of
Process Plant Disasters, Gulf Publishers (1988)
21. Lees, F.P., Loss Prevention in the Process
Industries, Buttrworth-Heinemann (1980, 1996)
One Hundred Largest Losses: A Thirty Year Re-
view of Property Damage Losses in the Hydrocar-
bon-Chemical Industries, (Chicago: M&M Pro-
tection Consultants, 1986)
The Phillips 66 Company Houston Chemical Com-
plex Explosion and Fire: A Report to the Presi-
dent, OSHA Report, April (1990)

1. Crowl, D.A., and J.F. Louvar, Chemical Process Safety: Fun-
damentals with Applications, Prentice Hall, Englewood
Cliffs, NJ (1990)
2. Bodurtha, F.T., Industrial Explosion Prevention and Protec-
tion, McGraw Hill, New York, NY (1980)
3. Ashar, S.M., and B.K. Harrison, The ASTM Computer Pro-
gram for Chemical Thermodynamic and Energy Release
Evaluation CHETAH 7.2, American Society for Testing
and Materials (1997)
4. Gay, D.M., and D.J. Leggett, "Enhancing Thermal Hazard
Awareness with Compatibility Charts," J. Testing & Evalu-
ation, 477 (1993)
5. Braton, J., and R. Rogers, Chemical Reaction Hazards, Gulf
Publishers (1997)
6. Johnston, H.M., The Dow Chemical Company, personal com-
munication, May 27, 1998
7. Welker, J.R., University of Arkansas, personal communica-
tion, March 23, 1998
8. Stull, D.R., Fundamentals on Fire and Explosion, AIChE
Publishers, New York, NY (1976)
9. Hoefflich, T., "Orchestration of Techniques in Reactive
Chemicals Problem Solving," unpublished paper available

from the author via e-mail (1996)
10. Welker, J.R., and C. Springer, Safety, Health, and Loss
Prevention in Chemical Processes, AIChE CCPS (1990)
11. Willey, R.J., Problem Set Heat Transfer, AIChE SACHE
12. Willey, R.J., Problem Set Kinetics, AIChE SACHE (1996)
13. Willey, R.J., Properties of Materials, AIChE SACHE (1996)
14. Willey, R.J., Seminar on Nitroaniline Reactor Rupture,
(Sauget Illinois 1969), AIChE SACHE (1994)
15. Willey, R.J., Seveso (Italy) Release Accident, AIChE SACHE
16. Willey, J.R., (ed.), SACHE Faculty Workshop: Characteriza-
tion and Control of Chemical Process Hazards, May 17-20
17. Ainsworth, S., "Arco Agrees to Pay a Record Fine Following
Blast," C&EN, Jan. (1991)
18. Shanley & Melhem, G.A., "On the Estimation of Hazard
Potential for Chemical Substances," International Sympo-
sium on Runaway Reactions and Pressure Relief Design
Proceedings, Boston, MA (1995) 0

Book Review: Chemical Processes
Continued from page 268.

The next major section of the book concentrates primarily
on heuristics, both for operating conditions and design pa-
rameters. Much of this material is summarized in tabular
form, providing students with a quick reference that they report
was heavily used as they worked on their design projects.
In the third section of the book, the main focus is on
getting the most out of existing processes and equipment.
Again, there is a wealth of valuable information presented
here, much of which cannot be found in other design texts.
This material will probably find most use in a two-term design
course, or in connection with some types of design projects.
The next section covers process synthesis and optimiza-
tion, including heat integration (pinch technology), as well
as use of process simulators. The chapter dealing with pro-
cess simulation provides a particularly good introduction to
this topic, emphasizing the need to start simple and warning
of the various pitfalls students are likely to encounter.
The final section of the text deals with various other issues
often covered in the design course, including ethics, safety
and environmental issues, and communications (written and
oral reports). This last material is especially noteworthy and
includes a detailed report writing "case study" in which a
report is critiqued and a checklist of common errors pro-
vided. This is by far the strongest and most detailed treat-
ment of technical communications issues that I am aware of
in a design textbook.
In addition to the costing data noted above, appendices
also provide details needed in some of the examples and
back-of-chapter problems, and present three design project
pairs (one part of the pair a "grass roots" project, the other
focusing on improvement of an existing facility). Additional
design projects are available from the authors.
Perhaps more so than any other required course in the
Chemical Engineering Education

chemical engineering curriculum, there seems to be no com-
plete agreement as far as just what topics should be covered
in the design course. The emphasis of the course can be
significantly different at one school compared to another.
This book does a very good job of covering what I, and I
think probably most others, would consider the real core
material, while providing a good framework for launching
into most other types of material the instructor might want to
cover. For example, I like to teach students a little about
corporate financial analysis, culminating in them being asked
to analyze a real company's financial data as reported to the
SEC. This follows quite easily from the material on eco-
nomic analysis in the text. I also try to provide students with
some additional material on the mathematical modeling con-
cepts and numerical methods underlying process simulation
and optimization. Again, this follows very nicely from the
material on simulation and optimization in the text. Overall,
my feeling is that the material presented in this text provides
a very sound core around which to teach a design course.
In addition to being a good text, this book also has features
that should make it a particularly useful reference book for
young and inexperienced engineers. These features include
the compilation of cost data and correlations, and the tabula-
tion and discussion of many rules-of-thumb for design and
for process operation and improvement. Seniors who use the
book in their process design class may well find continued
use for it as they begin their careers in engineering.
The availability of this book represents an important new
option when it comes to choosing a text for the process
design course. As discussed above, I have foundthat it is a
very well-written text providing excellent coverage of the
core material in process design, and that it goes well beyond
this to provide valuable features not found in other current
texts. Many instructors may also appreciate the inclusion of
some excellent design projects in the text (and the availabil-
ity of additional projects from the authors). Faculty who
teach process design should give this book strong consider-
ation as a textbook for their course. C

book review

Phase Equilibria: Measurement and Computation
by J. David Raal and Andreas L. Muhlbauer
Published by Taylor & Francis, 1900 Frost Road, Suite 101, Bristol,
PA 19007-1598; 461 pages, $89.95 (1998)

Reviewed by
Scott W. Campbell
University of South Florida

I have long been looking for a book that would give equal
attention to the experimental and computational aspects of
Fall 1998

phase behavior. Lately, this has become even more desirable
since some of the most recent experimental methods are
indirect and require substantial calculation in the reduction
and analysis of data. Therefore, it was with a great deal of
interest that I reviewed Phase Equilibriq: Measurement and
Computation by Raal and Muhlbauer.
The coverage in the text is not as broad as one would
expect from the title; only vapor-liquid (VLE) and binary
liquid-liquid (LLE) equilibria are covered, and methods for
mixtures containing polymers or electrolytes are not de-
scribed. But this is more a criticism of the title than of the
work since there is more than enough material in the au-
thors' scope to fill a book.
The book is comprised of 18 chapters distributed in four
parts: Low-Pressure Phase Equilibrium Measurements (Part
1, Chapters 1-4); High-Pressure Phase Equilibrium Mea-
surements (Part 2, Chapters 5-9); Low-Pressure Phase Equi-
librium Calculations (Part 3, Chapters 10-13); and Computa-
tion and Thermodynamic Interpretation of High-Pressure
Vapor-Liquid Equilibrium (Part 4, Chapters 14-18).
Chapter 1 covers the thermodynamic fundamentals needed
for an understanding of the experimental methods and com-
putations. Definitions of partial molar properties, activity
coefficients, fugacity coefficients and the like are given, and
useful relations such as the coexistence equation and Gibbs-
Duhem equation are derived. The level of detail is appropri-
ate to the purpose of the book and only a small amount of
extraneous information is given.
Chapter 2 covers the standard techniques for low-pressure
VLE measurement and also includes discussions of semim-
icro techniques and the measurement of infinite dilution
activity coefficients. Chapter 3 describes methods for mea-
surement of LLE. These chapters, along with Chapters 6
and 7 (discussed later), are among the strongest in the book.
The methods are described with ample line drawings and
photographs and with textual detail that any experimenter
will appreciate. The advantages and disadvantages of each
technique are noted.
Gas chromatography is often used to determine composi-
tions of coexisting phases, and the authors devote Chapter 4
to this subject. In addition to discussing response factors and
composition analysis, the authors describe a technique de-
veloped by one of them to prepare standard gas mixtures for
detector calibration. Enough detail is given that the reader
could build the device.
After giving some background about high-pressure behav-
ior in Chapter 5, the authors survey techniques for high-
pressure phase equilibrium measurement in Chapters 6 (dy-
namic methods) and 7 (static methods). The coverage here is
as detailed as in Chapters 2 and 3. In Chapter 8, the authors
step the reader through the design, construction, and opera-
tion of an experimental facility for the measurement of high-
Continued on page 285

Random Thoughts...



North Carolina State University Raleigh, NC 27695

Scene. Saturday, December 20, 8:45 p.m. The annual
Bilgewater University Materials Engineering Depart-
ment Christmas party is in full swing at the home of
Ray, the department head. This year is unusual be-
cause Don (the Provost) and Harry (the Dean of
Engineering) have shown up. Also attending are Eddie
(the Assistant Dean for Academics) and most of the
department faculty. Several people have commented
that the punch tastes a bit strange.

Don: "Hello there, young man-I don't believe we've met."
George: "George Murchison, joint in biomedical and mate-
rials-just finished my dissertation at Berkeley and started
here this fall."
D: "Ah yes-aren't you the one who's been working on
truth serum? Fascinating stuff-oubt that it works, of
G: "Oh, it works-I've got some graduate students testing
sodium pentathol and I could tell you some stories that
D: "Yes yes, I'm sure you could-anyway, I'm pleased you
chose to join the Bilgewater family."
G: "It wasn't exactly a choice-I sent resumes to 65 schools
and this was the only offer I got."
D: "I know what you mean-I send my resume out when-
ever a chancellorship opens up but I never make it past
the first interview."

Eddie: "Hey, Janet. There's something I just learned about
E.C. 2000 that I'd like to run by you."
Janet: "E.C. 2000. That's when all the computers crash,
E: "No, that's Y2K-E.C. 2000 is the new system they're
going to use for..."

J: "I'm spellbound, Eddie, I truly am, but I need to grab
Harry while he's here so I can snatch some of his slush
funds for a lab renovation-catch me up on it later."

Harry: "So, Al, how's life going for the assistant professors
these days?"
Al: "What life-this semester I taught two new courses and
sent out three proposals and Ray told me I'd better get
more focused if I want to make tenure. Last week my
wife mailed me a postcard at work to ask me how I've
been and whether I have any spots on my calendar for
her next year."
H: "Cute. So Ray doesn't think you're working hard
A: "He has no clue about what I'm doing-I said some-
thing about it the other day and he told me to stop
H: "Has a unique management style, doesn't he...anyway, I
heard you got great teaching evaluations last semester
and-oh, hello, Bill. I was just commenting on Al's
teaching evaluations. Right up there, they tell me."
B: "Yeah, they're almost obscene. If he keeps doing that,
he'll end up with an Outstanding Teacher Award."
H: "Whoa, Al, don't want that! I'd hate to see you wind up
like most of the other untenured professors who won one
of those things."
B: "Yeah, take it from me, kid-if you're spending that
much time on teaching, you're not paying enough atten-
tion to the important stuff."
A: "But...."

Irv: "Nice party, Ray-better than your usual snoozers.
What made Harry and Don decide to show up?"

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Ray: "They know I'm looking to get out of this dump, and if
I go my $12 million scanning microscope project goes
with me and the overhead from that grant is keeping
them alive."
I: "I thought all the grumbling about you at the last depart-
ment review might have had something to do with it-
aren't the assistant professors on your case about dump-
ing all those new course preparations on them and push-
ing them to crank out ten proposals a year?"
R: "Ah, they're just a bunch of whiners. Besides, who
would Harry get to replace me? Jones? Frobish? Daffy
Duck? Bring Eddie back from the Dean's office?"
I: "How about me? I could get drunk every day before
breakfast and do a better job than you."
R: "Well, you've been practicing the first part for 15 years-
my only question is whether they'd catch you first for
mismanaging funds or working your way through the
secretary pool."

Eddie: "Hey, Gene-I was just looking at E.C. 2000 the
other day, and I noticed that it..."
Gene: "E.C. 2000-that's the Web site with those porno
pictures of movie stars, isn't it?"
E: "No, it's where they define the criteria for..."
G: "Oh yeah...good talking to you, Eddie, but right now I
need to check with Don about something."

Don: "Larry, my boy. How are things in the microelectron-
ics business these days?"
Larry: "Couldn't be better-I brought in another $2 million
in industry money last month with a couple of well-
placed phone calls."
D: "Splendid, splendid-I don't know how you keep doing
it with those same old results year after year."
L: "Nothing to it-the companies don't do real research
any more and I just have to mention 'superconductivity'
and 'profits' in the same paragraph and they toss me the
keys to the just burs me that most of it gets
skimmed by that army of beancounters you keep em-
D: (chuckling) "Now now, you know how hard it would be
to get anything done on this campus without my hard-
working staff-besides, at the Council of Provosts in
Maui last month I was the envy of all those bozos when I
told them Bilgewater has a bigger administrative staff

than Harvard."
* *

Charlie: "So the blond says, 'I don't know-which one was
the horse?'"
Joe: (laughing) "That's a good one. You hear the one about
the blond, the undertaker, and the rabbi who..."
Eddie: "Hi, guys. Did either of you know that E.C. 2000
J: "E.C. 2000-that's the real estate company I listed our
house with three years ago...let me tell you, that agent I
had couldn't sell flies to a frog-she was so..."
C: "No, dummy-E.C. is that hospital the
other day this woman was rushed in with a liver infec-
tion caused by contaminated tofu that some beef produc-
ers were slipping into health food stores, and then the
phony doctor who was really an escaped sex offender
E: "No, no-E.C. 2000 is this outcome-based assessment
system where you have to..."
C: "Oh right-I remember now...excuse me, I need to
freshen my drink."
J: "Hey, Harry is free-I better grab him before Ray
snatches him. Later, Eddie."

Harry: "How's it going, Ray?"
Ray: "Just peachy. I've got a bunch of total losers on my
faculty-tiny grants, no release time, constantly whin-
ing-and undergraduates complaining about huge lec-
ture sections and absent professors and incomprehen-
sible TAs, and 40 hours a week of useless committee
meetings, most of which you, my wife is giv-
ing me a hard time about my working nights and week-
ends, my kid just got her nose and God knows what else
pierced, and I've got killer hemorrhoids. Thanks for
H: "Boy, I'd really find all that distressing if it weren't for
the fact that I couldn't care less about it. I've got my own
problems, you know."
R: "Oh yeah-like what? They won't let you add three new
associate deans to the five you already have?"
H: "Well, for starters the trustees are talking about institut-
ing post-tenure review, including for administrators."

A sudden silence envelops the room, and no tenured faculty
member says another wordfor the rest of the evening. 0

Fall 1998

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

Sfi class and home problems

The object of this column is to enhance our readers' collections 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 and that
elucidate difficult concepts. Manuscripts should not exceed ten double-spaced pages if possible
and should be accompanied by the originals of any figures or photographs. Please submit them to
Professor James O. Wilkes (e-mail:, Chemical Engineering Depart-
ment, University of Michigan, Ann Arbor, MI 48109-2136.



University di Padova 1-35131 Padova PD, Italy

Understanding the principles of batch distillation is a
hard task for undergraduate chemical engineering
students because it belongs to the class of dynamic
processes, which are usually only a small fraction of the
processes analyzed during a basic unit operations class. An
undergraduate student who runs across this process for the
first time usually has insufficient background in process
modeling and process dynamics to deal with it. This is why
the classical approach for teaching binary batch distilla-
tion['l21 still relies on the quasi-steady-state assumption, which
enables the operation to be analyzed by means of McCabe-
Thiele diagrams. The students are very familiar with these
diagrams since batch distillation is normally taught after
continuous distillation. But while a McCabe-Thiele analysis
indeed helps to get an idea of how the fractionation is pro-
ceeding in a given batch column, it only gives a "static" (i.e.,

instantaneous) view of the process. As a result, the use of the
Rayleigh equation is necessary in order to provide an overall
picture of the fractionation (i.e., in order to "integrate all the
McCabe-Thiele diagrams over time"). Therefore, a purely
graphical approach is impossible for analyzing a batch distil-
lation operation.
The students' troubles even increase when they are intro-
duced to the concept of batch distillation optimization, be-
cause they believe that optimization is a tough subject that
involves high-level mathematics and requires outstanding
programming skills. While this is somewhat true for many
optimization problems, it is also true that simple case studies
requiring a moderate-to-low programming ability can be
The problem considered in this paper is the project assign-
ment given to my undergraduate students during a unit op-
erations class. The project has two basic aims:
- Teaching the principles of batch distillation operation
and optimization by means of a case study.
Teaching the students how to develop, write down, and
use a simple (yet meaningful) computer program for
solving a chemical engineering problem.
The project was assigned after the necessary prerequisite
of batch distillation was taught using the classical approach.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Massimlliano Barolo is Assistant Professor of
Chemical Engineering at the Universiti di
Padova (Italy). He received his PhD from the
same University in 1994. His teaching respon-
sibilities include the development and class-
room teaching of tutorial exercises on the de-
sign and rating of chemical process equipment.
His main research interests are in the field of
dynamics, optimal operation, and control of dis-
tillation columns.

As an additional required background, a basic knowledge of
numerical methods was requested.

Problem Statement
Money & Money, Inc., is a small company producing
intermediate chemicals for the chemical process industry. A
total amount of 45,000 kg of a boiling methanol/ethanol
liquid mixture of composition 59 wt% methanol is obtained
as a by-product from the batch production cycle. The sales
department believes that it could sell the methanol, provided
that its concentration is increased to at least 89 wt%. The
plant manager has asked you to make a preliminary study in
order to check the possibility of obtaining the methanol at
the desired purity.

The methanol can be puri-
fied in the same batch recti-
fier where other separations
are performed during the pro-
duction cycle at atmospheric
pressure. Clearly, before the
column can be used for "your"
separation, it must be prop-
erly purged. Currently, it takes
about 2.5 h to manually per-
form the charge/discharge/
purge operations. An auto-
matic system could be bought
and installed if necessary, so
that the changeover time
could be reduced to about 0.5
h. The column is normally
operated with a constant
boilup rate of 50 kmol/h. The
total fractionating capacity of

the column is equivalent to
16 theoretical stages (including the reboiler). The column is
a randomly packed one, so you will not endanger your
calculations very much by assuming that the column holdup
is negligible; the reflux drum holdup is also negligible.
The plant manager recommends keeping the column op-
eration easy. "Operators are reluctant to change," he used to
say. "They get used to operating the column at a constant
reflux ratio; they need to operate the column that way!" He
also wonders whether revamping the column (perhaps with a
highly efficient structured packing) could bring any advan-
tage. The sales office suggests that you evaluate the column
productivity, P, through a simple formula:

P = Dw (kg / h) (1)
tdist + change
where D, is the weight amount of methanol on specification,
td,s, is the distillation time, and change is the changeover time.
Now it's up to you! Money and Money, Inc., expects that
Fall 1998

you can determine the operating conditions that guarantee a
maximum in the column productivity. Be convincing!

Problem Analysis
Solution of the problem requires the determination of the
optimal reflux ratio for the operation of a given batch col-
umn (the batch rectifier is shown in Figure 1). The objective
function, P, (column productivity) is actually the same as the
capacity factor CAP introduced by Luyben.131 The only dif-
ference is that CAP is defined on a molar basis, while P is
defined on a weight basis. The solution assumes that no
costs are associated with by-product disposal. The feed mix-
ture is also assumed to be at its boiling point.

The column productivity di-
rectly accounts for two differ-
ent effects of the reflux ratio
on the column performance:
A high reflux ratio in-
creases the amount of dis-
tillate product, D, eventu-
ally collected, but it also
increases the distillation
time (i.e., the energy costs).
A low reflux ratio results in
a faster operation, but the
amount of collected prod-
uct decreases.
The column and reflux drum
holdups are assumed to be neg-
ligible in such a way that no
differential equations need to
be integrated for simulating the
column (quasi-steady-state as-

sumption). This allows the stu-
dents to focus mainly on the characteristics of the distillation
process rather than on the numerical aspects related to the
integration of the system differential equations. Therefore,
the computational time is minimized, allowing the project to
be worked out easily on a PC. But in no way does this
simplification relax the problem to a trivial one, as will be
shown below.
The productivity function must be calculated at different
values of the reflux ratio. For each value of the reflux ratio,
the initial distillate composition, xD, is first to be searched.
The correct value of x,i is the one that enables 16 stages to
be stepped off exactly between D,i and the feed composition,
XF. Then, the problem can be tackled in different ways, some
of which are illustrated below. A discussion on the ease of
implementation of these approaches is given later.
Approach #1
As a consequence of the quasi-steady-state assumption,
the total amount, B, of by-product accumulated in the reboiler


Figure 1. Sketch of the batch rectifier.

at the end of the batch can be calculated through the Rayleigh

fn B fdXB (2)
F J xD -XB
where XB,,e represents the reboiler composition at the end of
the batch. The other symbols are listed in the Notation to this
paper. This final composition is not known a priori, and
must be searched numerically, a possible procedure for which
a) Choose one value r of the reflux ratio.
b) Determine the relevant value of x,,,
c) Guess a value for xB,,,.
d) Run the simulation at a number of "slowly" decreas-
ing values of the distillate (i.e., overhead vapor)
composition x,; for each value of XD, step off 16
stages in order to get the reboiler composition x,.
e) Stop the simulation when xB = xend is found.
f) Evaluate B numerically through Eq. (2).
g) Estimate the accumulated distillate product composi-
tion XD by combining the total material balance
F=D+B (3)
with the more volatile component balance

FXF = DXD + BXBend (4)
h) If the value of x, is not the desired one (i.e., it is not
equal to the product specification XD,spec within a
specified tolerance), make a new guess for X,,end and
go back to d) until convergence is reached.
i) Calculate the distillation time, tdit:

(r +l1)D
tdist (r+ (5)

j) Choose another value of the reflux ratio and restart
from b).

Modified Approach #1
Basically, the calculations are the same as before, but the
search of XB,end is not performed by trial and error. Rather,
after step b) has been executed, the calculation proceeds as
c) Decrease the overhead composition by a "small"
d) Starting from this new value of the overhead composi-
tion, step off 16 stages and get the reboiler composi-
e) Evaluate B numerically through Eq. (2).
f) Estimate the accumulated distillate product composi-

tion XD by combining the total material balance and
the component balance.
g) If the value of xD is not the desired one, go to c) in
order to get a new overhead composition.
h) Calculate the distillation time through Eq. (5).
i) Choose another value of the reflux ratio, and go to b).

Approach #2
As a consequence of the quasi-steady-state assumption,
the following balances hold at any time step Ik:

Bjk-1 = Blk + ADIk (6)

(XBB)|k-1 = (XBB)Ik + (XDAD)k (7)

where ADIk is the amount of distillate product collected
between time steps Ik and Ik. By combining Eqs. (6) and

ADk XBk- XBjk
XDIk XBk B (8)

The calculation procedure is:
a) Choose one value for the reflux ratio.
b) Determine the relevant value of x,, and set
XDik-1 = XD,i-
c) Consider a "small" decrease AxD of the overhead
vapor composition, and set xDok = XDik AxD.
d) Starting from xDik, step off 16 stages in order to get the
bottom composition xBIk.
e) Using Eq. (8), calculate the corresponding increase
ADIk of the amount of distillate product.
f) Calculate the corresponding increase A(XDD)Ik of the
more volatile component amount: A(XDD)Ik = (DAD)Ik.
g) Calculate the total amount Dik of distillate product
(Dik = Dik_1 + ADIk ) and the total amount of bottom
product ( Bik = BlkI ADk).
h) Calculate the total amount (XDD)Ik of the more volatile
component in the distillate product:
(XDD)Ik = (XDD)Ik_1 + A(XDD)Ik.
i) Evaluate the average distillate product composition:
(x D)1
D|k = Dk (9)
j) If the value of RD is not the desired one, go to c)
proceeding to time step 1k +I.
k) Calculate the distillation time through Eq. (5).
1) Choose another value of the reflux ratio, and go to b).
A modification of this approach involves differentiating
the Rayleigh equation (Eq. 2) and then considering the fol-
Chemical Engineering Education

lowing finite difference approximation of the resulting dif-
ferential equation:
AB AxBIk (10)
Blk XDk-1 XBik-

Therefore, at any time step |k, the bottom product is de-
creased by the "small" amount AB; so, the value of AxBlk is
calculated by means of Eq. (10), and the reboiler composi-
tion is calculated as xBk = XBlk-1 + AXBIk. Then, the overhead
composition xDlk is calculated by trial and error by ensuring
that 16 stages are stepped off between xBik and XDik at the
relevant value of r. Finally, the value of XDIk is determined
by applying Eq. (4) to the current time step. The knowledge
of XDIk allows stepping further to time Ik +1.

No further information was given to the students other
than that in the problem statement. No suggestions were
made about which kind of solution procedures would keep
the computational load to a minimum. The students were
only informed that they could find me in my office at any
time if they wanted to discuss any matter about the project
development. About a dozen students asked for some kind of
help (typical questions and comments were "Where can I get
the Antoine equation coefficients for these components?" or
"My Newton-Raphson loop does not converge!" and even "I
believe you have not supplied all the data to me.")
Listed below are some computational aspects that might
be taken into account when giving the project assignment.
Some of them can also be discussed with the students in the
classroom when they have submitted their reports.
Thermodynamics The procedures outlined in the previ-
ous section are independent of the thermodynamic model
used to describe the vapor-liquid equilibria (VLE) of the
mixture. In the proposed project, a methanol/ethanol mix-
ture was chosen. No thermodynamic data were supplied to
the students. Thus, first they had to identify the general
behavior of this system (nearly ideal mixture), and then they
had to determine the system relative volatility a. Most stu-
dents derived a from the pure component vapor pressures; a
few of them looked for some experimental VLE data for this
system in the literature and fitted a to these data by using a
popular spreadsheet software. Depending on the time avail-
able and on the student preparation, one may complicate the
problem by considering the separation of a non-ideal binary
mixture, using any thermodynamic model for the description
of the liquid-phase activity coefficients.
Determination of the product amount and composition *
At least one Newton-Raphson (or similar method, for ex-
ample the bisection method) convergence loop needs to be
Fall 1998

implemented to calculate the initial distillate composition
XD, at each value of the reflux ratio. Depending on the
solution approach considered, a second convergence loop
may or may not be needed.
Approach #1 was definitely preferred by the
students. About 80% of them resorted to this
approach or to some modification of it. Clearly, the
disadvantage of this approach is that a convergence
loop must be used to determine the final reboiler
composition XBend.
Modified approach #1 was implemented by about
15% of the students. According to this approach,
only one convergence loop (the xi-search loop)
needs to be executed. Thus, this solution procedure
is significantly easier than the previous one.
Approach #2 is computationally very easy. In fact,
after XD. has been determined at the relevant value
of r, the calculation proceeds in a straightforward
manner (no iterative calculations are necessary).
Only a few students considered this approach,
probably because the analysis phase is somewhat
harder in this case.
Note that the suggested modification to Approach #2 does
require a further convergence loop to be implemented, be-
cause the calculation of xDok is not direct.
Numerical integration When Approach #1 or Modified
Approach #1 are considered, Eq. (2) needs to be integrated
numerically, most simply by approximating the integral as a
summation. This approximation is quite rough, but never-
theless many students found it convenient because it is very
simple to implement. But it gives sufficiently accurate re-
sults only if the integration interval is partitioned in very
small steps, which may be time consuming. Some students
preferred to use slightly more sophisticated methods (the
trapezoidal method, or even Simpson's rule).
Conversely, numerical integration is not required when
Approach #2 is considered.
Finding the optimal reflux ratio The easiest way to find
the optimal reflux ratio is by graphical inspection of the
productivity curves. One simply stores the values of the
distillation time and of the product amount in a file at differ-
ent reflux ratios, and then plots the productivity functions for
the two different changeover times by appropriate software.
If the students have a sufficient background on optimization
algorithms, they may be requested to try and implement a
numerical search algorithm. Although the graphical method
may not seem too "brilliant," it allows a somewhat more
critical analysis of the results, as will be shown in the next

Figure 2 shows the results obtained for the two values of

the changeover time changee = 2.5 h and tchange = 0.5 h). A
number of considerations can be drawn.
In general, the distillation time grows with the reflux
ratio at a higher rate than the accumulated distillate
product does. This would not justify the existence of a
maximum in the productivity curve unless one recog-
nizes that the total operation time depends also on the

ever the value of the reflux ratio.
* If the column charge/discharge is performed automati-
cally, operating the column close to the optimum re-
sults in much off-cut being stored in the reboiler at the
end of the batch. Therefore, one needs to recycle this
product to the next batch. A simple procedure similar
to the one developed by Luyben[31 could be considered

time needed to
charge, discharge,
and possibly purge
the column. There-
fore, the optimal re-
flux ratio depends on
* Reducing the
changeover time in-
creases the column
productivity by about
12% for this problem.
Whether or not a sys-
tem enabling auto-
matic column charg-
ing, discharging, and
purging should be in-
stalled depends on
the cost of this sys-
tem and on the sell-

600 28
570 tchng 05 -24
S480 change 16
450 -----12
0 420- -8
S390- -st "(h)
D360 .1000 (kg) 4
1.55 1.60 1.65 1.70 1.75 1.80 1.85 1.90 1.95
reflux ratio

Figure 2. Column productivity, distillation time, and amount of
distillate product for the two values of the changeover time.

ing price of the recovered methanol.
* When a large changeover time is necessary (tchange=2.5
h), a high reflux ratio should be employed (r = 1.82),
resulting in a prolonged distillation (td,, = 19.8 h), with
a large amount of product recovered (D, = 11620 kg).
Conversely, when a short changeover time is allowed
changee = 0.5 h), the optimal operation is attained at a
lower reflux ratio (r = 1.68), in a much shorter distilla-
tion time (tdist = 10.1 h) and with less product recovery
(D, = 6215 kg). In practice, when the changeover time
is large, the optimal operating conditions are such that
frequent charges/discharges/purges of the column are
* The sensitivity of the optimal operation to the reflux
ratio is somewhat larger when the changeover time is
short, which means that operating the column at condi-
tions close to the optimum is a bit harder when change =
0.5 h. In fact, small deviations of the reflux ratio from
the optimal value may cause a significant decrease of
the column productivity. Conversely, when change = 2.5
h, a productivity close to the optimal one can be ob-
tained by operating the column within a quite wide
range of reflux ratios (r e [1.75; 1.9]). This is because the
productivity curve is less steep; in this case, the total
product recovered ranges between 9000 and 14000 kg.
Productivity is always larger when thange = 0.5 h, what-

need to be performed in

for off-cut recycling; a
steady-state chain of
batches would be reached
within a few cycles. This
analysis is beyond the pur-
pose of the project assign-
ment, however.
When manual charge/
discharge is necessary, op-
erating the column close to
the optimum requires 15 to
24 h (distillation time) plus
2.5 h (changeover time).
Therefore, the column
would be dedicated to this
separation for at least 17.5
h per batch. One should
check whether this time in-
terval is compatible with
the other separations that
the same column during the

production cycle. Conversely, with a changeover time
of 0.5 h, only 10.6 h need to be allocated for each
methanol/ethanol separation in this column. It might be
easier to allocate this time within the whole production
The current separating capacity of the column (N = 16
ideal stages) is very close to the separating capacity of
a column having an infinite number of stages. In fact,
increasing the number of ideal stages (for example,
letting N = 40) leads to Pma = 529 kg/h when tchange =
2.5 h, and to Pma = 595 kg/h when tange = 0.5 h; these
productivities are virtually the same as those repre-
sented in Figure 2. A simple analysis by means of a
McCabe-Thiele diagram shows that the column is
pinched at the bottom for every value of the reflux ratio
when N = 16. Therefore, revamping the column in
order to increase column productivity brings no advan-
tage in this case. Incidentally, since the column is bot-
tom-pinched throughout the batch, the integral term in
Eq. (2) can be computed analytically for a constant
relative volatility system, leading to'[4

n l = l-In F x(1- xF) (11)
F r+l LI-xB 1-XI XF(i-XB)J

Some students actually recognized the existence of the pinch
Chemical Engineering Education

point, so they avoided performing the numerical evalu-
ation of the integral term.

The feedback from the students on this assignment was
good. They liked working on a project that resembles a real-
life problem. The time needed for writing the program and
studying the results was not too long. Many of the findings
outlined in the previous section were spotted by students. In
any case, the alternative solution procedures and the results
were finally discussed in the classroom.
The project proved to be very instructive, but it appears to
be oversimplified for use at a graduate level. In this case, one
may have a look at Diwekar's textbooks51 in order to get
useful instructive material on advanced batch distillation
design, simulation, and optimization.

I would like to thank Prof. Gian Berto Guarise for helpful
discussions. Support by MURST (fondi 60%) is gratefully

B bottom product (kmol)
CAP column capacity factor (kmol/h)
D distillate product (kmol)
Dw distillate product (kg)
F feed charge (kmol)
P column productivity (kg/h)
P maximum column productivity (kg/h)
r reflux ratio
tcae changeover time (h)
tdist distillation time (h)
V vapor boilup rate (kmol/h)
x, reboiler composition (mole fraction)
X.Bend final reboiler composition (mole fraction)
xD overhead vapor composition (mole fraction)
XD average distillate product composition (mole fraction)
oD,spec specification on the average distillate product composition
(mole fraction)
x, feed composition (mole fraction)

a relative volatility
A increase
Ik k-th time step

1. Coulson, J.M., and J.F. Richardson, Chemical Engineering,
Vol. 2, Pergamon Press, Oxford, United Kingdom (1991)
2. McCabe, W.L., J.C. Smith, and P. Harriott, Unit Operations
of Chemical Engineering, McGraw-Hill Book Co., New York,
NY (1993)
3. Luyben, W.L., "Multicomponent Batch Distillation. 1. Ter-
nary Systems with Slop Recycle," Ind. Eng. Chem. Res., 27,
4. Guarise, G.B., Distillation, Absorption and Liquid-Liquid
Extraction for University Students (in Italian), CLEUP,
Fall 1998

Padova, Italy (1996)
5. Diwekar, U.M., Batch Distillation: Simulation, Design, and
Control, Taylor & Francis, London, United Kingdom (1995)

Book Review: Phase Equilibria
Continued from page 277
pressure VLE and, in Chapter 9, they describe equilibrium
cells for multiphase, high-pressure systems. The focus of
both of these chapters is on the authors' own work, and it is
written with a level of detail that one would typically find in
a dissertation. Particular attention is given to experimental
difficulties (usually related to sampling) and techniques to
address them. For this reason, these chapters will be useful
to anyone who plans to build a high-pressure apparatus-
even one based on a different design.
Part 3 of the book covers low-pressure phase equilibria
computations and includes chapters that cover correlative
methods for activity coefficients (Chapter 10), flash calcula-
tions (Chapter 11), predictive methods for activity coeffi-
cients (Chapter 12), and liquid-liquid calculations (Chapter
13). Part 4 covers calculations at high pressure, including
background information (Chapter 14), equation-of-state meth-
ods (Chapter 15), gamma-phi methods (Chapter 16), and
mixing rules (Chapter 17). The book concludes with Chapter
18, which discusses thermodynamic consistency testing.
Parts 3 and 4 of the book do not have the breadth of Phase
Equilibria in Chemical Engineering by Walas or of Proper-
ties of Gases and Liquids by Reid, Prausnitz, and Poling. But
they are written at the depth of the former, and the breadth is
consistent with the experimental techniques that are described.
The discussions of mixing rules and consistency tests, and
the relation between VLE and heat of mixing, are more
detailed than are generally found in other books. Topics that
aren't covered in extreme detail, such as methods for inte-
grating the coexistence equation, are accompanied by enough
references that the reader may find the details elsewhere.
This book will be of limited use as a textbook because it
contains no exercises. But I believe that its coverage of
experimental methods will be highly useful to anyone who
measures phase equilibria. I know that it would have saved
me months of work had it been available a number of years
ago. The sections on computations will probably not (due to
limited breadth) be a primary source of information on this
topic, but they will be a useful supplement since they are
current and cover several topics that are given only passing
mention in other references.
I have additional criticism about the organization of the
book, and I noticed a number of typographical errors and a
missing figure. But the first criticism is probably a personal
prejudice and the second should be fixed in the next printing.
None of these minor complaints will prevent my copy of the
book from receiving a lot of use. n

S SOclassroom




University of Porto 4099 Porto Codex, Portugal

he experiment described in this paper deals with the
study of single-component gas diffusion across a
meso/macro-porous membrane. This work allows not
only for the introduction of important mass-transfer con-
cepts, but also for an initiation to separation membrane
This experiment reflects the basic philosophy of our de-
partment at the University of Porto concerning undergradu-
ate laboratories: it is pedagogically interesting, setup and
maintenance are relatively inexpensive, and the operation is
safe and environmentally friendly.

In single-component gas flow across a porous membrane,
three transport mechanisms can be considered:1'14] Knudsen
flow, viscous flow, and surface diffusion along the pore
walls. The pore diameter mainly determines whether trans-
port within the gas phase is characterized by Knudsen or
viscous flow. Surface diffusion is dependent on the tendency
of the diffusing species to adsorb on the pore walls.
Knudsen flow is dominant when the pore diameter is of
Fernio Magalhfes is Assistant Professor of
Chemical Engineering at the University of Porto,
Portugal. He received his PhD from the Univer-
sity of Massachusetts in 1997. His research
interests involve mass transport and sorption in
porous solids and membranes.

Adelio Mendes received his licentiate and PhD
from the University of Porto, Portugal, where he
is currently Assistant Professor. He teaches
chemical engineering laboratories and separa-
tion processes. His main research interests in-
clude membrane and sorption gas separations.

Copyright ChE Division of ASEE 1998

the order of the molecular mean free path. Collisions be-
tween molecules and the pore walls are therefore very fre-
quent, more so than molecule-molecule collisions. This
mechanism therefore accounts for all the axial momentum
loss. Based on basic gas kinetic theory, the following result
can be derived for the molar flux of a species diffusing
across a membrane under a pressure difference Ph P:1151
EN, (Ph-p) (1)

2 8 RT
where Dk= 2p rp 8
3 nM
From Eq. (1), the ratio of the single-component fluxes of
two different gases (for the same membrane and pressure
drop) is given by

Nk,j (2)

It is interesting to note that relationship 2 is not exclusive
to the Knudsen regime, but can be generalized to bulk diffu-
sion under isobaric conditions-Graham's relation."1
On the other hand, for wider pores, inter-molecular colli-
sions predominate. Molecule-wall collisions still take place
and lead to loss of axial momentum (i.e., pressure drop), but
they are a consequence of a series of molecule-molecule
collisions. This can then be treated as a Poiseuille flow
problem, leading to'[51

2 r (P +P) E r 2
N, r (P (Ph P= rp (- P2 (3)
Sf 8 gRT 2 e e 16 pRT
Knudsen diffusion presents a selective character that is
absent in viscous flow. The Knudsen diffusivity (Dk) is
inversely proportional to the square root of molecular weight,
while on the other hand viscous flow depends on the viscos-
ity (I), which is much less sensitive to the nature of the gas.

Chemical Engineering Education


Both transport processes, Knudsen and viscous, may have a
significant contribution for a certain system. This situation can
be described by combining Eqs. (1) and (3) in a parallel ar-

N= D (P, -P,)+ (P -P (4)
ft RT 16 p f)

The ratio of fluxes for two different gases now becomes

N, Dk, + Dv.i(l+ Ph/P)
Nj Dk,J + Dv,j(1 + Ph IP)

where Dv is defined as D, = Pfrp /16 i.

A membrane permeation module was built (see Figure 1). Six
PVDF (poly(vinylidene fluoride)) hollow fibers (7.3x10-4 m ID,
1.2x10-3 m OD, and 2.6x101' m length), provided by GKSS
Geesthacht (Germany), were assembled inside a transparent
tubular shell. These fibers have no selective dense film. They
present macro and mesoporosity and are used as a mechanical
support layer in gas separation membranes.
The gas feed enters the bore side of the fibers at one end of
the module. Since the other end is closed, all gas permeates

I Epoxy glue
1C ^Hollow fiber

/ Hollow fiber Shell
FESTO fitting (length = 26 cm) (plastic tube
| 024mm)

Figure 1. Sketch of the membrane module.

Figure 2. Sketch of the experimental setup.

through the walls and flows along the shell counter cur-
The entire experimental setup is shown in Figure 2.
Three gases were studied: He, N,, and Ar. Other possible
choices, like 02, H2, or Xe, were eliminated due to their
hazardous nature or higher cost. A metering valve was
used to regulate the pressure at the module inlet, which
was measured with a pressure sensor (0 Ixl05 N/m2
relative pressure range, 103 N/m2 resolution). The perme-
ate flow rate was measured with a J&W Scientific
ADM2000 flow meter (0 1000 seem).
The outlet pressure at each gas bottle was set at 1.5x105
N/m2 (relative), and the metering valve was used to select
the module inlet pressures between 0.1xl05 and 1x105 N/
m2 (relative) in 0.1x105 N/m2 intervals.
The experiments reported in this paper were performed
at T = 200C and P, = 1.011x105 N/m2.

For this experiment, students were asked to
Determine the single-component permeabilities
of the three gases
Identify the intra-porous flow regime
Attempt to estimate the average pore radius of
the membrane
The surface diffusion can be neglected in this system since
the gases adsorb very little on the polymeric membrane.
Thus, only Knudsen and viscous flow must be considered.
According to Eqs. (1) and (3), only for Knudsen flow is
the flux a linear function of the inlet pressures, or in other
words, the effective permeability (defined as L=N/( Ph Pt)
is independent of pressure. Plots of the measured flow
rates (F) as a function of Ph P, are shown in Figure 3.

0 0.2 0.4 0.6 0.8 1 1.2
( P,- P,) x 10"5 (N/m2)
Figure 3. Measured flow rates as a function of
pressure drop for Helium, Nitrogen, and Argon.

Fall 1998

The permeabilities for each gas can be computed directly
from the slopes of the straight lines and, as expected, helium,
the smallest molecule, shows the highest value.

Since N vs. (Ph P() follows a linear relationship, students
are led, at this point, to conclude prematurely that viscous
flow does not play a role in this system. But the conclusive
test consists in checking whether Eq. (2) is verified. The
computed flux ratios (averaged over the studied pressure
interval) are shown in Table 1 and are compared to the
values predicted for Knudsen flow by Eq. (2).
Any error higher than 10% must never be dismissed a
priori by an engineer. Therefore, students should find it
worthwhile to check for the existence of a not-negligible
viscous flow contribution. This can be done efficiently using
Eq. (4), rewritten as 51

NRT E E rp2 (Ph+Pr)
(Ph P), P vk rk 16 LVk
(Ph ~ e e kft 16 Clvk

2 L8RT
where Vk ~ M
3 itM

A plot of NRT/(Ph -PP)vk as a function of (Ph +P,)/ItVk
will be a straight line, with the slope and the intercept de-
pending only on the unknown parameters rp and e/t. In
other words, the data points corresponding to different gases
all fall on the same line. Notice also that in the absence of
viscous flow, the plot would be a horizontal line.

Figure 4 shows the plot of our data using Eq. (6). The
necessary gas properties data is given in Table 2. From the
plot, one sees that an analysis based on the results for N2 and
Ar alone would be inconclusive due to data dispersion and to
the similar properties of the two gases. Once he is consid-
ered, however, it becomes evident that the line is not hori-
zontal, and therefore viscous flow must be considered in
addition to Knudsen flow.
The value of rp computed from the linear regression is
846 nm (for a 95% confidence interval). Using this value in
Eq. (5), the theoretical flux ratios can be recomputed. The
results, shown in Table 3, confirm that the model with com-
bined Knudsen and viscous flows describes the experimen-
tal data significantly better.
Finally, the relative contributions of each flow process can
be computed for each gas

Nk Dk 7
Nk+Nv Dk +Dv(1+Ph /P,)

The results are shown in Table 4.
We expected that helium would show the greatest contri-
bution from Knudsen flow since this molecule has the low-
est molecular weight and hence the highest Knudsen
diffusivity (Dk).


The experiment described here is simple to set up and not




2.0-- He
S. N2
Z 1.5- Ar

1.0 . .
0 1 2 3 4 5 6
(P,+ P,)/(vk) x 10- (m 1)

Experimental (Average) and
Theoretical (Eq. 5) Flux Ratios

Gases N/N exp. N/Nj Error%

2.32 2.2
2.79 1.8

Figure 4.
data points
to Eq. (6).

Chemical Engineering Education

Experimental (Average) and
Theoretical (Eq. 2) Flux Ratios

Gases N/JN exp. N/NJ Knudsen Error%

He/N2 2.27 2.65 17
He/Ar 2.74 3.16 15

Molecular Weights and Viscosities at
200C['6 for the Studied Gases

Gases M (g mor') g(Kg m-' s-')

He 4.003 1.98 x 10-
N2 28.02 1.76 x 10-5
Ar 39.94 2.22 x 10-5

He/N, 2.27
He/Ar 2.74

Average Percentage of Knudsen
Flow in the Pressure Interval
Studied (from Eq. 7)

Gases % Knudsen flow
He 92
N, 80
Ar 81

expensive. Data collection is also quite unelaborate and stu-
dents can easily perform the necessary calculations during
class time. Besides allowing for a first contact with separa-
tion membrane technology, the experiment provides insight
into intra-pore gas-phase transport mechanisms. The sequence
of calculations necessary for the result analysis should be
quite intuitive for the students and help in the assimilation of
the theoretical concepts.
For the particular system studied here, evidence was found
of viscous flow coexisting with Knudsen diffusion. This
provides an excellent example of how the transport mecha-
nism affects membrane selectivity. Helium is the molecule
with higher permeability, but the He/N2 and He/Ar flux
ratios decreased by more than 10% in relation to Knudsen
transport alone, due to the intrusion of viscous flow. Another
membrane, with smaller pores, would show no viscous flow
and therefore maximize selectivity.
Finally, the experiment illustrates an effective method for
measuring the membrane average pore size.
A possible improvement to this setup would allow for the
mean pressure in the module to be set in a range above or
below atmospheric pressure. It can be seen from Eq. (4)
that, say, for a constant pressure drop across the mem-
brane at higher mean pressures, the transport becomes
predominantly Poiseuille flow. This would imply adding
a pressure sensor, a valve, and a vacuum pump at the
module's outlet stream.


The authors wish to thank Professor Carlos A. Costa for
his careful review of the manuscript, and the Department of
Chemical Engineering for providing financial support for
the setup of this experiment. We are also grateful to Dr.
Kneifel from GKSS-Geesthacht (Germany) for kindly pro-
viding the hollow fibers.


2 j8 RT
Dk p"R-- Knudsen diffusivity (m2 s-')
D 3 PV cM
D PP (m2 s-1)
v 16 [


Fall 19\


ilumetric flow (m3 s ')
embrane thickness (m)
rrmeability (m3 s- N m2)

molecular weight (g mol')
molar flux (mol m-2 Sl)
molar flux due to Knudsen diffusion (mol m-2 s')
molar flux due to viscous flow (mol m' s ')


gasdynamics chemical equilibrium chemical kinetics



* Education
* Combustion
* Pollution
* Explosions

* Graphical Interface
* Integrated Plotting
* Batch Calculations
* On-line Help

p a .. -
F_ i^ I_ J

7.- 3- Dm V si

Combustion Dynamics Ltd.
fax: 403-529-2516

Ph inlet pressure (N m-2)
P, outlet pressure (N m-2)
r membrane average pore radius (m)
R ideal gas constant (8.314 J K-' mol-')
T temperature (K)

2 8RT
Vk 3 tM
E membrane porosity
i gas viscosity (Kg m' s')
t membrane tortuosity

1. Jackson, R., Transport in Porous Catalysts, Elsevier, New
York, NY (1977)
2. Cussler, E.L., Diffusion: Mass Transfer in Fluid Systems,
Cambridge University Press, Cambridge (1984)
3. Mulder, M., Basic Principles of Membrane Technology,
Kluwer Academic Publishers, Dordrecht, Netherlands (1991)
4. Krishna, R., and J.A. Wesselingh, "The Maxwell-Stefan Ap-
proach to Mass Transfer," Chem Eng. Sci., 52, 861 (1997)
5. Datta, R., et al., "A Generalized Model for the Transport of
Gases in Porous, Non-Porous, and Leaky Membranes. I.
Application to Single Gases," J. of Membrane Sci., 75, 245
6. Reid, R.C., J.M. Prausnitz, and B.E. Poling, The Properties
of Gases and Liquids, 4th ed., McGraw-Hill, New York, NY
(1987) 0

ME, -curriculum




Lessons Learned

Clemson University Clemson, SC 29634-0909

ne of the complaints heard most often from engi-
neering undergraduates is that they do not see any
thing practical until late in the curriculum. To the
students, course material becomes a seemingly endless stream
of apparently useless equations and concepts. This situation
is typical for most students in a traditional curriculum be-
cause the early years are filled with engineering science,
with the final year addressing process design, including a
capstone design project in the last semester or quarter. There
has been considerable discussion for a number of years
about integrating design throughout the chemical engineer-
ing curriculum. Much of the discussion has been driven by
faculty trying to find a better way to do things and ABET
criteria that encouraged the introduction of design elements
before the senior year in engineering programs.
The objectives of design integration are for the students to
apply engineering science earlier in the curriculum and, if
the integration is set up properly, for the students to realize
the interconnections between various courses in a difficult
curriculum. The latter objective relates to the students get-
ting a broader view of a process design. For example, a
portion of a chemical plant may consist of a reactor, distilla-
tion column, condenser, reboiler, and pump. The students

Douglas E. Hirt is an Associate Professor of
Chemical Engineering at Clemson University.
He received his BS and MS degrees from
Virginia Tech and his PhD from Princeton Uni-
versity. His research interests are in the areas
of interfacial phenomena and polymer films.
He has been involved with the SUCCEED coa-
lition since 1992. He received a Dow Out-
standing New Faculty Award from ASEE in
1995 and the 1998 Ray W. Fahien Award.

need to realize that they cannot view something in isolation
("I learned about heat exchangers in my heat-transfer class. I
don't need to know that for distillation."); if there is a distur-
bance in one part of the process, that disturbance may affect
other parts of the process.
Many departments have integrated design elements into
single, lower-level courses or throughout an entire curricu-
lum, and many more departments are thinking about it.'1
This paper discusses the lessons learned from a program of
design integration that involves a significant number of
courses over a period of five semesters. The program is
similar to work being performed at West Virginia Univer-
sity,121 the major exception being how to handle the design
integration when a significant fraction of the students are co-
op students.

Perhaps the most popular method of integrating design
into a curriculum is through the use of case studies, which is
analogous to problem-based learning.11 In our program, we
refer to a case study as an evolving design project because
the students work on the same project as they proceed through
the curriculum and the level of detail increases with time.
The intent of this early-design experience is for the students
to continue to work on the same case study through a critical
portion of the curriculum. Using our program as an example,
the courses and the average number of students that are
involved in the case-study approach are summarized in Table
1. Note that several courses are offered twice a year to
accommodate the schedules defined for co-op students.
The concept of an evolving design project is best under-
stood by describing, in general terms, a particular case study.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

Figure 1. Schematic diagram of one portion of an
ethanol case study. (Exercise: ask the students if there
are any problems associated with pumping river water
directly through the condenser.)

Fall 1998

Typical Number of Students in the Case-Study Program
(numbers vary depending on the cyclic nature of ChE enrollment)

Course #of students, #of students,
Number Course normal-sequence off-semester
offering offering*
First-Semester Sophomore
ChE 211 Introduction to Chemical Engineering 45 25

Second-Semester Sophomore
ChE 220 Thermodynamics I 40 20

First-Semester Junior
ChE 311 Fluid Flow (including pumps) 35 20

Second-Semester Junior
ChE 312 Heat and Mass Transfer 35 20
(including heat exchangers and humidification)

First-Semester Senior
ChE 413 Separations (including distillation) 35 **
ChE 450 Reaction Engineering 35 **
ChE 431 Process Design I 35 **

* Off-semester offerings are provided primarily for students in the
co-op program, although other students also enroll in these
offeringsfor various reasons.
** Off-semester section not offered.

A program of this type offers advantages because
it is flexible. It has the ability to handle any
number of students, and the evolving design
concept can be integrated into a single
course, a sequence of courses, or
across an entire curriculum.

The case study used as an example here is taken from Felder
and Rousseau14' and involves the fermentation of grain to
produce ethanol. It is given to the students as first-semester
sophomores in the first chemical engineering course, and it
initially consists of a written description of a chemical pro-
cess. Groups of students sift through the description and then
develop a flowsheet of the process as they envision it.
This first activity gives the students an opportunity to see
how the various pieces of equipment can come together to
form a successful design. The students then perform detailed
material and energy balances around the entire process and
around selected pieces of equipment (material- and energy-
balance problems associated with this case study may be
found in the reference cited above).
Depending on the timing of the case study during a semes-
ter, the material and energy balances either replace or rein-
force homework problems. We emphasize that, at this early
stage, the students are not expected to know every detail of
the design, but that by the end of the curriculum they will be
able to understand all of the design elements.
The students working on this case study then move on to
Thermodynamics I and perform a thermodynamic analysis
of a refrigeration system that uses CO2, which was produced
during the fermentation process, as an environmentally safe
refrigerant. This homework assignment demonstrates how
thermodynamics can be valuable in understanding a "real
world" process, and it serves to maintain the continuity of
the case study as the students move through the curriculum.
This particular case study involves separating ethanol from
water in a distillation column. As Figure 1 shows, a mixture
of ethanol and water vapors exits the top of the distillation
column and passes through a heat exchanger where the va-
por is condensed by cooling water. In the class discussion
we emphasize that the cooling water somehow has to be
delivered to the condenser. Although not part of the original
process description, we explain that the cooling water is
pumped from a river, through the condenser, and then to a
cooling tower that is used to reduce the temperature of the
water before it is discharged back into the river. In the Fluid
Flow course, the students are given additional specifications
about the process, and (using the actual pump curves) they
select the most cost-effective pump and piping for the flow
system. In the Heat and Mass Transfer course, we give the
students more information and ask them to "design" the



condenser (we also discuss the fact that it is not good prac-
tice to pump water from a river directly into the condenser
because of the likelihood of fouling). After studying humidi-
fication in the Heat and Mass Transfer course, the students
size the cooling tower. In the first semester of their senior
year, the students size the distillation column and a reactor in
their Separations and Reaction Engineering courses, respec-
tively, and in the Process Design course they perform an
economic analysis of one portion of the process or the entire
process, depending on the complexity of the case study.
The objectives of these open-ended assignments are to

course. When those students retake a course, they work on a
subsequent case study and lose some of the continuity pro-
vided by the evolving design project. Continuity is main-
tained, however, for the vast majority of the students.
A major logistical difficulty arises when there is a large
number of co-op students involved, because they are con-
tinually rotating on and off job assignments. Funded by the
NSF SUCCEED coalition, this program was intentionally
set up to determine how to accommodate a relatively large
number of co-op students (120, or 60%, of our students are
in the co-op program). Since we have students who start the
Introduction to Chemical Engineering course in either the

Reinforce the classroom
lectures with real-life
1 Have the students perform
calculations and make
engineering judgments
> Show the students that
changes in one part of a
process may affect the
performance of other pieces
of equipment

For example, we address the issue
of what happens if the feed flow
rate to the distillation column in-
creases. Will the cooling-water
pump be able to handle the increased
load in the condenser? Should the
pump be oversized to handle the
possibility of an increased load? By
how much? What happens if the

Year 1 Soph.1 Soph.2

Year 2 Junior 1 Junior 2

Year 3 Senior 1

Figure 2. Five-semester sequence in which stu-
dents work on their evolving design project. A
certain percentage of students follow this se-
quence and the logistical problems are mini-
mal. Co-op students typically do not follow
this sequence, and logistical difficulties have
to be managed.

temperature of the river water changes? Does that affect the
performance of the pump, the condenser, or the cooling
tower? It is important to have the students wrestle with these
questions; otherwise, they perceive the design calculations
as just another homework exercise, and they do not realize
the interactions that exist between process components.

Each year a case study is initiated with the sophomores.
The students work on that evolving case study from their
sophomore year through the first semester of their senior
year (five semesters altogether, as shown in Figure 2). All of
the assignments and solutions for a given case study are kept
in a three-ring binder, which is available for the faculty.
Likewise, the students are strongly encouraged to maintain a
portfolio of their work. This program works well for stu-
dents who proceed through the curriculum outlined in the
course catalog. In fact, a similar program is in place in the
chemical engineering department at West Virginia Univer-
sity.121 Minor problems arise when students drop or fail a

fall or spring semester, there are two
options: 1) initiate a different case
study in each semester, or 2) use the
same case study for the fall and spring
semesters in a given academic year.
The first option is a theoretically
reasonable approach, but from a prac-
tical standpoint it is difficult to carry
out because of what happens in the
upper-level courses. As co-ops go to
and return from job assignments, they
are not in the natural progression of
the curriculum as depicted in Figure
2, so there are students in junior- or
senior-level courses who began with
different case studies. In fact, it is
possible that an instructor would have
to be aware of three or four different
case studies in a senior-level course,
which is impractical. Therefore, it was
determined that the best way to in-

corporate design integration with a large fraction of co-op
students was to use the second option. The same chemical
process is used in the fall and spring semesters of the sopho-
more year, but some of the specifications (flow rates, tem-
peratures, pressures) may be changed slightly from one se-
mester to the next. In this way, only one "type" of case study
is initiated each year, and instructors in upper-level courses
do not have to keep track of as many types of projects.
But, even this is not a panacea. Depending on how many
co-op students took the maximum five work assignments, a
faculty member in a senior-level course could still be con-
fronted with up to three different case studies (the situation
with three case studies is rare; two is more common). In this
instance, the instructor may either develop similar but sepa-
rate assignments for each of the case studies (which can be a
daunting task) or identify the case study used by most of the
students and focus on it. We have found that the latter option
is a good compromise because it reduces the time pressure
placed on the faculty member while still affording the stu-
dents an opportunity to apply what they are learning in a

Chemical Engineering Education

design situation. To minimize the logistical difficulties in
upper-level courses even further, it is possible to introduce
the same case study to the incoming sophomores for many
semesters in a row, but a different case study should be
initiated at the sophomore level every two or three years.
There are several things that must be considered when
using evolving design projects.

The selection and development of a case study is
crucial. It should be simple enough so that sopho-
mores do not feel overwhelmed by the experience, but
it should contain all of the elements that are desir-

*Sources abound for ideas for case studies.14-71 Once
the main idea for the case study is fixed, then it takes
a little creativity and time to integrate the various
design elements,1'" but the students will benefit from

An ideal plan with which to assess the effectiveness of the
case studies would involve forming parallel sections of each
course, with some of the students enrolled in the "tradi-
tional" sections and the remaining students enrolled in the
sections that incorporate a case study. Also, the same profes-

Student Responses Before and After Working on Case Studies
(1=strongly disagree; 2=disagree; 3=neutral; 4=agree; 5=strongly agree)


I am an effective problem solver.
I would rate my written communication skills as good.
I would rate my oral communication skills as good.
I would consider myself an effective team member.
I would consider myself a capable leader.

though they may be working on different case studies
(e.g., each process should contain a distillation
column so that the students are all sizing a distilla-
tion column regardless of whether the student is
working on case study X, Y, or Z).

Tell the students why this case-study approach is
being used. Explain to them that this is not simply
another homework assignment. Explain that the main
goal is for them to be able to understand the inter-
connections between design elements and that they
will be better engineers because of that understand-
Have the students work in groups. Most of the
assignments should be developed so they can be
tackled by student groups ranging in size from two to
five people, depending on the size of the class and the
complexity of the task. A vast majority of the students
say that they value the group interactions.

If possible, assign projects at different times during a
semester. Try to avoid dumping an open-ended
design project on the class near the end of a semes-
ter. Students do not appreciate that and it can easily
kill all of the momentum that has been accumulated
up to that point.

Fall 1998

Mean (Std. Dev.)
Before After
(N=89) (N=148)
3.8 (0.6) 3.9 (0.7)
4.0 (0.7) 4.1 (0.7)
3.7 (0.8) 3.8 (0.8)
4.4 (0.6) 4.3 (0.6)
4.1 (0.7) 4.1 (0.8)

sor should teach both
sections of a given
course, the cohorts of
students should be simi-
lar in terms of defined
criteria (e.g., GPA, frac-
tion of commuting stu-
dents, fraction of co-op
students), and the stu-
dents should not cross
over from the traditional
to the modified sections
and vice versa. Once the
two groups of students
finished the curriculum,
they would be evaluated

through various means to determine if there is a difference
between them. This ideal scenario is extremely difficult, if
not impossible, to achieve. Therefore, we decided to inte-
grate the case studies in many of our core courses (shown in
Table 1) without forming special sections and to assess the
program using several questionnaires.
To evaluate the effectiveness of the case-study approach, a
survey was given to sophomores before they worked on their
particular case study. The survey was then given to these
same sophomores after they worked on their case studies in
the Introduction to Chemical Engineering course and to jun-
iors who had worked on several assignments related to their
case studies. The survey consisted of a set of five statements
(shown in Table 2) and the students were asked to indicate
the extent to which they agreed with the statements
(l=strongly disagree, 2=disagree, 3=neutral, 4=agree,
5=strongly agree). These statements were used because they
represent the major attributes that industry looks for in new
engineering graduates. Results for the mean and standard
deviation for each statement are presented in Table 2.
The total number of responses was about 90 before the
implementation of the design project (encompassing three
semesters of Introduction to Chemical Engineering). The
total number of students responding after they had worked
on case-study assignments was about 150, including stu-

able in order to
demonstrate the
connections between
topics in a course or
a curriculum. If an
instructor chooses to
have several case
studies used at the
same time in an
upper-level course,
the projects should
have common
elements so students
can be working on
the same types of
problems even

dents from the same three sections of Introduction to Chemi-
cal Engineering and from one section each of the Fluid Flow
and the Heat and Mass Transfer courses. The surveys were
staggered so that different students were contributing to the
150 "after" responses (e.g., the juniors in the Fluid Flow and
the Heat and Mass Transfer courses were different groups of
students). As shown in Table 2, the mean and standard
deviation remained roughly the same for "before" and "af-
ter" responses to each survey statement. These data indicate
that the case-study approach does not seem to significantly
influence the students' perception of their capabilities with
regard to the five attributes referred to in Table 2, but their
responses to other questions provide additional information.

To further investigate the potential effect of the case stud-
ies on the comprehension of topical material in a given
course, we asked an additional question of the juniors in the
Fluid Flow and the Heat and Mass Transfer courses: "Do
you think that you learned course content better from home-
work problems in the textbook or from the case study?
Explain your answer." As shown in Table 3, about 40% of
the students thought that the case studies were more effec-
tive in terms of learning the course material, and 35% felt
that the combination of conventional homework problems
and case-study projects helped them learn the material bet-
ter. The latter result was particularly interesting because the
question was phrased as an "either/or" question, yet a signifi-
cant number of the students responded that both types of
assignments are needed. To supplement the statistics, a repre-
sentative sample of student comments is included in Table 3.

We then asked the students to list what they liked most
about the case studies and what they liked least. Although
not comprehensive, a listing of many of the comments is
provided in Table 4. The responses concerning what the
students liked about the case studies affirmed most of our
expectations. Many of the comments about the dislikes were
also expected, and some of these comments can be used to
take corrective actions and improve subsequent case-study
assignments. Most of the student dislikes, however, can be
viewed as desirable results as indicated by the faculty replies
shown in parentheses in Table 4. Finally, even though it can
be a little more work for the students (and the faculty), the
case-study approach has a definite influence on how stu-
dents relate to course material, as evidenced by the positive
written comments from students. For example

3 "The design project has been a real pain in the
neck, but after each part gets done, I really do feel
like I've learned something and actually been able
to apply it."
3 "I think the project was very instructive. It gave
life to the problems in the book and gave a feel for
how it might be in the real world."
3 "The design project is an excellent tool in this

Student Responses Concerning
Homework Problems vs. Case Studies


"Do you think that you learned course content better
from homework problems in the textbook or from the
case study? Explain your answer."


Case Study

No. of Responses
(% of Total Responses)
12 (24%)
21 (41%)
18 (35%)

Student Comments:

"Homework problems. By the time we did the case study, most
of the concepts that we used were common sense."
"The homework problems were good for learning the course
content. However, the case study allowed for more than one
topic at a time and helped me to learn the application of course
"Better from homework problems, but learned real-life
applications from case study."
"I felt better doing homework (as if I accomplished more) but I
believe I learned more through the case study."
"In order to do the case study effectively, you had to do the
required homework problems. I don't think I learned more from
one approach over the other. I do think that both approaches
helped each other out and tied everything together that we
learned over the course of a semester."
"The case study was more like the test questions in that it was
much more complicated than the homework problems.
Therefore, it provoked more in-depth [thought] and helped me to
learn more."
"I learned more about pumps from the case study than from the
homework problems. The case study seemed to be more detailed
than the homework problems."
"The case study and homework each tend to be very helpful.
The homework narrow down specific problems while the case
study combines all problems together to see how they are used
"Both are required, but I will retain more from the case study. If
we were assigned a case study first, we would have no
foundation on which to start work, so homework is necessary to
build up our skills before delving into a case study."
"I feel I have truly learned from the case-study approach. Case
studies required me to analyze situations and to use judgment
skills. Sometimes homework can be 'follow the formula,' 'right-
or-wrong' type questions that make it easy for one to just go
through the motions rather than understand why."

Chemical Engineering Education

class. It increases the student's knowledge by
incorporating ideas learned throughout the
semester into one unit."

Student Likes and Dislikes About Case Studies

What did you like most about the case study?

* Real situation
* Thought provoking
* Learned a lot
* The math involved
* It requires analysis and judgment skills
* Learning how to actually use what I learned to solve a real-life
* Gaining new insights on particular strategies
* Struggling for hours and resolving the problem
* Allows students to get more in-depth than "usual" homework
assignments allow
* Enjoyed the group work
* Allows us to see the big picture
* Putting the ideas of several people together to come up with
better combined efforts

What did you like least about the case study?
(Faculty comments are in parentheses)

* Had to make too many assumptions
(develops judgment)
* Had to learn some of the topics at the last minute
(plants the seed for life-long learning)
* Writing the report
(develops communication skills)
* Not knowing where to start
(develops judgment)
* Time consuming; getting stuck for hours
(perseverance is an attribute)
* Working with those incapable of problem solving
(develops personnel-management skills)
* Finishing the project and finding out some small error that
affected the overall outcome-this probably happened at least six
(promotes more careful work habits)
* Some people like to take charge and do all the work without
conferring with others or giving them a chance
(develops teamwork skills)
* Having to arrange convenient times for everyone to meet
(develops teamwork skills)
* Unsure of answers
(real-world experience)

These case studies are not meant to replace the capstone
design project. The program is formulated to teach the stu-
dents to think about design and to prepare them for the
capstone design experience in the second semester of their
senior year. The use of evolving design projects can have a
significant positive effect on the education of chemical engi-
neering undergraduates. The students discuss the various
case studies, they exchange ideas, and they even look for-
ward to open-ended projects. Evolving design projects cre-
ate a unique mode of learning for the students.
A program of this type offers advantages because it is
flexible. It has the ability to handle any number of students,
and the evolving design concept can be integrated into a
single course, a sequence of courses, or across an entire
curriculum. Logistical difficulties can be encountered if the
department has a substantial co-op program in place. Never-
theless, in terms of student development, an evolving design
project offers a common thread through the curriculum that
can lift the students to the next level of understanding engi-
neering design and its consequences.

This project was funded by the National Science
Foundation's SUCCEED coalition. The author would like to
thank the chemical engineering faculty at Clemson for their
coordinated effort in implementing and evaluating this inte-
grated design program.

1. Technical session on "Integrating Design Across the Cur-
riculum" at the AIChE Annual Meeting, Miami Beach, No-
vember 12-17, 1995
2. Bailie, R.C., J.A. Shaeiwitz, and W.B. Whiting, Chem. Eng.
Ed., 28, 52 (1994)
3. Woods, D.R., Problem-Based Learning: How to Gain the
Most from PBL, W.L. Griffin Printing Limited, Hamilton,
Ontario, Canada (1994)
4. Felder, R.M., and R.W. Rousseau, Elementary Principles of
Chemical Engineering, 2nd ed., John Wiley & Sons, New
York, NY, Chap. 14 (1986)
5. Turton, R., R.C. Bailie, W.B. Whiting, and J.A. Shaeiwitz,
Analysis, Synthesis, and Design of Chemical Processes,
Prentice Hall, Englewood Cliffs, NJ (1998)
6. Peters, M.S., and K.D. Timmerhaus, Plant Design and Eco-
nomics for Chemical Engineers, 3rd ed., McGraw-Hill, New
York, NY (1980)
7. Douglas, J.M., Conceptual Design of Chemical Processes,
McGraw-Hill, New York, NY (1988)
8. Hirt, D.E., "Designing a Pumping System: Why Worry About
Other Process Elements?" proceedings of ASEE National
Meeting, Milwaukee, WI, June 15-18, 1997 n

Fall 1998




McMaster University Hamilton, Ontario, Canada L8S 4L7

G reat strides have been made in understanding learn-
ing, but few teachers have paid much attention to
them. Teachers have continued to teach the way
they were taught, namely, by using the lecture system.
There is growing dissatisfaction, however, by both stu-
dents and taxpayers about the quality of student learning.
Learning has now become more and more the focus, and
there have been a number of published research findings
on how to improve it.1'1-31

A number of principles can be employed to improve learn-
ing. A teacher can change his or her approach to
B Show concern for the students as persons
B Demonstrate concern for the students' success
B Include activities that prepare students for the real
world or for their profession, placing emphasis on
what and how the students learn and not on how the
teacher talks
O Empower students with parts of the learning process
Publish clear goals and criteria for learning
B Make assessment processes consistent with the goals
and criteria
Include activities to help students create a useful
knowledge structure
B Give prompt feedback
E Motivate the students

Students can also embrace change to improve their learning.
They can

U Be active
U Work cooperatively
B Have a clear understanding of the task and the time

allotted for the task, and apply themselves diligently
to the task
E Be aware of their own learning preferences and
know how to use them effectively
B Be aware of the need for knowledge structure and
the importance of cues
U Have ownership of elements in the learning process
An interesting sidenote is that many of these principles for
improving learning are also required in the second trend on
"what to teach," featured in a later section of this paper.
So what can we, as teachers, do? The easiest, least expen-
sive, and most effective change is a change in attitude, as
outlined above. Identify ways to demonstrate your positive
attitude to the students. (A questionnaire that gives you a
chance to check your attitude is available from the author.)
If you currently lecture, try including some student activ-
ity in the lecture. One of the easiest methods to incorporate is
"Turn to your neighbor..." When you see that glazed look in
students' eyes, or when the twenty-minute boredom sets in,
use this technique. The conditions determine the activity. I
may use the comprehension-check activity described below
several times in a single 50-minute period. I use it if I have
just explained a difficult topic, if the students seem bored, if
a student asks a question that might be representative of the
whole class, or when I want to change from just me talking.
In this activity, you might ask the students to

Don Woods is a professor of chemical engineer-
ing at McMaster University. He is a graduate of
Queen's University and the University of Wis-
consin. His teaching and research interests are
in surface phenomena, plant design, cost esti-
mation, and developing problem-solving skills.

Copyright ChE Division of ASEE 1998
Chemical Engineering Education

- Talk about their understanding
> Explain the idea to each other
> Answer the question, "So what?"
> Perform a task that requires that they "know the
subject," This could be a calculation, a simplified or

complex problem, or a small section of the pr<
solving process. For this question, identify
the pertinent knowledge you think they
need to solve the problem and tell them to
write out what they think they are asked to
It is useful to obtain feedback on how well the
teaching and learning is progressing by using
"ombudspersons." Ombudspersons are volunteer
members of a team from the class who provide
you with feedback throughout the year. As teacher,
on the first day of class you can greet your stu-
dents with
Welcome to this course in chemical
engineering. My role is to help you learn. I
see learning as a two-way street, with you
and me working together to help you get an A
in the course. To help me in my role as
teacher, I need periodic feedback from you.
Here's what I suggest that we do: I would
like three volunteers to be ombudspersons.
Your role would be to relay to me, on behalf
of the class, any comments, suggestions, or
ideas on how I can help you the most. Any
questions? Are there volunteers?
Often there are no volunteers, especially if the
idea is a new one to them. So, continue

class, or perhaps to meet somewhere for coffee. The time
needed to capture the ombudspersons' feedback varies. It
has been my experience over the past twenty years that most
ombudspersons come with a written list, and our time is
spent clarifying my understanding of the feedback, usu-
ally about three or four minutes. At times the feedback is


There is
however, by
both students
and taxpayers
about the
quality of
Learning has
now become
more and more
the focus, and
there have
been a number
of published
findings on
how to improve

No volunteers? It'll look good on your resume. In
terms of time commitment, I have found it useful if I
meet with the ombudspersons formally for about ten
minutes twice a semester, at about week four and
again during week seven. This is not a large time
commitment, and look at the skills you will be
developing-communication skills, political skills,
succinct summarization of any frustration from your
colleagues and expressing it so that I understand.
Now, do I have volunteers?
You will probably have more than enough volunteers at this
point. I use about three for classes of fewer than 60 and
about six for classes of 400. Try to select volunteers who
represent a good cross-section of the class. Once they have
been identified, ask them to stand so everyone in the class
knows their names and can recognize them outside of class.
At about week three, suggest that it is time to get feedback
from the ombudspersons. Check with them to see if it is
more convenient for them to meet ten minutes before or after

verbal, and in those cases ten minutes could be
required. I prefer to take the feedback in front
of the class to demonstrate to the students that
I take it seriously.
When meeting with the ombudspersons, make
an accurate record of their comments, remember-
ing that you are not responding to them but to the
other students, and that your response will be to
the class as a whole. At the beginning of the next
class, read the ombudspersons' comments and
make comments such as:
"We think you wear lousy ties." Thank you. I
have a limited wardrobe, but I will try to select
better ones in the future.
"We have to spend too much time doing the
journal writing in this class-last week it was 15
hours. It takes too much time!" OK, here is
why I have found journal writing to be vital.
Without the reflection and elaboration needed in
the journal writing, we find that the skills are not
developed as well as we hoped for. An alumnus,
Kyle Bouchard, said : "...writing those journals
every week was a pain. But without that I don't
think I ever would have learned the skill!" OK,
that's why I think the writing is important. Now,
about the time: fifteen hours is expecting way too
much. Is three hours more reasonable? Five

hours? Ok, let's try for two hours a week. Now, let's gather
some ideas about how to actually do that. Please form small
groups and brainstorm fifty ideas in five minutes about how
we can gain the benefits from journal writing and yet take
only two hours a week doing it. Then, from each group I
want the top three ideas.
To summarize, collect the information, then solve the
problem together with the students. Bring them in on the
solution. Ombudspersons-don't hold a class without them!
It is also useful to pose practical problems near the begin-
ning of each new subject so students can see the application
and learn the new knowledge within the context of a profes-
sional problem to be solved.

Teaching students chemical engineering is not enough.
Since the 1970s, a number of professors, departments, pro-
grams, and colleges have felt that graduates should know
more than the subject knowledge-they should be able to

Fall 1998

apply that knowledge to solve problems
and should have a wide variety of pro-
cess skills and attitudes needed in pro-
fessional practice. The process skills in-
clude communication, teamwork, self-
assessment, self-awareness, changing
management, coping with change, lead-
ership, and "professional attitude."
To achieve this goal, the skills and
attitudes are described as valued abili-
ties, competencies, or outcomes of
courses or programs, and an effort is made
to have specific activities within the pro-
grams to develop the skills. They then
assess how well the students can display
the skills. In the literature, these pro-
grams are called "outcome-based" pro-
grams. Some examples of early efforts in
this direction are:
Alverno College, Milwaukee, Wis-
consin (1972):"[4-151 all graduates of ev-
ery program must show at least four lev-
els of competence in each of eight abili-
ties: 1) communication, 2) analytical
thinking (critical thinking), 3) problem
solving, 4) making value judgments and
independent decisions, 5) effective so-
cial interaction, 6) taking responsibility
for the global environment, 7) effective
citizenship, and 8) responding to and ap-
preciating the arts and humanities. The
college set up a separate assessment de-
partment, separate "departments" for each
of the competencies, and then left it to
the individual instructors of each course
to identify and develop certain levels of
skills. Their approach has attracted world-
wide attention and numerous awards for
their innovation. They run annual work-
shops on assessment.
McMaster Medical School,
Hamilton, Canada (1969):[12'161 graduates

What Outcomes are Valued in Your

laws, concepts

Process Skills
problem solving
team work
lifetime learning
time management
change management
communication (written/oral)

intellectual curiosity
cope with ambiguity
willing to risk
restrain impulsiveness _
manage distress
meet deadlines
document activities
plan ahead
anticipate difficulties

or tne medical program must display a
variety of knowledge, skills, and profes-
sional attitudes. Table 1 illustrates these
outcomes. The approach has attracted worldwide attention
and has spread to many health science programs. In particu-
lar, variations on the approach are now used in medical
schools at Maastricht in the Netherlands, Newcastle in Aus-
tralia, and in many schools of nursing, occupational therapy,
and physiotherapy, and more recently in pharmacy. The
program runs annual workshops and uses a small group,
self-directed, self-assessed problem-based format.

ing environmer
of the ideas to
velopment con
effective use o
that both the s
ity to generate

West Virginia University,
Morgantown, West Virginia (1972);[17-2]
graduates of engineering programs are
provided with communication, team, and
problem-solving skills using a small-
group, problem-based approach called
"Guided Design."
McMaster Chemical Engineering,
Hamilton, Canada (1978);121] graduates
are required to show skill in communica-
tion, team, problem solving, self-assess-
ment, change management, and self-di-
rected, lifetime learning.
On the one hand, these approaches have
increased in popularity because of the
positive student response to the programs,
the positive response from alumni, and
the extremely positive response from
employers and those in private profes-
sional practice. On the other hand, em-
phasizing the development of these "pro-
cess skills" and attitudes is not easy. The
traditional methods of teaching do not
seem to be effective in developing the
skills. For example, asking students to
solve problems, asking teachers to dem-
onstrate how they solve problems, ask-
ing students to work in groups, asking
students to solve open-ended problems,
asking students to demonstrate how they
solve problems, rarely develop the stu-
dents' skill or confidence. We should se-
lect explicit programs that have been
proven to be effective in the past.
The trend is linked to research on im-
proving learning, as discussed earlier.
Figure 1 illustrates the links. Ideas to
improve learning lead to actions in the
classroom and provide at least three spe-
cific ideas that are needed to develop
process skills: the need for explicit goals,
the importance of using active workshops
(instead of lectures), and the need for
self-assessment (to develop student's con-
fidence in the process skills). One learn-
it (problem-based learning) implements most
improve learning: active, student ownership,
t, cooperation, and knowledge structure de-
nplete with cues and prompt feedback. The
f problem-based learning, however, requires
students and the tutor possess process skills.
learning requires and develops student's abil-
Slearning issues, to teach each other new

Chemical Engineering Education

Figure 1. Linking ideas to improve learning to the development of
process skills and to the creation of knowledge structure with cues.

Contact Time for Developing Skills and Attitudes

References and subject context Conventional programs or courses

Albanese/Mitchell 12'
medical schools
MPS program-211

100% knowledge

knowledge, and to self-assess.
From another viewpoint, alumni, business, and industry
have identified a range of skills they hope graduates possess:
process skills, subject knowledge that can be readily used to
solve unique problems, and lifetime learning skills. One
effective learning environment to develop the latter skills is
problem-based learning. Connecting both the valued out-
comes and problem-based learning is the explicit develop-
ment of the process skills. The next section elaborates on the
development of knowledge structure.

We can develop process skills in several ways:
Determining the valued outcomes or skills you want
graduates to possess. Table I list some possible
For each course in the program, identify both the
outcomes and where in the program the skills are

Fall 1998

Small group, self-directed, self-assessed
problem-based learning that requires
process skills
80% knowledge because 20% is spent
applying and using process skills
82% for chemical engineering courses with
18% spent explicitly developing process skills

developed and practiced. Developing process skills
effectively takes time, however. Table 2 summarizes
published evidence that suggests that the skill
development needs about 20% of the available time.
This means that educators who want to move in this
direction have to expand the semester, increase the
number of semesters in the program, or reduce the
amount of subject knowledge in the current curricu-
lum by at least 20%. We chose the last. We retained a
core of fundamentals and streamlined the number of
- Select proven methods to develop the students'
confidence and skill. Options that have proved to be
effective are to give students clear objectives and
examples of examinations to test such skills,1241 to
give students a chance to try the skill and to receive
immediate feedback, and to provide students with


target skills and opportunities to practice, practice,
practice. Bandura1251 offers explicit suggestions about
how we can develop students' confidence by care-
fully structuring activities to help them succeed.
Sternberg1261 and Beyer127] offer general criteria for
selecting programs. The option we used is based on
the research of Sternberg, Beyer, and Bandura.

Effective problem solvers initially need to have both ge-
neric problem-solving skills and subject knowledge. The
subject knowledge should be readily accessible for the con-
texts in which it is needed to solve problems. How we learn
knowledge affects our ability to use it in the future.
There is some interesting research to support this conten-
tion. Godden and Baddeley required some underwater divers
to memorize lists of words on land and others to memorize
lists while under water. Later, they were asked to recall the
list of memorized words. Recall was hampered if it was
done in the opposite environment, that is water instead of
land and vice versa.[281
Zhu, et al.[291 and Zhu and Simon3"' worked with high
school students who were learning to factor quadratic equa-
tions and to solve problems about buoyancy. Three different
approaches were taken: 1) the conventional teacher-talk about
the fundamentals and sample applications, 2) students worked
individually with carefully structured and sequenced ex-
amples in the form of "if-then" rules ("if" the equation has
this form, "then" take this action), and 3) students were
given problems and answers to the problems. They found
that students who learned by working examples focused on
acquiring the "conditions" to identify when to apply the
knowledge. The students noticed cues about when condi-
tions were satisfied and when to apply a given set of
knowledge. They spent time elaborating on what they
found to be effective in solving the problems. From each
success, they created checklists of the relevant condi-
tions; they learned to ignore the irrelevancies. The re-
searchers concluded that how students learn knowledge
affects their future ability to apply it.
Eylon and Reif1311 showed that teaching or helping stu-
dents learn the material using deductive methods helped
them perform better on future deductive tasks (compared
with a control group). Those who learn or were taught knowl-
edge following a historical or developmental organization
performed better on future historical tasks. They also con-
cluded that how students learn the knowledge affects their
future ability to apply it.
Schmidt3.41 provides more details of other research and
relates it to problem-based learning. He emphasized the
importance of giving students a chance to elaborate on the
knowledge as they learn it.

Some of the techniques we might use to good advantage
Consider using problem-based learning in which the
problem is posed to drive the learning. In other
words, students learn knowledge because they need
it to solve a problem.
Include activities such as "Turn to your neighbor
and..." to have students reflect and elaborate on the
ideas just presented in class.
Ask students to create concept maps of the knowl-
edge; emphasize the importance of "cues." Figure 1
is an example of a concept map. Each bubble is
meant to be a key issue, and the bubbles are joined
by lines with arrows that show the relationship
between concepts. Words should be added to each
line to explicitly show the connection. Details about
concept maps and how to assess them are given by
Woods[12.241 and Novak and Gowin.1321
Require that the students complete checklists, such
as those developed by Larkin133 describing the
terms in the equations, limitations, conditions, and
cues as to when to use them.
Provide "knowledge structure" charts. Examples of
such charts are given by Woods and Sawchuck,1341
Felder,"351 Bird,'361 Fogler,1371 and O'Connell.381
Use the didactic model suggested by Bagno and
Eylon1391 with an added emphasis on the addition of

This paper reviews three recent trends in education. The
first trend is that much of the research on the learning pro-
cess provides teachers with a rich resource of proven ideas to
improve student learning. Research has shown us about a
dozen ideas on how to improve student learning. Teachers
should focus on facilitating learning rather than on teaching.
A teacher's task is to help students succeed. Seven ideas on
the roles of students in the learning process were given. Four
specific suggestions for teachers are to conduct an "atti-
tude check," to bring activity into the classroom by tell-
ing the students to "Turn to your neighbor and...," to
solicit feedback about the quality of the learning by us-
ing ombudspersons, and to introduce new topics by pos-
ing practical problems.
A second trend is that alumni and employers want gradu-
ates to have more than just a sound grounding in the funda-
mentals of the subject discipline. They should also be skilled
in problem solving, in communication, in teamwork, and in
a host of generic skills called "process skills." This paper
describes four programs that produce graduates with this
combination of knowledge and process skills. The programs
are at Alverno College, McMaster University Medical School,
Chemical Engineering Education

and at the chemical engineering programs at West Virginia
University and McMaster University. The difficulties in de-
veloping these process skills are reviewed, example effec-
tive interventions are outlined, and suggestions are given as
to their implementation.
A third realization is that the context in which students
learn new knowledge affects future retrieval and use of that
knowledge. In other words, the structure of knowledge in
memory is important. That structure is affected by the con-
text in which the knowledge is learned. For example, if
students learn new knowledge in a problem-based context,
the knowledge is learned in the context of having to learn
knowledge to solve a problem. Cues relate the knowledge to
the problems. Suggestions are given to help students acquire
some of that structure when they learn in the more traditional
lecture- or subject-based format.
These three trends are connected. Problem-based learning
is suggested as an integrating vehicle.


1. Chickering, A.W., and Z.F. Gamson, "Seven Principles for
Good Practice in Undergraduate Education," AAHE Bulle-
tin, March 3-7 (1987)
2. Chickering, A.W., and Z.F. Gamson, "Applying the Seven
Principles for Good Practice in Undergraduate Education,"
New Directions for Teaching and Learning, 47, Jossey-Bass,
San Francisco, CA (1991)
3. Schmidt, H., "Problem-Based Learning: Rationale and De-
scription," Med. Ed., 17, 11 (1983)
4. Schmidt, H., "Foundations of Problem-Based Learning: Some
Explanatory Notes," Med. Ed., 27, 422 (1993)
5. Wankat, P.C., "Learning Principles: A Quick Guide to What
Works," Chem. Eng. Ed., 27(2), 120 (1993)
6. Wankat, P.C., and Frank S. Oreovicz, Teaching Engineer-
ing, McGraw-Hill, New York, NY (1993)
7. McKeachie, W.J., et al., Teaching Tips, 9th ed., D.C. Heath
and Co., Lexington, MA (1994)
8. Felder, R.M., "Things I Wish They Had Told Me," Chem.
Eng. Ed., 28(2), 108 (1994)
9. Ramsden, P., "How Academic Departments Influence Stu-
dent Learning," HERDSA News, 4, 3 (1985)
10. Ramsden, P., and N.J. Entwistle, "Effects of Academic De-
partments on Student's Approaches to Studying," Brit. J. of
Ed. Psychol., 51, 368 (1981)
11. Stone, H.L., "Data Gathering Instruments for Evaluating
Educational Programs and Teaching Effectiveness in the
Center for Health Sciences," University of Wisconsin-Madi-
son, WI (1991)
12. Woods, D.R., Problem-Based Learning: How to Gain the
Most from PBL, Woods, Waterdown ON, distributed by
McMaster University Bookstore, Hamilton, Ontario, Canada
13. Woods, D.R., "PS Corner," J. College Science Teaching, 19,
109, Dale's cone of learning (1989)
14. Mentkowski, M., and A. Doherty, "Abilities that Last a
Lifetime: Outcomes of the Alverno Experience," AAHE Bul-
letin, 36(6), 5 (1984)
15. Edgerton, R., "Abilities that Last a Lifetime: Alverno in
Fall 1998

Perspective," AAHE Bulletin, 36(6), 3 (1984)
16. Barrows, H.S., and R. Tamblyn, Problem-Based Learning,
Springer, New York, NY (1980)
17. Wales, C.E., and R.A. Stager, "Educational Systems De-
sign," West Virginia University, Morgantown, WV (1973)
18. Wales, C.E., and R.A. Stager, "Teaching Decision-Making:
Guided Design," West Virginia University, Morgantown,
WV (1977)
19. Wales, C.E., "Does How We Teach Make a Difference?" Eng.
Ed., 69, 394 (1979)
20. Bailie, R.C., J.A. Shaeiwitz, and W.B. Whiting, "An Inte-
grated Design Sequence: Sophomore and Junior Years,"
Chem. Eng. Ed., 28(1), 52 (1994)
21. Woods, D.R., et al., "Developing Problem Solving Skills: The
McMaster Problem-Solving Program," J. Eng. Ed., 86(2), 75
22. Albanese, M.A., and S. Mitchell, "Problem-Based Learning:
A Review of Literature on Its Outcomes and Implementa-
tion Issues," Academic Med., 68, 52 (1993)
23. Winslade, N., "Large Group PBL: A Revision from Tradi-
tional to Pharmaceutical Care-Based Therapeutics," Am. J.
ofPharma. Ed., 58, 64 (1994)
24. Woods, D.R., "Problem-Solving Objectives," in "Problem-
Based Learning: Resources to Gain the Most from PBL,"
available on the WWW at un-
der Innovations: Problem-Based Learning (1997)
25. Bandura, A., "Self-Efficacy Mechanism in Human Agency,"
Am. Psychologist, 37, 122 (1986)
26. Sternberg, R.J., "Criteria for Intellectual Skills Training,"
Ed. Researcher, p. 6 Feb (1983)
27. Beyer, B.K., Developing a Thinking Skills Program, Allyn
Bacon, Boston, MA (1988)
28. Godden, D.R., and A.D. Baddeley, "Context-Dependent
Memory in Two Natural Environments: On Land and Un-
der Water," Brit. J. of Psychology, 66, 325 (1975)
29. Zhu, X., Y. Lee, H.A. Simon, and D. Zhu, "Cue Recognition
and Cue Elaboration in Learning from Examples," Proc.
Natl. Acad. Sci. USA, 93, 1346 (1996)
30. Zhu, X., and H.A. Simon, "Learning Mathematics from Ex-
ample and by Doing," Cognition and Instruction, 4(3), 137
31. Eylon, B.-S., and F. Reif, "Effects of Knowledge Organiza-
tion on Task Performance," Cognition and Instruction, 1, 5
32. Novak, J.D., and D.B. Gowin, Learning How to Learn, Cam-
bridge University Press, Cambridge, UK (1984)
33. Woods, D.R., "PS Corner," Larkin's Checklist, J. College
Sci. Teach., 13, 467 (1984)
34. Woods, D.R., and R.J. Sawchuk, "Fundamentals of Chemi-
cal Engineering," Chem. Eng. Ed., 27(2), 80 (1993)
35. Felder, R.M., "Knowledge Structure of the Stoichiometry
Course," Chem. Eng. Ed., 27(2), 92 (1993)
36. Bird, R.B., "The Basic Concepts in Transport Phenomena,"
Chem. Eng. Ed., 27(2), 102 (1993)
37. Fogler, H.S., "An Appetizing Structure of Chemical Reac-
tion Engineering for Undergraduates," Chem. Eng. Ed.,
27(2), 110 (1993)
38. O'Connell, J.P., "Thermodynamics," Chem. Eng. Ed., 27(2),
96 (1993)
39. Bagno, E., and B.-S. Eylon, "From Problem Solving to a
Knowledge Structure: An Example from the Domain of Elec-
tromagnetism," Am. Assoc. of Physics Teachers, 65, 726
(1997) D

l classroom



Tulane University New Orleans, LA 70118

Teaching separations processes is a central part of the
undergraduate chemical engineering curriculum. The
most important separations processes, including dis-
tillation, gas absorption, stripping, and liquid-liquid extrac-
tion, are based on driving forces determined by phase equi-
librium and rates determined by diffusion (mass transfer).
These separation processes are carried out with either plate
columns or packed columns. The teaching of plate-column
and packed-column design usually proceeds from different
viewpoints: plate columns are taught from a phase-equi-
librium perspective, while packed columns are taught
from a mass-transfer perspective."1 Due to the different
approaches, the relationships between plate and packed
columns are not always clear.
This paper will show how the design of packed columns
can be approached from the phase-equilibrium perspective
used for plate columns. This approach is based on the physi-
cal phenomena underlying the processes: a packed column is
shown to be equivalent to a plate column with an infinite
number of infinitesimally-efficient plates. This approach
contrasts with correspondences between plate and packed
columns based on the concept of equivalent theoretical
plates,[21 which does not have physical significance for a
packed column.13'41
The philosophy underlying this approach is based on two
premises. First, a unified approach to teaching plate and

DanielJ. Lacks is Associate Professor of Chemi-
cal Engineering at Tulane University. He re-
ceived his BS in chemical engineering from
Cornell University and his PhD in chemistry from
Harvard University. After doing postdoctoral re-
search at MIT, he joined the Tulane faculty in
1994. His research interests involve the applica-
tion of molecular simulations to chemical engi-
neering problems.

Copyright ChEDivision of ASEE 1998

packed columns will elucidate their relationship; second,
plate-column concepts (e.g., plates and efficiences) are physi-
cally intuitive and easily visualized with graphical methods
such as McCabe-Thiele diagrams, while packed-column con-
cepts (e.g., transfer units) are much more abstract. An under-
standing of packed-column concepts in terms of plate-col-
umn concepts will therefore enhance student intuition for the
packed-column concepts.
This teaching strategy has been implemented in the unit
operations course taught at Tulane. In this course, the design
of equilibrium plate columns is taught first, followed by the
design of inefficient plate columns. The design of packed
columns is then introduced as an extension of inefficient
plate-column design as detailed below, allowing packed-
column concepts to be understood and visualized in terms of
plate-column concepts. After this introduction to packed
columns, the traditional mass-transfer approach to packed
columns is taught, motivated by the fact that the present
approach yields only the number of transfer units but not the
height of a transfer unit.
In particular, this paper addresses the underlying physical
relationship between plate and packed columns, which leads
to a definition of transfer units for plate columns that is
consistent with the usual definition of transfer units for packed
columns. This new definition of transfer units is based on
concrete and intuitive quantities and elucidates the relation-
ship of transfer units to other plate- and packed-column
concepts. First, however, the fundamentals of plate-column
design necessary for the present analysis of packed columns
are briefly reviewed. Note that the present focus is on dilute
systems, to simplify the analysis.

The vapor and liquid streams flow counter currently through
the column (vapor flows up and liquid flows down), with
flow rates of V and L, respectively. The vapor and liquid

Chemical Engineering Education

phases on an equilibrium plate n are in equilibrium; i.e., the
vapor concentration yn = Yn and the liquid concentration
x, = xn, where y* and xn are the concentrations determined
by the equilibrium relation yn = mx*. On real plates, how-
ever, equilibrium is not reached in the time that the streams
are in contact, and therefore y, # yn and xn # x,. The frac-
tional progress toward equilibrium can be described by the
Murphree plate efficiencies for plate nl'51

n) Yn Yn+l or
Yn -Yn+i

(n) Xn -Xn-_
Mx --= n
"n n-

(where plate numbering starts from the top of the column)
which are related byI '

1Mx mV + ( mV) (2)
--My +L ---L-- ) (2)
For dilute systems, m, V, L, 7Mx, and rly are approximately
constant throughout the column, which allows the number of
plates required for a separation, Npiates, to be obtained ana-
lytically with the Kremser equation for inefficient plate col-


yin-Ymxin ( mV mV
LYout mxin L ) L
_n[1+m (m'yV )]

(an analogous equation based on nlMx also exists). The over-
all column efficiency, ro, is defined as

0io = (4)

where NETP is the number of equivalent theoretical plates
(i.e., the number of plates when the plate efficiency equals
one). The relationship between the overall efficiency and the
Murphree efficiency is

n[ (TIMY L .rnv
n10 = m (5)

Note that all of these concepts for plate columns are readily
visualized in terms of McCabe-Thiele diagrams.111


Physical Basis
A packed column is presently considered as follows: At a
given height in the column, Z, the vapor stream flows up to
Z with concentration y(Z), and the liquid stream flows down
to Z with the concentration x(Z*), where Z and Z' refer to
the heights just below and above Z. If equilibrium were
reached between the vapor and liquid at Z, then the vapor

concentration at Z is given by y(Z)=y'(Z)=mx(Z). But the
liquid and vapor streams are flowing in opposite directions
and are in contact at the height Z for only an instant; there-
fore, y(Z) will not be at the equilibrium value but will lie in
between y(Z) and y*(Z). This situation is equivalent to a
plate column, with the height Z corresponding to plate n, the
height Z corresponding to plate n+l, the height Z* corre-
sponding to plate n-1, and the change in concentration y(Z)-
y(Z) described by the Murphree efficiency of plate n. These
effective plates in a packed column are only infinitesimally
efficient: since the liquid and vapor streams are in contact
only instantaneously at a given height Z, the change in
concentration y(Z)-y(Z-) is infinitesimal. But there exists an
infinite number of these effective plates, corresponding to
all heights Z. This relationship of packed and plate columns
is physically justified, in contrast to the relationship based
on equivalent theoretical plates.

Concept of Transfer Units
The concept of transfer units is abstract and thus difficult
for students to fully understand. To address this difficulty,
the physical relationship between plate and packed col-
umns is used to define transfer units in terms of more
concrete plate column concepts. For dilute systems, the
number of transfer units contributed by a plate is defined
here to be equal to the Murphree efficiency of the plate.
The total number of transfer units for a plate column,
NOte co., is obtained as the sum of the contributions from
each plate

N plates
late col. y Y (6a)
= Yn Yn+l
= NplatesflMy (6b)

The number of transfer units of a packed column is now
obtained in the limits liMy 0 and Nplates --> For the very
inefficient effective plates that comprise a packed column,
the quantity yn-Yn+1 approaches zero; thus yn Yn+ = dyn,
and yn yn+l = Yn yn + dyn = Yn ,n. The number of trans-
fer units for a packed column is obtained by integration
(rather than summation) over the infinitely many effective
plates between the top and the bottom of the column
Noy f d (7)

(note that the subscripts have been dropped because they are
the same for all terms). This expression for the number of
transfer units in a packed column for dilute systems is the
same as that obtained from the standard mass-transfer analy-
sis,[1'41 which demonstrates the validity of the present ap-
proach. Additionally, this approach shows that the number
of transfer units can be understood physically to be equal to

Fall 1998

the sum of the plate efficiencies of a column.

Derivation of the Colburn Equation
The present approach leads to a simple derivation of the
Colburn equation, which is the analytical expression for the
number of transfer units required for packed column separa-
tions with dilute systems. The derivation begins with the
number of transfer units of a plate column (Eq. 6b) with
Nplates given by the Kremser equation (Eq. 3). The number of
transfer units for a packed column is then obtained in the
limit that the Murphree efficiency approaches zero. As TIMy
approaches zero, the denominator of the Kremser equation
fn[l + ]]My(mV/ L 1)] = ]My (mV / L 1), and thus

n Yin -min mV)
LYou.-mxin, L) L
L )
y iN
Oy mTJ

This equation is the Colburn equation and is identical to the
result obtained from mass-transfer analysis.",6] The present
approach shows clearly the relationship between the Kremser
and Colbum equations.

Relationship of Transfer Units
to Equivalent Theoretical Plates
A usual source of confusion for students is the relationship
between transfer units and equivalent theoretical plates; the
present approach elucidates this relationship. Combining Eqs.
(4) and (6b) yields the results for plate columns

col. My (9)
Oy ^My (9)

To obtain the result for packed columns, the ratio rMy / To is
evaluated in the limit that riMy approaches zero. Using the
relationship between r1My and Tio given in Eq. (5), and
noting again that n[l+lMy(m V/L-1)] = TMy(mV/L-1) as
TMy approaches zero, the following result is obtained for
packed columns:

NOY en[ L]
Ny =_ mV (10)
L )

This result is the same as that obtained from the standard
mass-transfer analysis. But the present derivation demon-
strates that the relationship between NETP and Noy is the same
as the relationship between ro and 1lMy, which is readily
understood in terms of McCabe-Thiele diagrams, as shown
in Figure 1.

Other Definitions of Transfer Units
Another source of confusion for students is the different
definitions of transfer units, based on either liquid and vapor

concentrations and either bulk or interfacial equilibrium
concentrations (i.e., Noy, Nox, Ny, and Nx); again, the present
approach elucidates this concept. The analysis described
above could analogously be carried out based on liquid
concentrations to give N'atecol. = NplatesMx, and thus

NOte ol. TMx (11)
N plate col.
NOy C My

As the Murphree efficiencies approach zero, the ratio
TMx / TMy = mV / L (see Eq. 2), and thus for a packed column

(a) m > L/V
0.2 NETP > NplateMy
110 > 11My
0.15 .

0.2 (b) m=L/V
NETP = NplatceTMy
S1o = TlMy
>"0.15 NETP = Noy
0.15 -

0.25 ~(c) m < L/V

NETP < NplateiMy
0.2 10 < 1oMy

S|NETP 0.15 -.-

0.05 0.1 0.15 0.2

Figure 1. McCabe-Thiele diagrams that show the origins
of the differences between lo and TnMy. The solid lines
represent the actual plates, and the dashed lines repre-
sent the equivalent theoretical plates; the operating lines
are the upper lines, and the equilibrium lines are the
lower lines. The analysis of these diagrams proceeds as
follows: For m>L/V, diagram (a) shows that more than one
equivalent theoretical plate is required to achieve the
same separation as two plates with nlMy=0.5, and so
NETP >NplatesrlMy for m>L/V. From Eqs. (4) and (6b), it follows
that no > 7nMy and NETP > Na"'"eo.for m>L/V. The analysis
for m=L/V and m

Chemical Engineering Education

Example Problem

Problem Statement
A packed column is being used to absorb acetone from
an air stream using water at 293K and 1 atm. The inlet
and outlet concentrations of acetone in the air stream are
0.6 mol % and 0.1 mol %, respectively, and the entering
air and water flow rates are 10 mol/hr and 40 mol/hr,
The separation given by this column must be improved
(at the same flow conditions), such that the outlet acetone
concentration in the air stream is reduced to 0.01 mol %.
By how much should the height of packing be increased
in order to achieve the desired separation?
The equilibrium relation for acetone at 293K and 1 atm
is given by y=1.19 x, where y is the molar concentration
in the air phase and x is the molar concentration in the
water phase.
Problem Solution
The solution to this problem is that the height of pack-
ing must be increased by the factor New /Nd (because
ing ~ ~ o mutbeN y (because

Nox mV
Noy,, L

Again, this result is the same as that obtained from the
standard mass-transfer analysis, but the present derivation
demonstrates that the relationship between N,, and Noy is
the same as the relationship between tlMx and 1My, which is
readily understood in terms of McCabe-Thiele diagrams.
Note that the number of transfer units based on interfacial
concentrations (Ny and N,) cannot be addressed with this

The present approach, of course, does not lead to the
complete design of packed columns-the height of a transfer
unit has not been addressed. This approach is therefore in-
tended to supplement, not replace, the usual mass-transfer
approach to packed-column design. Even so, some design
problems with packed columns can be addressed with just
the present treatment, as shown in Table 1.
The present analysis was described for dilute systems for
simplicity, but can be extended to concentrated systems.
This extension requires the use of a more complicated defi-
nition of the plate efficiency

q'y Yn Yn+l
(1- y I n+
1- yn )
in place of 1lMy. Note that yT = lMy in the dilute limit. The
sum of the efficiencies rly for all plates will generate the
correct equation for the number of transfer units of a packed
column for a concentrated system, in the limit that ily ap-
proaches zero and the number of plates approaches infinity.

the height of a transfer unit does not change). The values of NoyW and Nl can be obtained
from the Colbum equation, because the system is dilute:

o [ y.-mxin ( mV) mV
old yout-mXin L
oy- ( mV)

[Yin -mx, ( mV +mVl
y out-mx in L L
Oy =mV)

[0.006-1.19(0) 1.19(10) 1.19(10)
(ni 1--- + I-
[0.001-1.19(0) 40 ) 40
1 1.19(10)

F 0.006-1.19(0) 1.19(10) 1.19(10)
n 1- +
[0.0001-1.19(0) 40 40 J5.34
1 1.19(10)
40 )

NOy 5.34
= = 2.50
N 2 2.14
Thus, the height of packing must be increased by a factor of 2.50.

As noted above, this approach was used in the unit opera-
tions course at Tulane. The approach was received very well
by the students in the course. They seemed to enjoy thinking
through the steps in going from plate columns to packed
columns and did not have problems understanding the rela-
tionship between the columns based on this approach. The
students generally thought that this approach gave them a
physical understanding of the concept of transfer units and
provided a clear understanding of the relationships between
plate columns and packed columns.

1. E.g., (a) King, C.J., Separations Processes, 2nd ed., McGraw-
Hill, New York, NY (1980); (b) Sherwood, T.K., R.L. Pigford,
and C.R. Wilke, Mass Transfer, McGraw-Hill, New York,
NY (1975); (c) Perry, R.H., and D. Green, eds., Perry's
Chemical Engineers'Handbook, 6th ed., McGraw-Hill, New
York, NY (1984); (d) McCabe, W.L., J.C. Smith, and P.
Harriott, Unit Operations in Chemical Engineering, 4th ed.,
McGraw-Hill, New York, NY (1985); (e) Geankoplis, C.J.,
Transport Processes and Unit Operations, 3rd ed., Prentice
Hall, Englewood Cliffs, NJ (1993); (f) Wankat, P.C., Equi-
librium Staged Separations, Prentice Hall, Englewood Cliffs,
NJ (1988); (g) Treybal, R.E., Mass Transfer Operations, 3rd
ed., McGraw-Hill, New York, NY (1980); (h) Hines, A.L.,
and R.N. Maddox, Mass Transfer Fundamentals and Appli-
cations, Prentice Hall, Englewood Cliffs, NJ (1985); (i)
Seader, J.D., and E.J. Henley, Separation Process Prin-
ciples, John Wiley, New York, NY (1998)
2. Peters, W.A., Ind. Eng. Chem., 14, 476 (1922)
3. Sherwood, T.K., and E.R. Gilliland, Ind. Eng. Chem., 26,
1092 (1934)
4. Chilton, T.H., and A.P. Colburn, Ind. Eng. Chem., 27, 255
5. Murphree, E.V., Ind Eng. Chem., 17, 747 (1925)
6. Colburn, A.P.,. Ind. Eng. Chem., 33, 459 (1941)
7. Silver, L., Trans. Inst. Chem. Eng., 12, 64 (1934) 0

Fall 1998

E] M learning in industry

This column provides examples of cases in which students have gained knowledge, insight, and experience in the
practice of chemical engineering while in an industrial setting. Summer internships and co-op assignments typify
such experiences; however, reports of more unusual cases are also welcome. Description of analytical tools used and
the skills developed during the project should be emphasized. These examples should stimulate innovative ap-
proaches to bring real world tools and experiences back to campus for integration into the curriculum. Please submit
manuscripts to Professor W. J. Koros, Chemical Engineering Department, University of Texas, Austin, Texas 78712.



And How Does It Work?

University of Texas Austin, TX 78712-1062

me. My usual method of solving problems by memorizing a
I think this article should be read by every instruc- fixed procedure began to breakdown. I was blindly complet-
tor of introductory chemical engineering courses. The ing calculations on complicated problems by memorizing
author, a just-graduated BSChE, provides a view from the methods. I felt it would help to have a physical picture
the trenches of what it is like for a "practical learner" and understanding of the basic equipment before attempting
to translate the theoretical information we provide these calculations.
into meaningful knowledge. The student, Shanaya
Godiwalla, was more assertive than most and took Hoping to get some answers, I proceeded to one of the
responsibilityforfinding the missing "hands-on" com- chilling stations on campus. The plant manager gave me a
tour so I could actually see these mysterious pumps and
ponent to her education that she found necessary for tour so h could h ange eca see these mysterious pumps and
understanding. As Rich Felder has pointed out, stu- heat-exchangers. I became slightly frustrated when he showed a metal box and said, "This is a heat-exchanger," then
dent learning styles are diverse, and accommodating me a metal box and said, "This is a heat-exchanger," then
them will tend to maximize the efficiency of learning. proceeded to a smaller metal box, and said, "This is a pump."
SSha otes omtime the accommoaion can What does it look like inside? We sat down in the control
As Shanaya notes, sometimes the accommodation can
be quite simple. room with the owner's manual for the heat-exchanger and he
explained the components of the diagram, how they work
Bill Koros individually, and how they work as a unit. Things finally
began to come into perspective!
t all began in the first chemical engineering class "Mass I wanted and needed to know what I was calculating and
and Energy Balances." We had just learned to size pumps, how these calculations would be useful to me at a later date.
calculate the heat duty required for heat exchangers, and
draw process flow diagrams. I questioned exactly what a Shanaya Godiwalla is currently a chemical en-
gineering and dance major at The University of
pump looked like. How does it actually move fluid? What Texas at Austin. She has two years of industrial
are its internal components, and how would changing its experience and has accepted a full-time position
in process engineering at The Dow Chemical
internal components modify its capabilities? What does a I Company.
heat-exchanger really look like? And process flow diagrams-
does a chemical plant really look like that?
This class brought up an infinite number of questions for

Copyright ChE Division of ASEE 1998
306 Chemical Engineering Education

I have always been a "practical" learner. To learn something
effectively, the first few questions that need to be answered
consist of, "Will I need to do this someday? When, where,
and why?" Once these questions have been answered, my
interest and motivation develop because I have an under-
standing of the big picture. Then the "how" can be put into
Classroom learning usually skips straight to the "how."
There are many different types of learning styles, and my
learning ability is not maximized in a classroom. Touring the
chilling station allowed me to see the practical uses of heat-
exchangers, pumps, and process flow diagrams. This simple
experience allowed me to understand that these
calculations would become some of the basic
building blocks in my education ahead. As I CO-(
continued through the chemical engineering cur-
riculum, the same problem began to reemerge. I
found myself at a higher level, but again resort- e
ing to memorization rather than understanding
and analysis. Since I couldn't place where the
memorized techniques fit into a chemical
engineer's job, I again began to think I was edli
learning useless information. [It l
I decided co-oping was the answer to my
frustration. My thought was that the pieces Ir.e
would fall into place by my being in a plant too1l
environment, getting practical experience, and
developing a basic knowledge of process equip- :be -s
ment. As it turned out, the benefits of co-oping
greatly outweighed graduating a bit later. I re- the I
searched several companies and chose Dow
Chemical because of their high standards of
safety, intense training opportunities, and esprit UnJ
de corps of the employees.
My first term at Dow Chemical was in manu-
facturing. I was given a list of projects, which
varied from installing and automating pumps
and valves to demolishing old process units.
The experience provided hands-on opportunities and vast
exposure to basic process equipment. In addition, I learned
further how to work with people and how to "get work" from
My second term at Dow Chemical was in Epoxy Process
Research. My primary project was to develop a neural net-
work model for a gas phase reactor. This process model
would ultimately be used as an on-line sensor for process
control. The assignment gave me an opportunity to gain a
thorough understanding of Process Insights, a neural net-
work software. I also became more familiar with the VAX
and DEC term systems and completed a research and devel-
opment report that detailed the entire project. I also acquired
Fall 1998

a taste of the world of research and the unlimited opportuni-
ties of discovery it has to offer.
Engineering Sciences and Market Development represented
my third term at Dow Chemical. My project entailed running
co-polymerization experiments to better define the reaction
kinetics of a free radical initiated suspension polymeriza-
tion. This information helped to optimize the process, thereby
increasing production. My project used a reaction calorim-
eter to measure the heat flow released in combination with
an on-line mass spectrometer and Fourier transform infrared
spectrometer. The experimental data yielded conversion and
concentration versus time data to improve mathematical re-
action models. The project was completed
successfully, and I drafted a research and
g-iaOS development report prior to returning to
campus for my senior year.
R nan
Co-oping has been an essential part of
my my education. I am a hands-on type of
student who learns best and thrives in the
4Oll --* work environment rather than in the aca-
jed demic one. The co-op program has allowed
me to look at chemical engineering in a
B FhOt different light. Initially, it was very diffi-
Scult to imagine how I could apply some of
eed to, the things I was learning in school to the
ssf in_ outside world. Before co-oping, I learned
the academic material for the moment, never
fkplFa planning on using it again. Now I have a
better understanding of which material is
d to useful background information and which
stand will actually be useful in the workplace.
This experience was ideal for me because
1o "rea6 it forced me to use the information I learned
in school and to apply it immediately in the
W4iU-m d workplace. For example, after completing
_-e. thermodynamics, I worked on a reaction
calorimetry project at Dow.
Co-oping also promoted my personal
growth. I finally had the opportunity to experience living on
my own outside of a university environment and having a
job, just like a real adult. It led me to thinking about what it
would be like in the future to have that particular eight-to-
five job. I began to consider whether I would be happy
completing those types of tasks, working with those types of
people, and for that kind of company, day in and day out.
Working full-time and living on my own forced me to
learn more about myself and to establish some sort of career
path to prevent future misery. Co-oping helped me realize
what tools I need to learn to be successful in the workplace
and to understand what the "real world" would be like. I
would highly recommend the co-op program to anyone! 0

y i




] ,

SN lIaboratory




Filling Vessels

University of Newcastle New South Wales 2308 Australia

his paper describes an undergraduate laboratory ex-
periment that was developed to illustrate some of the
important concepts of compressible flow. A com-
pressible flow can be classified as one that demonstrates a
significant change in fluid density with a change in either the
pressure or the temperature, and is therefore usually ob-
served for gases. Indeed, compressible gas flows are en-
countered in many chemical engineering situations, e.g., the
transport and storage of high-pressure gases representing
feed or product streams from a chemical reactor, hence the
importance of this subject in the undergraduate curriculum.
A schematic of the experimental apparatus is shown in
Figure 1. It consists of an insulated pressure vessel con-
nected to the air mains via a globe valve. The first stage of
the experiment involves pressurizing the vessel (via the mains)
from initial atmospheric conditions up to a predetermined
elevated pressure (of maximum value 750 kPa) and measur-
ing the corresponding temperature change within the vessel.
The second stage of the experiment involves discharging the
vessel to the atmosphere via a nozzle and recording the
pressure versus time relationship for this process.
This paper considers only the first, or pressurizing, stage
of the experiment. The principal objectives are
To measure the temperature change within the vessel for a
number of different final pressures.
To develop a theory for this temperature change for comparison
with the experimental data.
Of particular significance are the concerns associated with
development of the theory for the vessel temperature rise
during pressurization. Theoretical analysis of the compress-
ible flow can be divided into three stages: 1) definition of the
"system" and the "surroundings" for the problem geometry;
2) application of an energy balance (the first law of thermo-

dynamics) for the defined system; and 3) algebraic manipu-
lation of the energy balance using the appropriate simplify-
ing assumptions and established thermodynamic relation-
ships. This paper illustrates the importance of the first step
(that of defining the correct system and surroundings) in the
model development, and in particular how a slightly incor-
rect definition can produce significant errors in the resulting
model predictions.

The experiment described in this paper involves pressuriz-
ing an insulated vessel (as shown in Figure 1) from a starting
pressure, P0, up to a final equilibrium pressure, PE, before
discharging the vessel to the atmosphere through a nozzle. In
this section, the theory for the first, or pressurizing, stage is
considered, with the objective of determining an expression
for the change in temperature of the system, (TE To), which
accompanies the pressurizing process.

Stephanie E. Forrester is a Research Associ-
ate in Chemical Engineering at The University
of Newcastle, Australia. She obtained her edu-
cation from the Universities of Edinburgh (BE
Hons) and Cambridge (PhD). Her research and
teaching interests include fluid mechanics and
computational modeling.

Geoffrey M. Evans is a Senior Lecturer in
Chemical Engineering at The University of
Newcastle, Australia. He obtained his education
from the University of Newcastle (BE Hons and
PhD) and has been a faculty member since
1985. His teaching and research interests are in
the areas of fluid mechanics, thermodynamics,
and multiphase systems.

Copyright ChE Division ofASEE 1998

Chemical Engineering Education

Figure 1. Schematic representation of the
experimental apparatus.

Figure 2. Definition of the system and surroundings, the
initial conditions, and the final conditions.

air supp
from mai

Fall 1998

In developing a theory for the temperature change,
ly the first stage is to correctly define the "system," or
S control volume, on which to perform the thermody-
namic analysis. Referring to Figure 2, the control vol-
ume is bounded by the outer surface of the stainless
steel vessel and the air in the mains pipe that enters the
pressure vessel during pressurization. Thus, the system
consists of the air initially inside the vessel, the air in
the mains that enters the vessel, and the stainless steel
wall of the pressure vessel.
It is important to consider the air in the mains as part
of the system because the system can be treated as a
closed system and Eq. (1) applies. Otherwise, when the
globe valve is opened, mass will cross the system bound-
ary and an open system would apply, and Eq. (1) would
need to be modified.
The initial and final conditions for this system are
also shown in Figure 2, while the principal assumptions
required in the analysis presented below are: 1) the
insulation on the pressure vessel allows the vessel to be
regarded as adiabatic, and 2) the air behaves as a per-
fect gas. Applying the first law of thermodynamics to
the closed system outlined above gives['
Qs-Ws=AUs (1)
where Ws is the work done by the system on the surround-
ings, Qs is the heat transfer between the system and the
surroundings, and Us is the internal energy of the system. As
discussed above, the system is assumed to be adiabatic, i.e.,
Qs =0. The work done by the system on the surroundings is
given by
Ws =PMAV (2)
where PM is the mains pressure (the pressure at which the
volume change is taking place), and AVs is the change in
volume of the system, given by
AVs =V, -(V +VM)=-VM (3)
where V, is the volume of the vessel and VM is the volume
of gas in the mains that enters the vessel.
Combining Eqs. (2) and (3) and applying the perfect gas
law to gas initially inside the mains gives
M(mMRTO / (4)
Ws _-M = -mMRTo (4)

where To is the initial temperature of the system, R is the
specific gas constant, and mm is the mass of gas in the mains
that enters the pressure vessel (see Figure 2).
The change in internal energy of the system, AUs, is given
AUs = AUG + AU = -W, = mMRTo (5)
where AUG is the change in the internal energy of the gas

within the system, and AUy is the change in internal energy
of the stainless steel within the system.
The change in internal energy for the gas is given by

AUG = mGC (TE T) (6)

where mG is the mass of gas in the system and CVG is the heat
capacity at constant volume. Equation (6) indicates that AUG
is related to the change in system temperature only (and not
to the change in system pressure).
The change in internal energy for the stainless steel vessel
wall is given by

AUv = mvAhv = mvCv (TE -To) (7)

where Ahv is the change in enthalpy of the stainless steel
vessel, my is the mass of the stainless steel vessel, and Cp is
its specific heat capacity.
Substituting Eqs. (6) and (7) into Eq. (5) yields

AU = mGCvG (TE T,)+ mvC(TE To)= mMRTo (8)
Equation (8) can be modified using the perfect gas law,

PEVV (9)
mG (9)
mM PEVV (10)

Substituting Eqs. (9) and (10) into Eq. (8) and rearranging
results in the following quadratic equation for the final equi-
librium temperature, TE:

a2T2 + aTE + o =0 (11)

a2 = mvCvT0o

S= VvT +(PoVvTo)- mvCpvT

"ao(""~")I JT

Equations (11) through (14) can be solved to give the final
equilibrium temperature of the system, TE, for a specified
value of the final equilibrium pressure, PE.

Equipment Description A schematic representation of
the experimental apparatus is shown in Figure 1. It consists
of a 102-liter insulated stainless steel pressure vessel, the
inlet of which is connected to the high-pressure air mains via
a globe valve, while the vessel is discharged to the atmo-
sphere through a nozzle attached via a second globe valve.
The temperature of the stainless steel pressure vessel wall

and that of the air within the pressure vessel are measured
using Chromel-Alumel thermocouples connected to a chart
recorder for continuous recording, while the vessel pressure
is displayed on a bourdon pressure gauge.
Experimental Procedure The preliminary stages in con-
ducting this experiment are to check that both the inlet and
exit valves to the pressure vessel are closed and to record the
ambient temperature and pressure conditions for the vessel.
The chart recorder can then be switched on, set to the correct
scale and speed (0.1 mV cm-' and 20 cm hr', respectively, in
our case), and the recording pens zeroed. Temperature re-
cording is commenced and the inlet valve connecting the
vessel to the mains air supply is opened. The pressure within
the vessel is visually observed to increase and when it has
reached the desired value, the inlet valve is closed and the
system left to reach equilibrium, i.e., the chart recorder indi-
cates a constant temperature for both the vessel wall and
internals, which typically takes around five minutes. The
final equilibrium conditions of vessel pressure and tempera-
ture are also recorded.
The vessel is then discharged by placing a muffler over
the exit nozzle (to reduce noise pollution) and opening the
outlet valve. The pressure is recorded at regular intervals
during the discharging process, which typically takes around
two to three minutes, and ends when the vessel has reached
atmospheric pressure. The vessel wall and internal tem-
peratures are allowed to reach steady state, taking ap-
proximately five minutes, before the process is repeated
for a different final pressure.

Figure 3. Schematic representation of the variation in the
vessel temperature and pressure during pressurization.

Chemical Engineering Education

air inlet
supply valve equilibrium
opened closed conditions







Important Note: There are a number of inherent safety
implications associated with this experiment that should be
noted. First, although air is a benign material under atmo-
spheric conditions, it can become hazardous at elevated pres-
sures (4-10 atmospheres in this case). At 10 atmospheres,
the pressurized air has enough force to shoot a water jet 100
meters into the air! Second, the experiment should only be
attempted using a properly certified pressure vessel fitted
with the appropriate pressure relief system. This is abso-
lutely essential when undertaking any experiment under pres-
surized conditions, but is especially important when gases
are involved since any rupture could result in a catastrophic
explosion of the vessel.

Variation in the Vessel Temperature and Pressure Dur-
ing Pressurization The first important stage in discussing
the experimental results is a qualitative understanding of the
variation in vessel wall temperature, internal air tempera-
ture, and vessel pressure as a function of time during pres-
surization. A schematic plot showing the typical time depen-
dency of these three parameters is given in Figure 3. Consid-
ering initially the vessel pressure, at t=0 the inlet valve to the
vessel is opened and the air from the mains starts to enter the
vessel, causing the pressure to rise steadily with time. When
the desired final vessel pressure, P,, is attained, the inlet
valve is closed, thereby isolating the vessel. The vessel is
then left until all three parameters have reached equilibrium;
during this equilibratingg' period, the vessel pressure drops
slightly to a final equilibrium pressure PE. This drop in

* experimental
* theoretical

*U ***

0 I Ii
200 400 600 800
PE (kPa)

Figure 4. Experimental and theoretical results for the varia-
tion in the system temperature difference as a function of
the final vessel pressure (data points collected from lower
to higher pressure).

pressure is a result of the drop in the internal air temperature
during the equilibrating period, for reasons discussed below,
as can be illustrated by the perfect gas law at constant vol-

PE FFj (15)

The variation in internal air temperature with time shows
an initial rapid rise, resulting in a peak temperature shortly
after the inlet valve is opened, followed by an exponential
decay to the final equilibrium value, TE. In contrast, the
temperature of the stainless steel vessel wall shows a steady
increase from the initial temperature, To, to the final equilib-
rium temperature, TE (the same as that of the internal air).
During the pressurization process, the work done on the
system is converted into heat, as defined by Eq. (8), which
produces the initially rapid rise in the internal air tempera-
ture. As the air temperature rises, then the temperature dif-
ference between the air and the stainless steel wall increases,
resulting in an increasing rate of heat transfer from the air to
the wall. The maximum in the air temperature represents the
point where the rate of heat transfer is sufficiently high that
the air temperature must start to fall. The air temperature
continues to drop and the wall temperature continues to rise
until they both reach the same equilibrium temperature, TE,
and no more heat transfer can take place.
Temperature Increase During Pressurization Versus Fi-
nal Pressure Typical experimental and theoretical results
for the system temperature change, (TE To), as a function of
the final vessel pressure, PE, are shown in Figure 4 (in which
the theoretical results are based on the input parameters
given in Table 1). For this case the lowest pressure was
selected first and the consecutively higher pressures were
chosen. It can be seen that the experimental and theoretical
data both show an approximately linear trend of increasing
temperature difference with increasing final pressure. The
good overall agreement between the two sets of data for the
first few pressures confirms that the theory provides an

Input Parameters for the Theoretical Calculations

Parameter Value Notes
VV 0.102 m
mv 43.0 kg
C G 718 J kg-' K' for air'21
R 287 J kg K-
CpV 503 J kg K-' for stainless steel'2'
Po 101 kPa in Figure 4
T, 295 299 K in Figure 4
P0 101 kPa in Figure 5
T, 295 K in Figure 5

Fall 1998

acceptable description of the experimental process. Although
the theory predicts the correct order of magnitude and varia-
tion in the results, however, there are differences between
the absolute experimental and theoretical results. The reason
for the discrepancy is outlined below.
Considering initially the results in Figure 4, clearly there is
good agreement between the experimental and theoretical
data at the lower final equilibrium pressure ([TE T] = 1 K
at PE = 300 kPa), but the experimental rate of increase in [TE
- To] with increasing PE is significantly higher than predicted
by the theory such that by the highest pressure employed, PE
= 700 kPa, the experimental temperature rise is about 6 K
compared to a theoretical value of about 3 K. This discrep-
ancy between the experimental and theoretical values is a
direct result of the limitations in the experimental procedure.
For the first experimental data point at PE = 300 kPa, the
pressure vessel has been unused for several hours or days
and is therefore in true equilibrium with the ambient sur-
roundings. Time constraints during a typical experimental
run are such that following discharge of the pressure vessel
to the atmosphere it is only feasible to allow of the order of
five to ten minutes for the system to return to ambient
equilibrium conditions before the next pressurization stage
is commenced. Inevitably, some residual heat remains in the
vessel wall for the subsequent pressurizing run, which re-
sults in a slightly higher final equilibrium temperature com-
pared to that which would have been achieved had this been
the first point to be measured. This trend continues as the
successive data points are collected, resulting in a pro-
gressively higher experimental temperature rise, com-
pared to that which would have been obtained under
ideal conditions. This represents the principal reason for
the discrepancy between the experimental and theoreti-
cal data points shown in Figure 4.
The argument presented above to describe the observed
difference between experimental and theoretical results for
[TE To] in Figure 4 is further reinforced by examining the
data in Figure 5 for a different experimental run. In this
separate experiment, the first data point to be collected was
that for the highest pressure, PE = 750 kPa, and the value of
the final equilibrium pressure was progressively reduced as
the experiment proceeded. As a result, the best agreement
between theory and experiment in the Figure occurs at the
highest value of PE, i.e., for the first data points collected,
and as the experiment proceeds, the experimental tem-
perature difference becomes progressively higher than
that predicted by the theory. The largest discrepancy
between the experimental and theoretical values of [TE -
To] is observed to occur at the lowest value of PE, i.e., for
the final data point collected.
In summary, the variation between theoretical prediction
and experimental measurement is determined by the amount
of time allowed for the system to equilibrate between each

experiment. Although not included in this paper, a much
closer agreement can be obtained by accounting for the
residual heat remaining in the steel shell. This is achieved by
assigning different initial (measured) temperatures, To, to
the steel shell and the mains air in the theoretical analysis.

Experimental Error The possible sources of error in this
experiment are: 1) the accuracy of the thermocouples used to
monitor the temperature of the vessel wall and internal air;
2) the accuracy with which the vessel pressure can be read
from the bourdon pressure gauge; 3) the accuracy with which
the chart recorder output can be read and converted to a
temperature. Typical errors associated with converting the
thermocouple signal to the chart recorder (or any other form
of datalogger) and pressure gauge readings are negligibly
small and do not affect the observed trend in the experimen-
tal and theoretical data.
By far the greatest source of error lies in the measurement
of a small temperature difference of a few degrees Celsius
between two relatively large numbers. In practice, the accu-
racy of the thermocouple system is very much a function of
the construction of the thermocouple and the method of
installation. The absolute error could be as high as C,
which represents a percentage error of order 100%. There is
also the added difficulty of obtaining or using a single point
to obtain a "representative measurement" of the average
temperature over the entire stainless steel shell. These are all
points that students should consider when discussing their
results or experimental methodology. But the main point of


8 theoretical



2 U *

0 I
200 400 600 800
PE (kPa)

Figure 5. Experimental and theoretical results for the varia-
tion in the system temperature difference as a function of
the final vessel pressure (data points collected from higher
to lower pressure).
Chemical Engineering Education

the experiment is to highlight the difference in temperature
predictions when the steel shell is not included in the analy-
sis, and to show that this assumption can lead to incorrect
temperature predictions of up to a hundred degrees!

Effect of Control Volume Selection on the Theoretical
Temperature Results To highlight the importance of cor-
rectly defining the system in the theoretical development, it
is of interest to examine the values of the temperature differ-
ence, [TE T0], based on the common error of neglecting the
stainless steel pressure vessel wall from the control volume
in Figure 2. This can be achieved by simply substituting
my=0 in Eqs. (11) through (14), which yields the following
expression for the temperature difference:

TE -T0=To IP1- +P (16)

Using the input parameters given in Table 1 with P0 = 101
kPa and To = 293K, Eq. (16) predicts that at PE = 300 kPa,
[TE To] = 70K, while at PE = 700 kPa, [TE To] = 95K.
Comparison of this data with those presented in Figures 4
and 5 illustrates that such a system selection results in almost
two orders of magnitude over prediction in the system
temperature rise. This clearly shows how a simple and
common error in the theoretical analysis of the com-
pressible flow can produce highly significant errors in
the resulting model predictions.


We have discussed an undergraduate experiment on com-
pressible flow, based on the pressurization of an adiabatic
pressure vessel. In particular, the vessel is pressurized from
ambient conditions up to an elevated value and the corre-
sponding temperature change of the vessel wall and internals
is measured; this process is then repeated for a number of
different final pressures of between 250 and 750 kPa. Dis-
cussion of the experimental results has involved: 1) a quali-
tative description of the variation of internal and wall tem-
peratures and of the vessel pressure as a function of time,
and 2) the development of a theoretical model for the pro-
cess and a comparison of the resulting model predictions
with the experimental data. We found that the time de-
pendency of the internal temperature and pressure can be
successfully explained based on simple thermodynamic
The equilibrium temperature rise versus final vessel pres-
sure was measured experimentally over the range PE = 250
kPa to 750 kPa and showed an approximately linear increase
in [TE To], from about 1 K up to about 6 K, with increasing
PE. The corresponding theory predicts a similar linear trend
in the data and good agreement with experiment over the
first few data points collected, thereby confirming its appli-
cability. But the theory progressively underpredicts the tem-
Fall 1998

perature rise as the experiment proceeds. This observation
has been well explained by considering the residual heat that
remains in the stainless steel wall of the pressure vessel as
the experiment progresses.
The importance of defining the correct system and sur-
roundings in the development of the theoretical model has
been identified. For the adiabatic pressure vessel connected
to the air mains, the control volume must include the air
initially in the vessel, the air in the mains that enters the
vessel, and the vessel wall. The common error of neglecting
the stainless steel vessel wall from this system control vol-
ume produces an almost two order of magnitude overesti-
mate in the temperature rise, which could obviously produce
significant industrial consequences.

CV specific heat capacity for the stainless steel vessel [J kg'
CvG heat capacity of the gas at constant volume [J kg-' K-']
Ahv change in enthalpy of the stainless steel vessel [J kg']
m mass of gas in the system [kg]
mm mass of gas in the mains that enters the vessel [kg]
my mass of the stainless steel comprising the pressure vessel
PE final (equilibrium) pressure in the vessel [N m-2]
P, vessel pressure when the inlet valve is closed [N m ]
P mains supply pressure [N m2]
Po initial pressure in the vessel [N m-2]
Qs heat transfer between the system and surroundings [J]
R specific gas constant [J kg K-]
TE final (equilibrium) temperature of the system [K]
TF vessel internal temperature when the inlet valve is closed
To initial temperature of the system, [K]
UG internal energy of the gas within the system [J]
Us internal energy of the system [J]
U, internal energy of the stainless steel wall of the pressure
vessel [J]
VM volume of gas in the mains that enters the pressure vessel
Vs volume of the system [m3]
Vv volume of the pressure vessel [m3]
Ws work done by the system on the surroundings [J]
ao constant given by Eq. (11) [J K2]
a, constant given by Eq. (11) [J K]
a2 constant given by Eq. (11) [J]
1. Kay, J.M., and R.M. Nedderman, Fluid Mechanics and
Transfer Processes, Cambridge University Press, Cambridge,
U.K. (1985)
2. Perry, R.H., and D. Green, Chemical Engineers' Handbook,
6th ed., McGraw-Hill, Singapore (1984) 0

[ Ui^ laboratory




Control Technology Centre University of Manchester

he difficulties that are faced when teaching process

control as part of a chemical engineering degree
course have been highlighted by a number of au-
thors.[13] These difficulties range from a lack of time avail-
able to teach even the most fundamental principles of pro-
cess control, to providing sufficient engineering practice of
the theory taught in the classroom. Consequently, the effec-
tiveness of many university process-control courses, judging
from industrial feedback, is poor.[4]
A number of recent publications have suggested the use of
process-control computer-simulation packages to overcome
these difficulties.I51 These simulations typically allow the
student to progress through a number of practical process
control problems ranging from simple PID controller tuning
to the development of complex model-based control strate-
gies. Although these software applications do provide valu-
able experience for students, they unfortunately fail to pro-
vide the real engineering practice that is essential for stu-
dents who wish to pursue a career in control engineering.
This experience can only be gained through development
and analysis of process control applied to real industrial
equipment. It has further been suggested"I that students find
control engineering courses much more interesting and use-
ful when they are practice oriented.
A simple and relatively cheap method of providing stu-
dents with some practical experience of process control is to
set up a laboratory experiment that regulates the level of
liquid in a tank using a control valve. Installing a single-loop
controller to such a system will even allow the student to
become familiar with simple forms of industrial controllers.
Unfortunately, this type of experiment tends to trivialize the
role of modem control engineers in industry where, typi-
cally, complex chemical processes are controlled with the

* Address: Monash University, Melbourne, Australia

*Manchester, United Kingdom

aid of sophisticated process-control hardware.
The Department of Chemical Engineering at Monash Uni-
versity has recently made moves to provide a link between
university teaching and the industrial application of process
control. This has been achieved by investing a significant
amount of resources in the development of a process-control
experiment that incorporates both industrial process equip-
ment and an advanced control platform. This paper describes
both the equipment and the experiments that have been
developed at Monash University.

The system was set up to meet three requirements. The
first was to provide all students in the chemical engineering
course with exposure to an industrial control problem. The
second was to provide a limited number of final-year stu-
dents with increased process-control exposure by way of a
six-month project, and the final requirement was that
each of the experiments could be used for demonstration
purposes throughout the course. The experiment that was
developed to fulfill these requirements is detailed in the
following section.

Barry Lennox received his BEng in Chemical Engineering (1991) and
his PhD in Process Control (1996) from the University of Newcastle-
upon-Tyne. He subsequently transferred to Monash University where
he worked as a Research Fellow in the Department of Chemical
Engineering. and in 1998 became a lecturer in Control Engineering at
the University of Manchester. His research interests lie in the advance-
ment and application of process control and monitoring techniques to
batch and continuous process systems.
Michael Brisk obtained his PhD in Chemical Engineering from the
University of Sydney in 1965. He worked for six years for ICI in the
United Kingdom before becoming a senior lecturer in chemical engi-
neering at the University of Sydney, where he spent the next eleven
years before transferring to ICI Engineering in Melbourne and estab-
lishing an Advanced Process Control Group. In 1994 he became
Adjunct Professor of Process Control and became Dean of Engineer-
ing at Monash University in 1995.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education



Secondary G
CV1 (GV1)

HX1 Steam

TK1 P1 Condensate


Figure 1. Heat exchanger process diagram.

Figure 2. Process control block diagram.

Figure 3. Comparison of single loop and cascade

Within the ever-present financial constraints that beset a university
department, the decision was made to modify and use an existing
experimental rig. The apparatus, originally designed to study heat-
transfer processes, was chosen for the following reasons:
l The time constant of the system is approximately 5-10 minutes. This
means that during a typical laboratory period lasting three hours, a
student has sufficient time to tune and modify a number of control
loops and strategies.
The equipment used in the system is relatively large (2-inch pipes and
2-meter-long heat exchangers) and has the appearance of an indus-
trial plant, which is essential if the students are to gain familiarity
with industrial process control.
0 Only minor modifications, such as the installation of control valves
and thermocouples, were required for this piece of equipment to be
transformed into a process-control experiment.
Figure 1 shows a schematic of the modified process. The steam
heater, HX1, heats the circulating water stream, fed from tank TK1,
by condensing steam on the tube walls. The outlet water from this
heat exchanger is then fed to HX2, a horizontal shell-and-tube heat
exchanger, which cools the water using a secondary coolant. Once
cooled, the water returns to the circulating tank. To provide the
control in the process, a control valve, CV1, was installed on the
supply line of the secondary coolant.
The objective of the experiment is to design a controller capable of
regulating the temperature of the circulating water as it exits HX2.
Once designed, the controller is tested for its effectiveness at
regulating this temperature in the presence of set-point changes
and process disturbances.
The overall process-control block diagram of this process is illus-
trated in Figure 2. In this diagram, TC 1 represents a conventional PID
feedback controller that can either regulate the temperature, T inde-
pendently or can be configured to form the primary loop of a cascade
controller, which incorporates FCl as the secondary loop. GMS and
GMT1 measure the secondary coolant flow rate and the outlet tem-
perature, respectively. GV1 and GP1 represent the dynamics of the
valve and heat exchanger, respectively. GD1, GD2, and GD4 repre-
sent the disturbances to the system, which are the secondary coolant
supply pressure, circulating-water flow rate, and the temperature of
the water as it exits the heat exchanger.
During the experiment, the students are asked to develop and com-
pare a conventional feedback controller with cascade control. In
doing so, they are expected to make a number of important decisions
alone; for example when tuning the controller, the decision must be
made as to what size step test should be performed to enable a step
response of the system to be analyzed for tuning purposes.
If the experiment is performed correctly, the result is that the
feedback controller and the cascade controller operate similarly for
servo control. But due to the relatively fast dynamics of the secondary
control loop, cascade offers significant improvement for regulating
the process during process disturbances. This is confirmed in Figure
3, which shows the temperature control of both the single loop and the
cascade controllers during a process disturbance. The top graph in

0 50 100 150
Sample No.

Fall 1998

200 250

this figure shows the control obtained using single-loop feed-
back control, and the graph below shows the control result-
ing from cascade control. In each case, there are two process
disturbances, a reduction in upstream coolant supply pres-
sure at sample number 45, and an increase in upstream
coolant supply pressure at sample number 160 (each sample
represents five seconds).

The control interface for the experiment was an ABB
MOD300 distributed control system (DCS), which is typical
of systems widely used in the chemical and process indus-
tries. By using this system, the students become aware of the
features of a control system that they might expect to en-
counter after leaving the university. Standard remote input/
output blocks were used to transmit the on-line measure-
ments to the DCS. As with all the process-control equip-
ment, these blocks were placed in such a way that the stu-
dents were able to view them.

Particular disadvantages when using many types of dis-
tributed control systems are
The graphical and data analysis capabilities of the systems
are restricted.
The systems were designedfor conventional control struc-
tures, such as PID, cascade, ratio, and simple forms of
model-based control. Although these control structures will
be all that is requiredfor the student experiments, the system
will also be used in postgraduate research, where more
advanced forms of model-based control will be required.
It is difficult to give a demonstration of the systems, since to
do so the demonstration has to take place in the laboratory,
where space is limited.
To combat the first of these drawbacks, a novel software
package has been developed at Monash University that en-
ables two personal computers (PCs) to communicate di-
rectly with the DCS. Figure 4 shows schematically how the
hardware has been configured for this system.
A short piece of code was written on the DCS in Taylor

Control Language (a programming language similar to PAS-
CAL). This program operates in the background and peri-
odically (currently every five seconds) sends the collected
process data, such as setpoints, values of the measured vari-
ables, and controller modes, via a serial link to a PC posi-
tioned beside the DCS. A software package running continu-
ously in the background on the PC receives the data and
makes it available to any other Windows-based software
application via a dynamic data exchange (DDE) link. At
present, the data is transferred to an open MATLAB
session. MATLAB allows the data to be collected in the
background, thus allowing the advanced graphical and
data-analysis functions, available in MATLAB, to be run
in the foreground in real time.
Furthermore, the system has been configured so that at the
same time as receiving data from the DCS, the PC is also
able to transmit data to the DCS. The DCS has a second
program running in the background that continually moni-
tors one of its serial ports and detects any signals that are
sent from the PCs. It is therefore possible to change param-
eters such as process set points and loop modes from the PC.
This feature enables complex control systems to be devel-
oped and operated from the PC, with the DCS acting as an
interface between the PC and the process hardware.
To enable the system to be demonstrated in the classroom,
a final piece of software has been developed that can be
operated by any PC connected to the Internet. This software
package allows the user to access identical displays to those
available on the DCS. Periodically, the software package
transmits a signal via the standard TCP/IP protocol to the PC
positioned locally to the DCS. The local PC replies to this
signal by transmitting the most recent process data. At present,
the signal that must be sent to the local PC, before data is
transmitted, is an ASCII string. But, if required this signal
could be encrypted to increase the security of the system.
Once the data is received, the relevant displays on the PC are
subsequently updated. It is also possible for any computer
connected to the Internet to change process setpoints and
controller modes remotely. This feature allows the experi-
mental systems to be operated via any computer connected

DCS Local PC

Control Internet
H ao cale Intemet
Field Bus

Visual Basic Communications -
TCL Procedures allowing on-line data analysis
using MATLAB.

Figure 4. DCS hardware set-up.

Chemical Engineering Education

to the Internet. In particular, it allows the process-control
hardware to be operated from the lecture room, thus allow-
ing the system to be used for lecture theater demonstrations.
Although it is possible to use a dynamic simulator for dem-
onstration purposes in lectures, we feel that the students
take more interest and consequently gain greater appre-
ciation of the application of control by operating the real
equipment through the remote interface. Figures 5 and 6
show two example screens that are provided by the PC
software package. The graphics are almost identical to
those available on the DCS.

Y. r25 YmU94j

-T 6 .0,




SLoop- A Block Diagram W


OP 1. T

MV 1. =4
SOC.. CodoG



This software package may in the future provide an oppor-
tunity to improve the distance education courses offered by
the university, since it would be possible to conduct the
experiment well away from the university campus.

A further advantage that the system offers is that it is also
possible for the students to run other applications in the
foreground because all the software on the local PCs has
been designed to operate in the background. For example, all
the process control course notes, developed in Lotus
Freelance, are available for the students to read through
while they are conducting the experiment.

siB S


1. 10.0

Figure 5. Example of a group graphic.





T Figure 6. Example of an overview display. FET

Figure 6. Example of an overview display.


The experimental system that has been designed and
commissioned at Monash University provides students
with the opportunity to gain familiarity with some of the
process-control structures and hardware commonly
used in industry. This experiment will lead the stu-
dents to a greater understanding of what they may
expect to see if they become involved with process
control after graduation.

The advanced communications package developed to
allow PCs to communicate directly with distributed con-
trol systems greatly enhances the flexibility of the sys-
tem for teaching. It enables the user to receive and
transmit data to the control system from anywhere in the
world, provided the user is connected to the Internet and
has TCP/IP compatibility. This facility has many appli-
cations, including distance education and remote con-
tact with process equipment.


The authors would like to acknowledge the help and
support of the Chemical Engineering Department at
Monash University. Particular thanks are due to Gill
Atkin for supplying the expertise in modifying and in-
stalling most of the process equipment. The support of
ABB is also gratefully acknowledged, as is the encour-
agement provided by Ming Tham.

1. Jovan and J. Petrovcic, "Process Laboratory: A Neces-
sary Resource in Control Engineering Education," Comp.
Chem. Eng., 20, 1335 (1996)
2. Lewin, J. Rockman, and R. Lavie, "Teaching Advanced
Process Control to Undergraduates," Comp. Chem. Eng.,
20, 1347 (1996)
3. Chung and R.D. Braatz, "Teaching Antiwindup,
Bumpless Transfer and Split-Range Control," Chem. Eng.
Ed., 32(3), 220 (1998)
4. Merrick and J.W. Ponton, "The Ecosse Control
Hypercourse," Comp. Chem. Eng., 20, 1353 (1996)
5. Bequette, K.D. Schott, V. Prasad, V. Natarajan, and
R.R. Rao, "Case Study Projects in an Undergraduate Pro-
cess Control Course," Chem. Eng. Ed., 32(3), 214 (1998) 1

Fall 1998



S3=!l laboratory





Manhattan College Riverdale, NY 10471-4099

Membrane process experiments involving ultrafil-
tration (UF) of dairy products have been devel-
oped and classroom tested. These experiments
provide an effective method for integrating membrane sepa-
ration processes into the curriculum while at the same time
offering learning experiences of broad-based interest to
chemical engineering students through familiar applica-
tions such as cheese production and environmental
cleanup. UF experiments can be used in conjunction with
the required unit operations laboratory and mass transfer/
separation processes courses, or as a component of elec-
tive courses in membrane process technology or bio-
chemical engineering.
Although there has been considerable discussion about
revamping the chemical engineering curriculum, topics cov-
ered in required courses remain fairly uniform throughout
academe. Required separations or mass transfer courses fo-
cus heavily on traditional equilibrium-staged operations, in-
cluding distillation, absorption, and extraction.111 These unit
operations are very important in the chemical and petroleum
industries, but chemical engineering students also need to
study separation processes that have become increasingly
important in other areas, such as food processing, biotech-
nology, and environmental engineering. A 1995 undergradu-
ate survey of chemical engineering programs indicates that
there has been little change in topics covered in mass trans-
fer courses other than slight increases in time spent on
adsorption (2.5%) and membranes (2.6%).1" The impor-
tance of including instruction on membrane process tech-

'Address: Tosco Refining Company, Linden, NJ 07036
2 Address: Clemson University, Clemson, SC 29634-0909
3 Address: Rowan University, Glassboro, NJ 08028-1701

nology in the chemical engineering curriculum has been
addressed in recent articles.12-8
Membrane processes are mass transfer unit operations
used for liquid- or gas-stream separations. The family of
processes covers a number of operations, including reverse
osmosis (RO), nanofiltration, ultrafiltration, microfiltration
(MF), dialysis, electrodialysis, gas permeation (GP), and
pervaporation (PV).[91 These processes are not really new,
but are often unfamiliar to the typical chemical engineer due
to lack of exposure during his or her education.
RO is in its fourth decade, having been developed in the
early 1960s for industrial operations such as production of
potable water from seawater. UF has been used on a com-

Thomas D. Conlee is currently employed as an environmental engineer
at Tosco Refining Company, where he is working on ground water compli-
ance. His job entails assuring that the refinery is in compliance with local
ground water regulations. He earned his BS degree in chemical engineer-
ing from Manhattan College and completed a senior honors project on
membrane process technology.
Helen C. Hollein is Professor and Chair of the Chemical Engineering
Department, and Director of the Biotechnology Program, at Manhattan
College. She earned her MS and DEngSc degrees from the New Jersey
Institute of Technology and her BSChE from the University of South
Carolina. Her research and teaching interests are in biochemical engi-
neering and bioseparations, and she is active in curriculum and laboratory
Charles H. Gooding is Professor and Chair of the Department of Chemi-
cal Engineering at Clemson University. He earned his BS and MS de-
grees from Clemson University and his PhD from North Carolina State
University. His primary research interests are membrane separation pro-
cesses, including nanofiltration, ultrafiltration, pervaporation, andgas sepa-
ration systems; mathematical analysis of membrane processes; and envi-
ronmental applications of membrane technology.
C. Stewart Slater is Professor and Chair of the Department of Chemical
Engineering at Rowan University. He received his BS, MS, MPh, and PhD
degrees from Rutgers University. His research and teaching interests are
in separation and purification technology, laboratory development, and
investigating novel processes for interdisciplinary fields such as biotech-
nology and environmental engineering.

Copyright ChE Division of ASEE 1998

Chemical Engineering Education

mercial scale since the early 1970s, with
dairy processing representing one of the larg- .
est applications of membrane technology in
the world.10"l The earliest commercial appli- -
cation of UF was the concentration of pro-
teins from whey, a milk by-product gener-
ated during the traditional coagulation pro-
cess for cheese production."0' Similarly, UF
is used to remove lactose and milk pro-
teins from whey wastewater streams prior -
to disposal, i.e., for environmental treat- -
ment. In newer processes for manufacture
of soft cheeses, yogurt, ice cream, and
other dairy products, UF is used to
preconcentrate whole or skim milk prior
to entering the production chain. 015]
This paper discusses laboratory experi- -
ments on ultrafiltration of dairy products that
may be used to increase coverage of mem-
brane processes in the chemical engineering
curriculum and gives typical results from
experiments run by students at Clemson Uni-
versity (CU) and Manhattan College (MC).
At CU, a spiral-wound UF system manu-
factured by Koch Membrane Systems, Inc.,
has been used to concentrate various dairy
products in the experimental component of i
an elective course on membrane separation
processes and in the senior unit operations
laboratory.161 The CU experiments focus on
application of the UF process for environ-
mental cleanup of rinse waters from dairy
operations, e.g., a whole-milk pasteurization
system, an ice-cream freezing unit, and a
chocolate-milk production system."161 At MC, a Millipore
Model TCF10 thin-channel UF system (formerly manufac-
tured by Amicon) is used in the unit operations laboratory
course for an experiment on milk concentration.
Based on the authors' experiences, the milk experiments
have an intrinsic appeal to all of the chemical engineering
students and instructors, not just the biochemical engineers
and membrane specialists. Interest is heightened by employ-
ing a typical local application; for example, UF is one of the
steps used in cheese production at Kraft Foods Inc. in
Tarrytown, New York (an industrial site near MC).
The UF experiments discussed in this paper are but one
component of an array of membrane process experiments
that can be employed to enhance the chemical engineering
curriculum. RO, GP, MF, and PV experiments are also in-
cluded in the courses at CU and MC. {5.617-19' Several mem-
brane process experiments will be integrated into the re-
quired mass transfer and separation processes courses at


Fall 1998

Rowan University. In addition, a hand-held
i RO system is used in the freshman engineer-
ing course at Rowan University.'2021]
The authors have previously outlined UF
experiments for separation and concentration
of protein and enzyme mixtures, experiments
that are used to supplement an elective course
in biochemical engineering at MC.15'22'23] Other
professors have reported on UF experiments
for concentration and separation of polyeth-
ylene glycol and of protein mixtures, adapted
to the unit operations laboratory and a bio-
technology laboratory, respectively.'24'25

Ultrafiltration is one of a group of mem-
brane filtration processes that depends on pres-
sure as the driving force for separation. The
average pore size in UF membranes varies
from 300 to 300,000 Daltons (ranges vary
depending on source), which will retain mac-
romolecules such as milk proteins while pass-
ing inorganic salts and small organic mol-
ecules such as lactose (C,2H220,) through
the membrane. The key proteins in milk are
caseins, which form large micelles (10 to 300
nm), a -lactalbumin and p -lactoglobulin. The
ca-lactalbumin has a molecular weight of
18,300, while p-lactoglobulin exists as a
dimer with a molecular weight of 36,000;
therefore, a membrane with a molecular
weight cutoff (MWCO) of 10,000 is needed
to insure complete retention of milk pro-
teins.l"0 A polysulfone membrane with a
MWCO of 20,000 is generally used for milk ultrafiltration,
with an elevated operating temperature (52-54) and trans-
membrane pressures of 170-310 kPa.11'" The membrane origi-
nally used for this application was cellulose acetate.141
The molecular weight cutoff of a membrane may be de-
fined as the molecular weight that is 90% rejected by the
membrane (some manufacturers use a different percent, e.g.,
95%), which indicates that a 10,000 MWCO membrane will
reject 90% of solutes having a MW more than 10,000. Re-
jection is actually a function of the size, shape, and surface-
binding characteristics of the hydrated molecule, as well as
the pore-size distribution of the membrane; therefore, mo-
lecular weight cutoff values can be used only as a rough
guide for membrane selection.
System performance is usually defined in terms of perme-
ate flux, J, with dimensions of (volume/area-time), i.e., typi-
cal units are (L/m2-h). Flux can be determined by measuring
each incremental volume of permeate, AV, collected in time

period, At, and dividing by the effective surface area of the
membrane or the transfer area.

J= AV/At (1)
transfer area
Effective surface area for the TCF 10 thin-channel system is
less than the total membrane surface area because por-
tions of the membrane are blocked by ridges along edges
of the spiral channel. In the spiral-wound membrane mod-
ule in the Koch system, the total membrane area is avail-
able for transfer.
Theoretically, flux of a pure solvent (Jsolv) through a po-
rous membrane is directly proportional to the pressure gradi-
ent across the membrane and inversely proportional to the
membrane thickness, tm, as follows:191

A A P (2)
Jsolv =Kv (AP A,) = (2)
tm Rm
where AP is the pressure drop across the membrane (trans-
membrane pressure) and Ai is the difference in osmotic
pressure across the membrane. Osmotic pressure is rela-
tively low for macromolecular solutions (the type separated
by UF processes), so the An term can be generally neglected
in Eq. 2.[9' Osmotic pressures for whey proteins become
more significant at higher concentrations."'1 The permeabil-
ity constant, Ko.v, accounts for factors such as membrane
porosity, pore-size distribution, and viscosity of the solvent
at a given temperature. The membrane thickness, tm, is not
readily measured for an asymmetric membrane; however,
the resistance to flow through the membrane, Rm, may be
used in place of the quantity (tm/Ksov). The value of Rm can be
determined at a given temperature by running water-flux
experiments at various operating pressures.
Concentration polarization, gel formation, and fouling are
important factors that need to be considered in UF separa-
tions. As shown schematically in Figure 1, a concentration
gradient or boundary layer of increased solute concentration
forms near the membrane surface during UF. This gradient,
which is called concentration polarization, results from coun-
teracting effects of convective flow of solute towards the
membrane and diffusion of solute back toward the bulk
fluid. Concentration polarization is regarded as a reversible
boundary-layer phenomenon that causes a rapid initial drop
in flux to a steady-state value, whereas fouling is categorized
as an irreversible phenomenon that leads to a long-term flux
decline.1"0 Concentration polarization may occur with or
without gelling (Figure 1 depicts gel formation), and gel
formation may be reversible or irreversible. If the gel is
difficult to remove from the membrane (irreversible), the
membrane is said to be fouled.[9-101
In UF processes, many of the solutions being filtered form
gels, cakes, or slimes at the wall because convective trans-
port toward the membrane is relatively high compared to

diffusivities of macromolecules.191 Membrane fouling is re-
ported to be common during milk ultrafiltration.110' A gel
layer was formed in the milk experiments reported in this
paper, verified visually at the end of the TCF10 experiments
as a pale yellow spiral of solids deposited on the white
membrane surface. Gel formation was deemed to be revers-
ible (non-fouling) since a long-term decline in solvent flux
was not observed and the gel came off the membrane quite
easily when rinsed with distilled water.
As mentioned above, concentration polarization with gel
formation is observed for many UF separations. The gel
layer frequently has far more resistance to flow than the
membrane and thus controls solvent flux.191 Equation (2) can
be modified to include the effect of the gel layer by adding a
term for resistance to flow through the gel,

AP AP (3)
(t /Ksv +tg/ Kg) (Rm+ Rg)

In Equation (3), tg is the gel thickness and Kg is the gel
permeability constant, which are generally not known, and
Rg is the resistance to flow through the gel, which can be
measured experimentally. The value of Rg varies with pres-
sure, bulk concentration, and cross-flow velocity at lower
transmembrane pressures, but tends to become pressure in-
dependent at higher transmembrane pressures."10
Another approach to solving this mass-transfer problem

Figure 1. Schematic illustrating flux of permeate through
an asymmetric ultrafiltration membrane (membrane plus
support) in the presence of concentration polarization with
gel layer formation. Symbols Cb, Cg, and C, represent bulk,
gel, and permeate concentrations, respectively.

Chemical Engineering Education






starts with the equation of continuity and assumes a stagnant
film of thickness 8. The derivation, as detailed by Zeman
and Zydney,[10] results in

D fcn(Cw -C
Jsol v=- ij (4)
8o Cb l -Cp

where k is the mass-transfer coefficient, Cw is the solute
concentration at the wall, Cp is the solute concentration
in the permeate, C, is the bulk concentration, and D is the
solute diffusivity.
In cases where the gel layer controls mass transfer and
Cp=0, the wall concentration in Eq. (4) can be replaced by
the gel concentration, C., which results in a simplified equa-
tion referred to as the gel-polarization model[19'"

Jsolv =k fn

where k =

The mass-transfer coefficient, k, increases as cross-flow ve-
locity increases; the gel concentration, C,, however, is gen-
erally regarded as a constant even though different values
are reported depending on the type of equipment used.1101
Also, diffusivity changes with temperature and viscosity.
Thus, at constant temperature and cross-flow velocity, ex-
perimental data can be measured for flux as a function of
bulk concentration, then graphed to determine values for k
and C9. This method can be used to estimate an experimental
value for k from the data collected in milk experiments run
with 10,000 to 18,000 MWCO membranes since C = 0
with respect to milk proteins. When the permeate is not a
pure solvent (Cp>0), which occurs with a 30,000 MWCO
membrane, equations for solvent flux must account for Cp
and are therefore somewhat more complicated.191
The mass-transfer coefficient can also be calculated from
empirical equations.9',"10 Fully developed turbulent flow in
UF devices appears to occur at Reynolds numbers around
2,000. In the TCF10 system, fluid flows through a spiral
channel of width, w, height, 2h, and length, L; thus, the
Reynolds Number (Re) is calculated using the equivalent
diameter, deq, of the channel:
cross- sectional area 2hw (
wetted perimeter 2w + 4 h

Similar equations are used to determine the equivalent diam-
eter in the spiral-wound Koch system.
In the equations that follow, u, is the average linear veloc-
ity through the channel, D is diffusivity, p. is viscosity, and
p is density. Physical properties are based on the fluid on the
retentate/feed side of the membrane, taken at the average
concentration from the beginning to the end of each run.
Viscosities of milk solutions as a function of concentration
and temperature are available in the literature."31 For turbu-
lent flow through a channel, the following equation can be
Fall 1998


k = 0.023 D Re0.83 Sc/3 (7)

where Re = q and Sc =
u Dp

where Sc is the Schmidt number, Re is the Reynolds num-
ber, and the other terms are previously defined. Additional
empirical correlations are given in the literature to determine
mass transfer coefficients for UF systems under both turbu-
lent and laminar flow conditions.[9'101

Experimental Methods

Manhattan College System
A Millipore TCF10 thin-channel ultrafiltration system (for-
merly manufactured by Amicon) was used at Manhattan
College for UF experiments with skim milk solutions. Pho-
tographs of this system and schematics of the thin channel
are given in the literature.110'12222 The TCF10 is a 0.6-L bench-
scale system that is designed to use 0.090-m diameter flat
membranes. Pressure is supplied to the top of the feed cham-
ber from a nitrogen cylinder. A peristaltic pump that is
provided with the UF system generates high-velocity flow
across the membrane by pumping a feed solution parallel to
the membrane surface through a thin spiral channel. Cross
flow is designed to minimize concentration polarization and
gel formation and subsequently to increase permeate flux
through the membrane. According to the manufacturer, the
spiral channel has dimensions of 0.0095 m in width (w),
0.00038 m in height (2h), and 0.414 m in length (L). The
effective membrane surface area or transfer area is re-
ported to be 4.0 x 10-3 m2. Experiments were run at room
temperature (22-230C).
The TCF10 unit was operated in a batch mode where the
retentate leaving the spiral channel was recycled back to the
feed chamber, while the permeate was separated from the
feed solution and collected in a separate container. This
mode of operation causes the bulk concentration, Cb, to
increase with time. All experiments used 0.4 L of feed solu-
tion and limited permeate collection to about 10% of the
initial feed solution in order to limit changes in bulk concen-
tration. Skim milk was prepared by dissolving powdered
milk in distilled water following instructions on the package
(10.54% solids). Different milk-water feed solutions were
prepared in ratios varying from 1:0 to 1:8 by diluting skim
milk with distilled water. Although higher milk concentra-
tions would occur in the industrial process, diluted solutions
can be used to demonstrate concentration effects in a reason-
able period of time. The most time-consuming step in the
milk experiments occurs between runs when the students

have to open the system, clean the membrane, and reas-
semble the unit.
Millipore markets four membranes that can be used for the
milk experiments; with water fluxes in decreasing order,
these are PM30>YM30>PM10>YM10. The symbols are PM
for polysulfone membrane, YM for cellulose acetate mem-
brane, 30 for 30,000 MWCO, and 10 for 10,000 MWCO.
The YM membrane is treated so as to be hydrophilic with
low protein binding properties, while the PM membrane has
the advantage of high throughput. These membranes have an
asymmetric or anisotropic structure, i.e., a very thin poly-
meric skin with an extremely fine, controlled, pore structure
supported by a much thicker (and stronger), highly porous
substrate. A YM30 membrane was selected for the experi-
ments run at Manhattan College in 1996 and 1997 because
this membrane permits a number of runs to be executed at
different concentrations, cross-flow velocities, and trans-
membrane pressures during a single laboratory period. The
PM10 and YM10 membranes were also used in 1998 for
comparison with the YM30 membrane.

A Proto-Sep IV portable spiral-wound UF system manu-
factured by Koch Membrane Systems was used at Clemson
University for UF experiments. The Koch bench-scale sys-
tem includes an HFM 180 (polyvinylidene fluoride) Abcor
spiral-wound UF membrane with an 18,000 MWCO and a
nominal surface area of 0.28 m2. Experiments were run in a
batch mode with the retentate returned to a 72.5-L stainless
steel feed tank. Transmembrane pressure and tangential flow
were both supplied by a Wilden MI-Champ air-operated,
double-diaphragm pump. Dairy solutions (whole milk,
milk-water solutions, and dairy wastewater streams) were
analyzed for total solids content, pH, chemical oxygen
demand (COD), total carbohydrates, fat content, and pro-
tein content as described by Steinbeck."l61 Experiments
were run at 30 to 500C.
The Koch Proto-SEP IV system is actually better suited
for qualitative feasibility studies than for obtaining quantita-
tive data for scale-up. The standard system contains a single-
pressure gauge and flow-control valve located in the concen-
trate line. The air-actuated diaphragm feed pump is rather
noisy, and it delivers a pulsatile flow rate that is dependent
on the air pressure. The concentrate pressure and flow rate
cannot be set in a completely independent fashion, and the
concentrate pressure is not the same as the average trans-
membrane pressure that is more commonly used as a charac-
teristic operating variable. Despite these limitations, the Koch
unit can be used to demonstrate basic UF operating prin-
ciples and certain relationships with a variety of feed streams.
For example, runs can be made to determine the effect of
pressure on flux at constant composition by returning the
permeate and concentrate to the feed tank. With the perme-




0.00 0.05 010 015
Time (h)

020 025 0.30

Figure 2. Effect of volumetric flow rate through the thin
channel of a TCF10 system on flux. Data taken with a 1:8
(skim milk to water ratio) solution at 207 kPa transmem-
brane pressure and room temperature, using a YM30 mem-



E 15
E 150

0.00 0.05 0.10 0.15
Time (h)

0.25 0.30

Figure 3. Effect of solids concentration on flux in a TCF10
system with a YM30 membrane. Data recorded at 207 kPa,
35.2 L/h, and room temperature.

Chemical Engineering Education

A 35.2 L/h
S13.3 L/h

a A

A 1:8 (milk:water)
+ 1:3 (milk:water)
o 1:1 (milk:water)

A A A Aa

i-+++ + + + + ++

13 [ O 13
3 0 0 E] [ 13

ate diverted to a different collector, the effect of feed con-
centration on flux can be observed. The feed tank can be
heated or cooled to demonstrate the effects of temperature.

The first step that needs to be performed in UF, MF, or RO
membrane experiments is measurement of the flux versus
pressure behavior of the membrane when filtering pure wa-
ter. This measurement may be used to determine whether the
equipment is set up properly by comparing water fluxes with
expected values and to calculate the membrane resistance
(Rm) from Eq. (2). Water fluxes should also be calculated
between runs (after removing, rinsing, and reinstalling the
membrane) in order to make sure that the membrane is clean
enough for reuse. If water flux is lower than expected, the
asymmetric membrane may be upside down or it may need
additional cleaning. If water flux is higher than expected, a
tear or other defect in the membrane is possible. The slope of
the water flux/pressure data for the YM30 membrane in the
TCF10 system is about 1.4 L/(m2-h-kPa).
The effect of cross-flow velocity on flux in the TCF10



20 \

1 10 100
Cb (wt%)

Figure 4. Experimental determination of mass transfer
coefficient, k, and gel concentration, Cv, using a Koch
spiral-wound UF system with an Abcor 18,000 MWCO
membrane. Data recorded at 500C, 1400 L/h, and 67 kPa
transmembrane pressure by Steinbeck.'116

system is graphed in Figure 2. As shown, flux of pure water
is relatively high and constant with time. Prior experiments
demonstrate that the characteristics of the flux curves are
quite different for the YM30 and PM30 membranes.123J Un-
der the same operating conditions, the flux of the polysulfone
membrane (PM30) drops gradually throughout the experi-
ment, while the flux of the cellulostic membrane (YM30)
drops immediately to a steady-state value and remains rela-
tively constant thereafter. 231 This observation can be attrib-
uted to the fact that the YM30 membrane is treated to mini-
mize protein adsorption and thus minimize fouling. At the
lower cross-flow velocity (Figure 2), a few high-flux points
are observed in the first few seconds of the run as the flux
drops due to concentration polarization and gel formation,
but the flux remains stable and constant thereafter, indicat-
ing that the gel, once formed, controls flux. At the higher
cross-flow velocity (Figure 2), the transition period is very
rapid, so that the flux appears to drop immediately to the
steady-state value. As expected, a higher cross-flow ve-
locity significantly increases solvent flux by improving
mass-transfer conditions at the membrane surface. Since
Rm was previously calculated from the water-flux data,
values of Rg for both the high and low cross-flow cases
can be determined using Eq. (3).
The students were also asked to determine the effect of
milk concentration on flux. Typical data are graphed in
Figure 3 for a YM30 membrane in the Millipore system. As
in the previous figure, milk fluxes (at high cross-flow veloc-
ity) drop immediately from pure-water values to steady-state
values. Comparing trials with milk-water ratios increasing
from 1:8 up to 1:0 (undiluted skim milk), it is evident that
flux decreases as concentration increases. The solvent fluxes
remain relatively constant at the steady-state values in Fig-
ures 2 and 3; specifically, there is no evidence of irreversible
fouling and its associated long-term flux decline. Since a
30,000 MWCO membrane was used for the data in Figure 3,
the milk concentration in the permeate is greater than zero,
which must be taken into account in determining experimen-
tal values of k and Cg.
Using data from the Abcor 18,000 MWCO membrane,"[6'
a meaningful graph of Eq. (5) can be generated as shown in
Figure 4, although there is room for speculation on the
location of the "best" straight line and the resulting slope and
intercept, or k and Cg values. Based on reports that the
maximum concentration levels are seven-fold for skim milk
and five-fold for whole milk (or about 65% solids),"" the
value of Cg determined in Figure 4 appears to be reasonable.
As evidenced in Figure 4, additional data points are needed
at higher bulk concentrations (closer to Cg) if more accurate
experimental values for k and Cg are desired.
After graphing and analyzing the data from the milk UF
experiments, the students were asked to calculate a mass
transfer coefficient using Eq. (7) or other appropriate em-

Fall 1998

pirical equations. Finally, they were expected to compare the
calculated coefficient with the experimental coefficient ob-
tained from the graphical analysis. For a more theoretical
approach, the students could be asked to derive Eq. (4) from
the equation of continuity. Based on three years of MC
student data, milk-flux data generally followed the expected
trends, but agreement between experimental and calculated
values for the mass-transfer coefficient was highly variable.
This variability results from the fact that high milk con-
centrations were not run, which in turn leads to errors in
the experimental values of k and from the approximate
nature of the empirical equation. Higher concentrations
would require more time for individual runs, but should
improve accuracy. It is questionable whether the extra
time per experiment is justified to tie down a single
number, because UF principles can be demonstrated quali-
tatively using lower concentrations, thus allowing time
to study other operating conditions.

Experiments involving ultrafiltration of dairy products have
been developed at Clemson University and Manhattan Col-
lege and tested in lecture and laboratory courses. These
experiments appeal to the students (like the "got milk?"
advertisement) and are an effective method of introducing
membrane separation processes into the chemical engineer-
ing curriculum. Whether UF experiments are run in an open-
ended or structured format, better results are generally ob-
tained when a "resident expert" who is familiar with mem-
brane processes is available to the students for consultation.
Based on the authors' experiences, learning is enhanced if
the experiments are used to enrich lecture courses where the
instructor can close the feedback loop with classroom dis-
cussion. On the other hand, respectable results were ob-
tained by students in the required senior laboratory course
when neither the instructor nor the teaching assistant claimed
any special expertise in membrane separations. The small
Millipore TCF10 system seems to be more user friendly than
the larger Koch Proto-Sep system; however, both systems
have been used effectively to demonstrate membrane sepa-
ration principles.


1. Griffith, J.D., "The Teaching of Undergraduate Mass Trans-
fer," AIChE Annual Meeting, paper 245a, Miami Beach, FL
2. Wankat, P.C., R.P. Hesketh, K.H. Schulz, and C.S. Slater,
"Separations: What to Teach Undergraduates," Chem. Eng.
Ed., 28, 12 (1994)
3. Slater, C.S. "Education on Membrane Science and Technol-
ogy," in Membrane Processes in Separation and Purifica-
tion, J.G. Crespo and K.W. Boddeker, eds., Kluwer Aca-
demic Publishers, Dordrecht, The Netherlands, p 479 (1994)
4. Slater, C.S., "Teaching the Engineering Aspects of Mem-
brane Process Technology," J. Membr. Sci., 62, 239 (1991)

5. Slater, C.S., and H.C. Hollein, "Educational Initiatives in
Teaching Membrane Technology," Desalination, 90, 291
6. Gooding, C.H., "A Course on Membrane Separation Pro-
cesses," AIChE Topical Conference on Separations Technol-
ogy, Vol. II, Miami Beach, FL, 515 (1995)
7. Slater, C.S., "A Graduate Course on Membrane Technol-
ogy," Int. J. ofEng. Ed., 8, 211 (1992)
8. Slater, C.S., "Instruction in Membrane Separation Pro-
cesses," J. Membr. Sci., 44, 265 (1989)
9. Wankat, P.C., Rate-Controlled Separations, Chapman and
Hall, New York, NY; Chap. 12-13 (1990)
10. Zeman, L.J., and A.L. Zydney, Microfiltration and
Untrafiltration, Marcel Dekker, New York, NY; Chaps 5-10,
12 (1996)
11. Kosikowski, F.V., "Membrane Separations in Food Process-
ing," in Membrane Separations in Biotechnology, W.C.
McGregor, ed., Marcel Dekker, New York, NY; Chap. 9
12. Cheryan, M. Ultrafiltration Handbook, Technometric,
Lancaster, PA; Chap. 8 (1986)
13. Maubois, J.-L., "Recent Developments of Membrane Ultra-
filtration in the Dairy Industry," in Ultrafiltration Mem-
branes and Applications, A.R. Cooper, ed., Plenum Press,
New York, NY; 305 (1980)
14. Harper, W.J., "Factors Affecting the Application of Ultrafil-
tration Membranes in the Dairy Food Industry," in Ultrafil-
tration Membranes and Applications, A.R. Cooper, ed., Ple-
num Press, New York, NY; 321 (1980)
15. Garcia III, A., B. Medina, N. Verhoek, and P. Moore, "Ice
Cream Components Prepared with Ultrafiltration and Re-
verse Osmosis Membranes," Biotech. Prog., 5, 46 (1989)
16. Steinbeck, M. "Reduction of Dairy Plant Waste Water
Strength by Ultrafiltration," MS Thesis, Clemson Univer-
sity, Clemson, SC (1993)
17. Slater, C.S., and J.D. Paccione, "A Reverse Osmosis System
for an Advanced Separation Process Laboratory," Chem.
Eng. Ed., 21, 138 (1987)
18. Slater, C.S., C. Vega, and M. Boegel, "Expriments in Gas
Permeation Membrane Processes," Int. J. of Eng. Ed., 7,
368 (1992)
19. Hollein, H.C., C.S. Slater, R.L. D'Aquino, and A.L. Witt,
"Bioseparation via Cross-Flow Membrane Filtration," Chem.
Eng. Ed., 29, 86 (1995)
20. Slater, C.S.,"A Manually Operated Reverse Osmosis Ex-
periment,"Int. J. ofEng. Ed., 10, 195 (1994)
21. Hesketh, R.P., and C.S. Slater, "Demonstration of Chemical
Engineering Principles to a Multidisciplinary Engineering
Audience," Ann. Conf. Proc. of ASEE, Milwaukee, WI; Ses-
sion 2513 (1997)
22. Slater, C.S., H.C. Hollein, P.P. Antonecchia, L.S. Mazzella,
and J.D. Paccione, "Laboratory Experiences in Membrane
Separation Processes," Int. J. of Eng. Ed., 5, 369 (1989)
23. Slater, C.S., P.P. Antonecchia, L.S. Mazzella, and H.C.
Hollein, "Applcation of Ultrafiltration to the Purification
and Concentration of Bovine Intestinal Alkaline Phosphatase
(BIAP) Enzyme," in Chemical Separations: Vol. II, Applica-
tions, C.J. King and J.D. Navratil, eds., Litharvan Litera-
ture, Denver, CO; 25 (1986)
24. Waters, J., R.D. Averette, Jr., and M. Fleishman, "Ultrafil-
tration Experiments for the Unit Operations Laboratory,"
Ann. Conf. Proc. ofASEE, Washington, DC; 178 (1984)
25. Davis, R.H., and D. S. Kompala, "Biotechnology Laboratory
Methods," Chem. Eng. Ed., 23, 182 (1989) p
Chemical Engineering Education

Graduate Education in Chemical Engineering

aft Faculty and Research Areas


of slty


Teaching and research
as well as
industrially sponsored fellowships
up to $17,000.

In addition to stipends,
tuition and fees are waived.

PhD students may get
some incentive scholarships.

The deadline for assistantship applications is
March 15th.

Digital Control, Mass Transfer, Multicomponent

Multiphase Processes, Fluid Flow, Interfacial Phenomena,
Filtration, Coalescence

Nanocomposite Materials. Sonochemical Processing,
Polymerization in Nanostructured Fluids, Supercritical
Fluid Processing

Catalysis, Reaction Engineering, Environmentally Benign

Molecular Simulation. Phase Behavior, Physical
Properties, Process Modeling

Materials Processing and CVD Modeling

Fixed Bed Adsorption, Process Design

Blochemical Engineering, Environmental Biotechnolog)

L. K. JU
Biochemical Engineering. Environmental Bioengineering

BioMaterial Engineering and Polymer Engineering

Nonlinear Control, Chaotic Processes


Professor Emeritus 2 Adjunct Faculty Member

For Additional Information, Write *

Chairman, Graduate Committee
Department of Chemical Engineering The University of Akron Akron, OH 44325-3906

Fall 1998 32

Chemical Engineering

at the




A dedicated faculty with state-of-the-art facilities
offer research programs leading to Master of
Science and Doctor of Philosophy degrees.

Research Interests:

Biomass Conversion, Catalysis and Reactor
Design, Energy Conversion Processes,
Environmental Studies, Interfacial Transport,
Magnetic Storage Media, Mass Transfer, Metal
Casting, Polymer Rheology, Process Dynamics
and Control, Reservoir Modeling, Suspension
and Slurry Rheology, Thermodynamics,
Transport Process Modeling

For Information Contact:
Director of Graduate Studies
Department of Chemical Engineering
The University of Alabama
Box 870203
Tuscaloosa, AL 35487-0203
An equal employ
Phone: (205) 348-6450 opportur

ment/equal e
lity institution

G.C. April, Ph.D. (Louisiana State)
D. W. Arnold, Ph.D. (Purdue)
E. S. Carlson, Ph.D. (Wyoming)
P. E. Clark, Ph.D. (Oklahoma State)
W. C. Clements, Jr., Ph.D. (Vanderbilt)
R. W. Flumerfelt, Ph.D. (Northwestern)
R. A. Griffin, Ph.D. (Utah State)
I. A. Jefcoat, Ph.D. (Clemson)
P. W. Johnson, Ph.D. (New Mexico Tech.)
A. M. Lane, Ph.D. (Massachusetts)
M. D. McKinley, Ph.D. (Florida)
SR. G. Reddy, Ph.D. (Utah)
L. Y. Sadler III, Ph.D. (Alabama)
V. N. Schrodt, Ph.D. (Penn. State)
.J. M. Wiest, Ph.D. (Wisconsin)
Chemical Engineering Education


The University of Alberta is well
known for its commitment to excel-
lence in teaching and research. The
Department of Chemical and Materi-
als Engineering has 34 professors and
over 100 graduate students. Degrees
are offered at the M.Sc. and Ph.D.
levels in Chemical Engineering, Ma-
terials Engineering, and Process
Control. All full-time graduate stu-
dents in the research programs re-
ceive a stipend to cover living ex-
penses and tuition.

For further information, contact
Graduate Program Officer WCM
Department of Chemical and Materials Engineering
University of Alberta
Edmonton, Alberta, Canada T6G 2G6
PHONE (403) 492-5805 FAX (403) 492-2881
Fall 1998

P. CHOI, Ph.D. (University of Waterloo)
Statistical Mechanics of Polymers Polymer Solutions and Blends
K. T. CHUANG, Ph.D. (University of Alberta)
Mass Transfer Catalysis Separation Processes Pollution Control
I. G. DALLA LANA, Ph.D. (Univ. of Minnesota) EMERITUS
Chemical Reaction Engineering Heterogeneous Catalysis
JA.W. ELLIOTT, Ph.D. (University of Toronto)
Thermodynamics Statistical Thermodynamics Interfacial Phenomena
D. G. FISHER, Ph.D. (University of Michigan) EMERITUS
Process Dynamics and Control Real-Time Computer Applications
J.F. FORBES, Ph.D. (McMaster University)
Real-Time Optimization Control of Sheet Forming Processes
M. R. GRAY, Ph.D. (California Inst. of Tech.) DEAN OF GRADUATE STUDIES
Bioreactors Chemical Kinetics Bitumen Processing
M. GUAY, Ph.D. (Queens University)
Nonlinear Process Control Statistical Modeling Multivariate Statistics
R. E. HAYES, Ph.D. (University of Bath)
Numerical Analysis Reactor Modeling Computational Fluid Dynamics
B. HUANG, Ph.D. (University of Alberta)
Controller Performance Assessment Multivariable Control Statistics
S. M. KRESTA, Ph.D. (McMaster University)
Turbulent & Transitional Flows Multiphase Flows CFD
S. LIU, Ph.D. (University of Alberta)
Fluid-Particle Dynamics Transport Phenomena Mass Transfer
D. T. LYNCH, Ph.D. (University of Alberta) DEAN OF ENGINEERING
Catalysis Kinetic Modeling Numerical Methods Polymerization
J. H. MASLIYAH, Ph.D. (University of British Columbia)
Transport Phenomena Colloids Particle-Fluid Dynamics Oil Sands
A. E. MATHER, Ph.D. (University of Michigan)
Phase Equilibria Fluid Properties at High Pressures Thermodynamics
W. C. MCCAFFREY, Ph.D. (McGill University)
Reaction Kinetics Heavy Oil Upgrading Polymer Recycling Biotechnology
P. A. J. MEES, Ph.D. (University of Alberta)
Computational Fluid Dynamics Transport Phenomena Pulp and Paper
K. NANDAKUMAR, Ph.D. (Princeton University)
Transport Phenomena Distillation Computational Fluid Dynamics
F. D. OTTO, Ph.D. (University of Michigan) EMERITIS
Mass Transfer Gas-Liquid Reactions Separation Processes
M. RAO, Ph.D. (Rutgers University)
AI Intelligent Control Process Control
S. L. SHAH, Ph.D. (University of Alberta)
Computer Process Control System Identification Adaptive Control
S. E. WANKE, Ph.D. (University of California, Davis) CHAIR
Heterogeneous Catalysis Kinetics Polymerization
M. C. WILLIAMS, Ph.D. (University of Wisconsin)
Rheology Polymer Characterization Polymer Processing
Z. XU, Ph.D. (Virginia Polytechnic Institute and State University)
Surface Science & Engineering Mineral Processing Waste Management


ROBERT ARNOLD, Professor (Caltech)
Microbiological Hazardous Waste Treatment, Metals Speciation and T
JAMES BAYGENTS, Associate Professor (Princeton)
Fluid Mechanics, Transport and Colloidal Phenomena, Bioseparations
MILAN BIER, Professor Emeritus (Fordham)
Protein Separation, Electrophoresis, Membrane Transport
WILLIAM P. COSART, Associate Professor and Associate Dean (Oregi
Heat Transfer in Biological Systems, Blood Processing
WENDELL ELA, Assistant Professor (Stanford)
Particle-Particle Interactions, Environmental Chemistry
JAMES FARRELL, Assistant Professor (Stanford)
Sorption/desorbtion of Organics in Soils
EDWARD FREEH, Adjunct Professor (Ohio State)
Process Control, Computer Applications
JOSEPH GROSS, Professor Emeritus (Purdue)
Boundary Layer Theory, Pharmacokinetics, Microcirculation, BiorheoJ
ROBERTO GUZMAN, Associate Professor (North Carolina State)
Protein Separation, Affinity Methods
ARTHUR HUMPHREY, Visiting Professor (Columbia)
ANTHONY MUSCAT, Assistant Professor (Stanford)
Kinetics, Surface Chemistry, Surface Engineering, Semiconductor Proc
KIMBERLY OGDEN, Associate Professor (Colorado)
Bioreactors, Bioremediation, Organics Removal from Soils
THOMAS W. PETERSON, Professor and Head (CalTech)
Aerosols, Hazardous Waste Incineration, Microcontamination
ALAN D. RANDOLPH, Professor Emeritus (Iowa State)
Crystallization Processes, Nucleation, Particulate Processes
THOMAS R. REHM, Professor Emeritus (Washington)
Mass Transfer, Process Instrumentation, Computer Aided Design
EDUARDO SAEZ, Associate Professor (UC, Davis)
Rheology, Polymer Flows
FARHANG SHADMAN, Professor (Berkeley)
Reaction Engineering, Kinetics, Catalysis, Reactive Membranes,
RAYMOND A. SIERKA, Professor Emeritus (Oklahoma)
Adsorption, Oxidation, Membranes, Solar Catalyzed Detox Reactions
JOST 0. L. WENDT, Professor (Johns Hopkins)
Combustion-Generated Air Pollution, Incineration, Waste
DON H. WHITE, Professor Emeritus (Iowa State)
Polymers, Microbial and Enzymatic Processes
DAVID WOLF, Visiting Professor (Technion)
Fermentation, Mixing, Energy, Biomass Conversion

For further information, write to

Graduate Study Committee
Department of Chemical and
Environmental Engineering
University of Arizona
Tucson, Arizona 85721
The University of Arizona is an equal
opportunity educational institution/equal
opportunity employer.
Women and minorities are encouraged
to apply.









The Chemical and Environmental Engineering Department
at the University of Arizona offers a wide range of research
opportunities in all major areas of chemical engineering and
environmental engineering, and graduate courses are offered in
most of the research areas listed here. The department offers a fully
accredited undergraduate degree as well as MS and PhD graduate
degrees. Strong interdisciplinary programs exist in bioprocessing
and bioseparations, microcontamination in electronics manu-
facture, and environmental process modification.
Financial support is available through fellowships, government
and industrial grants and contracts, teaching and
research assistantships.
Tucson has an excellent climate and many

recreational opportunities. It is a growing modern city of
450,000 that retains much of the old Southwestern atmosphere.

Chemical Engineering Education





Beaudoin, Stephen P., Ph.D., North Carolina State *
University Transport Phenomena and Surface Science o *
concerning Pollution Prevention, Waste Minimization, and
Pollution Remediation V 1
Beckman, James R., Ph.D., University of Arizona Crystalli- a
zation and Solar Cooling *
Bellamy, Lynn, Ph.D., Tulane Process Simulation
Berman, Neil S., Ph.D., University of Texas, Austin Fluid
Dynamics and Air Pollution
Burrows, Veronica A., Ph.D., Princeton University Surface
Science, Semiconductor Processing
Garcia, Antonio A., Ph.D., U.C., Berkeley Acid-Base Interactions,
Biochemical Separation, Colloid Chemistry
Kuester, James L., Ph.D., Texas A&M University Thermochemical
Conversion, Complex Reaction Systems
Raupp, Gregory B., Ph.D., University of Wisconsin Semiconductor Materials
Processing, Surface Science, Catalysis
Rivera, Daniel, Ph.D., Cal Tech Process Control and Design
Sater, Vernon E., Ph.D., Illinois Institute of Tech Heavy Metal Removal from Waste
Water, Process Control
Torrest, Robert S., Ph.D., University of Minnesota Multiphase Flow, Filtration, Flow in
Porous Media, Pollution Control


0 0
0 *
* AR'c'o,4 6

6II s"80"

r.*o oO-.o,,.Am



Research in a

High Technology


Dorson, William J., Ph.D., University of Cincinnati Physicochemical Phenomena, Transport Processes
Guilbeau, Eric J., Ph.D., Louisiana Tech University Biosensors, Physiological Systems, Biomaterials
He, Jiping, Ph.D., University of Maryland Biomechanics, Robotics, Computational Neuroscience, Optimal Control, System Dynamics and Control
Kipke, Daryl R., Ph.D., University of Michigan Computation Neuroscience Machine Vision, Speech Recognition, Robotics Neural Networks
Pizziconi, Vincent B., Ph.D. Arizona State University. Artificial Organs, Biomaterials, Bioseparations
Sweeney, James D., Ph.D., Case-Western Reserve University- Rehab Engineering, Applied Neural Control
Towe, Bruce C., Ph.D., Pennsylvania State University- Bioelectric Phenomena, Biosensors, Biomedical Imaging
Yamaguchi, Gary T., Ph.D., Stanford University Biomechanics, Rehab Engineering, Computer-Aided Surgery

Adams, James, Ph.D., University of Wisconsin, Madison Atomistic Simulation of Metallic Surfaces Grain Boundaries Automobile Catalysts *
Polymer-Metal Adhesion
Alford, Terry L., Ph.D., Cornell University Electronic Materials Physical Metallurgy Electronic Thin Films Surface/Thin Film
Dey, Sandwip K., Ph.D., NYSC of Ceramics, Alfred University Ceramics, Sol-Gel Processing
Krause, Stephen L., Ph.D., University of Michigan Ordered Polymers, Electronic Materials, Electron X-ray Diffraction, Electron Microscopy
Mahajan, Subhash, Ph.D., University of Michigan Semiconductor Defects, Structural Materials Deformation
Mayer, James, Ph.D., Purdue University -Thin Film Processing Ion Bean Modification of Materials
Stanley, James T., Ph.D., University of Illinois Phase Transformations, Corrosion

pla al(02 6-33ori1iet:Dr rc6ibeo hi lth rdae( nmteDeateto heiaBl, n ltra


Fall 1998




Robert P. Chambers University of California, Berkeley
Harry T. Cullinan Carnegie Mellon University
Christine W. Curtis Florida State University
Steve R. Duke University of Illinois
Mahmoud El-Halwagi University of California, Los Angeles
James A. Guin University of Texas, Austin
Ram B. Gupta University of Texas, Austin
Gopal A. Krishnagopalan University of Maine
Jay H. Lee California Institute of Technology
Y. Y. Lee Iowa State University
Glennon Maples Oklahoma State University
Ronald D. Neuman The Institute of Paper Chemistry
Stephen A. Perusich University of Illinois
Timothy D. Placek University of Kentucky
Christopher B. Roberts University of Notre Dame
A. R. Tarrer Pardu, UL ii t, /n
Bruce J. Talarchuk t ii i, ,tn at it,, i ,won

's "

.- .., .,}o

it, .' ;

-' '*. '* .- '!._.. -_ L -

r $ -

AjdbjrntU ngriiys~h qri~&k01oi

Biochemical Engineering Biotechnology
Pulp and Paper Process Control
Catalysis and Reaction Engineering
Computer Aided Process Synthesis,
Optimization and Design
Environmental Chemical Engineering
Pollution Prevention Recycling
Materials Polymers Surface Science
Colloid and Interfacial Phenomena
Thermodynamics Supercritical Fluids
Separation Electrochemical Engineering
Fluid Dynamics and Transport Phenomena
SFuels and Energy

6dn u
:__ '-_ : .. .' -- -_- -

.... ... --. .

.- -w

-_ _

-ra _,-.-z r-- .~on

_ U

c- --

~ :_':
tIY~. .
t' :ljy
K;_ --L-L1~ i_

I .~~~ ~*
, -- i.

: ,:




R. G. Moore, Head (Alberta)
J. Azaiez (Stanford)
H. Baheri (Saskatchewan)
L. A. Behie (Western Ontario)
C. Bellehumeur (McMaster)
P. R. Bishnoi (Alberta)
R. A. Heidemann (Washington U.)
C. Hyndman (Ecole Polytechnique)
A. A. Jeje (MIT)
A. Kantzas (Waterloo)
A. K. Mehrotra (Calgary)
S. R. Mehta (Calgary)
B. J. Milne (Calgary)
M. Pooladi-Darvish (Alberta)
B. B. Pruden (McGill)
W. Y. Svrcek (Alberta)
M. A. Trebble (Calgary)
H. W. Yarranton (Alberta)
B. Young (Canterbury, NZ)
L. Zanzotto (Slovak Tech. Univ., Czechoslovakia)

The Department offers graduate programs leading to the M.Sc. and Ph.D.
degrees in Chemical Engineering (full-time) and the M.Eng. degree in Chemical
Engineering, Petroleum Reservoir Engineering or Engineering for the
Environment (part-time) in thefollowing areas:
Biochemical Engineering & Biotechnology
Biomedical Engineering
Environmental Engineering
Modeling, Simulation & Control
Petroleum Recovery & Reservoir Engineering
Polymer Processing & Rheology
Process Development
Reaction Engineering/Kinetics
Transport Phenomena
Fellowships and Research Assistantships are available to all qualified applicants.

SFor Additional Information Write *
Dr. A. K. Mehrotra Chair, Graduate Studies Committee
Department of Chemical and Petroleum Engineering
The University of Calgary Calgary, Alberta, Canada T2N 1N4

The University is located in the City of Calgary, the Oil capital of Canada, the home of the world famous Calgary Stampede and the
1988 Winter Olympics. The City combines the traditions of the Old West with the sophistication of a modern urban center. Beautiful
Banff National Park is 110 km west of the City and the ski resorts of Banff, Lake Louise,and Kananaskis areas are readily accessible. In
the above photo the University Campus is shown with the Olympic Oval and the student residences in the foreground. The Engineering
complex is on the left of the picture.



Fall 1998







. offers graduate programs leading to the
Master of Science and Doctor of Philosophy.
Both programs involve joint faculty-student
research as well as courses and seminars within
and outside the department. Students have the
opportunity to take part in the many cultural
offerings of the San Francisco Bay Area and
the recreational activities of California's north-
ern coast and mountains.

Biochemical Engineering
Electrochemical Engineering
Electronic Materials Processing
Energy Utilization
Fluid Mechanics
Kinetics and Catalysis
Polymer Science and Technology
Process Design and Development
Separation Processes
Surface and Colloid Science













332 Chemical Engineering Education

University of California Davis

Department of Chemical Engineering & Materials Science
Offering M.S. and Ph.D. degree programs in both Chemical Engineering and Materials Science and Engineering


David E. Block, Assistant Professor Ph.D., University of Minnesota, 1992 Industrialfennentation, biochemical processes in pharmaceutical
Roger B. Boulton, Professor Ph.D., University of Melbourne, 1976 Fennentation and reaction kinetics, crystallization
Stephanie R. Dungan, Associate Professor Ph.D., Massachusetts Institute of Technology, 1992 Micelle transport, colloid and interfacial
science in food processing
Bruce C. Gates, Professor Ph.D., University of Washington, Seattle, 1966 Catalysis, solid superacid catalysis, zeolite catalysts, bimetallic
catalysts, catalysis by metal clusters
Jeffery C. Gibeling, Professor Ph.D.. Stanford University. 1979 Defonnation and fatigue of metals and metal matrix composites
Joanna R. Groza, Professor Ph.D., Polytechnic Institute. Bucharest, 1972 Plasma activated sintering and processing of nanostructured
Brian G. Higgins, Professor Ph.D., University of Minnesota, 1980 Fluid mechanics and interfacialphenomena, sol gelprocessing, coatingflows
David G. Howitt, Professor Ph.D.. University of California, Berkeley, 1976 Forensic and failure analysis, electron microscopy, ignition and
combustion processes in materials
Alan P. Jackman, Professor Ph.D., University of Minnesota, 1968 Protein production in plant cell cultures, bioremediation
Marjorie L. Longo, Assistant Professor Ph.D., University of California, Santa Barbara, 1993 Hydrophobic protein design for active control,
surfactant microstructure, and interaction ofproteins and DNA with biological membranes
Benjamin J. McCoy, Professor* Ph.D., University of Minnesota, 1967 Supercritical extraction, pollutant transport
Karen A. McDonald, Professor *Ph.D., University of Maryland, College Park, 1985 Plant cell culture bioprocessing algal cell cultures
Amiya K. Mukherjee, Professor D.Phil.. University of Oxford, 1962 Superplasticity of intermetallic alloys and ceramics, high temperature
creep deformation
Zuhair A. Munir, Professor Ph.D., University of California, Berkeley, 1963 Combustion synthesis, multilayer combustion systems, functionally
gradient materials
Alexandra Navrotsky, Professor Ph.D., University of Chicago, 1967 Thermodynamics and solid state chemistry; high temperature calorimetry
Ahmet N. Palazogln, Professor Ph.D., Rensselaer Polytechnic Institute, 1984 Process control and process design of environmentally benign
Ronald J. Phillips, Associate Professor Ph.D., Massachusetts Institute of Technology, 1989 Transport processes in bioseparations, Newtonian
and non-Newtonian suspension mechanics
Robert L. Powell, Professor Ph.D., Johns Hopkins University, 1978 Rheology, suspension mechanics, magnetic resonance imaging of
Subhash H. Risbud, Professor and Chair Ph.D., University of California, Berkeley, 1976 Semiconductor quantum dots, high T, superconducting
ceramics, polymer composites for optics
Dewey D.Y. Ryu, Professor Ph.D., Massachusetts Institute of Technology, 1967 Biomolecular process engineering and recombinant bioprocess
James F. Shackelford, Professor Ph.D., University of California, Berkeley, 1971 Structure of materials, biomaterials, nondestructive testing of
engineering materials
J.M. Smith, Professor Emeritus Sc.D., Massachusetts Institute of Technology, 1943 Chemical kinetics and reactor design
Pieter Stroeve, Professor Sc.D., Massachusetts Institute of Technology, 1973 Membrane separations, Langmuir Blodgettfilms, colloid and
surface science
Stenhen Whitaker. Professor Ph.D.. University of Delaware, 1959 Multiphase transport phenomena

Scrment: 17 mles
San Frano: 72 miles
Lak Tahoe 90 milrs

Davis is a small, bike-friendly university town
located 17 miles west of Sacramento and 72 miles
northeast of San Francisco, within driving dis-
tance of a multitude of recreational activities in
Yosemite, Lake Tahoe, Monterey, and San Fran-
cisco Bay Area.

For information about our program, look up our web
site at
or contact us via e-mail at
On-line applications may be submitted via

Graduate Admission Chair
Professor Ronald J. Phillips
Department of Chemical Engineering & Materials Science
University of California, Davis
Davis, CA 95616-5294, USA
Phone (530) 752-2803 Fax (530) 752-1031

Fall 1998

The multifaceted graduate study experience in the Department
of Chemical Engineering and Materials Science allows students to
choose research projects and thesis advisors from any of our faculty
with expertise in chemical engineering and/or materials science and
Our department faculty provide excellent access to the scientists
and facilities at nearby National Laboratories (LBL and LLNL) and
industry in the Silicon Valley and San Francisco Bay Area,




Graduate Studies in
Chemical and Biochemical Engineering
Materials Science and Engineering
for Chemical Engineering, Engineering, and Science Majors

Offering degrees at the M.S. and Ph.D. levels. Research in frontier areas
in chemical engineering, biochemical engineering, biotechnology and materials
science and engineering. Strong physical and life science and engineering groups on campus.

Nancy A. Da Silva (California Institute of Technology)
James C. Earthman (Stanford University)
Steven C. George (University of Washington)
Juan Hong (Purdue University)
Enrique J. Lavernia (Massachusetts Institute of Technology)
Henry C. Lim (Northwestern University)
Martha L. Mecartney (Stanford University)
Farghalli A. Mohamed (University of California, Berkeley)
Frank G. Shi (California Institute of Technology)
Joint Appointments:
G. Wesley Hatfield (Purdue University)
Roger H. Rangel (University of California, Berkeley)
William A. Sirignano (Princeton University)

The 1,510-acre UC Irvine campus is in Orange County, five miles from the Pacific Ocean and
40 miles south of Los Angeles. Irvine is one of the nation's fastest growing residential,
industrial, and business areas. Nearby beaches, mountain and desert area recreational
activities, and local cultural activities make Irvine a pleasant city in which to live and study.

For further information and application forms, please visit
or contact

Department of Chemical and Biochemical Engineering and Materials Science
School of Engineering University of California
Irvine, CA 92697-2575

* Biomedical
* Bioreactor
* Bioremediation
* Ceramics
* Combustion
* Composite
* Control and
* Environmental
* Interfacial
* Materials
* Mechanical
* Metabolic
* Microelectronics
Processing and
* Microstructure
of Materials
* Nanocrystalline
* Nucleation,
and Glass
* Polymers
* Recombinant
CeU Technol-
* Separation
* Sol-Gel Process-
* Two-Phase
* Water Pollution

chemicall Engineering Education





* Molecular Simulations
* Thermodynamics and
* Process Design, Dynamics, and
* Polymer Processing and Transport
* Kinetics, Combustion, and
* Surface and Interface Engineering
* Electrochemistry and Corrosion
* Biochemical Engineering
* Aerosol Science and
* Air Pollution Control and Environ
mental Engineering

Jane P. Chang
Panagiotis D. Christofides
Y. Cohen
M. W. Deem
T. H. K. Frederking
S. K. Friedlander
R. F. Hicks
E. L. Knuth
(Prof Emeritus)
James C. Liao
V. Manousiouthakis
H. G. Monbouquette
K. Nobe
L. B. Robinson
(Prof. Emeritus)
S. M. Senkan
W. D. Van Vorst
(Prof Emeritus)
V. L. Vilker
(Prof. Emeritus)
A. R. Wazzan


UCLA's Chemical Engineering Department offers a pro-
gram of teaching and research linking fundamental engineer-
ing science and industrial practice. Our Department has strong
graduate research programs in environmental chemical engi-
neering, biotechnology, and materials processing. With the
support of the Parsons Foundation, The National Science
Foundation, and the U.S. Department of Education, we are
pioneering the development of methods for the design of
clean chemical technologies, both in graduate research and

engineering education.
Fellowships are available for outstanding applicants in
both M.S. and Ph.D. degree programs. A fellowship in-
cludes a waiver of tuition and fees plus a stipend.
Located five miles from the Pacific Coast, UCLA's
attractive 417-acre campus extends from Bel Air to
Westwood Village. Students have access to the highly
regarded science programs and to a variety of experiences
in theatre, music, art, and sports on campus.


Fall 1998 335



L. GARY LEAL Ph.D. (Stanford) Fluid Mechanics, Physics and Rheology of Complex Fluids, including Polymers, Suspensions, and
ERAY S. AYDIL Ph.D. (University of Houston) Microelectronics and Plasma Processing.
SANJOY BANERJEE Ph.D. (Waterloo) Environmental Fluid Dynamics, Multiphase Flows, Turbulence, Computational Fluid Dynamics.
BRADLEY F. CHMELKA Ph.D. (U.C. Berkeley) Inorganic-Organic Hybrid Materials, Zeolites and Molecular Sieves, Polymeric Solids, Liquid
Crystals, Solid-State NMR.
GLENN H. FREDRICKSON Ph.D. (Stanford) (Chair) Statistical Mechanics, Glasses, Polymers, Composites, Alloys.
JACOB ISRAELACHVILI Ph.D. (Cambridge) (Vice-Chair) Surface and Interfacial Phenomena, Adhesion, Colloidal Systems, Surface
Forces, Biomolecular Interactions, Friction.
EDWARD J. KRAMER Ph.D. (Carnegie-Mellon) Microscopic Fundamentals of Fracture of Polymers, Diffusion in Polymers, Polymer
Surfaces and Interfaces.
FRED F. LANGE Ph.D. (Penn State) Powder Processing of Composite Ceramics, Liquid Precursors for Ceramics, Superconducting Oxides.
GLENN E. LUCAS Ph.D. (M.I.T.) Mechanics of Materials, Structural Reliability.
DIMITRIOS MAROUDAS Ph.D. (M.I.T.) Theoretical and Computational Materials Science, Microstructure Evolution in Electronic and
Structural Materials.
ERIC McFARLAND Ph.D. (M.I.T.) M.D. (Harvard) Biomedical Engineering, NMR and Neutron Imaging, Transport Phenomena in Complex
Liquids, Radiation Interactions.
DUNCAN A. MELLICHAMP Ph.D. (Purdue) Computer Control, Process Dynamics, Real-Time Computing.
PHILIP A. PINCUS Ph.D. (U.C. Berkeley) Theory of Surfactant Aggregates, Colloid Systems.
DAVID J. PINE Ph.D. (Cornell) Polymer, Surfactant, and Colloidal Physics, Multiple Light Scattering, Photonic Crystals, Macroporous
ORVILLE C. SANDALL Ph.D. (U.C. Berkeley) Transport Phenomena, Separation Processes.
DALE E. SEBORG Ph.D. (Princeton) Process Control, Monitoring and Identification.
T. G. THEOFANOUS Ph.D. (Minnesota) Multiphase Flow, Risk Assessment and Management
W. HENRY WEINBERG Ph.D. (U.C. Berkeley) Surface Chemistry, Heterogeneous Catalysis, Electronic Materials, Materials Discovery
using Combinatorial Chemistry
JOSEPH A. ZASADZINSKI Ph.D. (Minnesota) Surface and Interfacial Phenomena, Biomaterials.

The Department offers M.S. and
Ph.D. degree programs Finan-
cial aid, including fellowships,
teaching assistantships, and re-
search assistantships, is avail-
One of the world's few seashore
campuses, UCSB is located on
the Pacific Coast 100 miles
northwest of Los Angeles. The
student enrollment is over
18,000. The metropolitan Santa P
Barbara area has over 150,000
residents and is famous for its N'-
mild, even climate.

For additional information
and applications, write to
Chair Graduate Admissions Committee Department of Chemical Engineering University of California Santa Barbara, CA 93106
Chemical Engineering Education

Chemical Engineering at the





"At the Leading Edge"

Frances H. Arnold
John F. Brady
Mark E. Davis
Richard C. Flagan

George R. Gavalas
Konstantinos P. Giapis
Julia A. Kornfield
John H. Seinfeld

Aerosol Science
Applied Mathematics
Atmospheric Chemistry and Physics
Biocatalysis and Bioreactor Engineering
Chemical Vapor Deposition

David A. Tirrell
Nicholas W Tschoegl
Zhen-Gang Wang

Colloid Physics
Fluid Mechanics
Materials Processing
Microelectronics Processing
Microstructured Fluids
Polymer Science
Protein Engineering
Statistical Mechanics

For further information, write
Director of Graduate Studies
Chemical Engineering 210-41 California Institute of Technology Pasadena, California 91125
Also, visit us on the World Wide Web for an on-line brochure:





Fall 19'


?V? Id4To


V o saJ5 10o Orj, 5ef
N|,'Je of f6^ a-c? A a
9rad* afe .delf\ tir Carn\el;e
.e I, ? iepa~rfh ,f oe f
Clhi, al <, gieerir yoe cao
fter(or jroeu.dbreCa i j researJs
is l'iicer\'i er;'i eruiru lei nal
( tr\9' iieer'ig Proo.e9 9, fe-;?
eq 'ecr'i h, **o;S ,?afe en;'
or o.p ,l e,. ua;ds ci 4iTeeTri7v\.
kd ir e ,ef ored b o.ur

".. ." *, "..' *.
O 0."... rt a "j- .- .

; Carnegie-.
.' u .. "

1.3-* 3.


\or i'r orh ta 'o, pla't ,, rie
l/irce, or af C~ra P~parte^ af C isaI Enginer;?
Car* |i't \ef/0 U~ijcri/
Pitt nr4, P. 1.^13-3^0


Full Text

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd