Unsolved Problems in Chemical Engineering

Editors:
Liang-Shih Fan
Martin Feinberg
Geoffrey Hulse
Thomas Sweeney
Jacques L. Zakin
Copyright © 2004 by The Ohio State University Department of Chemical and
Biomolecular Engineering. Copying without fee is permitted provided that copies are not
made or distributed for profit or commercial advantage and credit is given to the source.
Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers
or to distribute to lists, requires prior specific permission.

2
 Table of Contents

A Short History of the Chemical Engineering Department ................................................................ 5

Biographical sketch of L. E. ‘Skip’ Scriven ..................................................................................... 10

When Chemical Reactors Were Admitted and Earlier Roots of Chemical Engineering.................. 11

Biographical sketch of Howard Brenner .......................................................................................... 30

Unsolved Problems in Fluid Mechanics: On the Historical Misconception of Fluid Velocity as
Mass Motion, Rather Than Volume Motion” ................................................................................... 31

Biographical sketch of Michael L. Shuler ........................................................................................ 42

Unsolved Problems in Biomolecular Engineering............................................................................ 43

Biographical sketch of Arthur W. Westerberg ................................................................................. 52

Unsolved Problems in Process/Product Systems Engineering ......................................................... 53

Biographical sketch of Matthew Tirrell............................................................................................ 62

Unsolved Problems in Nanotechnology: Chemical Processing by Self-Assembly .......................... 63

3
4
 A Short History of the Chemical Engineering Department

hemical Engineering at The Ohio State University traces its

roots back over 100 years. Offered initially as a “B.S. in Chemistry from the College of
Engineering,” the first degree was awarded to Samuel Vernon Peppel in 1899. In the
1902-1903 University Bulletin, The Outline of the Course in Chemical Engineering was
listed for the first time, although the degree conferred remained a “B.S. in Chemistry from the College of
Engineering”. In 1904 the name of the degree changed to “Bachelor of Science in Chemical Engineering.”
Its first two recipients were Arno Fieldner and Lewis Benjamin Case in 1906. The first M.Sc. was issued
in 1910 to Orlando Sweeney, the first Ph.D.s in 1918 to Herbert Spencer Coith and James Howard Young.

From 1902 to 1924, chemical engineering was a division of the chemistry department, a common practice
among universities at that time. (The beginning of chemical engineering education in the United States is
usually attributed to M.I.T. where the first degrees in chemical engineering were awarded in 1881.) Dr.
James Withrow headed the Division until 1924 when it became the Department of Chemical Engineering.
He was Department Chair until 1948. In 1925, the degree program was one of the first 10 to be accredited
by the American Institute of Chemical Engineers.

Groundbreaking for the current chemical engineering building took place in 1958; it was later named to
honor Dr. Joseph Koffolt, the Department Chair from 1948-1968. Plans are currently under way to initiate
construction of a new facility within the next 10 years, as the department enters its second century of
service to its students, to the University, and to the profession.

Qualitative Chemistry Lab, ca. 1900

5
 Dr. Robert Brodkey, Professor Emeritus (center, with camera) and Dr. Joseph Koffolt (right,
with shovel) at the 1958 groundbreaking ceremony.

In 1968, Dr. Aldrich Syverson succeeded Joe Koffolt as Chair. When his health failed in 1976, Dr. Edwin
Haering served as Acting Chair until Dr. Jacques Zakin took over in 1977 and held the position until 1994
when Dr. Liang-Shih Fan became Chair. The current Chair, Dr. Stuart Cooper, joined the faculty in
January 2004. The faculty, all of whom came to Ohio State since 1977, is relatively young and has both
increased the intensity and broadened the scope of research in the Department, while retaining close
personal contact with students, a long-standing tradition.

On April 24-25, 2003, a two-day program celebrated the Centennial of the Department’s founding. The
event was divided into two parts:

1. One day was devoted to a look backward at the chemical engineering profession and at the history of the
Department, including reminiscences by alumni from the ‘30s to the ‘70s and by faculty of the last 25
years. A video power-point presentation depicting the 100-year development of the department was
presented by Geoffrey Hulse. Professor L.E. (Skip) Scriven described the development of reaction
engineering in his talk, “When Chemical Reactors Were Admitted and Earlier Roots of Chemical
Engineering.”

2. A symposium was held in which four distinguished chemical engineers addressed the future in their
areas of research under the general title “Unsolved Problems in Chemical Engineering.” These written
presentations, and that of Professor Scriven, are provided here for the benefit of the chemical engineering
community. Their live presentations can be found at www.che.eng.ohio-state.edu. The topics are:

6
 Arthur Westerberg, Howard Brenner, Michael Shuler and Matthew Tirrell (April 24, 2003)

Howard Brenner – Massachusetts Institute of Technology

“Unsolved Problems in Fluid Mechanics: On the Historical Misconception of Fluid Velocity as Mass
Motion, Rather than Volume Motion”

Michael L. Shuler – Cornell University

“Unsolved Problems in Biomolecular Engineering”

Arthur W. Westerberg – Carnegie Mellon University

“Unsolved Problems in Process/Product Systems Engineering”

Matthew Tirrell – University of California at Santa Barbara

“Unsolved Problems in Nanotechnology: Chemical Processing by Self-Assembly”

The Department looks forward to continuing its contributions to the education of students at the
undergraduate and graduate level, expanded research on both theoretical and technological problems, and
serving the University, our community, the nation, and the chemical engineering profession in the years
ahead.

Liang-Shih Fan
Martin Feinberg
Geoffrey Hulse
Thomas L. Sweeney
Jacques L. Zakin

7
8
 When Chemical Reactors Were Admitted
And Earlier Roots of Chemical Engineering

9
 Biographical sketch of L. E. ‘Skip’ Scriven

L. E. 'Skip' Scriven is Regents' Professor and holder of the L E Scriven Chair of Chemical Engineering &
Materials Science at the University of Minnesota. He is a Fellow of the Minnesota Supercomputer Institute,
founded the Coating Process Fundamentals Program, and now co-leads it with Professor Lorraine F.
Francis. He is distinguished for pioneering researches in several areas of fluid mechanics, interfacial
phenomena, porous media and surfactant technologies, and the recently emerged field of coating science
and engineering. He promoted close interactions with industry by showing how good theory, incisive
experimental techniques, and modern computer-aided mathematics can be combined to solve industrial
processing problems. He graduated from the University of California, Berkeley, received a Ph.D. from the
University of Delaware, and was a research engineer with Shell Development Company for four years
before joining the University of Minnesota. He received the AIChE Allan P. Colburn Award four decades ago,
the William H. Walker Award two decades ago, the Tallmadge Award in 1992, and the Founders Award in
1997. He has also been honored by the University of Minnesota and the American Society for Engineering
Education for outstanding teaching. He has co-advised or advised many undergraduate, graduate and
postdoctoral research students, including over 100 Ph.D.’s. Elected to the National Academy of Engineering
in 1978, he has served on several U.S. national committees setting priorities for chemical engineering and
materials science research. In 1990-92 he co-chaired the National Research Council's Board on Chemical
Sciences and Technology, and in 1994-97 he served on the governing Commission on Physical Sciences,
Mathematics, and Applications.

10
 When Chemical Reactors Were Admitted and Earlier Roots of Chemical Engineering

Professor L. E. ‘Skip’ Scriven

Department of Chemical Engineering & Materials Science
University of Minnesota
Minneapolis, MN 55455
[email protected]

In the Beginning in the industrial background today: chemical engineers

at large pay them little heed, though sulfuric acid and
The Industrial Revolution’s first major innovations in sodium carbonate are indispensable.
chemical manufacture were the chamber process for
sulfuric acid in the mid-18th century in England and Ernest Solvay’s 1872 Ammonia-Soda process was a
the Leblanc process for sodium carbonate. The latter breakthrough. He divided the process into distinct
originated in France in the early 19th century and soon operations of gas-liquid contacting, reaction with
diffused to England and throughout the Continent. cooling, and separations; he invented new types of
Both stimulated other technological developments, equipment for carrying them all out continuously on a
some stemming from what would today be called their large scale; and he himself dealt with the chemistry, the
environmental impact. Both drew competition: the materials handling, the process engineering, and the
one from vapor-phase catalytic processes for sulfuric equipment design. In short, the Belgian with no
acid (invented by Phillips in 1831 but not university education performed as what would come to
commercialized for more than 50 years), which finally be called a chemical engineer. Though this was not
replaced it early in the 20th century; the other from the evident to his contemporaries, his performance did
marvelously inventive Solvay process, which more catch some attention in England, and it surely
rapidly replaced the earlier Leblanc technology. Both impressed the aggressive Americans. They soon
of the newer technologies, heavily metamorphosed, are licensed the process, integrating it and its principles
into a fast-developing inorganic chemicals industry that
would be invading European markets around the turn
of the century.

In these two heavy-chemical fields are the roots of

chemical engineering. They brought need for chemists
and for engineers in chemical manufacture, and then
for individuals versed in chemistry and engineering.
They gave rise to George E. Davis, the British inventor
of chemical engineering.

Davis left the Royal School of Mines — not a

University — around 1868, at 18 or 19 years of age.
Starting in the Midlands as an analyst of benzene in
coal gas, he conceived of a benzene recovery plant,
contrived to finance and build one, then cashed out
when the price of benzene quartered. He invented and
sold an ammonia still. He started a company to sell
ammonium sulfate to farmers and instruct them in
using it. And so on. He was a dynamic, competent
young chemical entrepreneur and inventor. He was
noticed by some of Manchester’s intellectual and
scientific elite, among them Angus Smith, a
University-educated chemist dedicated to the cure of
nuisances — by which he meant chemical pollution.

Chemical plants, crudely conceived, poorly built, and

Ernest Solvay
often badly managed, had become dreadful polluters.

11
Hydrogen chloride emissions brought Parliament’s first
“Alkali Works Act” in 1863. Angus Smith was
appointed Chief Inspector. It took him several years to
organize the Alkali Inspectorate, for which he chose
four professionals more competent in analytical and
industrial chemistry than most of the manufacturers.
The Inspectors could get compliance by supplying
money-saving or even money-making advice. Indeed,
the Inspectors became welcome visitors to most works.
Smith chose George Davis for the plum district:
Lancashire, Yorkshire, and Cheshire, heartland of the
chemical industrial revolution.

Davis, the leading light among the four Inspectors,

came to know “everything that went on in everybody’s
works” in the 1870’s. That was a time when everyone
kept proprietary secrets but most of the secrets were
well known. Davis, it soon transpired, saw much more
than “secrets” and thought penetratingly and creatively
about all of his experiences. He also began his lifelong
striving to find ways to abate pollution.

An Idea Takes Root

By 1880 Davis was back in business for himself as

consultant and entrepreneur. He was talking with a George E. Davis
few like-minded colleagues about a new idea: the idea
of chemical engineering and chemical engineers, which building his case. He was elected Vice Chairman of
they thought fit themselves. Their vision was clear: “a the Manchester Section. The Chairman was his five-
chemical engineer is a person possessing knowledge of year older friend and client, Ivan Levinstein, the
chemistry, physics, and mechanics and who employed influential proprietor of the largest dyestuff works in
that knowledge for the utilization of chemical reactions Britain, later President of the Society of Chemical
on the large scale” (as Davis recalled in 1901). Industry, prime driver in reform of English patent law,
Formation of a national society of technical chemists, and Du Pont’s instructor in dye manufacture after the
or industrial chemists, to parallel the young, First World War broke out.
academically oriented Chemical Society, was being
widely discussed. When Davis and his colleagues as Levinstein, born in Germany, by age 20 had studied
attendees at an organizing meeting in Manchester chemistry at Berlin Technical University, emigrated to
proposed a Society of Chemical Engineers, they were Manchester, and begun manufacture of dyes from coal
turned down: most of the rest of the attendees did not tar aniline. He prospered. His interests were wide. He
think of themselves as chemical engineers or did not became a mover and shaker. He spoke and wrote on
know what a chemical engineer should be. Some were challenges facing the chemical industry. In an 1886
leading industrialists, founders of companies that article on the international competitiveness of the
eventually merged into Imperial Chemical Industries, British chemical industries, he defined chemical
or ICI. Others were prominent chemistry professors engineering as the conversion of laboratory processes
with consulting practices. They called the society they into industrial ones, called Ernest Solvay a chemical
formed in 1881 the Society of Chemical Industry, and engineer, and proclaimed that professors who
that is its name to this day. Of its first 297 members, combined scientific attainments, practical knowledge,
14 described themselves as chemical engineers. and industrial contacts were needed to train such men
so that Britain could, among other things, meet the
George Davis, thirty-one years of age, was elected first rising German competition in coal tar dyes. Levinstein
General Secretary. At the Annual Meetings and local was particularly interested in education. He was a
section meetings he gave technical papers based on his governor of the University of Manchester’s forerunner.
experience in trouble-shooting and what later would be He backed the Manchester Technical School, which he
called process development, pilot-planting, and design, later helped upgrade into the Manchester Institute of
all in the course of a thriving partnership with a brother. Science and Technology of today. Then it was chiefly
He called himself Chemical Engineer and began an evening school. Levinstein suggested to Davis that

12
he develop his ideas on chemical engineering into a make the entire process profitable. He recognizes mass
course of lectures. action effects on reaction equilibrium and rate. His
principle for dealing with reaction problems is a
Technical Laboratory for scaling up from the chemist’s
The Invention of Chemical Engineering gram-scale benchtop procedures to the chemical
engineer’s development scale of a few kilos and
apparatus more like that used in manufacture. He
In 1887 Davis delivered his course at the Technical implies that further ton-scale trials may be needed —
School. Many of his lectures appeared one-by-one over what came to be known as semi-works or pilot plant
the next six years in Chemical Trade Journal, a weekly operations. He states that new chemical processes
newsmagazine he founded the same year. He collected often require new combinations of apparatus and newly
additional material and published some of it — even devised “appliances.” He notes that in developing a
useful formulas like one for head loss in pipe flow. He new process “privacy is often of the greatest
had advised Professor Henry Edward Armstrong on importance,” and that privacy is afforded by a secure
setting up a diploma course called “chemical purpose-built Technical Laboratory. Davis and his
engineering” at the City and Guilds College in London, brother had such a laboratory, where they did process
a combination of chemistry and engineering instruction development for clients and themselves.
that enrolled 11 sophomores in 1886 but had its name
changed to chemistry the next year. He became aware The principles for dealing with engineering problems
of a similar combination of industrial chemistry and Davis recognized could be organized around basic
mechanical engineering courses, including a new one operations common to many: fluid flow, solids treating,
on “chemical machinery from an engineering point of heat or cold transfer, extraction, absorption, distillation,
view,” set up as a chemical engineering curriculum in and so on. These were covered chapter by chapter in
1888 at Massachusetts Institute of Technology by his book, the forerunner of the unit operations texts on
Lewis Mills Norton (an MIT graduate and 1879 which chemical engineers cut their eye teeth from 1923
Göttingen Ph.D. in chemistry). Norton died young two onwards. There was also an important chapter on
years later and the new curriculum languished, though materials of fabrication. And, as W. K. “Doc” Lewis
it kept its name. In the United States chemistry and noted in 1952 in an eloquent, long overdue
chemical manufacture were seen as frontiers. The idea acknowledgment of Davis’s impact on the pioneers at
of chemical engineering was vague yet attractive to Massachusetts Institute of Technology, the
chemistry professors. For example, at Minnesota the development was as quantitative as the resources of the
first curriculum in chemistry, established in 1891, was end-of-century had allowed.
named chemical engineering, and the first four
graduates (in 1897) received the degree Chemical Very few people were ready to act on George Davis’
Engineer, but that designation was not repeated for a vision of a discipline of chemical engineering, either in
decade. The University of Pennsylvania in 1892 was 1887, or in 1901, or in 1904. It fit no university
apparently second in establishing permanently a curriculum of that time. Davis had no close ties with
curriculum called chemical engineering, Tulane in technical schools, much less universities in the United
1894 third, and Michigan in 1898 fourth. Kingdom, though he certainly knew George Lunge, an
international authority in industrial chemistry with
Davis in 1901 transformed his lectures into the first similar views of the basic operations, who was a
book on the discipline of chemical engineering. His professor at the Swiss Federal Institute of Technology
preface highlighted the mounting competition from (ETH) in Zurich. But the vision did not go unnoticed
America (the British viewed the United States then in the United States, where there were others who were
rather as the Americans viewed Japan in recent times) calling themselves chemical engineers and there were
in heavy chemicals and the “wonderful developments university curricula called chemical engineering. Just
in Germany of commercial organic chemistry.” An how that term gained currency is not clear, but it is
expanded, two-volume second edition of A Handbook likely that the Chemical Trade Journal crossed the
of Chemical Engineering was published in 1904. It Atlantic and it is certain that the Journal of the Society
departed radically from the earlier textbooks and of Chemical Industry had avid readers here. Davis died
handbooks on industrial chemistry, which covered each in 1907 at 57, deprived of seeing his invention, once it
chemical industry separately. Davis had recognized was reduced to practice by Walker, Lewis, and
that the basic problems were reaction management and colleagues, grow into a flourishing profession and
engineering issues. discipline, first in the U.S. Perhaps he would have
fanned to flame their smoldering growth in Britain
Reaction management he does not name but illustrates ignited by wartime pressure on that nation’s chemical
by examples of choosing reaction feeds and conditions industry.
to reduce reactor volume or shorten reaction time,
reduce by-products, facilitate product purification, and

13
 which was dedicated to the memory of Lewis Mills
Organic Chemicals and Industrial Chemistry Norton, late Professor of Industrial Chemistry; Allan
Roger’s Industrial Chemistry (1902, 1912, 1915, 1920,
1925, 1931; 1942 edited by C. C. Furnas); and Emil
The second round of innovations in chemical Riegel’s Industrial Chemistry (1928, 1933, 1937) from
manufacture were batch processes for small-volume University of Buffalo (the last edition of which was
production of high value-added dyes and other coal tar one of the writer’s burdens as an undergraduate student
derivatives. These began with Perkins’ mauve in a few years before the admission of reaction
England. Apart from Levinstein’s firm in England this engineering). Until 1923 such texts were the only ones
field was very soon dominated by German research unique to chemical engineering curricula.
prowess in organic chemistry. The rise of that prowess
was aided by university-industry research cooperation,
an innovation triggered by British example; and it was
integrated into a German industrial juggernaut that Electrochemicals, Niagara, and Outgrowths
before long controlled international markets. Early
pharmaceuticals followed the same route. Here, The third round of innovation was electrochemical
though, the research chemist’s laboratory methods processing, which rose in England, Germany, and
were turned over to mechanical engineers to scale up France in the decades before 1900 but diffused to
directly, and many of them became skilled enough that America, where cheap electricity generated from cheap
in 1886 Levinstein thought some to be scientific coal, and mass production in the new tradition of the
chemical engineers. Still, this was no harbinger of iron, steel, copper, nickel, and tin industries, enabled
chemical engineering as a discipline, nor did it lead the the Americans to compete successfully, even invading
way to continuous processing and the economies that international markets — for example, with electrolytic
that could bring when markets expanded and caustic soda and chlorine. Though not much American
competition demanded. science came up to European standards of the era,
“Yankee ingenuity,” especially in improving on
Meanwhile the sugar industries, distillation industries, European inventions and technologies, was very much
and many others were evolving in Europe. Process in evidence. Charles Hall invented in 1886 the most
engineering and industrial chemistry curricula were successful process for producing aluminum. Such
being installed in the technical universities (Technische innovations and commercializations were plentiful
Hochschule) that had appeared in Germany, before a U.S. scientist had received a Nobel Prize.
Switzerland, Austria, and even Hungary. A few of the
professors and their counterparts, the chief engineers in In 1895 came the stupendous hydroelectric
certain companies, highly educated in science and development at Niagara Falls; by 1910 that was the
mathematics, took to analyzing common, constituent location of “the world’s greatest center of
operations like heat transfer (Peclét in France), electrochemical activity,” not only of production but
vaporization, condensation and drying (Hausbrand in also of process research and product development.
Germany), and distillation (by Hausbrand and by Sorel Outstanding was Frederick Becket and his Niagara
in France). Peclét’s seminal monograph went through Research Laboratories, where he invented processes for
several editions earlier in the century; monographs by making carbon-free chromium, tungsten, molybdenum,
the rest of these authors were published between 1890 and vanadium by direct reduction of their oxides, and
and 1899. These too were important forerunners of the other important processes as well. This was one of the
discipline of chemical engineering. So also were very first industrial research labs in America, and it and
atlases of chemical manufacturing equipment Becket were soon bought up by Union Carbide to be
organized by basic operations rather than industries. theirs.
Consulting professors wrote about technologies of
specific chemical industries. German-British-Swiss Electrochemistry was the glamour science and
George Lunge’s successive editions on sulfuric acid emerging technology of the era. The Electrochemical
were magisterial. These fed into the discipline Society was the meeting ground for leaders of the new
emerging in the U.S., as did the compendious texts on science of physical chemistry like Ostwald, Nernst,
industrial chemistry — purely qualitative descriptions Bancroft of Cornell, and Whitney — who moved from
of industry upon industry — that began appearing. Arthur Noyes’ circle at the Massachusetts Institute of
Examples were Samuel Sadtler’s Industrial Organic Technology to Schenectady to head General Electric’s
Chemistry (1891, 1895, 1900, 1912, 1923) from the brand-new research laboratory, the country’s first
University of Pennsylvania and Philadelphia College of corporate research establishment. It was the meeting
Pharmacy (in 1908 Sadtler was elected the first ground for educated inventor-entrepreneurs like Elmer
president of the American Institute of Chemical Sperry, then of National Battery, and Herbert Dow of
Engineers); Frank Thorp’s Industrial Chemistry (1898, Midland; and for prominent industrial chemists and
1908), from M.I.T. in Boston, a textbook for students chemical engineers like Samuel Sadtler of Philadelphia,

14
William Walker of the partnership of Little & Walker
in Boston, and Fritz Haber of Karlsruhe Technical
University in Germany.

Emergence of the Discipline

With the mention of William Walker, who led the

implementation of George Davis’s invention, it is
appropriate to go back to the scene in America’s
colleges and universities in the 1880’s.
Industrialization of the country was accelerating, and
with it the need for engineers and, to a lesser extent,
chemists. From decades earlier there was popular
demand for relevant college education, and this had
been answered by the appearance of engineering
schools like Rensselaer Polytechnic Institute in Troy,
New York, Brooklyn Polytechnic, and the
Massachusetts Institute of Technology; by scientific
schools at Yale, Harvard, Dartmouth, Columbia, and so
on; and by the Morrill Act of 1862, which enabled the
states to establish land-grant colleges, one purpose of
which was education in “the mechanic arts,” which William H. Walker
turned out to be engineering — civil, mining, and
mechanical; then electrical, metallurgical, and finally the chemical engineering curriculum. Those ideas can
chemical. be glimpsed in his still diffuse 1905 article, “What
Constitutes a Chemical Engineer,” in Richard Meade’s
By the 1880’s, strong curricula in science and partisan magazine, The Chemical Engineer, which was
engineering had sprung up in many land-grant schools, addressed to “technical chemists and engineers in the
among them Pennsylvania State College. There young laboratories and supervision of the operations of the
William Hultz Walker enrolled in 1886 and graduated great chemical and metallurgical industries” of the
in chemistry in 1890. Having been a good student, he United States. From what happened it’s plain that
set off for graduate study in Germany — as did some Walker’s program was to incorporate the new sciences
1,000 other top graduates in chemistry between 1850 of physical chemistry and thermodynamics, the new
and World War I and perhaps 9,000 more in other engineering sciences — as they could have been called
fields. Germany then was the center of freedom of — of heat transfer, distillation, evaporation, and fluid
learning, freedom of teaching, academic research, and mechanics that had emerged in Europe, and his own
chemistry. Returning from Göttingen with a Ph.D. in specialty, corrosion; to organize a chemical
1892, Walker taught for a couple of years at Penn State, engineering course including laboratories around
moved to M.I.T., then resigned in 1900 to join M.I.T. - Davis’s (and his colleague George Lunge’s) conception
chemistry graduate Arthur Dehon Little in an industrial of what would come to be known as “the unit
chemistry consulting partnership. (Little had lost his operations”; and to develop research through student
original partner Griffin in a laboratory explosion. theses and industrial interactions. As it happened, in
the same department his chemist colleague Arthur
In 1902 William Walker not only became a charter Amos Noyes, an 1890 Ph.D. with Ostwald in Leipzig,
member of the American Electrochemical Society; he had just developed a course in physical chemistry that
also accepted an appointment in the Chemistry emphasized, among other things, problem solving.
Department at M.I.T. to take charge of the curriculum Noyes urged, and William Walker insisted, that every
called chemical engineering, though he continued his chemical engineering student take it.
partnership with Arthur Little until 1905. To his new
position he brought his recollections of Norton’s ideas The success of Walker’s program was assured when he
from 1888-1890; whatever he had picked up from such arranged for a 1905 graduate, Warren Kendall Lewis,
writings as Ivan Levinstein’s published lecture in the to go to Germany for a Ph.D. in chemistry at Breslau.
Journal of the Society of Chemical Industry in 1886 For upon returning in 1908, Lewis joined Walker and,
and George Lunge’s influential one in the Journal of besides teaching, embarked on a series of analyses of
the American Chemical Society in 1893; his own distillation, filtration, fluid flow, countercurrent
annotated copy of George Davis’s magnum opus on contacting, heat transfer, and so on. Most of them
Chemical Engineering; and his ideas for transforming involved bachelor’s and master’s theses and many were

15
published in the young Journal of Industrial and Whether plates actually behave as equilibrium stages
Engineering Chemistry, which in 1909 the American was not broached. Current versions of all of these
Chemical Society started along with the Division of things are embedded in the basic education of every
Industrial Chemistry and Chemical Engineers, as an in- chemical engineer.
your-face counter to the A.I.Ch.E.!. These researches
developed principles of the physical side of chemical The other chapters, on fuels and power, combustion,
engineering, which were expounded in mimeographed furnaces and kilns, gas producers, crushing and
notes for classes and then in the famous 1923 book, grinding, and mechanical sizing, were as purely
Principles of Chemical Engineering, by Walker, Lewis, descriptive as George Davis’s book or anything else of
and the younger McAdams, who joined them in 1919. the era. There was nothing, not even rules of thumb,
about capacities, dimensions, or shapes. And not a
This book, the first textbook of chemical engineering, word about reaction kettles, converters, or other sorts
was organized along the lines of Davis’s 1904 magnum of reactors. But in many places the text pointed to
opus, but omitted the managing of chemical reactions needs for experimental data, correlations, and
and overall view of a chemical process. The first quantitative design procedures. From the M.I.T.
chapter was on simple industrial stoichiometry. A later faculty other textbooks preceded slightly or followed:
one opened with manometers, then dealt with pipe flow Robinson’s on Elements of Fractional Distillation
— both straightline (laminar), citing Lamb’s (1922), Lewis and Radasch’s on Industrial
Hydrodynamics, and turbulent, laying out the Stoichiometry (1926), Haslam and Russell’s on Fuels
engineers’ version of Bernoulli’s equation and and Their Combustion (1926), McAdams’ on Heat
Fanning’s correlation of viscous losses with the newly Transmission (1933), and later Sherwood’s on
accepted Reynolds modulus. Another chapter put Absorption and Extraction (1937). Fluid flow
forward a crude bundle-of-capillary-tubes model of textbooks were left to civil and mechanical engineers;
filter cake, developed equations of constant-pressure- mechanical comminution and separations, to mining
difference and constant-rate filtration from Poiseuille’s and mechanical engineers. Chemical engineering
equation, and sketched how to optimize a batch authorities from other institutions contributed books to
filtration. A short chapter told about putative stagnant the campaign which was orchestrated, for a time
fluid films next to solid surfaces and their posited role exclusively, by the New York publisher McGraw-Hill.
in heat transfer, citing Langmuir’s work on hot wires Then John Wiley launched a series with a new kind of
(as in recently developed light bulbs) but omitting the introductory text, a combination of stoichiometry,
relevance of gases’ viscosity rising with absolute applicable physical chemistry and thermodynamics:
temperature. It also described colleague W. G. Olaf Hougen’s and Ken Watson’s Industrial Chemical
Whitman’s brand-new hypothesis of stagnant films Calculations (1931) from the University of Wisconsin.
next to liquid surfaces — the embryonic “two-film
theory.” A later chapter returned to this as the basis of The first textbook, Principles of Chemical Engineering,
interphase transfer coefficients in equations to correlate a fine one for its day (apart from sparse references to
performance of dehumidifiers and cooling towers. the literature, finally rectified by the third and last
Another chapter treated drying simply in terms of an edition in 1937), shaped the discipline and helped mold
evaporation coefficient. The wet bulb thermometer the profession. Follow-on textbooks, beginning with
was analyzed by postulating that heat and humidity Badger and McCabe’s 1931 Elements of Chemical
must diffuse across the same stagnant film at its surface. Engineering, were organized along the same lines — to
A chapter covered rudimentary heat conduction and the exclusion of the chemical reactors that are the heart
radiation, then drew on mechanical engineering papers of a chemical process and dictate the needs for the unit
and McAdams’s recent data from M.I.T. undergraduate operations, as Davis had pointed out.
theses to treat convective transfer to flowing fluid,
condensing vapor, and boiling liquid in and on tubes. Incidentally Warren McCabe as one of the earliest
A companion chapter dealt with evaporation — even Ph.D. candidates at M.I.T. had teamed with fellow
basic analysis of a four-effect multiple evaporator. The graduate student Ernest Thiele around 1925 to devise
distillation chapter, after a lot of qualitative the neat McCabe-Thiele diagram for graphically
descriptions, got to relative volatility, Rayleigh’s 1904 analyzing the simplest sort of binary separation by
equation for simple batch distillation of binary distillation in a column of distillation plates. The
solutions, Sorel’s 1893 equations for binary distillation diagram has of course been a staple of chemical
in a column of ideal plates that are equilibrium stages, engineering since the 1927 second edition of the
and the critical design choices: a reflux ratio greater Principles and the 1931 first edition of the Elements.
than would require an infinite number of plates, a
column diameter large enough that the rising vapor From around 1910 Walker and Lewis, along with
flow does not splash and entrain downflowing liquid Columbia University’s Milton C. Whitaker, a
excessively, and liquid depth on each plate great Coloradan who alternated between industrial and
enough to accommodate the liquid flow across it. academic careers, were the vanguard of those

16
developing the laboratory course for the emerging board of governors (the Corporation), he coined a name
discipline. What they wrought was virtually devoid of that stuck. The name was “unit operations” for those
chemical reaction and came to be known later as the basic physical operations in chemical manufacture that
“unit operations laboratory,” the cornerstone of Davis had written about decades earlier (Little nowhere
chemical engineering education. mentions being aware of Davis’s writings though he
must have been reading the Journal of the Society of
Chemical Industry before the turn of the century).
Founding of the Profession and Naming of Unit George Lunge too had spotlighted them in an
influential 1893 speech to the World’s Congress of
Chemists at America’s first great international
Operations exposition, the Columbian Exposition in Chicago, and
his speech had appeared in the fledgling Journal of the
Walker’s program was further reinforced by his former American Chemical Society. Some professors and
partner, the well-connected consulting industrial other members of the newly organized profession were
chemist Arthur D. Little. In 1908 they together took talking and writing about those operations. Many in
part in an ad hoc meeting provoked by Richard Meade, the chemical and metallurgical industries were of
a young industrial chemist with a cause, and his four- course practicing them in the course of their work.
year-old magazine The Chemical Engineer for a pulpit.
The outcome was a recommendation for organizing an What nobody got around to identifying were the basic
American Institute of Chemical Engineers, which was kinds of chemical operations in chemical manufacture,
done in Philadelphia that year. It appears that no one although by 1905 Walker, like Davis before him, had
with a degree in chemical engineering was among the epitomized the chemical engineer as one “who can
established practitioners and professors trained as devise, construct, and operate industrial plants based on
industrial chemists and mechanical engineers who chemical reactions,” and in Germany the rudiments of
founded the A.I.Ch.E. chemical reaction engineering were being uncovered
by Knietsch, Haber, Bosch, and others. In Germany,
Surprisingly, neither Walker nor Little were founders. there was neither a discipline nor an organized
They became instead the Chairman and Secretary of profession to follow up (only in the 1930s did Eucken
the Division of Industrial Chemistry and Chemical and Damköhler take up the line — in a research
Engineering formed that year by the American university). In the United States a young metallurgical
Chemical Society in direct opposition to the A.I.Ch.E. engineer (Columbia University) turned physical
When attitudes tempered, the new division took its chemist (1906 Ph.D. with Nernst at Göttingen) named
present name, Division of Industrial and Engineering Irving Langmuir, who was teaching at Stevens Institute
Chemistry. It catered particularly to those concerned of Technology in New Jersey, authored a pregnant
with industrial applications of chemical reactions. It paper in 1908. It was an astute theoretical analysis of
developed vigorous technical sessions at A.C.S. the continuous flow reactors that had come into vogue
meetings and a quality journal for industrial chemists in Germany for studies of reaction rates in gases — and
and chemical engineers. for Knietsch’s commercialization at BASF of Phillips’
old invention of vapor-phase oxidation of sulfur
The A.I.Ch.E. in its meetings, transactions, and dioxide to make sulfuric acid. Mixing by convection
curriculum accreditation activities seems tacitly to have and diffusion was the key issue and it led Langmuir to
backed away from the reaction operations in chemical equations and solutions for what later was called a
manufacture. (Although chemical engineering continuous flow stirred tank reactor, the completely
benefited greatly, on balance, by developing in close stirred limit, and for the latter-day plug flow reactor,
connection with chemistry, the association was not a the totally unstirred limit. The next year appeared a
uniformly positive one.) Some chemical engineers clear, thorough, equation-grounded exposition of
joined both organizations. Both Walker and Little chemical kinetics basic to reaction engineering:
joined the A.I.Ch.E. within a few years and Little, opposing reactions, side reactions (concurrent
A.C.S. president from 1912 to 1914, was elected reactions), sequential reactions (consecutive reactions),
A.I.Ch.E. president for 1919. (He was also president catalysis, effects of temperature and pressure; the book
of the British-based binational Society of Chemical was J. W. Mellor’s Chemical Statics and Dynamics
Industry for 1928-1929.) From 1908, Arthur Little (1909). An indication that these events were not
took increasing interest in chemical engineering entirely missed at M.I.T. is that W. K. Lewis and E. D.
education and by 1915 was the chairman of a Visiting Ries later published on temperature-controlled sulfur
Committee for Chemical Engineering at M.I.T. Little dioxide converter design based on their analysis of
was famous as a fine speaker and writer (though windy Knietsch’s tubular reactor data. Would reactors have
by today’s standards) and was particularly eloquent been admitted to the emerging discipline had Langmuir,
regarding the need for industrial research in America. unhappy at Stevens, gone not to General Electric’s
In that year’s report to the M.I.T. administration and research laboratory but to M.I.T.’s department of

17
chemistry and chemical engineering, there to fall in continuous flow pipe stills replaced shell stills and
with Lewis as both were starting their careers ? fractionators came in after World War I. Mixer-settler
trains for treating gasoline gave way to continuous
Alas, there was no nascent subdiscipline of reactor countercurrent extraction in towers of perforated plates.
analysis and design, only the tradition of descriptive Abruptly, there was high demand for analysis and
industrial chemistry, to inform the A.I.Ch.E.’s 12-year- design of distillation and extraction equipment as well
long efforts to define chemical engineering more fully, as cracking furnaces. Petroleum refining became the
prescribe a basic curriculum, and establish an frontier. When Jersey Standard moved to set up a
accrediting scheme—the first in engineering—all of research and development department to apply
which succeeded ultimately. Arthur Little himself led chemistry, physics, and chemical engineering to the oil
the later stages of the effort to establish the curriculum, industry, it retained as advisors chemist Ira Remson,
the core of the discipline. The principles embodied in president emeritus of John Hopkins, physicist Robert
that curriculum, having later made room for chemical Millikan of Caltech, and Warren Lewis.
reactor engineering, have prevailed ever since.
At M.I.T., Lewis pushed combustion, the only sort of
reaction to draw concerted investigation by chemical
engineers of that era, and launched Robert Haslam and
Relationships with Petroleum and Chemical Robert Russell, whose researches were capped by their
1927 volume on Fuels and Their Combustion. Haslam
subsequently became development manager at
Industries Standard Oil Development Co. and later vice president
and a director of Jersey Standard. Clark Robinson,
By the turn of the century, major American industries who had graduated in 1909, spent five years in industry,
were being consolidated through mergers and and returned to do a master’s and join the staff,
acquisitions, partly because of the stress of answered demand in 1922 with the first physical
international competition. Frank A. Vanderlip chemistry-based text on The Elements of Fractional
observed in Scribner’s Magazine in 1905 that “as Distillation, a best-seller that went through several
combinations are made in the industrial field, the editions, later with younger Edwin Gilliland as
possibility of employing highly trained technical coauthor. In the same year Robinson also published a
experts rapidly increases. Technical training is book on The Recovery of Volatile Solvents. In the next
therefore becoming of vastly more importance than year, he coauthored a seminal little textbook with
ever before, and those nations which are offering the Professor Frank Hitchcock of the mathematics
best technical training to their youths are making the department based on a course they had begun in 1919.
most rapid industrial progress . . . the relative A remarkable outburst!
efficiency of nations was never before so largely
influenced by the character of their educational The little book, Differential Equations in Applied
facilities”. The transformation of chemical engineering Chemistry, had only isolated impact, though all through
that William Walker launched was remarkably an era when few engineers truly understood calculus it
successful, and he is widely regarded as the father of showed off powerful mathematical tools, among them
the discipline. Testimony to the success were the (in the 1936 second edition) numerical solution
impacts of graduates, ideas, approaches, and research methods. Not only did it connect with chemist J. W.
results on the chemical and petroleum industries of the Mellor’s 1902 Higher Mathematics for Students of
United States. That country had come out of World Chemistry and Physics—with Special Reference to
War I the world’s most powerful economy. Practical Work; it also displayed analyses of reaction
rate in simple reaction systems from Mellor’s Chemical
The relationship with petroleum refining was Statics and Dynamics (1909). It did inspire at least a
particularly significant. Refining, which had few scientifically-minded young chemical engineers to
developed outside the mainstream of chemical develop more accurate mathematical descriptions of
processing, was a growth industry of the motor age, a some of the basic physical and chemical operations that
large and expanding area of opportunity for chemical they encountered. Perhaps more influential were
engineering. W. M. Burton’s process of thermal textbooks in applied mathematics for engineers that
cracking, a crude kind of batch reaction processing, began appearing in the 1930’s (e.g. Sokolnikoff &
had been innovated at Standard Oil of Indiana in 1912; Sokolnikoff’s in 1934, Reddick & Miller’s in 1938,
J. A. and C. P. Dubbs’ competing patents led to von Kármán & Biot’s in 1940) to which small
continuous thermal cracking — and the Universal Oil numbers of chemical engineers were exposed in
Products Company, a nursery of reaction engineering. elective academic courses and rump study groups in
Catalytic cracking was a couple of decades in the industrial R & D organizations. The argument can be
future. Distillation separations, incredible as it may raised that chemical reaction engineering and
seem today, were a batch process in refineries until systematic process engineering, including

18
comprehensive methods of addressing control and course the vitals of a chemical process: they dictate
design optimization, could not develop until adequate what the separations and other unit operations must do
mathematical tools had been assimilated into the in the rest of a plant.
discipline, which took decades.
Only for mature processes like sulphuric acid
Petroleum refining was but one of the arenas of manufacture had rules of thumb of design and
innovation in that era that came to need chemical operation passed into the possession of consultants and
engineering. Courses were strengthened or installed in authors like George Lunge and George Davis. Davis,
universities across the United States, and war veterans for example, gave such figures for continuous sulfuric
swelled enrollments in the young discipline to 6000. acid manufacture as 16 cubic feet of lead-lined
The country’s output of B.S. chemical engineers rose chamber for sulfur dioxide oxidation and absorption
through the 1920’s. Ph.D. programs, an inheritance per pound of sulfur burned per 24 hours — a rough
from the chemistry side — not the mechanical stand-in for residence time or space velocity. (Two
engineering side of the early conglomerate curricula — decades later Riegel noted that higher-temperature
began to take hold. Apparently the first doctorates in operation and the use of circulating fans had brought
chemical engineering were awarded at the University the figure down to 12 cubic feet.) Otherwise the design
of Wisconsin starting in 1905. By 1920, thirty had and operation rules had to be learned by engineers after
been granted nationwide: the flow of chemistry they joined the plant, and then on a need-to-know basis.
graduates, pure and industrial, to Germany for Generally, the batch furnaces, ovens, kilns, autoclaves,
postgraduate study had been staunched. The total of kettles and pots, the continuous converters and coiled
U.S. Ph.D.’s and Sc.D’s in the following decade was tubes were chosen on the basis of experience and the
120. Many went to industry, others — the split is not ease of adding another to the “battery;” they were
clear — to chemical engineering faculties, which operated with proprietary recipes derived from
continued growing and which A.I.Ch.E. accreditation chemists’ benchtop trial-and-error and successive
standards expected to be engaged in research. scale-up experiments. Processing small-volume
batches of dyes, medicinals and the like might be
Petroleum refining consumed a lot of university adjusted on the production scale by the chemists
outputs, but so did the American chemistry industry, themselves, as in Germany. For larger volume
already in aggregate the world’s largest and still processes Davis seems to have advocated scale-up
growing lustily. Among new large volume processes from bench to “technical experiment” to see what
were those for carbon tetrachloride and chloroform, happens in plant-type equipment, to explore and
which required accurate temperature control and improve operating conditions, to assess impacts of
precise fractional distillation; for the intermediates impurities and by-products, to test materials of
phenol, monochorobenzene, phthalic acid, and construction, and so forth; then to semi-works scale for
diphenyl; for the alcohols from methanol to hexanols more of the same plus salable product; then to full
and the cellosolves; for acetone and acetates; for high- scale — three steps of scale-up, four levels. Writing in
strength hydrogen peroxide, for magnesium; and for a 1930 (in an Introduction to Badger and Banchero’s
variety of plastics precursors like styrene, vinyl esters, book), Arthur D. Little states that in his practice he has
polybasic acids, polyhydric alcohols, and acrylates. found it advisable, for an important process which may
Smaller volume batch processes appeared for a host of ultimately require heavy investment, to have five or six
synthetic dye stuffs, medicinals, and germicides as well steps between bench-top operations and final plant
as other specialty chemicals, organic and inorganic. design: “Thus the validity of the proposed reactions
[emphasis added] is confirmed, difficulties are
As had been true in Davis’s time and before, design disclosed and overcome [presumably he included
and performance of the reactors in these processes matters of recycling unreacted reactants and purging
were closely held by the companies that developed them of troublesome constituents], yields gradually
them. Any information was kept secret that related key improved, and any defects in materials of construction
performance factors — conversion of the limiting or limitations of units of equipment revealed. Not until
reactant, yield of the product, and by-product the operation of a plant of semiworks size has verified
distribution — to the operating parameters: the estimates of costs of production and the product has
proportions of reactants and diluents in the feed; been sold at a satisfactory price . . . can one feel that
unrecovered products, by-products, and diluents in he has reached the goal of the long road of process
recycle; volume and holding time in a batch reactor or development”. A quest expensive in money and time!
dimensions and nominal residence time (the reciprocal It’s no wonder that reactor design, operation, and
of what was called the “space velocity,” which was not performance were closely held within companies. It’s
a velocity) in a flow reactor; and pressure and also no wonder that investigators in universities could
temperature program in time or profile along the length. rarely afford to study realistic reactors.
Patents were vague, professional papers and technical
articles generally silent on these things, which are of

19
Unit Processes Cul-de-Sac

What the academics along with journal editors went

after instead were qualitative process flow sheets,
which eventually became a fashionable form of public
relations with such audiences as the readerships of
Industrial and Engineering Chemistry and Chemical
and Metallurgical Engineering. From the journals the
sheets went to a new variety of industrial chemistry
book. The idea of unit processes was put forward in
1928 in Chemical and Metallurgical Engineering by P.
H. Groggins, an industrial chemist in aniline and dyes,
in a paper on “Nitration — A Unit Process of Chemical
Engineering.” What he meant by a unit process was a
chemical conversion in contrast to the physical
transport and transformations called unit operations.
“Unit processes deal principally with chemical
reactions, whereas unit operations relate largely to
physical phenomena,” he wrote in the 1935 book he
edited and wrote a large part of: Unit Processes in
Organic Synthesis, a description of nitration,
halogenization, sulfonation, oxidation, alkylation, etc.,
etc. with flow sheets. There were only casual mentions Carl Bosch
of reaction conditions and losses and scarcely any
information on reactor design and performance. It was
updated industrial chemistry and it filled needs. the “processes” do not, he never grasped, although he
Generations of chemical engineering undergraduates did sense that the reaction engineering emerging
memorized portions of its contents in the industrial around him from industry and taking root in academia
chemistry courses that still survived. Only for its fifth was the future. That did not diminish the largely
edition 23 years later did Groggins commission stultifying effect of the book wherever it was inflicted
Thomas Corrigan (1949 Ph.D. with Kenneth Watson at on undergraduate chemical engineers. As one
Wisconsin) of Olin Mathieson Chemical Corporation distinguished chemical engineer in management wrote
and John McKetta of University of Texas to write short in that era, “The modern employer does not engage a
chapters on Chemical Kinetics and Chemical-Process graduating student because the student can describe for
Kinetics, i.e. reaction engineering. his employer how his product is made”.

The unit processes idea was indefatigably promoted by

Norris Shreve, an organic technology professor in the Ascendancy of Continuous, Catalytic, and High-
chemical engineering department — so large it was
called a School — at Purdue University in Indiana.
Dissatisfied with the absence of chemical reaction from Pressure Reactors
the unit operations paradigm, Shreve by 1933
organized lecture courses (“Flowsheets 101” to the In the decade before World War One, Fritz Haber,
irreverent) and lab courses (“Pots and Pans 102”), the Walther Nernst, and others in Germany had taken up
latter taking gram-scale syntheses and separations to gas-phase reactions — Haber published his instructive
100 gram-scale, that were fairly widely emulated in book on the thermodynamics of such reactions in 1906
university curricula. From 1937 he organized an — and were soon pursuing nitrogen fixation by
annual symposium on unit processes for the Division ammonia synthesis, a goal of tremendous economic
of Industrial and Engineering Chemistry at American and geopolitical significance. Haber got the
Chemical Society meetings; the symposia drew papers breakthrough by discovering a catalyst in 1908. A
and attendees engaged in industrial applied kinetics and leading German corporation, Badische Anilin-und-
nascent reaction engineering. His 1945 textbook, The Sodafabrik (BASF) took on the development, assigning
Chemical Process Industries, was organized around to it a self-made chemical engineer, Carl Bosch, a
more than 100 flow sheets that had been published in mechanical engineering graduate of Leipzig University
Chemical and Metallurgical Engineering and that with a Ph.D. in industrial chemistry, and to a
illustrated, he emphasized, the coordinated sequences methodical chemist, Alwin Mittasch, who reportedly
of unit processes and unit operations that constitute oversaw 20,000 trials with different catalyst
chemical processes. That the “operations” transcend formulations to find the best (the irreverent assert that
specific reactions, chemicals, and industries whereas though techniques of combinatorial chemistry have

20
speeded up catalyst design lately, the approach hasn’t and Thomas C. Chilton, a Columbia University
changed in nearly a century). Before 1900, Rudolf chemical engineering graduate with three years of prior
Knietsch of BASF had, by scientifically informed research experience, the program was a roaring success.
experimentation, systematically analyzed and designed It led to Du Pont’s high-pressure nitric acid process,
reactors for vapor-phase oxidation of sulfur dioxide. which was scaled up, commercialized, and licensed to
Building on this earlier pioneering, his colleague Bosch other chemical companies within four years, testimony
designed the first high-pressure, continuous flow to the astuteness of Vanderlip’s 1905 observation in
tubular reactor— actually a tube—and the rest of the Scribner’s Magazine. The great impact within Du Pont
plant. By 1913, they had the Haber-Bosch process for was heightened by contributions the seven or eight
synthetic ammonia in operation, a signal other chemical engineers (half of them postwar
accomplishment like Solvay’s 50 years earlier. graduates from M.I.T.) in the Research Department
made to process technology for producing ammonia,
At the war’s end the Allies took German technology as methanol, and other products. The research director,
a war prize. The war had also stimulated the U.S. Charles M. A. Stine, already a respected advocate of
chemical industry: explosives to be sure, but most fundamental research and spokesman for chemical
notably dyestuffs by Dow and Du Pont and nitrogen engineering, responded by organizing half his chemical
fixation and nitric acid manufacture by government- engineers into a group with the goal of getting their
financed plants at Muscle Shoals, Alabama. Gasoline discipline onto a sound scientific footing and
demand for proliferating motorcars provoked simultaneously serving as internal consultants. He put
improvements in thermal cracking. Easily refined young Tom Chilton in charge, who soon hired Alan
petroleum deposits began to seem quite limited. There Colburn, a fresh Wisconsin Ph.D. — thus the famous
were also wartime profits to invest in upgraded and Chilton & Colburn contributions to the discipline.
new technologies. So it was that in America in the
decade after World War I, process innovation turned to Preceding Du Pont’s cheaper nitric acid, the first
high pressure and often high temperature, capitalizing products from the high-pressure high technology of the
on German advances in vapor-phase catalytic reaction mid- and late- 1920s were methanol and other
processes and on advances in metallurgy and commercial solvents, setting the stage for what Arthur
fabrication — enormously important advances! D. Little in 1928 foresaw: the coming era of chemicals
–“new solvents and organic chemicals in great variety”
A few months after the war’s end, the U.S. government —from petroleum feedstocks. In particular there were
organized the Fixed Nitrogen Research Laboratory, the refinery off-gases that were being burned as fuel,
which pioneered high-pressure property and process rising amounts of ethylene and other olefins among
research and ammonia catalyst development for the them. More and more petroleum was being thermally
nation’s needs. Led in succession by Dr. A. B. Lamb, cracked to increase gasoline yield, and olefins were by-
Prof. R. C. Tolman, and Dr. F. G. Cottrell, outstanding products. The economic environment was ready for
and practical scientists all, it recruited talented the next high technologies to evolve.
chemists and engineers who were periodically raided
by companies assembling cadres for industrial research And evolve they did. Union Carbide went from light
and development—a classical mode of technology hydrocarbons to liquefying them (liquefied petroleum
transfer. Moreover, in 1925 and 1926, industry’s gas, LPG) to ethylene oxidation to Cellosolve solvent,
Chemical Foundation started financing high-pressure a glycol ether marketed in 1926; then “permanent
laboratories in the chemical engineering departments of antifreeze” in 1928 and synthetic ethyl alcohol in 1930.
Massachusetts Institute of Technology (Prof. W. G. The international Shell Group set up its development
Whitman), Yale University (Prof. B. F. Dodge), and Company in 1928 near Berkeley, California under the
University of Illinois (Prof. N. W. Krase) — each direction of Dr. E. Clifford Williams, who had just
laboratory a beautifully small center of engineering before established a chemical engineering curriculum
research! at the University of London. Shell Development
became the leading pioneer of petrochemical process
Among recipients of the research and personnel was development—alcohols, ketones, and glycerin, and
Du Pont, which had been growing fast and diversifying aviation gasoline to boot.
since its reorganization in 1902, the year it created its
corporate engineering department and first ventured On the other hand, the Shell Group as well as Standard
into corporate research as well. Du Pont actually Oil of New Jersey (which became part of Exxon) and
purchased ammonia synthesis technology from French Anglo-Iranian Oil declined in 1930 to back Eugene J.
and Italian interests. It launched a program of nitric Houdry, a French mechanical engineer pursuing his
acid from ammonia in its Research Department in 1924. vision of catalytic rather than thermal cracking of
Staffed by Guy B. Taylor, a 1913 physical chemistry heavy crudes. Losing French government financing
Ph.D. from Princeton, Fred C. Zeisberg, a chemist despite good progress in developing his process,
turned chemical engineer expert in acid manufacture, Houdry managed to interest Vacuum Oil Company

21
(which became part of Mobil) and then Sun Oil The profession itself altered its image. The early
Company in pilot plant work, moved his efforts to American Institute of Chemical Engineers had been
Paulsboro, New Jersey, and together with his partners small, elite, and prudent in dealing with other societies
achieved in 1937 the first commercial cat cracker, a that also catered to chemical engineers, in particular
fixed-bed, regenerative unit. This development those of chemistry, electrochemistry, mining and
spawned a lot of U.S. corporate research. Applicable metallurgy, and mechanical engineering. Most of those
scientific research on catalytic cracking was led in had superior journals that attracted chemical
Russia by petroleum chemists, but from 1930 the engineering research. The older American Society of
Universal Oil Products Company, stoked by Russian Mechanical Engineers in 1930 took aim on creating a
émigré Vladimir N. Ipatieff, had pulled ahead, first new Process Industry Division that would appeal to
with what was called olefin polymerization. Universal chemical engineers. This threat and Depression-
Oil Products was owned by seven major U.S. oil generated deficits propelled the 900-member A.I.Ch.E.
companies at that time. to liberalize its stringent membership requirements,
hire a half-time executive secretary, campaign to
Standard of New Jersey itself had been slowly recruit young graduates, strengthen ties to its Student
developing German technology for hydrogenation of Chapters in universities, and improve its meeting
heavy crudes, to which it had access through programming and Transactions publication. The latter
agreements that Robert Haslam helped negotiate with I. had begun to draw some of the quality technical papers
G. Farbenindustrie, the cartel forged by Carl Bosch, and research reports that had gone to the A.C.S.’s
who was by 1925 the top executive in BASF. The Journal of Industrial and Engineering Chemistry and
Houdry process and Kellogg Engineering Company’s elsewhere. But as late as 1943 the A.I.Ch.E. was still
concept of a moving-bed cracker spurred formation of “placing major stress on the unit operations and
another development combine, Catalytic Research declining to publish articles on reaction rates and
Associates, a multinational one that included I. G. reactor engineering”.
Farben, notwithstanding the rising war clouds. The
combine had competition from Socony-Vacuum The most significant event of all was publication, a
(Mobil), which was first to commercialize a moving- year later, in 1934, of the Chemical Engineer’s
bed cracker. Moreover, it was upstaged by Standard of Handbook, the prime reference work in the McGraw-
New Jersey’s own Standard Oil Development Hill series that had begun with Walker, Lewis, and
Company, still advised by Warren Lewis and larded McAdams’ textbook. Seven years in the making, the
with former students from M.I.T. There the concept of project had been entrusted to John H. Perry of Du Pont
gas lift was borrowed from grain transport practice, and as editor-in-chief of 62 specialists drawn from industry
the concept of fluidized beds was adapted from Fritz and academia (over one-quarter of them from Du Pont
Winkler’s 1921 development in Germany of high- and no more than a handful from or associated with the
pressure coal gasification to create fluidized bed petroleum and gas industry). The Handbook’s 2600
catalytic cracking, a momentous innovation. small pages were addressed to students as well as
practicing engineers (the latter were no doubt the
Like nitrogen fixation on the eve of World War I, targets of the section on patent law and of the 10 pages
catalytic cracking had tremendous economic and on arithmetic). It codified the discipline.
geopolitical significance. Through 1943, Houdry units
and after that fluidized bed units accounted for most of Thus Charles Stine’s goal at Du Pont of getting
the aviation gasoline available to the United States and chemical engineering onto a sound scientific footing
its allies in World War Two. led first to codification. Simultaneously, the center of
mass of unit-operations research shifted southwestward
Between the wars, the growing numbers of continuous from New England toward Wilmington, Delaware,
catalytic processes — in other manufactures as well as where the exceptional research engineer, Allan P.
petroleum refining and petrochemicals — absorbed Colburn, Olaf Hougen’s prize student at Wisconsin,
more and more chemical engineers. They brought had joined the others in Tom Chilton’s team. The
incentives to focus as much on selective reactions of remarkable output of incisive papers that followed was,
flowing fluids and suspensions as on the separations according to Chilton, a by-product of improving the
and particulate solids-processing methods that company’s existing technologies and engineering new
constituted the unit operations. They also became ones, with a premium on fast and accurate process
nuclei of all sorts of opportunities for the chemical design and scale-up. Younger talent attracted to the
engineering profession. team heightened its impact on the blending of unit
operations and chemical processing that had already
started in the high-pressure process for nitric acid; with
The Profession Begins Metamorphosing that blending, a discipline better based in science,
mathematical modeling, and rigorous computational
methods emerged — to be codified, too slowly perhaps,

22
and with too little attention to chemical reactions, in So it was that although chemical engineering was
succeeding editions of the Chemical Engineer’s practiced very well within segments of German
Handbook. industry, and the discipline had been compiled in fine
scientific form, it could not propagate without an
Though it was a bellwether, Chilton’s team was, academic base, nor could a profession emerge without
needless to add, far from alone. Competition came not an industry that wanted it, or journals for its voice, or a
only from other companies in the United States, but society for its organization.
also from Germany, where scientific engineering was
already especially well established in some universities
and had the attention of at least a few companies.
Colburn, for example, had built extensively on
Industry Drives Reaction Engineering
researches by Prandtl, Nusselt, von Kármán (when he
was in Germany), Schmidt, and Jakob (still in
Germany at that time). Just how well Colburn and his Profitability of a reaction process hinges on the
American colleagues kept up with developments in conversions of reactants, the yields of products, the
Germany is not clear. A chemical engineering distribution of by-products, the mixings of the input
discipline was being compiled there too. and separations of the outputs, the recyclings of
unconverted reactants. These depend on the rates of all
Chemische Ingenieur-Technik, a 2200-page treatise the reactions — and rare is the industrial reaction that
intended as something between a textbook and a proceeds virtually alone. Rarer, too, are the reactions
reference work, appeared in 1935. Its 24 authors were that proceed without liberating or absorbing heat, and
drawn mostly from industry, a handful from technical the reaction rates, which depend on temperature as well
universities, and a couple from the more prestigious as concentrations, set the rate of heating or cooling
universities. It is certainly more scholarly than Perry’s needed. A reactor is a volume where reactants react,
Handbook, and it avoids ancillary areas like patent law, and reaction rates determine its size and its design for
accounting and cost finding, safety and fire protection, heat transfer by conduction, convection, and radiation.
and report writing. Its editor, Ernst Berl, studied with They also determine the requirements for mixing,
George Lunge at ETH and became professor of separation, and transport. Catalysis is all about rates of
technical chemistry and electrochemistry at the reaction: catalysts that preferentially speed up
Darmstadt Technical University. Leaving Germany desirable reactions (or slow down undesirable ones)
like so many others, he had become professor at can improve profitability enormously. But catalysis at
Carnegie Institute of Technology, Pittsburgh, U.S.A. solid surfaces — heterogeneous catalysis — calls for
No champion remained: German chemical engineering high-surface solids, porous or finely divided, at which
research proceeded chiefly in engineering and reaction is typically slowed down by diffusional
Technischen Chemie. resistance to arriving reactants and departing products.
Moreover, gas-phase reaction processes have, from the
An encyclopedia about physical operations in chemical first, cried out to be continuous. They have also
and related process industries, Der Chemie-Ingenieur, invited pressure, which generally enhances reaction
was far more imposing. Twelve volumes were rates of gases and lowers reactor volume needed; it can
published between 1932 and 1940 under the editorship shift equilibria advantageously; and it can lower
of Professor Arnold Eucken of Göttingen and, initially, compression and pumping costs.
Professor Max Jakob of Berlin, who soon departed for
the United States. Eight volumes treated unit From the 1920’s through the heyday of petroleum
operations on a scientific footing, with abundant refining and chemical manufacture from petroleum and
citations of sources, frequent use of mathematical tools, natural gas feedstocks, process developers sought
and sophisticated accounts of the best practice. Then reliable ways of thinking about all of these things, of
came the add-on volumes, which treated analyzing benchtop and pilot-plant data, and of
physicochemical and economic aspects of chemical designing plants. At stake were financial resources to
reaction operations, i.e., chemical process principles; commercialize newly developed processes for creating
the chapters include one by Gerhard Damköhler that is new markets or advancing in established ones. No
now celebrated in chemical reaction engineering. The doubt the advances came first in large companies that
encyclopedia’s title apparently was inspired by Walker, had Research & Development arms, like Du Pont,
Lewis, and McAdams’ text, yet Eucken was a physical Mathieson, Union Carbide, Dow, Imperial Chemical
chemist who, with I. G. Farben support, championed Industries, Standard Oils of New Jersey, Indiana, and
Verfahrenstechnik as a field of engineering science. California, Shell, Dutch Staatsminen — and in the
Indeed, it was he who in 1934 steered Damköhler into unique process invention and development firm,
the study of flow reactors. Universal Oil Products, located near Chicago.

23
While industrial chemists were searching indefatigably
for new and better catalysts, academic chemists were
taking a broader view. In the first edition (1926) of
their influential book, Catalysis in Theory and Practice,
E. K. Rideal (Cambridge University) and H. S. Taylor
(Princeton U.) touched on catalyzed gas reactions,
mentioning the reactor concepts of space velocity and
space time yield. They noted that “No satisfactory data
have been published on the design of converters
[reactors].” A. F. Benton’s (U. Virginia) seminal
Kinetics of [adsorption] catalyzed gas reactions in flow
systems appeared in 1927 (Ind. Eng’g. Chem. 19 494-
7). G. M. Schwab’s (U. Munich) Catalysis from the
Standpoint of Chemical Kinetics, the 1936 translation
of his 1931 work, explicitly discussed competing
reactions. In their 1932 Catalytic Oxidation of Organic
Compounds L. F. Marck (M.I.T.) and D. A. Hohn (Mt.
Holyoke) summarized the few reactor designs and heat
removal schemes that had appeared in patents; they
also showed representative flow sheets, some of which
had integral recycle streams. Such were the early
public glimmerings of reaction engineering.

In 1930 N. W. Krause, an alumnus of the Fixed

Nitrogen Research Laboratory who had moved to
Chemical Engineering at the University of Illinois,
published a keynote article on high pressure, high
temperature technology (Chem. Met. Eng’g. 37 529- Kenneth M. Watson
24). Two years later appeared the magnum opus from
the Laboratory, Fixed Nitrogen edited by H. A. Curtis;
it covered conversion and yield versus space velocity phase nitration of benzene (Ind. Eng’g. Chem. 28 662-
in ammonia synthesis and ammonia oxidation — and 7 1936) R. H. McKee and Richard H. Wilhelm at
much more. In 1934 appeared Pressure synthesis Columbia U. had followed up on a tantalizing German
possibility for sulfuric acid manufacture (Chem. Met. patent in what proved to be the launch of Wilhelm, a
Eng’g. 41 571-5) by the distinguished emigré Ernst leading researcher in reaction engineering — a
Berl, who had taken his Ph.D. with George Lunge. In researcher well acquainted with Damköhler’s work.
1936 from a U. Michigan chemical engineering thesis
project sponsored by UOP came Laboratory cracking A singular achievement in 1935 was The theory of
data as a basis for plant design by R. L. Huntington short-circulating in continuous flow mixing vessels in
and George Granger Brown. (Ind. Eng’g. Chem. 27 series and the kinetics of chemical reactions in such
699-707). The same year saw Catalytic Reactions at systems by R. B. MacMullin and M. Weber of
High Pressures and Temperatures, the magisterial Mathieson Alkali Works at Niagara Falls (Trans.
book by the pioneer of high pressure catalysis, A.I.Ch.E. 34 409-58. They thoroughly worked out
Vladimir Ipatieff, whom UOP had induced to emigrate continuous flow stirred tank reactors five years ahead
from the Soviet Union. For the company he had of Kirillov, nine years ahead of Denbigh, but their
discovered inter alia olefin polymerization (“catalytic impact was not visible in the literature until the 1940’s.
condensation”) and alkylation reactions that were
commercialized. The next year (1937) brought Ernst W. Thiele’s celebrated Relation between catalytic
Pressure pyrolysis of gaseous paraffin hydrocarbons activity and size of particle — lucid and thorough on
(Ind. Eng’g. Chem. 28 324-32) by Hans Tropsch, a diffusional limitation of reaction on internal surface —
pioneer whom UOP had attracted from Germany, C. L. gave the world the effectiveness factor and Thiele
Thomas, and Gus Egloff, a UOP mainstay (Tropsch modulus. Thiele’s paper had been stimulated by that of
shortly died prematurely). UOP supported a project at Dutch consultant F. G. Laupichler the year before:
Armour Research Institute (later I. I. T.) by another Catalytic water-gas reaction: mass transfer and
prominent emigré from Germany, Eucken’s co-editor catalyst activity (Ind. Eng’g. Chem. 30 578-86 1938).
Max Jakob; in his paper on Measurements of the true
temperature and heat exchange in a catalytic reaction Signs of reactor design at UOP began to appear in the
(Trans. A.I.Ch.E. 35 563-86) Jakob called attention to literature. L. S. Kassel’s Application of reaction
Damköhler’s researches. Earlier in Catalytic vapor kinetics to process design, which contained theoretical

24
analysis indicating superiority of two reactors over one
and three in a certain case, was printed in 1939 (Ind.
Eng’g. Chem. 31 276-7). An unpublished paper by
UOP’s K. M. Watson presented before the Chemical
Engineering Division of the Society for the Promotion
of Engineering Education at Pennsylvania State
College in 1939 was liberally cited in Section 5 of the
Second Edition of the Chemical Engineers’ Handbook
in 1941. Watson also called attention to the lack of
information and training in applied kinetics and reactor
design, Olaf Hougen recorded later (Ind. Eng’g. Chem.
40 556-65 1948). From the Karpov Institute in
Moscow came another sign: Kinetics of ammonia
synthesis on promoted iron catalysts [in English]
(Acta Physicochim U.R.S.S. 12 327-356 1940) in
which M. Temkin and V. Pyzhev succeeded in
matching theoretical analysis with experimental data
on the Haber-Bosch process. And from Monsanto
Chemical Corporation’s R. R. Wenner came probably
the first textbook treatment of design of tubular flow
reactors, a chapter on “Converter Design” in his 1941
Thermochemical Calculations.

Industrial Reaction Rates, a Symposium of the

Industrial Engineering Division of the American
Chemical Society in December 1942, signaled the K. G. Denbigh
future admission of chemical reactors to the chemical
engineering canon. The papers, most of them
published in 1943 in Volume 35 of Industrial & flow reactors, tubular and multiple stage continuous
Engineering Chemistry, included Solid catalysts and flow stirred tanks (quite unaware of MacMullin and
reaction rates by O. A. Hougen and K. M. Watson (U. Weber’s 1935 paper), and advantages and
Wisconsin); Principles of reactor design by D. M. disadvantages of each in cases of competing reactions:
Hurd (Du Pont?); Conduction, convection, and heat Velocity [rate] and yield in continuous systems (Trans.
release in catalytic converters [with references to Faraday, Soc. 40 352-73). Similar in origin was the
Damköhler] by R. H. Wilhelm, W. C. Johnson, and F. original and definitive treatment of conditions for
S. Acton (Princeton U.). The chair of the organizers, multiple steady states in continuous flow stirred tank
Olaf Hougen, wrote that “despite the interest of reactors with exothermic reaction, by Carl Wagner in
chemical engineers in the release of industrial data on Germany (Technische Hochschule Darmstadt) in 1945:
the applications of reaction rate principles to reactor On the temperature control in highest performance
design, the committee soon found that contributions catalytic reactors [in German] (Die Chemische
were not readily forthcoming . . . [There were also] Technik 18 28-34). More academic in origin was H. M.
last-minute withdrawals . . . In an early postwar year Hulburt’s pair of 1945 papers (from Princeton U.)
another symposium on the same subject should be establishing general equations and their dimensional
timely . . . Many of the customary repeated and analysis for tubular reactors in general and for highly
extensive pilot plant investigations might be eliminated idealized instances of catalytic reaction — and due
by more complete understanding of the principles appreciation of Damköhler and Thiele: Chemical
involved.” A few slightly earlier industrial papers had reactions in continuous flow systems. Reaction kinetics.
released bits of data, for example two in 1940: Heterogeneous reactions (Ind. Eng’g. Chem. 36 1012-
Catalytic alkylation of isobutane with gaseous olefins 7 and 37 1063-9).
(Ind. Eng’g. Chem. 32 328-30) by F. H. Blunck and D.
R. Carmody (Standard Oil Co. of Indiana) recorded Industrial papers of the sort sought by Hougen’s
some data on conversion and distribution of products, committee finally appeared in 1948, notably Reactor
and Hydrogenation of petroleum (Ind. Eng’g. Chem. design for manufacture of toluene by catalytic
32 1203-12) by E. V. Murphree, C. L. Brown, and E. J. reforming (Chem. Eng’g. Progress [A. I. Ch. E.] 44
Gohr (Standard Oil Development Co. of New Jersey 195-200) by A. A. Burton et. al. (California Research
and Louisiana) listed some data on yields and quality. Corp. — Standard Oil Co. of California) and Catalytic
War production stimulated a 1944 masterpiece by K. G. dehydrogenation of ethylbenzene (Chem. Eng’g. Prog.
Denbigh in England (Ministry of Supply and 44 275-86). Both were unprecedentedly full reports of
Southhampton U.), in which he analyzed batch and pilot plant data, one including scale-up issues and

25
commercial plant data, the other discussing the method
of designing a commercial plant. Both drew interesting
published discussion. Both probably reported work
done some years before. As it happened, they were
preceded by seminal publications on reactor design.

In 1946 a series of four articles, Principles of reactor

design, authored by K. M. Watson (U. Wisconsin) and,
respectively, C. O. Reiser, P. S. Myers, R. H. Dodd,
and L. N. Johnson, that had appeared in the National
Petroleum News Technical Section, was reprinted and
circulated. It was soon quoted by Hougen & Watson
for their book and later by J. M. Smith for his, and so
had a strong influence. An industrial paper on design
of tubular flow reactors, or “coils,” with substantial
pressure drop was authored by J. H. Hirsch and
collaborators at Gulf Research & Development
Corporation: Projection of laboratory reaction
velocity [rate] data into commercial designs (Ind.
Eng’g. Chem. 38 885-90). In 1947 from Imperial
Chemical Industries in the U. K. came K. G. Denbigh’s
Continuous reactions. Part II. The kinetics of steady
state polymerization (Trans. Faraday Soc. 43 648-60),
which opened up the area of polymerization reactor
analysis and design. And from O. A. Hougen and K.
M. Watson at the University of Wisconsin came Part
III of their Chemical Process Principles volumes, Joseph Mauk Smith
namely Kinetics and Catalysis, the first textbook to
systematize analysis and design of plug-flow tubular Kinetics of gas-liquid reactions (Recueil des travaux
reactors. Though they totally missed continuous flow chimiques des Pays-Bas 67 563-99 1948). D. R.
stirred tank reactors, their book was without peer for Mason and Edgar L. Piret (U. Minnesota) with DuPont
instruction in applied kinetics and tubular reactor support inquired into transient behavior, in Continuous
design for nearly a decade, until 1956. stirred tank reactor systems . . . transient equations,
applications, and experimental confirmation (Ind.
In that interval the principles of reaction engineering Eng’g. Chem. 43 1210-19 1951). C. van Heerden
were better and better understood, as exemplified by a (Netherlands State Mines), apparently unaware of
remarkable monograph and four papers. The Wagner’s 1945 treatment, analyzed multiple steady
monograph was D. A. Frank-Kamenetzkii’s 1947 states in steady flow reactors with exothermic reactions,
Diffusion and Heat Exchange in Chemical Kinetics in what became a celebrated paper: Autothermic
from the Academy of Sciences in Moscow, finally processes: properties and reactor design (Ind. Eng’g.
translated into English in 1955 at R. H. Wilhelm’s Chem. 45 1242-7 1953). Another significant thing was
instigation. Focused on combustion, flames, and the appearance of A. A. Frost and R. G. Pearson’s
explosions, it covered basics of parallel and (Northwestern U.) textbook, Kinetics and Mechanism
consecutive reactions, oscillating reactions, (1953), which covered only homogeneous chemical
temperature rise and runaway in reactors, and reactions reactions but had a chapter on complex reactions and
at combustible and catalytic surfaces. Wilhelm et al. flow methods that appealed to some influential
(Princeton U.) published analog computer solutions of chemical engineers of the day.
governing equations and results of experiments, in
Reaction rate, heat transfer, and temperature In 1956 Chemical Engineering Kinetics, with chapters
distribution in fixed-bed catalytic converters (Chem. on reactor design, was published by Joe M. Smith, a
Eng’g. Prog. 44 105-16 1948). R. B. MacMullin young professor at Purdue University. The focus of
(Consultant) echoed the 1935 paper with one on Smith’s research and teaching had shifted from
consecutive reactions in multi-stage CFSTR’s: chemical engineering thermodynamics, in which he
Distribution of reaction products in benzene had coauthored a long-lived textbook, to reaction
chlorination: batch vs. continuous process procedures engineering, in which he had industrial experience at
(Chem. Eng’g. Prog. 44 183-8 1948). Dirk W. van Monsanto, Chevron, and Texaco going back to before
Krevelen, P. J. Hoftyzer and C. J. van Hooren 1945. His teaching impelled him to a comprehensive
(Netherlands State Mines) addressed, by experiment view of the subject. His was the first textbook of
and theory, gas absorption with chemical reaction in chemical reaction engineering. It emphasized reactor

26
design. It was widely adopted for undergraduate Admission of Chemical Reactors to the Canon
courses in “applied kinetics,” “reactor design,” or
“reaction engineering.” It arose in the same What signaled admission of chemical reactors and
department that housed R. Norris Shreve, who was chemical reaction engineering to the canon of chemical
stuck on “unit processes” — a striking coincidence. engineering? They became prominent in the Symposia
and recognition awards of the American Institute of
That the concept Smith developed was taking hold Chemical Engineers and the European Federation of
elsewhere was made plain by Chemical Reaction Chemical Engineering, which was formed after 1950.
Engineering: First European Symposium on Chemical They appeared in required courses of the chemical
Engineering, held in Amsterdam in 1957. It was a engineering curricula of more and more universities,
well-staged event with papers from the Netherlands, until finally the Accreditation Standards caught up and
Germany, England, and France that were afterward demanded them. They were integrated with applicable
printed as a special supplement to the initially chemical kinetics, explained in quantitative terms, and
Eurocentric young journal Chemical Engineering applied to realistic problems of design and operation in
Science. The Symposium and its sequels in 1960 and course-derived textbooks beginning with Olaf
1964 established chemical reaction engineering in Hougen’s and Kenneth Watson’s seminal Chemical
Europe and helped solidify it in the United States. Process Principles. III. Kinetics and Catalysis from
Wisconsin (and UOP) in 1947. Then came the
As just recounted, chemical reactors and reaction landmark: Joe Smith’s Chemical Engineering Kinetics
engineering took a couple of decades to work their way from Purdue in 1956. It was followed by Walter
on stage as applied chemical kinetics and practical Brötz’s Fundamentals of Chemical Reaction
catalysis, and thence toward the front and center of Engineering (in German, later translated into English)
technology, the discipline, and the profession. It is from Ruhrchemie AG and Technische Hochschule
worth noting that the ratio of significant publications Aachen in 1958, and Stanley Walas’s Reaction
by industrial authors to those by academic ones seems Kinetics for Chemical Engineers from Nofsinger Co.
about as in the foregoing selection. That the proportion and University of Kansas in 1959. A transatlantic
of outright chemists fell, and of chemical engineers flood ensued in the 1960’s — texts by Kramers,
rose with time is surely accurate. So is the indication Denbigh, Aris, Levenspiel, et al. The number of these
of shift of publication from the A.C.S. to the A.I.Ch.E. books that grew out of an author’s grappling with
journals. In 1942 the A.I.Ch.E. actually declined to analysis and design of industrial reactors is noteworthy.
hold the Industrial Reaction Rates Symposium that
proceeded instead under A.C.S. auspices. (MacMullin By the end of the 1960’s chemical engineering fully
and Weber’s 1935 article in Trans. A.I.Ch.E. was an accorded with George Davis’s eighty-year-old vision
anomaly.) Many of the academic authors had already of application of chemistry, physics, and mechanics to
served in process development trenches of industry. the utilization of chemical reactions on a large scale.
Some were actively encouraged by industrial people. Moreover the reactions’ reactors and their associated
Most notable of the lot was Ken Watson, who in 1932 recycles and purges had come to be clearly viewed as a
had left the University of Wisconsin for UOP, there chemical process’s hub on which turn all the
pioneered tubular reactor analysis, design, and scale-up, separations and other unit operations.
and — after his 1939 challenge to academics to take
these things up — in 1942 returned part-time to The ultimate imprimatur was the upgrading of reactors
Wisconsin for his wartime service in designing and and reaction engineering to the status of full section in
constructing plants for the U.S. government-driven the Fourth Edition of the Chemical Engineers’
synthetic rubber production. In those plants and many Handbook, published in 1963. They shared the section
others that were built under enormous time pressure for with not unrelated Thermodynamics, but they got top
war production, chemical and petroleum companies billing. One of the authors of their part was Robert H.
had to go from sparse laboratory data to full-scale Perry, coeditor of the Fourth Edition and son of Du
plants without the usual progression of pilot-plant and Pont’s John H. Perry, editor of the first three editions.
semi-works stages. Reactors had to be designed. A In the very first edition, in 1934, the older Perry had
by-product, as anticipated by Olaf Hougen, partly on himself coauthored a short subsection on Reaction
Watson’s experience, was greater willingness among Rates, including a paragraph on gas-phase flow
companies to publish results of kinetic studies and reaction and a couple on the prevalence of
pilot-plant trials — and to promote the new simultaneous and sequential reactions in practice.
subdiscipline. Perry had been certain that at least a peek at the heart
of chemical engineering had to appear in a handbook
dominated by the body of unit operations.

The fluidized bed catalytic cracker was invented and

developed at Standard Oil of New Jersey sites early in

27
World War Two. Already in the pages of the Acknowledgements
Handbook’s Third Edition, in 1950, it had appeared not
as the magnificent reactor that it is, but as a relative of This text is drawn partly from my article, “On the
continuously fed process furnaces and an example of Emergence and Evolution of Chemical Engineering,”
gas-solids contacting. In fact it epitomized the modern pp. 1-40 in Perspectives in Chemical Engineering,
chemical reactor. Neal Amundson on receiving the edited by C. K. Colton, Academic Press, New York,
National Academy of Engineering’s Founder’s Award 1991 (Advances in Chemical Engineering, Volume 16)
a decade ago pointed out that such a chemical reactor and partly from subsequent study and reflection. I am
“may be an extremely complicated device, indebted to Maribeth R. Mrozek (then a senior in
substantially more complex than a nuclear reactor . . . chemical engineering, since a development engineer,
The size of some . . . is awesome.” Indeed in 2003 and now a law student) for discerning assistance in
there are fluidized bed cat crackers processing 100,000 locating and assessing sources; and to John T. Lund
or more barrels per day, and circulating particulate (currently a senior) for unstinting help in tracking
catalyst between reactor and regenerator at rates of down materials.
3000 to 4000 tons per hour. The understanding, design
and operation of processes based on such reactors is a Note:
major success of our discipline, profession, and What I see as the basic questions about what
industries. constitutes the discipline of chemical engineering, what
maintains the associated profession, and how both
relate to the industries and technologies served are
stated in the closing of the 1991 article cited above.

28
Unsolved Problems in Fluid Mechanics
 Biographical sketch of Howard Brenner

Born and raised in New York City, Howard Brenner received his Bachelor's degree in Chemical Engineering
from Pratt Institute in 1950 and D.Eng.Sci. from New York University in 1957. His 48-year career as a
chemical engineering faculty member includes New York University (1955-1966), Carnegie-Mellon
University (1966-1977), the University of Rochester --- as Departmental Chair (1977-1981) and, since 1981,
MIT, where he is currently W.H. Dow Professor. Brenner has co-authored three fluid dynamics and transport
phenomena books, namely "Low Reynolds Number Hydrodynamics (1965)," "Interfacial Transport
Processes and Rheology (1991), " and "Macrotransport Processes (1993)." Honors include the American
Institute of Chemical Engineer's "Lewis," "Walker," and "Alpha Chi Sigma" Awards, the American Society for
Engineering Education's "Senior Research Gold Medal Award, " the American Chemical Society's Kendall
Award in "Colloid and Interface Science," the "Bingham Medal" of the Society of Rheology, and the
American Physical Society's "Fluid Dynamics Prize." Brenner holds membership in the National Academy of
Sciences, the National Academy of Engineering, and the American Academy of Arts & Sciences. Lifelong
research interests focus on hydodynamics and transport processes in fluid-particle systems --- addressing
both fundamentals and applications. He has published about 250 papers in these fields. Most recently, his
work has shown that the classical notions of fluid-mechanical transport processes in systems involving mass
density gradients, dating back to the founders of the subject, including Euler, Navier, Stokes and Fourier, are
patently wrong on principle. Concomitantly, Brenner's discovery and quantification of the notion that volume
can be transported diffusively, i.e. molecularly (above and beyond the purely convective transport of volume
accompanying the movement of mass) has been used to modify the basic momentum and energy transport
equations in fluid continua, thereby correcting the works of the above-cited authors. He is currently working
on a transport phenomena/continuum mechanics monograph encompassing these modifications.
 Unsolved Problems in Fluid Mechanics:
On the Historical Misconception of Fluid Velocity as Mass Motion, Rather than Volume Motion

Howard Brenner
Department of Chemical Engineering
Massachusetts Institute of Technology
Cambridge, MA 02139
[email protected]

Abstract
Existing experimental data on the thermophoretic velocity of a small, rigid, non-Brownian particle through an
otherwise quiescent gaseous continuum, when re-interpreted as representing the motion of a passive tracer entrained
in a moving fluid, reveals that macroscopic fluid movement (motion) can occur purely diffusively, by the movement
of volume, without a concomitant (convective) movement of mass1,2. This experimental fact negates Euler's 250-
year old generic, mass-based definition3 of the velocity field in fluid continua, undermining thereby the heretofore
seemingly rational foundations4,5 of fluid mechanics and derivative subjects. This, in turn, requires a fundamental re-
formulation of the basic equations of fluid mechanics6,7 as well as of molecular theories8,9 of transport processes10,11
in fluid continua. This detailed re-fashioning, which is effected elsewhere1, is based upon recognizing that volume
can be transported purely diffusively12, representing, inter alia, a previously unrecognized mechanism for
momentum and energy transport in fluids.
As a consequence of this unanticipated phenomenon,
Let me begin by congratulating the Ohio State coupled with my own tenacity and sheer brilliance (not
Chemical Engineering Department on the occasion of
its 100th anniversary. I am flattered to have been asked to mention an obvious lack of modesty), I stand before
to participate in the celebratory exercises surrounding you today as a failure, a man unable to honestly present
this happy occasion. May the Department continue to a talk entitled: "Unsolved Problems in Chemical
successfully serve the needs of its students, its faculty, Engineering." I will, therefore, instead, present a lecture
the University, the chemical engineering profession, entitled: "A Previously Unsolved Problem in Chemical
and the community of scholars at large. Engineering," more pretentiously subtitled: "On the
Historical Misconception of Fluid Velocity as Mass
Like several other invitees on this occasion, I was asked Motion, Rather than Volume Motion." The formal
to emote on the subject of "Unsolved Problems in lecture notes pertaining to this topic follow
Chemical Engineering," specifically in the general area subsequently, after the overview of the general topic
of Fluid Mechanics. However, having succeeded in given immediately below.
identifying a problem crying out for solution, how is an
indefatigable researcher expected to prevent his mind Overview of the Problem
from attempting to solve the very problem or problems
According to all standard works and research
that he/she has identified as being unsolved? And woe
publications on fluid mechanics, the velocity v at a
to the speaker who actually succeeds in solving these.
point of a fluid continuum is governed by the
For this very action will render him incapable of
movement of mass through a hypothetical surface fixed
presenting his lecture entitled "Unsolved Problems in
in space, as witness the first appearance of the symbol
Chemical Engineering," leaving him with the
v for velocity in the well-known continuity equation of
unpleasant alternative of identifying and writing about
fluid mechanics, ∂ρ / ∂t + ∇ • ( ρv m ) = 0 . On the other
yet another unsolved problem, and so on, ad infinitum,
ad nauseum. This is a logical conundrum, one that I hand, the velocity of a material object is measured
strongly suspect the planners of this lecture series failed experimentally by tracking its temporal movement
to anticipate. Obviously, they failed to reckon with through space. (Anyone who as ever thrown a ball to a
Heisenberg's uncertainty principle, whereby the mere catcher will recognize that this is, indeed, the way to
act of observing a system alters the state of that system measure velocity.) Adapting this scheme to the
--- the system here being the "unsolved problem" that measurement of fluid velocity, an object (a "tracer") —
one was asked to identify and discourse upon. which is sufficiently small such as to not disturb the
pre-existing fluid motion into which it is introduced —
can be visually tracked in time so as to the monitor the
undisturbed fluid movement serving to convect the

31
passive tracer through space; that is, the tracer merely
renders visible the otherwise invisible continuum fluid
motion existing its absence, without itself affecting that
very movement! But where is the experimental and/or
theoretical proof that the tracer and mass velocities are
indeed the same?

This is (or rather was) the "unsolved problem" with

which I began to prepare my talk. Based upon
seemingly unequivocal experimental and theoretical
grounds, I arrived at the surprising conclusion that these
two velocities are, in well-defined circumstances,
generally unequal. Indeed, in certain circumstances,
namely in the case of phoretic phenomena (e.g.
thermophoresis), involving the motion of a small, non-
Brownian particle through a bounded, single-
component fluid under the influence of a temperature
gradient (in the absence of gravity effects), no
continuum mass motion whatsoever exists at any point
of the fluid, as is easily verified from the Navier-Stokes,
continuity, and energy equations, together with the
thermal equation of state for the fluid. Nevertheless, a
thermophoretic particle, one animated by an externally-
imposed fluid temperature gradient, is observed to
move through the fluid from high to low temperature.
And this particle qualifies as a "tracer," since its
velocity is (observed experimentally to be) independent
of its size, thereby enabling us to regard it as an James Clerk Maxwell
effectively point-size object! This constitutes a
physically authenticated situation in which the mass
and tracer velocities of the fluid differ. The standard principles currently assumed to govern all fluid-
explanation for the phenomenon of thermophoretic mechanical transport phenomena.
motion (at least in gases), dating back constitutively to
Maxwell in 1879, and consistent with the particle's size- Has Fundamental Research in Transport
independence, argues that the phenomenon arises from Phenomena Been Ossified By Its Own Success
non-continuum fluid-mechanical effects existing near
the surface of the particle, and involving a violation of During the roughly 30-year period beginning in the
the no-slip velocity condition at the particle surface. mid-1950's, chemical engineering research moved
This, despite the fact that the Knudsen numbers (mean- heavily into the area of engineering science. This effort
free path to particle size) characterizing the fluid was largely characterized by an intense focus on
motion were vanishingly small during those transport phenomena, including mass, species, energy,
experiments for which particle-size independence of the and momentum transport processes. Much of the
thermophoretic velocity was observed. codification of knowledge in the field at the beginning
of that era was embodied in the classic 1960 textbook
We offer here an alternative, strictly continuum, non- by Bird, Stewart and Lightfoot, which, after a lapse of
slip explanation of phoretic phenomena, albeit one 40 years, was brought up to date by the recent
based upon a major modification of the Navier-Stokes publication of a second, thoroughly revised edition10,
equations governing the fluid's convective and diffusive co-authored by these same authors, a very singular
momentum transport processes in continua, wherein it accomplishment given the time lapse. The seemingly
is claimed that the mass velocity appearing in the classical status of this the subject would lead one to
continuity equation is not, in fact, the velocity of the believe that no outstanding problems of a truly
fluid continuum (as measured by a tracer)! The fundamental nature remain to be resolved in this field.
consequences of the proposed general velocity Undoubtedly, much is yet to be accomplished in terms
inequality are profound, in that they undermine the of actually solving the pertinent transport equations in
basic fluid-mechanical and chemical engineering the context of specific applications, an activity largely
transport principles that we all learned in school, delegated nowadays to computers. Even in the realm of

32
establishing the phenomenological coefficients entering
into relevant constitutive equations, the action has
largely moved from experiment towards purely
statistical-mechanical computation of these parameters.
The prevailing view is that, in terms of fundamentals,
basic research in transport phenomena, at both the
continuum and molecular levels, belongs to the ages.
This lecture aims to negate this impression by
identifying major problems in the field, as well as
providing a prescription for their resolution.

Specifically, a recent publication1 casts doubt upon the

comfortable perspective of a field ossified by its own
success. In particular, it appears to me, as well as to
several others with whom I have shared my concerns,
that a major flaw exists in the fundamental conceptions
underlying transport phenomena, traceable back to the
physical interpretation to be placed upon the velocity, v ,
existing at a point of the fluid continuum. Given that
this velocity serves to distinguish convective transport
from diffusive transport, any problems connected with
the proper identification of v automatically spill over
from the continuum into the molecular realm. Even
more generally, to the extent that a problem exists in the
transport phenomena field, comparable issues
necessarily obtain in the related fields of continuum and Leonhard Euler
statistical mechanics, including such subfields as
rheology and irreversible thermodynamics. The sole
publication1 claiming to undermine the foundations of
transport processes is too new to have been widely 250 years following its introduction by Euler in 1755.
studied, much less accepted by the engineering and At the same time, the lecture reinforces the general
scientific research communities. Nevertheless, need for performing critical experiments in the
irrespective of its present status, its arguments point up transport phenomena field, or indeed in any field of
non-trivial questions of a fundamental nature. Explicitly, research, an issue which in today's computer age would
the currently unresolved issues facing the transport and surely appear of little import to most young researchers
fluid-mechanical communities are two-fold: (i) Is the (although obviously still relevant in areas explicitly
proposed new theory correct; and (ii) if so, what are the recognized as a being in a state of flux). In any event, a
limitations of its validity, given that it incorporates quick study of the new theory would almost create the
several idealizations? Irrespective of the answer(s) to impression that such mid-course "corrections" to one's
these questions, the exercise of attempting to provide understanding of transport phenomena are important
rational answers thereto is bound to be informative, only in specialized circumstances, such as arise, for
certainly to its discussants, and hopefully to a wider example, in connection with phoretic phenomena, e.g.
audience as well. thermophoresis or diffusiophoresis, where the pertinent
Reynolds number is quite small owing largely to the
On the assumption that the first question is answered in relatively small sizes of the aerosol and hydrosol
the affirmative, in the sense of recognizing the particles involved. However, given the current focus on
existence of a basic flaw in our present understanding small-scale technological processes, including
of the continuum fluid velocity v , these written notes nanotechnology, microfluidics, and cellular biological
⎯ which are designed to supplement the main verbal transport phenomena, entailing low Reynolds number
lecture, which is more technical in scope ⎯ serve to flows resulting from the small particle and/or conduit
identify what would appear to be the historical source sizes encountered, we believe that this newly-altered
of the misconception. The latter deals essentially with view of transport processes will ultimately result in
understanding the intellectual background that gave rise novel applications to pertinent physical and biological
to the error in the first place, and which, subsequently, phenomena. It is with these and related fields in mind
prevented recognition of the logical flaw during the past that we formally address the following topic.

33
 centrally-symmetric forces13. In such circumstances,
and for strictly dynamical reasons arising from the fact
that, in the case of point masses, momentum is simply
mass times velocity, the velocity of choice for
characterizing the motion of a rigid body came to be
specified in terms of the movement of its center of mass,
a point within the body at which its entire mass is
regarded as effectively concentrated.

Thus began the intimate association of velocity with

mass, a fraternization subsequently adopted by Euler3,
the "father"4,5 of continuum fluid mechanics, when, in
1755, he derived the so-called continuity equation6,7,10,11,
a purely kinematical field relation, expressing the law
of conservation of mass at a point within a fluid
continuum in terms of the "velocity," v m , of the fluid's
mass movement. Fluids are, however, deformable
rather than rigid. As such, the physical identification of
fluid velocity with the movement of mass, so central to
rigid-body mechanics, loses its dynamical raison de
etre in the case of fluids, retaining only a kinematical,
purely mass-conservation, rationale for its appearance
therein. Failure to appreciate this fundamentally altered
role played by mass movement in fluids compared with
its role in rigid-body dynamics, along with the failure to
recognize that the fluid's Lagrangian or tracer velocity,
v l , may differ from its Eulerian or mass-based velocity,
v m , has led to a dynamical and energetic
Isaac Newton misinterpretation of fluid-mechanical phenomena1.
Demonstration, by experiment, of a difference between
v m and v l constitutes the focus of this lecture, with
The primitive notion of the (vector) velocity of a details provided elsewhere1 of the impact of this finding
material body, namely the directed distance traveled upon the correct equations of fluid mechanics and
through space divided by the time required to traverse derivative subjects. In the latter scheme, v l rather than
that distance, is one of the most important concepts in v m is found to constitute the fundamentally correct
our arsenal of theoretical tools required for a
continuum fluid velocity, not only in terms of the
quantitative description of physical phenomena. For
primitive notion of "motion," a purely kinematical
example, in the case of rigid bodies moving through
conception, but also, more importantly, in terms of the
empty space, the temporal rate of change of this
fluid's specific momentum density, kinetic energy, and
velocity, namely the body's acceleration, plays a
other attributes associated with the literal movement of
fundamental role in determining the forces acting upon
corporeal bodies through space, representing dynamical
the body, and conversely. As such, a clear
and energetic notions.
understanding of how velocity is to be measured
experimentally is central to any dynamical description
In creating the subject of continuum fluid mechanics
of the behavior of physical systems, rigid or otherwise.
using rigid-body mechanics as a model, Euler3,4
In the simplest case, involving the motion of a rigid
introduced the notion of mass per unit volume, namely
body moving in vacuo, any potential ambiguities
the mass density field, ρ ≡ ρ (x,t ) , at each point
regarding the definition of its velocity were removed at
the outset by Newton (actually by Euler), who began by x ≡ (x, y, z ) of the fluid continuum and at each instant
initially focusing attention on the motion of a of time t , an experimentally measurable fluid property.
hypothetical mass point, an abstraction. Subsequently, Additionally, he adapted to fluid continua the center-of-
rigid bodies of finite extent were addressed by mass velocity concept associated with rigid bodies, by
regarding them as being composed of a finite collection defining the fluid velocity field, v ≡ v(x, t) (our v m ),
of such idealized point masses, permanently joined such that v: = n m / ρ , where the mass flux density (or
together, with its members interacting through current), n m , represents the experimentally measured

34
mass of fluid per unit time instantaneously crossing a gravitational and inertial mass to be one and the same
space-fixed unit area centered at x at time t . (In fact, it entity). To the best of the author's knowledge, this view
would appear that n m cannot actually be measured has gone unchallenged for the past 250 years. Indeed, so
directly, although the truth of this statement does not pervasive is the universal acceptance of this equality
impact upon the issues addressed here.) This definition that but a single symbol, typically v , is employed to
of velocity, which involves experimentally monitoring denote both velocities. Nevertheless, on the basis of
the movement of mass at a point (and concomitantly both theory and experiment, this assumption was
measuring the fluid density at that point), appears, recently shown1 to be invalid in circumstances where
superficially, to be similar to, if not identical with, the density gradients, ∇ρ , exist within the fluid as a
Newtonian center-of-mass definition of the velocity of consequence of either composition gradients in
a moving body — the "material body"4,5 in this case multicomponent fluid mixtures undergoing mass
being the (differential) element of mass that one is transfer or temperature gradients in single-component
tracking as it crosses the (differential) area centered at fluids undergoing heat transfer — that is, in
x . This experimental protocol differs from the usual compressible, molecularly inhomogeneous gases and
scheme for measuring the velocity of a material object, liquids.
whereby one simply tracks the latter by monitoring its
trajectory through space, without regard to the mass of Internally consistent, tripartite evidence exists1 for the
the object being tracked. (Optically "tracking" the general velocity inequality, v m ≠ vl , based separately
statistically-averaged movement of, say, a small group and collectively upon quantitative arguments involving:
of photochromically-labeled or otherwise tagged (i) existing thermophoretic2 and diffusiophoretic15
molecules of the fluid does not qualify as being experimental data; (ii) a revised version of the current
isomorphic with a tracer velocity measurement. continuum theory of mass, heat, and momentum
Explicitly, a collection of molecules is not equipollent transport in fluids10,11, the revision thereto originating
with a material tracer, the latter being be a single, rigid, with the recognition that volume can be transported
corporeal entity.) diffusively12, independently of the movement of mass;
and (iii) molecular theory results available in the
It is with Euler's mass-based definition of fluid velocity literature, based upon the well-known Burnett non-
that we take issue. In particular, a differential mass continuum additions to the original Chapman-Enskog
element (explicitly a so-called differential material fluid continuum solutions of the Boltzmann equation9. In
particle) is neither isomorphic to nor equipollent with a support of the velocity inequality hypothesis we confine
corporeal tracer, in the sense that individual molecules, attention in this lecture exclusively to the experimental
each possessing its own individual properties (namely evidence furnished in connection with item (i).
mass and velocity), are free to enter and leave the mass
element as the latter moves deterministically through Thermophoresis is a phenomenon whereby a small,
space. As such, whereas the total amount of mass unrestrained, essentially weightless and otherwise
contained therein remains fixed during the material force-free particle suspended in a single-component
fluid particle's movement through space, this mass does isobaric fluid (usually a gas), within which a steady,
not consist permanently of the same molecules (i.e., the essentially homogeneous, temperature gradient exists,
same "matter"). Accordingly, despite the constancy of is observed to move from regions of high to low
its total mass, this differential material "particle" differs temperature16-19. Equivalently, a tethered particle
fundamentally from the mass point of Newton's rigid- experiences a force20 tending to move it towards the
body mechanics. Moreover, in an energetic sense, a low temperature region. Thermophoretic forces exerted
material fluid particle constitutes a molecularly "open" on aerosol particles were first recognized by Tyndall21
(rather than "closed") system as a result of the ability of in 1870 when he observed the presence of dust-free
individual molecules to freely cross its boundaries in regions proximate to hot surfaces in a dust-filled room.
either direction. As such, the First and Second laws of A non-continuum "explanation" of the basic physics
thermodynamics are not directly applicable to this underlying thermophoretic phenomena, one still
system without modification14. invoked today, was provided by Maxwell22 in 1879
(and, independently, by Reynolds23 in that same year)
Despite the fact that the mass-based definition of when he offered an analysis of the workings of
velocity, v m , differs from the primitive physical notion Crookes' radiometer24, the latter device first exhibited
of velocity, as embodied in the fluid's tracer velocity, v l , publicly in 1873. Maxwell's molecularly-based
Euler nevertheless implicitly assumed these two explanation invokes the hypothesis of a thin, non-
velocities to be one and the same entity (in, say, much continuum, Knudsen boundary layer (only later so-
the same spirit as Newton earlier hypothesized named) existing in the immediate neighborhood of a
solid body bathed by a gas of non-uniform temperature.

35
(This Knudsen layer is assumed to exist even when the thermophoretic particle immersed in the fluid
particle size-based Knudsen number is sufficiently continuum is observed to move with velocity16-18
small such as to expect purely continuum behavior.)
This non-continuum behavior is presumed to result in Cs
U=− υ∇ ln T , (1)
tangential "slip" of the fluid "velocity" v (i.e., v m ) 1 + (k s / 2k )
along the surface of the non-isothermal body in a
direction opposite to that of the surface temperature where υ and k , are, respectively, the fluid's kinematic
gradient, giving rise to thermal stresses of viscosity and thermal conductivity, and T is the
hydrodynamic origin exerted on the body, urging the absolute temperature of the gas; C s , which is of O(1) ,
latter towards regions of diminishing bulk fluid is Maxwell's (dimensionless) thermal stress slip
temperature. coefficient, whose "best fit"26 experimental value is
about 1.2 (slightly larger than Maxwell's original
Maxwell's "thermal stress," cum slip, explanation of the molecular theory-based estimate of 3/4); k s is the
general phenomenon underlies all contemporary
particle's thermal conductivity27. In this limit, such a
theories of thermophoresis in gases25. Elsewhere1,2,
zero-size particle serves as a "tracer" of the
however, we argue against this supposed non-
undisturbed, particle-free fluid movement through
continuum, Knudsen slip-layer explanation, suggesting
space (at least in circumstances where the particle's
instead an alternative, purely continuum, "no-slip"
thermal conductivity is small compared with that of the
velocity condition imposed upon v l , accompanied by
gas, k s /k << 1 , corresponding to an effectively
appropriate modifications of the basic equations of
thermally insulated particle, the latter property
continuum fluid mechanics, wherein v l replaces v m in
rendering the tracer "inert"1 with respect to its
all instances involving the explicit dynamical notion of physicochemical interaction with the fluid). Explicitly,
physical motion through space (as, for example, in the the tracer velocity, v l , of the undisturbed non-
case of the fluid's momentum density or kinetic energy).
isothermal fluid is simply equal to U . Inasmuch as
On the other hand, the mass "velocity," v m , is retained
v m = 0 and v l = U ≠ 0 , it follows, ipso facto, that
in those circumstances where the issue is purely
v m ≠ vl . That the intrinsic source of the fluid "motion"
kinematical, for example when following the contents
of a mass element (i.e., a material fluid particle) as it v l arises from a temperature gradient is irrelevant to
moves through space. the objective experimental measurement of the fluid's
velocity. In fact, it is actually the density gradient rather
Irrespective of the correct theoretical explanation of the than the temperature gradient that constitutes the
mechanism underlying thermophoretic motion, the animating force underlying such motion, as confirmed
existing correlation16-18 of empirical experimental data by comparable diffusiophoretic experiments15, as well
remains viable, and, as such, will be seen to provide an as by the general theory of phoretic phenomena1.
objective test of Euler's v m / v l velocity equality
hypothesis. The main attribute of these data impacting Written in a more objective, observer-invariant form,
upon the issue of the fluid's velocity (namely the the experimental tracer velocity, and hence the
velocity of the undisturbed, particle-free, fluid) lies in undisturbed fluid velocity, in the case of gases is thus
the experimentally-observed effect of particle size on found from eq. (1) to be
thermophoretic velocity in the case where the spherical
particle is sufficiently large such as to exhibit no v l − v m = −C sυ∇ ln T , (2)
sensible Brownian movement, and yet is sufficiently
small such that the temperature varies only with v l and v m each measured relative to the same
imperceptibly over its surface. Experiments16-18 reference frame. By way of comparison with the
performed under circumstances where the undisturbed experimental result (2), our theoretical expression1,2 for
fluid motion in the particle's absence is everywhere this velocity disparity, valid for both gases and liquids,
identically zero ( v m = 0 ) reveal that the particle's is
thermophoretic velocity, U , say, is independent of
particle size in the continuum region, where the gas's v l − v m = −αβ ∇T , (3)
mean-free path is small compared with the particle's
−1
radius (corresponding to an effectively zero Knudsen where α = k/ ρ cˆ p and β = vˆ (∂vˆ / ∂T ) p are,
number, continuum phenomenon). respectively, the fluid's thermometric diffusivity and
isobaric thermal expansivity, with cˆ p the specific-heat
Explicitly, in the hypothetical (extrapolated) limit of
"zero" particle size, a (non-Brownian) spherical capacity at constant pressure and vˆ = 1/ ρ the specific

36
volume. In the case of ideal gases8,10, β = 1/ T and density ρ , especially in the case of liquids. However,
α = (1/ 4)(9 − 5γ −1 )υ , where γ = cˆ p / cˆv is the specific like energy, momentum, or entropy, volume is not a
heat ratio (possessing the value 5/3 for monatomic material property that can, itself, be directly visualized.
gases and 7/5 for diatomic gases). Upon introducing Rather, volume is abstract rather than corporeal, and
these values into (3), the resulting expression is seen to therein lies the difficulty of conceptualizing its
agree, both constitutively and phenomenologically, movement through space.
with the experimentally-observed undisturbed fluid From a transport point of view, individual molecules
velocity result cited in (2). In addition to this accord for carry with them such extensive properties as mass,
the gaseous case, the theoretical formula (3) also agrees momentum, and (kinetic) energy, as they move about
satisfactorily1,2 with the limited experimental through space. But they do not themselves, as
thermophoretic particle data available for liquids28. individual entities, convey more abstract extensive
properties, such as volume (and, say, entropy), since
Whereas traditional continuum transport equations10 such extensive properties are collective in nature, rather
governing steady-state heat transfer indicate the than being associated with individual molecules.
particle-free fluid to be at rest, in the sense of the Addressing this distinction, by extending existing
absence of mass motion ( v m = 0 ), the fact remains that molecular theories8,9 so as to explicitly recognize and
hence take account of the diffusive transport of volume,
the undisturbed fluid is in motion ( v l ≠ 0 ), as
is clearly prerequisite to any comprehensive continuum
physically evidenced by the movement of a passive theory of transport processes that purports to derive
tracer inserted into the fluid to monitor the latter's rationally from subcontinuum models of the
motion through space! How can it be that the fluid is phenomenon.
both at rest and yet physically moving? The answer to
the paradox lies in the fact while there is no mass The pragmatic consequences of our remarks will likely
motion of the fluid, there nevertheless exists a be of interest only to practicing fluid mechanicians and
volumetric fluid motion driven by the fluid's thermal molecular theorists. However, since fluid mechanics
expansivity β acting in concert with the mass density often constitutes the first example6,7 of a non-trivial
gradient (engendered by the temperature gradient). field theory4,5 (i.e., a continuum theory) to which
However, since v m = 0 , this volumetric motion is applied scientists and engineers are exposed as
entirely non-convective, being purely diffusive in neophytes, the issues raised herein provide a classic
nature. Moreover, in the general case, without textbook example of the more generic philosophical
restriction to purely phoretic-type undisturbed fluid need for such novitiates to appreciate the intimate
motions, the fluid's Lagrangian velocity, v l , is relation existing between theory and experiment.
identical1 to its volume velocity, v v , the latter Explicitly, how is a given physical entity (appearing in
representing the flux density of volume (volume per a mathematical description of the phenomenon to which
unit time per unit area crossing a space-fixed surface), the student is being exposed) to be measured
with the diffusive flux density of volume, jv , defined experimentally. This lesson is especially needed in the
computer age, where, for example, molecular dynamic
generically by the expression1 jv = v v − v m . In the
simulations are often regarded as being isomorphic with
single-component heat transfer case, the constitutive the performance of actual physical experiments. That
equation for this volume current is1,12 jv = α ∇ln ρ . In view notwithstanding, in order for the consequences of
the case of steady-state heat transfer through a confined such theoretical computations to be physically
single-component fluid the preceding expression leads interpreted, the statistical predictions emanating
eventually1,2 to the velocity disparity quantified by eq. therefrom need to be expressed in terms of their
(3). With an appropriate transposition of symbols, a macroscale experimental counterparts. And if such
comparable theoretical result also applies to the case of equivalences are subject to possible ambiguities, such
binary convective-diffusive mass transfer phenomena as those resulting from an incomplete appreciation of
occurring in isothermal, isobaric fluid mixtures (the the experimental protocol underlying their definitions,
latter obeying the "law" of additive volumes), the fundamental physical misconceptions can result. Such a
resulting expression being in accord with experimental misinterpretation has arisen (at least in our view), for
diffusiophoretic particle data15. example, during attempts to understand the role of
literal, molecular-level "boundary conditions"
The initial strangeness of the notion of a purely prevailing at solid surfaces, in whose neighborhood the
diffusive, non-convective mechanism for volume macroscale continuum physics — described by the very
transport represents a failure of one's intuition, arising equations to which such boundary conditions are meant
from a psychological predisposition to associate volume to apply — is not itself applicable owing to obvious
with mass through the intermediary of the fluid's inadequacies in such macrosocale models of the

37
phenomenon at molecular distances from the surface29. sagacity of Einstein's advice. This object lesson is
Apart from its fluid-mechanical implications, this particularly striking when viewed in the context of the
lecture aims to emphasize, to a broader audience, the large number of students and professionals, scientists
general importance of experiments in the theoretical and engineers, prominent and otherwise (including, of
sciences. course, myself at an earlier, more naive stage of life),
who, despite their often extensive exposure to fluid
In this context of the preceding remarks, it is mechanics, failed to question Euler's implicit
informative to recite an anecdotal story related by the assumption positing equality of the fluid's Eulerian and
late, pioneering rheologist, Karl Weissenberg (1893- Lagrangian velocities.
1976) to a continuum-mechanical audience assembled
at Carnegie-Mellon University in the early 1970's, In summary, the work briefly reported upon here, if,
which lecture I had the good fortune to attend. independently substantiated by others, negates the
Prefacing his remarks (on Osborne Reynolds theory of foundations of 250 years of fluid mechanics, as well as
dilatation in sand-water mixtures, during which he derivative subjects, including continuum mechanics,
demonstrated a wonderfully simple illustrative transport processes, molecular theories of gases and
experiment), Weissenberg noted that just prior to World liquids, irreversible thermodynamics, and rheology.
War I he had served as a young applied mathematician
at the Kaiser Wilhelm Institute of Physics in Germany, References
which Einstein visited periodically as part of the his
"administrative" responsibilities as Director. On one of
these visits, Weissenberg collared Einstein, telling him 1. Brenner, H. "Fluid mechanics revisited." (Originally
how much he admired his work, and seeking the latter's submitted in November 2002 to Phys. Rev. E under the
advice and counsel regarding an appropriate course of title: "Inequality of Eulerian and Lagrangian velocities
self-study that would permit his metamorphosis from in molecularly inhomogeneous fluids".)
applied mathematician to theoretical physicist. Without
a moment's hesitation, Einstein suggested that his 2. Brenner, H. and Bielenberg, J. R. "A continuum
questioner learn how to design scientific instruments. approach to phoretic motions: I. Thermophoresis."
Not understanding this response, Weissenberg posited (Originally submitted to Phys. Rev. E in December
that perhaps Einstein had misunderstood his original 2002.)
query, since his wish was to become a theoretical
physicist, not an experimental one. Einstein assured 3. Euler, L. "Principes généraux du mouvement des
him that he had not misunderstood the original question, fluides." Mém. Acad. Sci. Berlin 11, 274-315 (1755)
and went on to elaborate that the only mechanism by [reproduced in Leonhardi Euleri Opera Omnia. Series
which one could truly and deeply understand the II, Vol. 12, pp. 54-92 (Füssli, Zürich, 1954)].
fundamental meaning of a physical entity appearing in Additional historical information can be found in the
the guise of a mathematical symbol in a theoretical "Editor's Introduction" to the latter volume by Truesdell,
equation was by going through the mental exercise of C. "Rational fluid mechanics, 1687-1765," pp. VII-
systematically and methodically identifying the CXXV.
sequence of steps prerequisite to an unambiguous
experimental determination of that entity. (Presumably 4. Serrin, J. "Mathematical Principles of Classical
underlying Einstein's response was the decisive role Fluid Mechanics." In Handbuch der Physik (Volume
played in his 1905 special theory of relativity by his VIII/1, Fluid Dynamics I (S. Flügge, ed.), pp. 125-263
gedanken experiment involving the measurement of (Springer-Verlag, Berlin, 1959)].
time, as well as by his then current efforts to understand
the theoretical implications of the seemingly empirical 5. Truesdell, C. and Toupin, R. A. "The Classical Field
equality of inertial and gravitational mass, an Theories." In Handbuch der Physik Volume IIII/1,
unequivocal experimental fact established earlier by Principles of Classical Mechanics and Field Theory (S.
Eötvös30 in 1889.) Flügge, ed.), pp. 226-793, (Springer-Verlag, Berlin,
1960).
In any event, the fact that the relationship between the
experimental measurement of fluid velocity and the 6. Landau, L. D. and Lifshitz, E. M. Fluid Mechanics,
symbol v appearing in fluid-mechanical equations, 2nd. ed. (Butterworth-Heinemann, Oxford, 1986).
posited by Euler some two-and-a-half centuries ago, has,
until now, never been properly subjected to rational 7. Batchelor, G. K. An Introduction to Fluid Mechanics
inquiry and subsequent critical experimental test (most (Cambridge Univ. Press, Cambridge, 1967).
simply, in compressible fluids at rest), points up the

38
8. Hirschfelder, J. O., Curtiss, C. F. and Bird, R. B., R. Soc. Lond. 170, 231-262 (1879); reprinted in The
Molecular Theory of Gases and Liquids (Wiley, New Scientific Papers of James Clerk Maxwell, Vol. 2, 681-
York, 1954). 712 (W. D. Niven, ed.) (Cambridge Univ. Press,
9. Chapman, S. and Cowling, T. G. The Mathematical Cambridge, 1890).
Theory of Non-Uniform Gases, 3rd. ed. (Cambridge
Univ. Press, Cambridge, 1970). 23. Reynolds, O. "On certain dimensional properties of
matter in the gaseous state. Part I. Experimental
10. Bird, R. B., Stewart, W. E. and Lightfoot, E. N. researches on thermal transpiration of gases through
Transport Phenomena, 2nd. ed. (Wiley, New York, porous plates and on the laws of transpiration and
2002). impulsion, including an experimental proof that gas is
not a continuous plenum. Part II. On an extension of the
11. Bird, R. B. "Viewpoints on transport phenomena." dynamical theory of gas, which includes the stresses,
Korean. J. Chem. Eng. 15, 105-123 (1998). tangential and normal, caused by a varying condition of
gas, and affords an explanation of the phenomena of
12. Brenner, H. "On the diffusive transport of volume. transpiration and impulsion." Phil. Trans. R. Soc. Lond.
Kinematics." (Originally submitted to J. Fluid Mech. 170, 727-845 (1879).
under the title: "On the molecular and convective
transport of volume" in March 2001). 24. Crookes, W. "On repulsion resulting from radiation.
Parts III & IV." Phil. Trans. R. Soc. Lond. 166, 325-376
13. Landau, L. D. and Lifshitz, E. M. Mechanics (1876).
(Addison-Wesley, Reading, MA, 1960).
25. Rosner, D. E. and Papadopoulos, D. H. "Jump, slip,
14. Haase, R. Thermodynamics of Irreversible and creep boundary conditions at nonequilibrium
Processes (Dover reprint, New York, 1990). gas/solid interfaces." Ind. Eng. Chem. Res. 35, 3210-
3222 (1996).
15. Bielenberg, J. R. and Brenner, H. "A continuum
approach to phoretic motions: II. Diffusiophoresis." 26. Talbot, L., Cheng, R. K., Schefer, R. W. and Willis,
(Originally submitted to Phys. Rev. E in October 2003). D. R. "Thermophoresis of particles in a heated
boundary layer." J. Fluid Mech. 101, 737-758 (1980).
16. Brock J. R. "On the theory of thermal forces acting
on aerosol particles." J. Colloid Sci. 17, 768-780 (1962). 27. Because of the assumption of small Knudsen
number, eq. (1) is valid only when k s /k = O(1) or less,
17. Waldmann, L. and Schmitt, K. H. "Thermophoresis as pointed out by Bakanov, S. P. "Thermophoresis of
and diffusiophoresis of aerosols." In Aerosol Science aerosols - A review." J. Aerosol Sci. 22, S215-S218,
(C. N. Davies, ed.), pp. 137-162 (Academic, London, Suppl. 1 (1991).
1966).
28. McNab, G. S. and Meisen, A. "Thermophoresis in
18. Derjaguin, B. V. and Yalamov Yu. I. "The theory of liquids." J. Colloid Interface Sci. 44, 339-346 (1973).
thermophoresis and diffusiophoresis of aerosol particles
and their experimental testing." In Topics in Current 29. Brenner, H. and Ganesan, V. "Molecular wall
Aerosol Research, pp. 2-200 (G. M. Hidy and J. R. effects: Are conditions at a boundary "boundary
Brock, eds.) (Pergamon, Oxford, 1972). conditions?" Phys. Rev. E 61, 6879-6897 (2000);
erratum, ibid 62, 7547 (2000).
19. Zheng, F. "Thermophoresis of spherical and non-
spherical particles: a review of theories and 30. Eötvös, R. V. "Über die Anziehung der Erde auf
experiments." Adv. Colloid Interface Sci. 97, 255-278 verschiedene Substanzen." Math. Naturw. Ber. Ungarn
(2002). 8, 65-68 (1890).

20. Davis, E. J. "A history of single aerosol particle Acknowledgements: I am grateful to my student,
levitation." Aerosol Sci. Technol. 26, 212-254 (1997). James R. ("Jim") Bielenberg, for his significant
contributions towards unraveling the mysteries of
21. Tyndall, J. "On dust and disease." Proc. R. Inst. 6, phoretic phenomena. I am also grateful to Dr. Sangtae
1-14 (1870). Kim of Eli Lilly and Company for the generous and
unconditional support of my fluid mechanics research
22. Maxwell, J. C. "On stresses in rarified gases program.
resulting from inequalities of temperature." Phil. Trans.

39
40
Unsolved Problems in Biomolecular Engineering

41
 Biographical sketch of Michael L. Shuler

Michael L. Shuler is the Samuel B. Eckert Professor of Chemical Engineering, Director of the School of
Chemical Engineering, and Director, Bioengineering Program at Cornell University He joined the faculty at
Cornell in 1974 after receiving his doctorate in Chemical Engineering from the University of Minnesota. He
holds a B.S. in Chemical Engineering from the University of Notre Dame. Shuler has held visiting
appointments at ETH (Zurich, Switzerland, 1995), Boyce Thompson Institute (1995/96), University of
Wisconsin (1988/89), and the University of Washington (1980/81). He has held short-term distinguished
visiting professorships at the University of Florida, Osaka University, and the University of Newcastle upon
Tyne. Shuler was the Director of the School of Chemical and Biomolecular Engineering from 1998 to 2002.

Shuler has current research interests in several aspects of bioengineering, particularly bioreactors. These
interests include production of secondary metabolites from plant cell cultures (e.g. Taxol), structured
mathematical models of cells (e.g. a "minimal" cell model to relate genomic structure to cell function),
microfabricated, heterologous protein expression systems, and in vitro toxicology using cell culture analogs
and physiologically-based pharmacokinetic models. He authored or co-authored over 200 refereed papers,
and 7 books, including a textbook, "Bioprocess Engineering: Basic Concepts", now in its second edition

He was elected to the National Academy of Engineering in 1989, an Inaugural Fellow for the American
Institute of Medical & Biological Engineers in 1992, and American Academy of Arts and Sciences (1996).
He has served on numerous taskforces and committees for the National Research Council, including the
committee responsible for "Beyond the Molecular Frontier, Challenges for Chemistry, and Chemical
Engineering". He also has been an active member of the ACS, American Society for Microbiology,
American Society of Pharmacognosy, and AICHE (Fellow, 1997). He is currently a member of AIChE's
board of directors and headed a recent taskforce on how to better integrate bioengineering into AIChE.

42
 Unsolved Problems in Biomolecular Engineering

Michael L. Shuler
School of Chemical and Biomolecular Engineering
Cornell University
Ithaca, NY 14853
[email protected]

Abstract:
The growing interface of biology with that of chemical engineering principles is a major challenge in chemical
engineering education (esp. curriculum) and research. The “problem” for the chemical engineering profession is
how do we best exploit these emerging opportunities. The most critical driver of change is the advent of molecular
biology, genomics, and related technology. Biology is evolving from a data poor to a data rich science where the
underlying molecular mechanisms of life are becoming clearer. Traditional strengths of chemical engineers are their
ability to think across length and time scales and integrate descriptions of molecular level phenomena into an
understanding of macroscopic systems. Chemical engineers are well positioned to contribute to biological discovery
because of their skills in systems integration across length scales. Chemical engineers will be major players in the
emerging field of systems biology. Further, chemical engineers are key contributors to the tools necessary for rapid,
accurate and cost effective analysis of biomolecules. Finally, chemical engineers will increasingly convert the basic
insights from the emerging understanding of biology into useful processes, diagnostics, therapies, and devices that
will be of broad benefit to human kind. Some key examples are improved, lower cost cell culture processes to make
therapeutic proteins that are more authentic, devices that allow rapid and more accurate assessment of disease,
methods to evaluate pharmaceuticals for toxicity to reduce the cost of drug approvals, and more accurate ways to
deliver drugs and genes to specific target tissues.

These intellectual drivers are also changing the nature of chemical engineering education. While the core chemical
engineering principles will remain intact, biology (esp. biochemistry) will become a foundational science for all
chemical engineers. Examples from biomolecular engineering as well as other emerging areas such as electronics
and materials need to permeate the core courses supplementing examples from petrochemicals. Ultimately, I predict
that within chemical engineering departments there will emerge an identifiable track for those students who wish to
focus on biomolecular engineering.

Introduction As it does so, the opportunities for chemical engineering

science to be relevant to biology and biology to
The grand challenge for the chemical engineering
chemical engineering increase significantly.
community is how to best exploit the emerging
Biomolecular engineering emerges from this interface.
opportunities at the interface of biology and chemical
Biomolecular engineering was defined by National
engineering principles.
Institute’s of Health for a December 1992 meeting as
“…research at the interface of chemical engineering and
The “Problem” or “Challenge”
biology with an emphasis at the molecular level.”
Biology is undergoing a revolution. The emergence of Why Are Chemical Engineering Principles Key to
molecular biology followed by high throughput Biology?
analytical techniques is generating data at rates that
were inconceivable 15 years ago. Biology is evolving Life obeys all the rules of chemistry and physics.
from data poor to a data rich science. This abundance Living cells are complex chemical plants. While
of data will drive biology towards formulation of mechanical and electrical forces (and their
theoretical constructs to summarize these data transformation into chemical signals) are important, the
efficiently and to form general principles and models. dominant science in cells is chemistry. Biology,
Biology is moving from an observational, descriptive especially at the molecular and cellular level, is
science towards a mechanistic and quantitative science. primarily the study of chemical systems. Cellular
regulation – information sensing and control of the
corresponding chemical reactions – is typically done

43
using chemical signals and signal transduction various lengths) combine with proteins in the cell to
pathways. block or encourage transcription of genes; these
combinations of letters constitute regulatory elements.
Analysis of cell function calls upon all of the principle The language of the processes in the Central Dogma is
components in a chemical engineering education: mass universal; the same language applies to humans, plants,
and energy balances, thermodynamics, reaction kinetics, and microbes. The universality of this language is
transport processes (esp. mass transfer), separations essential to modern biotechnology.
(esp. natural membranes), and process control. The
chemical engineering education is also unique in its Simple bacterial cells (or procaryotes) are small (ca.
emphasis on integrating from molecular to the macro 1µm) and have no internal organelles. A single, double
scale as in the design of a chemical plant. These stranded DNA (or chromosome) encodes the organism’s
principles and the ability to work quantitatively across genetic blueprint. In higher organisms (e.g. humans,
both length and time scales places chemical engineers in plants, insects, even yeast) cells are larger (ca. 10 µm)
an unique situation to contribute to both using biology and have distinct organelles. The nucleus is one such
to create devices or developing useful therapeutic organelle. Multiple double-stranded DNA molecules, or
strategies and to discover more clearly how living chromosomes, are present and maintained within a
organisms function. nuclear membrane. The minimal cell to be discussed
later is a simple bacterial cell while eucaryotic cells are
Basic Biological Concepts used in the “animal-on-a-chip” described at the end of
this article.
While some readers will have an in depth knowledge of
biology, others may not. Consequently I cover some Genomics and Systems Biology
elementary concepts and vocabulary to insure that the
remainder of the discussion is readily accessible. The major advances in biology have led first to
(Greater detail is available in Shuler & Kargi, 2002) understanding the molecular basis of life, then to
molecular techniques to manipulate cells (e.g.
Molecular biology is principally concerned with the recombinant DNA technology) in a predictable manner
storage, replication, and use of information. A living and now to genomics. “Genomics” is the set of
cell must process information from the external world experimental and computational tools which allow the
and act upon it to ensure survival and replication of the genetic blueprints of life to be read. By this statement
organism. The primary challenge to an intact cell is the we mean that we know the sequence of all the letters in
flow and control of information. The primary tenant of genome or the total inheritable DNA in an organism.
molecular biology is known as the Central Dogma. The We have full genomic sequences for over 100
information necessary for a cell is encoded in the double organisms. While most of these are for bacteria with
helix of DNA (deoxyribonucleic acid). DNA acts as only a single chromosome, full genomic sequences are
template for its own replication. Segments of DNA that available for representations of higher and more
encode information for the primary structure (i.e., complex organisms (e.g. insects, plants, mammals and
sequence of amino acids) in a protein are called genes. humans). As the technology for sequencing DNA
The information for production of the protein is first improves, many additional species will be sequenced.
transcribed from DNA through synthesis of a messenger
RNA (ribonucleic acid) molecule by the process of The emerging challenge is to understand this sequence
transcription. Regulatory elements on the DNA data in terms of how an organism works. The term
determine when and how often a particular gene is “functional genomics” is often used to suggest the range
transcribed. The information on the messenger RNA is of activities necessary to relate genetic blueprints
converted to a protein through a process known as (genomic sequence) to the structure and behavior of
translation. If the physiological conditions in the cell organisms.
are correct, the nascent protein folds into its proper 3-
dimensional shape. Proteins act as catalysts, To fully understand regulation and physiology, DNA
transporters, receptors of information, regulators, and information is insufficient. Interactions among the
structural elements in the cell. expression of different genes requires complete
knowledge of DNA sequence, mRNA expression, and
The language of biology consists of four letters (A, T, protein content. Microarrays for determining time and
C, and G corresponding to four different nucleotides) in situation dependent expression of mRNAs have been
DNA and three letter words. Each three letter word developed. A microarray is a high-density
specifies a particular amino acid (or start or stop oligonucleotide array. Each oligonucleotide hybridizes
translation). Other combinations of DNA letters (of with a corresponding m-RNA. For each known gene (eg

44
from sequence analysis) an oligonucleotide (eg 40 bases Why Is This Important?
in length) can be synthesized that binds uniquely the m-
RNA from that gene. Using the techniques of
The interface of chemical engineering with biology
photolithography, arrays for about 10,000 genes can be
provides the best basis to develop “systems biology.”
made that are less than 2 cm2 in area. Binding with a m-
One of the most fundamental questions asked by
RNA can alter a fluroscent signal allowing the array to
humans is: “What is life?” What is the essence of being
be read optically.
alive? Systems biology cannot ignore this question; a
natural consequence of building a systems model is to
A study of the total protein content (proteome) of
ask what are the essential functions for life. Those
biological systems provides a unique protein fingerprint
biomolecular engineers engaged in developing systems
for a system and permits an elucidation of the key genes
biology will be leaders in this aspect of scientific
involved in a phenotype (eg functional features
discovery.
displayed by a cell). Hatzimanikatis et al. (1999) have
shown that an understanding of the regulatory
Biomolecular engineers will also create the devices and
structure/response requires simultaneous data on DNA,
therapeutic strategies that arise from the “new” biology.
RNA, and protein. Proteome analysis is slow, difficult,
One aspect of creation will simply be the tools
and expensive. Consequently, DNA and RNA are often
necessary to more efficiently sequence DNA, to build
obtained in the absence of protein information resulting
microfabricated array systems to more effectively
in an incomplete understanding of the relation of genes
determine mRNA expression, or to better separate,
and cellular regulation.
purify and analyze all the proteins in a tissue or cell.
The more critical aspect will be to couple engineering
Even with complete knowledge of DNA, RNA, and
analysis and systems biology to create for example:
protein, the full analysis of cellular function requires a
cells that more effectively produce therapeutic proteins
knowledge of the small molecules in a cell or tissue, the
in large scales; to devise not only new vectors for gene
physical structure of the cell, and the structure of
delivery, but the overall predictive analysis to make
metabolic pathways. The nature of a living cell is also
gene therapy an useful clinical tool; to build new
dynamic (time-dependent) and requires a knowledge of
methods to separate different cells on a large scale to
current environmental state and past history to fully
diagnosis and treat cancer; to grow artificial organs that
describe.
can replace living tissues; and so forth.
It is also critical to understand how these components
Perhaps this point is a good place to ask what separates
work together. As an example, Bhalla and Iyengar
biomolecular engineering from biology. A key feature
(1999) have shown that when individual well-defined
of biomolecular engineering is the emphasis on design:
biochemical signaling pathways share common
the ability to predict the output from controllable inputs.
elements to form a network, new properties emerge
Another is to do the design in a macroscopic system
when the network is analyzed. As this example
using explicit links of the molecular scale to the
illustrates the full properties of a cell or organism are
macroscopic system. While engineering design,
more than a sum of identifiable components, but result
especially in the area of systems biology, may be lead to
from the complex non-linear interactions of the
biological discovery, the essential role of the
components. Understanding biology will mean
biomolecular engineer is to create what has never been
understanding both which components are involved and
(e.g. a device, a computer model, or a therapeutic
the non-linear dynamics of how they interact. It is with
procedure). Scientists are primarily interesting in
respect to systems integration and analysis that
discovering what exists (e.g. a new organism or novel
engineering approaches may prove critical to biological
protein).
discovery.
What Is Known?
These considerations lead to the concept of “systems
biology.” A goal of systems biology is the integration The origins of biomolecular engineering arise from the
of genomics, high throughput, organism-wide data on development of large-scale processes to produce
cell composition and observations of antibiotics during World War II. Companies began to
physiology/function into a comprehensive mathematical pair chemical engineers and microbiologists to gain the
model. Such a model of a living organism can predict integrated knowledge of process engineering and
the behavior of the organism in response to internal and microbiology necessary to produce these new
external perturbations. compounds. The rise of processes to produce
antibiotics induced students to seek combined training

45
in chemical engineering and microbiology. These • Biocatalysis concerns the generation of
students became the parents of biochemical engineering. protein catalysts (enzymes) with novel or
enhanced activities often in unusual
Somewhat in parallel, chemical engineers sought to environments, such as at high temperature or
apply the basic principles of chemical engineering to in organic solvents. These enzymes can be
human health (e.g. development of the artificial kidney). used to synthesize novel chemicals.
This second group worked as biomedical engineers.
From the 1950s to the 1980s these activities were In addition biomolecular engineers are key contributors
thought to be rather separate. In the 1980s and beyond to research activities such as:
it became increasingly apparent that common principles
from molecular and cellular bioengineering were • Drug delivery is an area in which chemical
foundational to both activities. For example, in both engineers have had a major impact,
medical and non-medical applications it became particularly for controlled delivery of
important to understand how cells responded to pharmaceuticals to specific target sites (e.g.,
different surface chemistries, or how they would a tumor).
respond to alterations in gene number of deletion, or
how fluid shear would alter cell physiology, etc. By • Gene therapy is really metabolic engineering
1990 biochemical engineering had come to mean combined with drug delivery. The right
bioprocess engineering and seemed to exclude genes need to be delivered to the desired
biomedical engineering. There was a need for a new tissues, and proteins from that gene need to
word that would recognize that common principles be made at the right time in the right amount.
supporting both medical and non-medical applications The lack of success with gene therapy is, at
at the chemical engineering and biology interface. In least in part, due to the inability of medical
1992 (as described earlier) the term biomolecular scientists to deal with these issues of well-
engineering was coined to include all activities at the controlled gene delivery and gene
chemical engineering and biological interface with expression. The solution lies in a
special emphasis on the emergence of molecular and quantitative analysis of the entire system.
cellular biology as the common linkage independent of
area of application. • Biomaterials (particularly for controlled
release of bioactive compounds or surface
Biomolecular engineering is perhaps best defined by modifications to become biologically
what biomolecular engineers do. There are four compatible or to actively direct biological
activities in which those with a chemical engineering activities).
background dominate:
• Cell and tissue engineering combines
• Bioprocesses (biochemical reaction biomaterials and broad concepts from
engineering; bioreactors with associated metabolic engineering and analysis of
bioseparations). Bioprocesses produce chemical signaling. The issues for
pharmaceuticals, chemical commodies and manufacturing (e.g., artificial skin) are very
energy products and are an integral part of similar to those for bioprocesses.
waste treatment.
• Systems biology, especially to computer
• Bioseparations support bioprocesses and models of cells and to rapid separation,
have important medical applications such as purification and measurement of cell
cell separations (e.g., stem cell recovery) and components such as proteins.
proteomics (e.g., total analysis of protein
content in a cell). • Nanobiotechnology (e.g., “lab-on-a-chip”
type devices; chemical engineers have
• Metabolic engineering is the analysis and played key roles in development of these
redirection or enhancement of the metabolic devices which are basically small-scale
activities of a cell. Chemical engineers are chemical plants).
currently playing a leading role in this area,
often for the purposes of supporting If the reader is interested in another listing of challenges
bioprocess development for more efficient to be solved by biomolecular engineers, they are
production of pharmaceuticals or chemicals. referred to the new book just released by the National

46
Research Council (2003) on “Beyond the Molecular We have previously developed a “complete” cell model
Frontier: Challenges for Chemistry and Chemical of E. coli that contains all of the functional elements for
Engineering” the cell to grow, divide, and respond to a wide variety of
environmental perturbations. This model serves as the
The next section details two programs our research basis for a minimal cell model, but does not qualify as a
group has undertaken that illustrate our response to the minimal cell model for two reasons. First, it is specific
challenges arising from the new biology-chemical to a particular species (E. coli). Second, chemical
engineering interface. species are lumped into a single model component so
that the dynamic function of the cell cannot be mapped
Examples of Biomolecular Engineering directly back to the genome. Basically, the model is a
good summary of the functionality required for a
The Minimal Cell Model
minimal cell, but it does not capture explicitly the
physical chemistry that supports those functions. We
We are constructing a mathematical model of a
described our first mathematical model of a single E.
"minimal cell" to provide a platform to better
coli cell in 1979 (Shuler, et al, 1979); at the time of its
understand the "design logic" of cellular regulation.
publication, it was the only model of an individual cell
The model, which contains kinetic and thermodynamic
which did not include artificially-imposed constraints on
constraints as well as stoichiometric constraints, will be
aspects such as mode of growth, timing of cell division
used as a tool to identify the organizing principles
(e.g. growth rate), and cell size. Also, it was unique in
which relate the static linear sequence information of
its ability to respond explicitly to concentrations of
the genome to the dynamic non-linear functioning of the
nutrients in the environment (Bailey, 1998). This base
cell. A minimal cell will be a prokaryote with the
model (Domach, et al 1984) has been embellished with
minimum number of genes necessary for growth and
additional biological details to allow prediction of a
replication in an "ideal" environment (i.e. a rich medium
wide-range of microbial responses to environmental and
with preformed precursors, constant temperature and
genetic manipulations (Shuler, 1999). The initial model
pH, and at a sufficiently low cell concentration that no
included only 18 components, and most components
metabolic by-product becomes inhibitory). Kooin
represented large groups of related chemical species; the
(2000) provides a review of the minimal cell concept.
mathematical description of cellular function, which is
the core of the model, is based on time-variant mass
The success of whole organism genome sequencing and
balances for each component. Each mass balance takes
high-throughput measurements provides an opportunity
into account the component's synthesis (as a function of
for system-level analysis of whole organisms or what
availability of precursors and energy, and in some cases,
has been termed "systems biology" e.g. (Ideker, T., et al
relevant enzymes), utilization, and degradation.
2001). The emphasis in our project is on modeling the
Stoichiometric coefficients for relating components
complete functionality of a cell and its explicit response
through mass balances were derived primarily from
to perturbations in its environment (Browning & Shuler,
published research, and in some cases, from our own
2001). Our attempt to generate a "complete" model that
experiments. It is important to note that the model was
predicts time-dependent responses in a cell
not developed by using adjustable parameters to fit
differentiates this project from others.
model predictions to experimental results, nor did the
stoichiometric mass balances assume a steady state (i.e.,
The model we are developing focuses on essential
the amount of each component was allowed to vary with
functions while finding examples of gene products that
time). Despite the simplifications that were made in
can perform those functions. While the set of minimal
describing cell composition and relationships, the model
genes we postulate may change (for example if a new
can accurately predict changes in cell composition, cell
multifunctional protein is found), we believe we can
size, cell shape, and the timing of chromosome
find a set of truly essential functions. Further, the
synthesis as a function of changes in external glucose
technical difficulties associated with generating an
and ammonium concentration.
experimental minimal cell and the ambiguities in
interpretation of comparative genomic data argue for the
While the minimal cell model differs from the E. coli
establishment of a theoretical computer model of a
model (primarily in that each component in the minimal
minimal cell that is explicit in regard to minimal
cell model will be directly related to an actual
function and includes a realistic set of gene products to
component in a living cell), the E. coli model has been
accomplish these functions. This is, in essence, the
an excellent starting point from which to construct a
primary objective of the project.
minimal cell model. The E. coli model is “complete” in
terms of function and is “modular”. By modular we
mean that we can “delump” a pseudochemical species

47
into individual components while still maintaining the Our ultimate goal is construction and validation of an in
essential connectivity to other functions in the cell. This vitro device (cell culture analog or CCA) that
allows us to add detail in parallel efforts on different realistically and inexpensively mimics uptake,
“modules” and then have confidence that they can be distribution, metabolism, and biological response of
recombined into a functional and functioning whole. humans to exposure to various chemicals and drugs
(Ghanem & Shuler, 2000). A cell culture analog (CCA)
We have tested the hypothesis that it is not the exact is a device that is a direct physical replica of a
values of parameters in the model that determine physiologically based pharmacokinetic (PBPK) model.
function, but that the relative values to one and another A PBPK model is a mathematical model that represents
is critical. We tested this hypothesis by varying all the body as interconnected compartments representing
kinetic rates by a scaling factor (or kinetic ratio). The the functions of selected organs or tissues. The CCA
growth rate scales directly with the kinetic ratio over uses living cells in “organ/tissue” compartments to
about two orders of magnitude. At low values of growth represent some aspects of metabolism in that organ.
rate the terms for maintenance due to membrane Coupling of a CCA and corresponding PBPK can
energization become important, and these factors are not provide insight into molecular mechanisms of toxicity.
directly influenced by the kinetic ratio and the linearity This device will impact biomedical research by
is lost. Cell composition (e.g. protein/cell, RNA/cell, providing a cost-effective, pre-clinical system to
etc.) remains constant for a wide range of kinetic ratios. estimate human response to a wide range of drug leads
Further, relative growth rate changes for models with and assist in risk assessment for environmental exposure
different kinetic ratios are essentially the same for a to chemicals while reducing dependence on animal
wide variety of perturbations to cell function. All studies.
together these findings suggest that “…a minimal cell
based on a hypothetical set of properties using While our initial experiments were on macroscale
dimensionless parameters should provide realistic systems, there are a number of advantages to greatly
insight into cellular regulation” (Browning & Shuler, reducing the size of the CCA device to microscale. Our
2001). design calculations for a prototype chip system have
shown that we can achieve physiologically realistic
A particularly important aspect of the original Cornell organ residence times and ratio of fluid to tissue in an
E. coli model was a model that mechanistically coupled “organ” compartment, and maintain fluid flow rates at
cell growth, chromosome replication, and cell division. values that create only physiologic values of shear stress
We have updated the model for control of chromosome on cells. Only by going to the dimensions of a
replication based on new experimental evidence over microsystem can we attain a physiologic design.
the last 25 years. While the revised model is Further, the microscale reduces the cost of the device
significantly different in terms of biological mechanism and makes effective use of expensive reagents or
(positive vs. negative control), the mathematical difficult cell cultures.
characteristics are quite similar. Indeed it may be that
any functional mechanism for control of replication We have constructed prototype microCCAs in silicon
must satisfy similar mathematical characteristics. with 3 (liver-lung-other tissue) compartments. The
constraints placed on the design were: (1) it should fit
We have nearly completed our model for nucleotide on a 2cm X 2cm silicon chip; (2) the ratio of the organ
metabolism and have demonstrated that a functional compartment sizes and the liquid residence times in
system can be achieved with significantly fewer gene- each compartment should be physiologically realistic;
encoded functions (13) than estimated previously. This (3) each compartment should have a minimum of
“module” has been integrated into the whole cell 10,000 cells to facilitate analysis of chemicals and
minimal model and the resulting cell functions in a enzyme activity; and (4) the hydrodynamic shear stress
biologically realistic manner confirming the modularity on the cells must be within physiologic values
of the model. We are currently extending the model to (2dynes/cm2 for body cells and 6-14 dyne/cm2 for
describe other “modules” with genetic level detail. endothelial cells). We have also extended this
technology to a 4-compartment system (liver-lung-fat-
While this project focuses on the cellular level, it is other tissues). Initial studies have been done with
possible to build useful models of complex, naphthalene as the model toxicant. These studies have
multicellular organisms, including humans. used two-dimensional cell cultures (i.e., monolayers) of
established cell lines. The models demonstrate the
An “Animal-on-a-chip” production of reactive, relatively stable metabolites in
the liver compartment that circulate to the lung
compartment and cause preferential cell death in the

48
lung compartment. When a fat compartment is present how to offer specialized education in biomolecular
the level of toxicity is reduced due to adsorption of engineering and how to integrate biomolecular
hydrophobic compounds into the fat compartment. engineering into traditional courses that teach the basic
Further we have identified napthaquione as the toxic principles of chemical engineering.
metabolite.
For those students who wish to focus on biomolecular
We are currently adapting this approach to evaluate engineering, they need not only basic biology courses,
chemicals for the potential usefulness as compounds to but also courses in molecular and cellular principles of
suppress multidrug resistance in cancer cells. If biomolecular engineering, quantative analysis of
successful we may find chemical mixtures, including physiological systems and bioprocess engineering
chemotherapeutic agents, that will allow treatment of (metabolic engineering, bioreactors, bioseparations). I
cancers that currently do not respond to chemotherapy. believe the biggest unmet need is course material for
molecular and cellular bioengineering. No textbooks
These two examples of work at the chemical are yet available to support such courses. It is material
engineering biology interface demonstrate the need for at this level that will define the intellectual foundations
biomolecular engineers to be expert in chemical for biomolecular engineering.
engineering with significant biological knowledge. In
the next section we discuss some implications for Even students who do not intend to work in
chemical engineering education. biotechnology or health are looking for inclusion of
biomolecular engineering into the curriculum. The
absence of biological based examples and context in
Implications for Chemical Engineering Education
textbooks is a barrier to the true integration of biology
into chemical engineering. All of the chemical
I believe that all chemical engineers (indeed all engineering principles can be developed in the context
educated citizens) should have a basic college level of biological systems. In addition to the need for
understanding of biology and biochemistry. Without a modified textbooks, many chemical engineering faculty
basic understanding of biology, chemical engineers will are uncomfortable with the intersection of chemical
be excluded from certain types of jobs, particularly in engineering and biology. While good textbooks would
the pharmaceutical and biotechnology sectors. More relieve some of the faculty anxiety, short courses
importantly they will be unable to understand the basis specifically targeted to this issue would be helpful.
for environmental and health regulations while they are Chemical Engineering as a sustainable unified discipline
responsible to meet those standards in the operation or will require a professorate that are comfortable with
design of any chemical facility, whether it is biology as a core science of chemical engineering.
biologically based or not. Further, many of the great
debates in our society will arise from advances in Chemical engineering education has undergone many
biology (e.g. stem cell therapies). It is also important changes over the last fifty years. Evolution to
for the unity of the discipline that all chemical engineers incorporate biology and biomolecular engineering into
have biology as one of the supporting sciences. chemical engineering education should be a natural
process. I believe that this evolution will be
It must be recognized that learning biology is not simply accomplished within less than a decade.
a matter of new vocabulary, but also understanding a
different culture. What constitutes “proof” of a concept Concluding Thoughts
in biology is different than the typical approach in
engineering. In particular, biologists have had to make
I believe that it is important to recognize that the
progress in the absence of a quantative theoretical
intellectual revolution in biology provides a basis for
framework. Thus biologists are particularly sensitive to
fostering biomolecular engineering. Chemical engineers
the use of appropriate experimental controls; a concept
who embrace the basic concepts in biomolecular
not stressed in traditional engineering education where
engineering are well prepared to become leaders in the
experiments are usually evaluated by comparison to
field of systems biology.
theoretical expectations. I believe chemical engineers
will become better chemical engineers as they broaden
In addition biomolecular engineers will continue to be
their perspective to incorporate both approaches to
leaders in bioprocess and bioreactor engineering,
testing hypotheses.
bioseparations, and metabolic engineering. Further
biomolecular engineers will be critical in the
As biomolecular engineering emerges all chemical
development of the molecular and cellular aspects of
engineering departments are faced with two problems:

49
biomedical engineering. Particularly important will be Browning, S.T., and Shuler, M.L. (2001) Towards the
opportunities in controlled drug delivery, gene therapy, development of a minimal cell model by generalization
tissue engineering, diagnostics (based on separations of a model of Escherichia coli: Use of dimensionless
technology), and nanobiotechnology. rate parameters Biotechnol. Bioeng. 76(3), 187-192.

For the health of the chemical engineering profession, Damach, M.M., Leung, S.K., Cahn, R.E. Cocks, G.G.,
biology must be integrated into the curriculum. and Shuler, M.L. (1984) Computer-
Chemical engineering principles should be developed Model for Glucose-Limited Growth of a Single Cell of
throughout the curriculum in the context of biology as Escherichia-Coli B/R-A Biotechnol. Bioeng. 26(3), 203-
well as in the context of more traditional areas. Finally 216.
the academic enterprise should provide opportunities for
specialization in biomolecular engineering which will Ghanem, A. and M.L. Shuler. 2000. Combining Cell
require development of new courses which integrate Culture Analogue Reactor Designs and
biology and engineering in the same way chemistry and PBPK Models to Probe Mechanisms of Naphthalene
engineering were integrated at the beginning of the Toxicity. Biotechnol. Prog. 16(3):33-345.
profession.
Hatzimanikatis, V. and K.H. Lee (1999). Dynamical
The opportunities and challenges to chemical Analysis of Gene Networks Requires Both mRNA and
engineering that arise from the biological revolution Protein Expression Information, Metabolic Engineering
will invigorate the profession and attract students who 1:275-281 (E1-E7).
otherwise might not select chemical engineering. Many
of humankind’s oldest problems may be solved by Ideker, T., Galitski, T., and Hood, L. (2001) A new
combining chemical engineering principles with approach to decoding life: Systems biology Ann. Rev. of
opportunities emerging from the new biology. Genomics and Human Genetics 2, 343-372.

References Koonin, E.V. (2000) How many genes can make a cell:
the minimal-gene-set concept Ann. Rev. of Genomics
and Human Genetics. 1, 99-116.
Beyond the Molecular Frontier: Challenges for
Chemistry and Chemical Engineering. National
Shuler, M.L., S., L., and CC., D. (1979) A mathematical
Academies Press, Washington D.C., 2003.
model for the growth of a single bacterial cell Ann. NY
Acad. Sci. 326, 35-55.
Bailey, J.E. (1998) Mathematical modeling and analysis
in biochemical engineering: past accomplishments and
Shuler, M.L. (1999) Single-cell models: promise and
future opportunities. Biotechnol. Prog. 14(1), 8-20.
limitations J. Biotechnology 71(1-3), 225-228.
Bhalla, U.S. and R. Lyengar. 1999. Emergent
properties of networks of biological signaling
Shuler, M.L. and F. Kargi. 2002. Bioprocess
pathways. Science 283:381-387.
Engineering: Basic Concepts (Second Ed.). Prentice
Hall, Englewood Cliffs, N.J.

50
Unsolved Problems in Process/Product Systems Engineering

51
 Biographical sketch of Arthur W. Westerberg

Arthur W. Westerberg is currently the Swearingen University Professor of Chemical Engineering at Carnegie
Mellon. He received his BS from Minnesota ‘60, MS from Princeton ‘61 and PhD from Imperial College,
University of London ‘64, all in chemical engineering. He spent two years at Control Data Corporation and
nine years as a faculty member at the University of Florida before joining Carnegie Mellon University in
1976. He has served as Director of the Design Research Center (78-80), Chemical Engineering Department
Head (80-83), and founding Director of the Engineering Design Research Center, an NSF funded
Engineering Research Center (86-89).

His research interests are in engineering design, specifically in design synthesis, analysis, optimization,
computer support environments for modeling and design collaboration.

A member of the National Academy of Engineering, he is the recipient of several awards from AIChE, ASEE,
the ACS and CMU.

52
 Unsolved Problems in Process/Product Systems Engineering

Arthur W. Westerberg
Dept. of Chemical Engineering and
the Institute for Complex Engineered Systems
Carnegie Mellon University
Pittsburgh, PA 15213
[email protected]
Abstract
We examine the process/product systems engineering (PSE) area of chemical engineering. A brief history of some
of the accomplishments of the past four decades follows and notes the strong coupling of PSE with developments in
computing. “Driving forces” in place today suggest where PSE research and development might move. The paper
ends with a discussion of some major outstanding problems in PSE.
Introduction about what molecules to make. At the next scale, we
often have to consider the aggregation and structuring of
Today the Chemical Engineering Department of Ohio
materials from these molecules. These first two types of
State University is celebrating its 100th year, a very
decision are more in line with thinking about products
proud accomplishment. As a part of that celebration the
rather than processes. We progress to our traditional
department has invited several faculty from throughout
scale, where we make decisions about production units
the country to prepare papers and presentations, each of
and then whole processes. Companies worry about the
whose theme is to be the unsolved problems in one of
proper designing and operating of multiple processes.
the research areas in this discipline. Mine is to examine
Finally, at the meso-scale, companies, interlinked into
the relatively young area of process systems
supply chains, make decisions about how they fit into
engineering. I will take the liberty of adding “product”
the whole of the manufacturing and delivery activity for
to the title as we are more and more interested in
an industry. Our move to the meso-scale is coming
delivering processes and products. While young, this
quickly. For example, in January of next year in
area has been active, and, in the conventional process
Kunming, China, we have organized the seventh
area, it is reaching maturity. Maturity means there are
conference on PSE around the topic of business decision
many others in the area, and all are competing to find
making (http://pse2003.chemeng.tsinghua.edu.cn/).
and develop new ideas and to obtain research funds.
With maturity comes the obligation to re-examine Our final definition for PSE became: “Process
accomplishments and to think about where we should be Systems Engineering is concerned with the improvement
heading. In this manner we are in the same situation of decision making processes for the creation and
that many believe chemical engineering as a discipline operation of the chemical supply chain. It deals with
is in at this time. In spite of our maturity, some very the discovery, design, manufacture and distribution of
difficult issues remain unresolved. We shall examine chemical products in the context of many conflicting
them in the final section of this paper. goals.”
What is process systems engineering? What is different about PSE?
Three years ago, Ignacio Grossmann and I prepared a We noted in our perspectives paper that much of
“Perspectives in Chemical Engineering” article for the chemical engineering research is science-based. Its
AIChE J on the process systems engineering area1. Our intellectual base is the study of nature, where the goal is
first task was to define this area. We noted “companies to seek and understand the “truths” of nature. We
must design and operate chemical processes effectively perform as engineers when we use this understanding to
and efficiently so they may survive in today’s highly create useful artifacts for the betterment of humanity.
competitive world. Providing the methods, tools and PSE resides within the activity of engineering, and its
people that allow industry to meet its needs by tying intellectual challenge is the discovery of concepts and
science to engineering is a compelling aspect of Process models that will allow us to improve all levels of
Systems Engineering (PSE).” I find it hard to improve decision making processes. What we create does not
on this view. have to be an exact representation but rather a useful
one that leads us with least effort to correct decisions.
We elected to broaden the more conventional view
that our domain was process design, control and
operation. We start at the nano-scale with decisions

53
What are some accomplishments in PSE? much better. There has been a lot of later work to
improve one’s ability to converge to solutions under
Computing
adverse numerical conditions, and throughout the 80s
PSE as an active research area started in the late 1950s, and 90s, we participated in the creation of systems to
at the same time that engineers could access computers aid engineers in writing and debugging process models.
that were more than adding machines. The first
Discrete decision-making
computer I used was a Univac 1101. It was a room full
of tubes (valves in UK English) that would have, had In the early 1980s, Ignacio Grossmann added binary
the air conditioning failed, led to the burning down of variables to our models. His first work showed how to
the building in which it was housed. It had 1000 words set up and solve mixed integer linear programming
of memory and a computer language that was one to one (MILP) problems. Binary variables allow one to choose
with machine instructions – and it was little more than among discrete alternatives while optimizing, such as
programming in 0s and 1s. By 1960, I was using a dual choosing between using a single- or a two-stage
drum IBM 650 at the Forrestal Research Center at compressor or the choosing of the number of trays in a
Princeton with twice 2048 words of rotating drum column. The solving algorithms involve the use of
memory and by the next spring a CDC 1604, a decision trees and bounds rather than just gradients, and
transistor-based machine with 32K words (at this time they take a lot longer. He and his students
words were of the order of four bytes). The first subsequently developed the DICCOPT algorithm, which
supercomputer was likely the CDC 6600 that appeared became a part of the commercially available modeling
in the early 1960s. In 1977 we saw the CRAY 1, a system GAMS and thus accessible to all of us to solve
super-supercomputer that, according to the folklore I mixed integer nonlinear programming (MINLP)
have heard, had about the capability of a 100 megahertz problems. Many in PSE have now contributed to the
Pentium. The IBM personal computer appeared in 1983 literature that abounds with many, many tricks in
– I had two. It had 640K words of fast memory and a formulating and strategizing about classes of problems
10 megabyte disk and cost about $4400. 1995 was the so one can solve them. Industry often attests to using
year of the internet explosion. Today we have giga- this type of method and saving millions of dollars
everything on our desk. My latest computer is dual annually. One example is to plan where and when to
multithreaded xeons rated at 2.66 gigahertz each, with 2 place oil rigs in the Caribbean. Another is to plan the
gigabytes of fast memory and a dual, auto backup, 200 investments for annual plant modifications while
gigabyte disk. My new Ethernet card has a gigabit simultaneously making monthly production decisions to
performance label. control inventories when one can only guess future
demand and technology improvement. These
Analysis and optimization techniques are routinely use to schedule batch/semi-
continuous, multi-purpose plants.
With this development in computing, PSE continually
attacked larger and larger problems. Three data points
Synthesis
are the solving of 19 equations in 19 unknowns in about
as many hours in the late 1950s, to the bragging about Another area in which PSE researchers contributed is in
solving 650 equations in the mid-1960s, to the solving developing so-called “process synthesis” methods to aid
of a million equations today in a few tens of minutes. engineers to be creative when designing new processes.
A very early paper on distillation by Lockhart in 1947
In the late 1960s, we could optimize a flowsheet,
gave guidelines on whether to separate an ideal ternary
while using the computer time required to perform
mixture by separating the most or the least volatile
about 2000 flowsheet simulations. It took the entire
component first. In the late 1960s, Dale Rudd and his
weekend and often failed to get to the answer. People
students really started others to look at this area. They
became skilled at saving intermediate results every hour
asked if one could automate the selection of equipment
so they would not have to start over when the inevitable
and its configuration first for heat exchanger networks
failure occurred. With the advent of the generalized
and then for a complete flowsheet. In 1971 Ed
reduced gradient algorithm, optimizing a flowsheet used
Hohmann, a student of Lockhart, showed in his PhD
the equivalent in time of 100 converged flowsheet
thesis that one could compute the least amount of hot
calculations. In the late 1970s, we participated in the
and cold utilities required for a process WITHOUT
development of the sequential quadratic programming
knowing the heat exchanger network that could
algorithm, which uses a Newton-based scheme to solve
accomplish it. He tried to publish twice, and peer
the model equations and the necessary equations for
review rejected his papers. He also showed how one
optimality together. We demonstrated that we could
could estimate the heat exchange area required.
optimize in about the time we took simply to solve the
Linnhoff in 1977 independently discovered the same
flowsheet equations alone. We believe it cannot get

54
ideas using a different formulation and brought these include aiding engineers to decide what to measure,
ideas into engineering practice. With these tools, how to structure the control system, and how to alter the
synthesis reached the goal of outperforming what design to improve one’s ability to control a process.
engineers could do without them.
We also see work, often based on artificial
In the 1970s, Gary Powers and many in chemistry intelligence approaches, for creating computer support
looked at doing what I remember doing as a student in systems to help operators, and now computers,
organic chemistry in the late 1950s: invent different determine what just went wrong and how to take safe
ways to synthesize a particular organic molecule. This corrective action in the running of a process.
automation showed that there were literally millions of
Control systems involve the interaction of
pathways, most of which unaided humans would never
operators, computer programs, PLC circuits, control
think to explore.
elements, and the process units. There is much
In addition, the early 1970s produced many papers opportunity in such a complex structure for there to be
on how to find the best sequence of distillation columns errors, many of which could be hidden catastrophes in
when separating relatively ideal mixtures. The real fun the making. Can one hope to “verify” the correctness of
was the work on designing distillation and decanter such systems? Gary Powers, continuing his earlier
systems for separating highly non-ideal mixtures, where work on safety, introduced the use of methods to verify
azeotropes and liquid/liquid behavior abound. This the correctness of computer chips (e.g, a Pentium
latter work appeared first in the late 60s in the ex-Soviet computer chip) to checking such systems. He is talking
literature (they had to be smart on paper because they about 1050 pathways, which makes Avogadro’s number
did not have computers – see a soon-to-appear book by looks small in comparison.
Petlyuk) and then with the work of Doherty and Perkins
and then Doherty and his students.
So what is next?
In the early 1980s, Tennessee Eastman (now
PSE has had a rich and exciting history. The academic
Eastman Chemicals) built a plant to manufacture methyl
community has also grown from only a handful of us in
acetate that required about one sixth (yes, 85% less) the
the 1960s to hundreds today (with a large percentage --
investment and operating expenditures of previous
often stated to be about half of the academics -- on the
plants. Their design had one extractive, reactive
Roger Sargent family tree). Engineers now routinely
distillation column in it, and it started many research
use our tools and methods to improve how they design,
efforts into understanding reactive distillation. Again,
control, and operate processes.
among the first were Doherty and his students. Several
of the simulation companies added reaction to their Our domain for the last four decades has been
column simulation packages. However, simulation and chemical processes, with an emphasis on commodity
much of the analysis did not suggest to anyone if one chemical processes. While far from a fully explored
should use reaction and, if so, where to place it in a problem area, it is becoming increasingly difficult to use
column. Work still progresses in this area, and we now problems from this domain to raise research funds. If
have a number of insights that can help engineers be we mention the word “distillation,” support agencies
creative in the design of these systems. decline our proposals almost out of hand. So are we
becoming a research topic without a problem domain?
Other areas I will suggest that the future of PSE is going to be in
many, very diverse research areas that the about-to-be-
Most of the above discussion relates to the design of
discussed drivers suggest. This moving of PSE research
conventional chemical processes. Major activities in
into other domains is already happening. We cannot
PSE also cover the control and operation of chemical
just move, however; we must become domain experts in
processes. Process control has moved from the using of
these new areas to feel comfortable in contributing to
frequency response methods and simple PID controllers
them. I see an analogy to many in computer science
in the late 50s, to computer supported optimal control in
who have proposed and developed excellent new
the 60s, to heavily computer supported model predictive
approaches to problem solving but have lacked
control in the last two decades and lately to less
interesting problems on which to apply them. To me it
complex but guaranteed stable control based on
has been much more comfortable to be an engineer who
passivity theory. Better analysis methods include
then learns computer science than the reverse.
creating more accurate dynamic simulators, creating
very fast simulators for training, showing how to But what will be the domains into which we should
analyze, optimize and control dynamic and distributed migrate or areas out of which future PSE researchers
systems, and providing methods to determine the will migrate into PSE? Let us explore the answer by
stability of a controlled process. Synthesis methods first looking at the world today to identify the “drivers”

55
that will push or pull us into what we will do. The Networking is creating another driver:
following are my thoughts. globalization. Companies now operate around the
globe. We can no longer operate in isolation.
Drivers for research
Globalization gives us more markets, but it also gives us
Let us assume (though the last two years could make more competitors.
this a dubious assumption) that people have increasing
Another driver is our progress in biology. Our
amounts of “extra” funds, that is, funds over and above
ability to know, to interpret and to modify gene
the amount they deem necessary to survive. On what
sequences is changing such things as medicine, our
will we spend our extra funds without questioning the
understanding of biology, and how we will grow our
price? I would suggest two: on our health and on our
food.
leisure. On most other things, we tend to question
price, but, on these two, we seem to accept the price To support life as we know it, governments build or
charged. If so, these two areas may be where business regulate the building of infrastructures such as
will put its money in the future. We should at least transportation systems, power generation and
review what we propose against this contention. distribution systems, gas pipeline systems, road
systems, phone systems, water distribution systems,
We are increasingly socially aware. We worry a
waste recovery systems and so forth. Many of the tools
lot more about harming the environment, about
we use in PSE are directly applicable for the design and
sustainability and about the total life cycle costs of
operation of infrastructures. Deregulating many of our
decisions. This worry must increase as we continue to
infrastructures has caused turmoil; this will continue as
harm our planet more and more. Thus, it will have to be
we explore more and more how best to manage these
a driver for the future.
large, often monopolistic systems. New technology is
We are learning how to manufacture very small constantly altering the rules for how to construct them
things, with the dimensions of molecules. NSF has also, as with cellular phone systems and power
programs almost requiring us to say the word “nano” in generation and distribution systems. Integrating them,
the title (with the consequence that a number of such as where the transportation system uses the phone
proposals now discuss 1300 nanometer sized objects). system to effect control, is impacting these systems.
Materials where we can maneuver atoms into desirable
These systems are becoming more integrated and
positions have truly remarkable properties. They are
much more complex. So handling complexity is a
often remarkably stronger, or they are able to detect and
driver, too.
handle minute amounts of material, and so forth. We
will be able to do for pennies what now cost hundreds
And yes, there are still new algorithms to discover
of dollars – and we can now make these nano-objects.
and tools to build
We are only scratching the surface. So “tiny” is a
driving force of the future. Related to this nano area For about four decades now, we have seen enormous
and by using thinner and thinner lines, computers are improvements in the speed of computer hardware.
still following Moore’s law. They will ultimately run There have been equally impressive improvements in
into the limits of physics, but not yet. Perhaps we have algorithms, many of which we just discussed and that
another decade. What can we hope to do with major we in PSE have contributed, such as when solving the
computing that is tiny, fast and virtually without cost for difficult problems we like to pose.
the hardware?
The improvements typically come from tailoring
But in computing, it is the network that is likely algorithms to fit the particular features of the problem.
the real contributor to our future. We are fast becoming The first method to solve simultaneous nonlinear
a wired planet of people. There are two implications: equations involved gridding the space and writing
information is unbelievably available, and we can work embedded DO loops. The approach found the best point
with people around the globe almost as easily as with in the grid and around it regridded with a finer mesh.
someone in the next office. For information, how many Next, we used iterative methods, such as successive
of us hesitate now to look up any topic using Google, substitution, in our flowsheet systems. We added
and with amazing success? One can even use it to accelerators like Wegstein’s method and Broyden’s
check the spelling of a word (unless it is commonly method. Then we implemented full Gauss elimination
misspelled). When we want to know about a company algorithms as the inner loop of Newton-Raphson
– we type in www.companyname.com and find methods, but we could only deal with a few tens of
everything we would ever want to know. We want to equations. Sparse matrix methods developed in
buy a digital camera, we find reviews everywhere and modeling electrical circuits migrated into chemical
get back lists of vendors sorted by price. engineering in the 1970s and dramatically improved

56
computation times for the linear equation solving inner program. Agent-based software systems have both
loop of Newton-Raphson based methods. We also these features. John has measured the impact of both
learned how and when to use iterative methods for diversity and collaboration, and both dramatically
solving linear equations for extra large problems. improve performance on a contrived but very difficult
Optimization went from pattern search to GRG to SQP, optimization problem. With five degrees of freedom,
as noted earlier. We have contributed codes for solving his problem has 108 local optima and with ten over 1017
mixed integer nonlinear programming problems that are (1017 microseconds equals 3200 years). He has also
widely used. We have also explored parallel computing investigated having two competing objectives where his
whereby we decompose problems into parts that we goal is to find the trade-off curve (Pareto surface) for
solve in parallel. them. Single agents seldom find the global optimum;
diverse agent systems typically do, and they find better
It is evident in all this progress that many of us
answers remarkably faster. In parallel computing ten
have been tool builders. Can we continue to be tool
computers might allow problem solution to be four
builders? Actually, I like the analogy to a tree falling in
times faster than for one computer; in agent-based
the forest. If no one is there to hear it, was there a
systems the time reduction could be by a factor of 100.
sound? If no one uses our methods, was there a
method? Yes, we will have to continue to build tools, We see in this approach the opportunity to alter
but fewer of us will get credit for their novelty. how we solve hard problems. It is a “softer” approach.
There are a few elusive proofs already available, but
Is there anything new out there that will change the
when tested this approach really works. It has a
effectiveness of computing? Let us assume that we can
“social” model underlying it, and we know this cannot
have unlimited amounts of high speed networked
be science – right? But it could open some new
computers for solving our problems. Can we find
research areas or allow us to revisit some old ones we
effective ways to exploit such a resource to solve very
could not tackle before.
large hard problems? Parallel computing is a
possibility. Unfortunately, in virtually all
decompositions to parallelize algorithms, one finds a Unsolved problems
coordination step for which all the parallel steps must
I will illustrate the kinds of problems that seem to me to
wait before carrying out the next iteration. If
be unsolved in the process systems area. The first
coordination takes ten percent of the time, then
example will take care of problems being too large.
Amdahl’s law states we are limited to a 10 fold decrease
Then we will consider problems that we do not know
in “clock” time, even if we have hundreds of computers.
how to solve yet, although many people have worked at
Agent-based systems: Another approach is the solving small examples of them.
research of John Siirola, who is currently doing his PhD
I cannot solve a problem with 10,000 binary
under the direction of Steinar Hauan and me. His topic
variables – yet
is the use of agent-based computing to solve hard
problems. This type of approach has been in the A trivial answer for noting an unsolved problem is to
literature for about two decades, but we have not been say that we always want to solve modeling,
exploiting it in chemical engineering. People have optimization, control and operating problems that are
generally only used the method without examining larger than our computers will allow us to solve at this
quantitatively its effectiveness. time. The implicit solution is that we need only wait,
and larger computers will solve the problem. History
We can use the following metaphor to appreciate
predicts we will solve problems about 10,000 times
the ideas here. Suppose you were to direct a new
larger in the next decade, and there are 10,000 seconds
research program at the NSF and were given a budget of
in a week. However, we should question if it is only the
$10 million per year. You could choose to fund the one
need for larger computers.
best, perhaps best by far, researcher in the US, or you
could choose to fund the top two dozen. Most of us This simple three column flowsheet will not converge
would agree that the latter approach is better. Now hold
back 5% of the budget and insist that these researchers In analysis and optimization, we all often fail to
converge our problems. Much work in convergence has
meet twice a year and share results.
to do with improving the numerical schemes. For
There are two features of this last approach that example, there are modified Newton schemes that take
appeal. We should like the diversity of having two bounded steps between a steepest descent and a Newton
dozen research programs looking at the topic. Second, direction. These really do make a difference, but I am a
the collaboration should aid researchers to see in the firm believer that we need physical insights to converge
results of others better ways to proceed in their own most of our problems. For example, we may well place

57
a specification on a separation device that device cannot contribution methods available. Also what makes a
deliver. In azeotropic separation, we may initialize a good solvent is a lot more than an activity coefficient.
column in the wrong “distillation region” from which it How much it costs, its toxicity, how easy it is to handle,
will fail to converge, while initializing in the right can we recover it, and so forth are all part of the
region, even if done poorly, will generally succeed. In a evaluation. Also, when one thinks of “order of
“four sided” distillation region, we can find our magnitude,” the number of components in our libraries
distillation problem to be so sensitive to the numbers for which group contribution methods work is not that
(changing the least significant bit of a double precision much larger than the number of groups. With gigahertz
number) that the solution will jump from one behavior computers, we can search over a lot of components
to a very different one. It is still an unsolved problem as directly. So we can often try all pure component
to how we can provide feedback that tells an engineer possibilities rather easily. Mixtures and structures are
why his problem is not converging and how he can another story. There will be mixed integer programs
overcome any such problem. We have made inroads on formulated to solve these problems. What we suffer
this problem in our ASCEND system, but we can only from more than anything though is a lack of good
provide “clues” that the engineer still must decipher. I models for predicting the properties we want. That puts
believe the approach will be to find hidden embedded this problem in part outside the PSE area.
near singularities and bifurcations and to trace their
meaning through the problem. Connecting such a trace Should I buy this tanker of crude oil?
to its physical meaning will be an enormous benefit, but
I pick this problem as it is a very large and very critical
it will be very difficult.
one for an oil company, and it illustrates the complexity
of making business decisions. Buying a wrong tanker
I need a good solvent
of crude can have a serious economic impact on a
A “cute” way to think of a reverse problem is that it is company, much more than solving its design, control,
one in which we supply the answer and ask for the and operating problems optimally. The problem is
question. Reverse questions are at the heart of product many faceted. First, how does the company estimate
and process design, and thus especially of process with any accuracy what is in the tanker? Can its
systems engineering. Assume our tools are models and refineries process it? Oil companies use information on
not experiments, a good assumption for PSE. Given the oil field from which it comes as a first indicator.
what we know and can model, what answers can we Several measurements the suppliers made on the crude
find that will aid us in engineering a solution to a oil aid the oil company to estimate what it is it might be
problem? The flowsheet synthesis problem is a reverse purchasing. However, getting representative samples on
problem of discovering a flowsheet that can which to make the measurements is a difficult problem.
manufacture a desired product. Another is to find a The simple question as to how much oil is actually in
good catalyst using computational methods. Reverse the tanker is also surprisingly difficult to answer, even if
questions always have many, many candidate solutions. one monitors the delivery. Suppose the company has a
good estimate of what is in the tanker. It next needs to
There are a few simple reverse questions that we
account for its forecasts as to what it needs to deliver as
keep thinking we should be able to solve, and we still
products for the next few months. What is currently in
cannot do them very well. One is to find components,
its storage tanks? What products can it make from these
mixtures of components and/or structures of
inventories when it considers having this tanker of crude
components that have some desired physical properties.
and when it does not? It will have to anticipate what
This type of problem is at the heart of designing a
crude it can buy in the near future, also. Where are
chemical product. For example, we use the Wilson
available tanks located into which it can unload this
method to predict the vapor-liquid behavior of many
tanker? What is in them? Will it mix or not? Should
common organic mixtures. Suppose that we supply the
it? Is there a “slot” available at the unloading facilities?
physical behavior we want – e.g., we want a new
Perhaps the company should wait for the next tanker of
solvent for our process. What solvent should we use?
crude? How will currency fluctuations affect this
There is a considerable literature on this simple
decision? Should it buy the tanker with the intention to
problem. Most of the time a paper on this topic will
trade it in a day or so?
suggest stating the activity coefficient one wants and
then seeking the groups (as in group contribution Can we hope to create and maintain the models
methods) that will give such a coefficient. The problem needed to respond to all these questions? One will need
is then one of constructing molecules from those to represent many of the variables probabilistically.
groups, and these papers typically find there to be many, Can we solve such an immense problem? What is the
including some good surprises. Unfortunately, only a right problem to solve? Are we interested in the low
few of the components we wish to consider have group

58
risk solution more than the high profit one? How will suppliers until the supplier company reverted to its old
our decision affect our competitor’s position? way of running its process.
Problem formulation – the oil tanker problem, Another anecdote is more personal. In 1986 we
second chapter won NSF funding for our Engineering Design Research
Center. Our goal was to improve the competitiveness of
We already have many tools for analysis and
US industry by improving how it did its designing. We
optimization. As the above problem should have made
argued that by bringing many engineering and other
clear, we now suffer from not knowing what problems
disciplines together, we would greatly accelerate the
we need to solve. Formulation is more important than
sharing of methods and tools. We envisioned supplying
solving. If we formulate the wrong problems, we get
much improved computer-based analysis and
answers experienced engineers will reject out of hand.
optimization tools.
While it may sound trite, it is better to solve the right
problem approximately than the wrong one optimally. Westinghouse brought us one of our first projects.
We were to watch them design a new control system for
To understand how formulation is critical, what are
a 2000 megawatt coal fired power plant, one in which
the main selling points for portable computers today?
the improved control would reduce emissions while not
Computer speed is not everything. Rather customers
impacting overall efficiencies of the plant. We quickly
are also very interested in battery life, reliability,
discovered that they seldom were running computer
weight, screen resolution, security, connectivity,
simulations and optimizations. Other researchers doing
ruggedness, and service. If one overlooks an important
ethnographic studies of such activities have estimated,
customer need, one could doom his product.
as did we, that the designers used these tools for only
Can there be strategies we can uncover to aid in about 15% of the time they spent on the design. Rather
formulation? Can there be tools we can develop to do the design team members were in meetings sharing
the same? information, discussing how to make decisions, what
decisions they had made and so forth. We started very
How can I get my company to be one that survives? quickly to question how we were to improve how they
(subtitled: “Designing as a research topic”) – and would do design if we only considered supplying
the saga started above continues improved analysis and optimization tools.
I also believe that designing is an important research
Since the spring of 1999, I have been directing the
domain but designing in a much broader sense than we teaching of a course on product design at CMU. It is
have generally defined it to date. There is no doubt that open to all juniors, seniors and graduate students at
the tools we discussed above are enablers for a company
CMU. Company sponsors supply us with problems and
to be good at designing. Companies interested in low expertise, as well as funding, to run the course. We
cost have to be skilled at modeling and optimizing. have discovered that diversity on these teams (having a
Companies should not ignore the available synthesis computer scientist, a chemical engineer, an industrial
tools whose positive impact many have now well designer, a mechanical engineer and a business major
documented.
on the same team) dramatically improves what they will
These tools enable a company to be good, but they create, something the students always find very
do not make a company great. We suggest that only the surprising and enjoyable.
great companies are likely to be here in the future.
A main theme of a very interesting book by
What makes a company great? Is there research open to Bucciarelli2 is that designing is a social as well as a
chemical engineers that speaks to this question? I technical process. Design teams carry out designs, and
suggest that the great companies are those that
the design it selects, of the many possible, is a strong
understand what they are about, are creative, and function of the personal values of its members. He
understand what their customers want. argues that design is social because of the process by
In commodity chemicals, price is likely the major which one carries it out and social by the impact that
issue. But is it? An anecdote will illustrate. A to-be personal values have on its outcome. To quote
unnamed, but real, chemical company delivered a Bucciarelli: “The realization that design is a social
commodity chemical to one of its customers. This process, that alternative designs are possible, and that a
supplier company discovered it could deliver a product design’s quality is as much a question of culture and
that went from being 98% pure to being 99% pure while context as it is of a thing in itself or of the dictates of
decreasing its costs. Of course, a purer feed should science or market forces -- all this is a prerequisite to
please its customer. Well, it did not. The customer had moving beyond simplistic images and myths about
tuned its process to use the less pure feedstock. Its technology and doing better as designers, as corporate
product became slightly discolored. It almost switched

59
strategists, as government regulators, as consumers, and contribute to it? Should they? Obviously I have
as citizens.” thought so and still do. I do not see how we can forsake
understanding what it takes to make companies great
This topic is fraught with problems. The most
rather than just good.
significant is that humans and their behavior are a major
part of what we must understand. Chemical engineering
researchers are usually not very interested in such a soft References
problem. However, how can we really aid designers if 1
Grossmann, I.E., and A.W. Westerberg, “Research
we ignore these issues? Only companies that deal with
Challenges in Process Systems Engineering,” Invited
these issues will be great at designing.
Perspective, AIChE J, 46(9), 1700-1703 (2000).
We have found our research niche in this area by
2
developing tools to support information management Bucciarelli, L, Designing Engineers , MIT Press
and to support collaboration. This area is far from (1994).
understood. Can chemical engineering researchers

60
Unsolved Problems in Nanotechnology

61
 Biographical sketch of Matthew Tirrell

Matthew Tirrell received his undergraduate education in Chemical Engineering at Northwestern University
and his Ph.D. in 1977 in Polymer Science from the University of Massachusetts. He is currently Dean of the
College of Engineering at the University of California, Santa Barbara. From 1977 to 1999 he was on the
faculty of Chemical Engineering and Materials Science at the University of Minnesota, where he served as
head of the department from 1995 to 1999. His research has been in polymer surface properties including
adsorption, adhesion, surface treatment, friction, lubrication and biocompatibilty. He has co-authored about
250 papers and one book and has supervised about 60 Ph.D. students. Professor Tirrell has been a Sloan
and a Guggenheim Fellow, a recipient of the Camille and Henry Dreyfus Teacher-Scholar Award and has
received the Allan P. Colburn, Charles Stine and the Professional Progress Awards from AIChE. He was
elected to the National Academy of Engineering in 1997, became a Fellow of the American Institute of
Medical and Biological Engineers in 1998, was elected Fellow of the American Association for the
Advancement of Science in 2000 and was named Institute Lecturer for the American Institute of Chemical
Engineers in 2001.

62
 Unsolved Problems in Nanotechnology: Chemical Processing by Self-Assembly

Matthew Tirrell
Departments of Chemical Engineering and Materials
Materials Research Laboratory
California NanoSystems Institute
University of California, Santa Barbara, CA 93106-5130
[email protected]

Abstract

The many impressive laboratory demonstrations of controllable self-assembly methods generate considerable hope
and interest in self-assembly as a manufacturing method for nano-structured products. The practical realization of
this hope will require several achievements in chemical engineering, including the precision manufacturing of
precursors, expansion of our ideas of molecular chemistry to super-molecular chemistry from a bonding, kinetics
and mechanistic point of view and a focus on self-assembly processes that can be scaled up to interesting levels.
This leads to a rich array of research and process development issues for chemical engineers.

Chemical engineers are playing very significant roles in and electronic devices, to name a few areas of current
developing a new science base in phenomena, structures engineering exploration. Self-assembly is one of the key
and potential applications where understanding at mechanisms by which nature builds products, from
nanometer length scales is crucial. Synthesis and biological molecules such as proteins to larger
demonstration of new catalysts, new multi-component structures such as cells and extracellular matrices, the
polymers and new biomaterials are all fine examples of spatial arrangement of atoms is determined, in large part,
areas to which chemical engineers are making major by information built into the assembling units.
contributions to nanoscale science, in particular but not Complexity, in the sense of development of emergent
exclusively, via the self-assembly route that is the properties of an assembly that cannot readily be
subject of this paper. Whether this new science base envisioned from the constituents (Ottino, 2003), can
leads to important technological developments depends arise spontaneously during self-assembly and often does,
even more on chemical engineers. The thesis of this especially in biological systems. We are only beginning
paper is that the opportunities for chemical engineers to to develop sufficiently sophisticated synthetic
advance nanotechnology demand an emphasis on assemblers to mimic biology in this way, as illustrated
developing suitably effective processing technologies. in Figure 1, from Simon, et al (2001).
Treating self-assembly as a serious chemical process is
rich with chemical engineering opportunities.

Self-assembly is a spontaneous process, guided by

information content intrinsic to the assembling units, in
which multiple levels of structural organization are built
into a product (Kuhn and Ulman, 1995). Owing to the
complexity of structures formed and the nominal ease of
spontaneous organization, self-assembly is increasingly
being examined as a practical chemical processing
method. It is akin to crystallization except that the
complexity of the structural elements involved is greater
in self-assembly and length-scales beyond a single
lattice constant emerge in the product. In some cases,
self-assembly can result in intricate, interlocking, useful Figure 1. Schematic diagram showing how
architectures of matter. The resultant structures are complexity of self-assembly may increase as the
being actively explored as new materials, surface number of building blocks in a copolymer
treatments, catalysts, membranes, photonic materials increases. (adapted from Simon, et al. 2001)

63
Indeed, other routes to self-organization, including those Brief summary and critique of the state-of-the-art of
far from equilibrium (Ottino, 2003), may also be of self-assembly processing
interest for nanotechnology.
Numerous processes for producing chemical products,
The ability to form structures spontaneously from the
surface layers or new materials by self-assembly have
bottom up, using principles akin to biological assembly,
been developed. Table1 gives a substantial but not
has inspired considerable research among chemists,
exhaustive list. A short discussion of what is included,
physicists and chemical engineers to develop useful
and what is omitted, and why, is warranted. Adsorption
products (Whitesides, et al. 1991). The high level of
of polymers consisting of identical units, that is
scientific activity in this field creates several kinds of
homopolymers (not on the list), is a widespread and
major opportunities for chemical engineering research.
effective practical method to modify materials surfaces.
The overall opportunity might be characterized as a shift
However, it is a random process without any particular
from product to processing research in self-assembly.
guidance to the assembly coming from the molecular
As self-assembly attempts to evolve from laboratory
structure, and therefore it lacks an important ingredient
techniques and provocative demonstrations of structure,
(a trajectory guided by information content of the
to practical, larger-scale, applied processing methods, a
assemblers) in the definition of self-assembly. On the
broad spectrum of important engineering research
other hand, in the adsorption of multi-component
problems is being generated.
polymers, such as block copolymers or hydrophobically
modified polymers, some pieces of the molecule are
Opportunities for new chemical engineering research in
driven by their chemical nature to stick to interfaces;
self-assembly are the focus of this article. What this
others are not. The resulting molecular arrangement is
requires, among other matters, is taking self-assembly
directed by information (internal structure) in the
seriously as a processing method and asking, “What are
adsorbing molecules. Amphiphilic molecules "know"
the barriers to its development and widespread
which end is up when placed in a surface or interfacial
implementation in the commercial sector?” This review
region.
takes a processing perspective, appropriate to motivate
this underdeveloped field of chemical engineering
research. A recent review (Schreiber, 2000), from a Table 1: Processes Incorporating self-assembly
surface science perspective, describes the current state Adsorption of multicomponent polymers
of laboratory growth methods and structural Cooperative supramolecular self-assembly of surfactant
characterization comprehensively, for the case of self- organic mesostructures
assembled monolayers. An up-to-date review on three- Grafting of polymers on interfaces
dimensional self-assembled structures has yet to be
written. This article aims at processing two- and three- Langmuir-Blodgett deposition
dimensional self-assembled structures. Layer-by-layer deposition or sequential adsorption
Micellar control of reactions
An important characteristic of self-assembly processing, Microcontact printing
which changes both the chemistry and chemical
Organizing colloids into arrays and crystals
engineering research perspective, is that the
fundamental objects participating in the process are, Self-assembly of monolayers
more often than not, neither simple atoms nor simple Spin and dip coating of supramolecular solutions and
molecules. Rather, they are modular, prefabricated, dispersions
complex molecules or objects with internal structure Surface directed ordering of molecules at interfaces (liquid
and external size, shape and functionality. The structure, crystals)
size, shape and functionality define the information Surface modification by monolayer or multilayer deposition
content of the assembler. Colloidal crystallization of Templating
structured particles typifies the process of developing Vesicle fusion
organized assembly of complex, multimolecular objects.
The rules of chemical bonding (interaction potentials) Some of the processes in Table 1 have been known for
and chemical kinetics for reactions among objects such quite some time (e.g. Langmuir-Blodgett deposition);
as these to form stable, useful products remain to be others have been developed in the last decade (e.g.
thoroughly explored to form a good basis for process microcontact printing). Note that by definition, in our
development. parlance, a self-assembly process, while being largely
controlled by the chemical architecture of the
assemblers, may have elements that are neither
spontaneous nor inherent in the assembling entities.
That is, some of these are directed assembly processes,
64
 interactive complementarity (usually of opposite
charge). Möhwald and co-workers have generalized this
method to LbL on particle surfaces (Caruso, et al. 1998;
Caruso and Möhwald, 1999) developing the capability,
after destruction of the templating particle (usually by
dissolution), to make self-assembled capsules with
controlled structure, contents and wall properties
(mechanical and transport). Destruction of an organic
template to leave a self-assembled inorganic material
has proven a very effective route to new porous
materials for catalysts, sorbents and membranes (Asefa,
et al. 1999; Melde et al. 1999; Freemantle, 2000).

Microcontact printing (MCP) is another burgeoning

laboratory-scale method incorporating self-assembly
(Kumar, et al. 1994). In the basic application of this
methodology, a stamp is “inked” with a self-assembled
layer. The stamp is then used in MCP processing to
transfer the self-assembled material to another surface.
Figure 2. Extension from molecular to microscale
systems. Selective self- organization of charged
The possibilities for making MCP an efficient, versatile,
colloid particles using patterned multilayers as even continuous, process seem very achievable, based
templates. Initial patterning done with microcontact as it is on classical printing technology. Clearly, many
printing, followed by layer-by-layer assembly, with of the processes in Table 1, MCP included, are not
the final step incorporating electrostatic assembly of purely self-assembly, or even directed assembly
nanoparticles. (adapted from Chen, et al. 2000) processes but involve the intervention of some other
processing technique in concert with self-assembly.
which may incorporate non-thermodynamic influences Self-assembly used as one method in a multi-step
(dipping, printing, spinning, quenching, etc) or effects process is a very likely outcome of much of current
of external influences (templates, surfaces, fields, etc.). nanoscale processing research. This would characterize
In fact, the effective use of these influences will be one methods such as micellar or microemulsion controlled
of the most important issues in developing practical reactions, as well as a range of “hybrid” self-assembly
self-assembly processes. This article does not make a processes. Figure 2 (Chen, et al. 2000) shows the
sharp distinction between purely self-assembly product of a hybrid LbL-MCP process, followed by the
processes and directed assembly processes. In both, the deposition of a layer of colloidal particles.
information content of the assembling objects plays a
major role. Other hybrid processes include ink-jet printing followed
by a self-assembly process such as vesicle fusion
The practical history of Langmuir-Blodgett (LB)
deposition as a processing method (Swalen, et al. 1987)
offers a sharp focus, and a sobering shadow, on the
future of self-assembly processing. LB processing has
never been an important practical technology, despite its
wide-spread utility and unique capability of controlling
structure in laboratory-scale materials assembly, owing
among other things to the difficulty in scaling-up the
process via continuous, parallel or repetitive processing.
As we shall see, continuous processing has been
achieved in few examples of self-assembly processing.

One promising self-assembly method that seems to lend

itself to scale-up and to continuous processing, and also
possesses great versatility in the range of components
that can be used in the process, is Layer-by-Layer Figure 3. Schematic diagram of the process used
growth (LbL), pioneered by Decher and co-workers, to generate nano-crystal alignment by the phage
and now being much more broadly explored (Decher, display method. (Reprinted with permission from
1997). LbL works by successive dipping cycles of Science, Lee, et al. 296:892-895. Copyright 2002
substrates into solutions of molecules with some AAAS.)

65
(Cremer and Boxer, 1999), and biological production of inventing combinations of new technologies for
self-assembled products (Cha, et al. 2000; Belcher, et al. nanoscale processing.
j2000; Seungju, et al. 2000; Wu, et al. 2002), one of
which is illustrated in Figure 3.
New issues in self-assembly processing

Most of the current self-assembly processing methods There are several identifiable directions in which
function primarily to create surface and thin film scientific and technological advances in the chemical
structures (LB, LbL, multicomponent polymer sciences must be made in order for self-assembly
adsorption,). This “two-dimensional” emphasis derives processing to become an important, widely used
both from the many applications for surface processing tool. The directions discussed below are each
modification and for the templating assistance that important; a prioritized list is not possible in the current
surfaces generally offer to self-assembly processes. state of the field.
Nevertheless, much more work is needed in the
direction of three-dimensional self-assembly methods. Precision synthesis of precursors.
As work in colloidal crystallization has shown, not only
is three-dimensional assembly of pre-formed, The ability to form a desired state of organization
supermolecular objects a route to materials with spontaneously depends on the internal structure of the
interesting new structures and properties (Velev, et al. assemblers. It is then clear that information must be
2000) but also, studying the crystallization of objects built into the assembling objects with some degree of
reveals considerable new insight into the inter-object fidelity. There are various avenues for achieving this.
potentials (Davis, et al. 1989). Size, shape (in rigid objects), conformation (in flexible
objects), sequence (in multicomponent macromolecules),
Processes competitive with self-assembly spatial relationships among subunits all can be “read” in
molecular recognition interactions. In synthetic
molecules, small amphiphiles such as surfactants or
An issue that must be addressed is that of competing lipids are classes of molecules that can be synthesized
processes: What competitive processes are available to with precision, that is, they can be made and isolated as
achieve similar targets? Structuring at the nanoscale is pure compounds. The ability to make micellar and bi-
not available exclusively via self-assembly. While the layer membrane structures by self-assembly is a direct
“bottom-up”, additive approach embodied in self- consequence of the precise architectures of these
assembly has much to recommend it, “top-down”, molecules. Impurity in assembling molecular structure
subtractive processes, such as nanolithography by x-ray, creates disorder or other degradation in quality in the
electron beam and UV also have considerable potential self-assembled product.
yet to be tapped. Furthermore, subtractive processes
have a track record of past accomplishment. Direct- Only biological systems currently achieve precise
write processes, which are automated layer-by-layer syntheses of molecules of more than a few hundred
depositions, including ink-jet printing (Sachs, et al. daltons. There are two broadly thought-provoking
1992), stereolithography or 3D printing (Griffith and points in this realization for current self-assembly
Halloran, 1996), are also advancing rapidly, based in research. One is that precision synthesis and fidelity of
part on advances in robotics (Lewis, 2000). Nano- information content is essential for effective self-
imprinting, a nano-scale throwback to the metal- assembly of intricate, functional systems. This should
stamping of, say, auto body parts or refrigerator motivate a broad array of useful chemical synthesis and
housings, is also showing serious promise in electronic process research to produce the starting materials for
device fabrication. (Chou, et al. 2002) More non-biological self-assembly. Protein folding, the
conventional processes, such as lamination and unimolecular self-assembly process by which proteins
multilayer extrusion (Cesarano and Calvert, 2000), are (multicomponent macromolecules with precise
also achieving a greater degree of spatial control. Sol- sequence distributions) take their functional forms, is far
gel processing and certain combustion/aerosol processes from fully understood. However, it is certain that the
also are achieving some success in creating controlled final state of a protein folding process is determined by
nanostructured products. the primary sequence of amino acids in the
macromolecular backbone. Precision synthesis of the
Thus, there are serious competitors to self-assembly sequence controls the assembly. Folding processes of
processing. Engineering research in self-assembly nucleic acids are also increasingly recognized to be
processing should be realistic about comparative functionally important, while information is built into
advantages of alternative processes, and creative about polysaccharides by control of architectural features such
as branching. Successful development of self-assembly
66
processes will place a premium on new chemical or
directed biological routes to pure, precise
macromolecular synthesis and production or particle
formation. Furthermore, it will be fruitful to understand,
in much more detail than we currently do, the sensitivity
of various synthetic self-assembly processes to the lack
of precision in the architecture of the assemblers. Figure 5. Micelles of peptide amphiphiles bear
many similarities to protein macromolecules. They
A second lesson that may be extracted arising from have hydrophobic cores, hydrophilic bio-active
exteriors, globular shapes, and sizes in the tens if
consideration of biological self-assembly is
nanometer range. Connectivity can be introduced
understanding of how nature achieves precision via polymerizable lipid tails.
synthesis. Biological systems do not synthesize
perfectly; rather, considerable effort is expended in the molecule is shown in Figure 5. In our group, we
detecting and correcting errors en route to a pure have been exploring the properties and applications of
product. The chemical processing analogs of this are on- synthetic peptide-lipid conjugate molecules (Tirrell, et
line measurement, control and post-reaction separation al. 2003). The idea of these molecules is to confer on
processes. To achieve the necessary control over unstructured, short, potentially biologically active
molecular architecture, it will be necessary to explore peptide chains, some self-assembly character that will
both chemical and biological synthesis routes with a organize and present the functional peptide in an
realistic eye toward process efficiency. effective way. We have found that micelles formed
from these peptide amphiphiles resemble proteins in
Expanding the idea of the molecule. several significant, and possible useful ways,
enumerated in Figure 5. We now know (Yu, et al, 1998)
Since self-assembly often, even usually, involves the how to use this self-assembly character to drive the
spontaneous organization of objects considerably larger peptide into ordered secondary structures as they are
than molecules, it is useful to begin to think about larger found in intact protein macromolecules. Protein
assembling objects in self-assembly processes as we do analogous micelles are objects that can mimic protein
about molecules in chemical reactions. The atomic and molecules and, in this way, expand the notion of how to
molecular notions of valence, bond strength, etc., have use protein functionality in self-assembly.
inexact but pragmatic counterparts when applied to
interactions among structured macromolecules, Scale-up of self-assembly processes.
nanotubes, particles and other important self-assembly
constituents. Figure 4 (from Manoharan, et al. 2003) As mentioned earlier in the context of Langmuir-
shows that composite, or structured, collodial Blodgett deposition, scale-up of self-assembly processes
nanoparticles can be formed by the controlled to technological significant scales has not been pursued
consolidation of primary particles. The symmetry of to an appreciable degree. That is perhaps natural given
these structured particles endows them with the the state of development of the field. Nevertheless, lack
potential for directional bonding, the molecular analog of ways to scale-up can be the demise of a promising
of the symmetry of electronic wavefunctions. processing route in a technological arena. Some of the
most innovative developments in self-assembly
Another relevant example of an enlarged definition of processing to date have been achieved by chemists,
where advances in assembly on non-planar surfaces
have led to the ability to assemble macroscopic, three-
dimensional structures spontaneously (Jacoby, 2000;
Gracias, et al. 2000; Whitesides and Grzybowski, 2002).
A great opportunity exists for some tools of process
systems research to be applied to self-assembly
processing, and nanoscale engineering in general.
Success in enlarging the capacity of self-assembly
processes by continuous, parallel or repetitive
processing would be very valuable.

Figure 4. Composite particles formed by Kinetics and mechanisms of self-assembly processes.

association of spheres confined within emulsion
droplets. The precisely formed aggregates confer Another important issue is the matter of kinetics. Self-
a selectable series of symmetries on the composite assembly science is grounded in processes that tend
particles (adapted from Manoharan, et al. 2003)

67
toward desirable equilibrium structures. Chemical
processes generally should go as fast as possible,
consistent with product quality. It is difficult to envision
the processing research called for above proceeding
fruitfully without data on the rates of self-assembly
processes. Our view is that, for design and development
purposes, we need to understand fully the trajectories,
through time and structural intermediates, of self-
assembly processes.
This need for research in self-assembly kinetics, and the
related issue of mechanisms, has several facets. The Figure 6(b) Kinetics of bilayer deposition via vesicle
acquisition of such data will stretch and eventually fusion as studied by ellipsometry. (A. Parra, D.
expand the tools we have to observe self-assembly in Stroumpoulis and M. Tirrell, unpublished data)
progress. Research on self-assembly kinetics requires
data in real-time, pushing the field toward methods that in its technological applications, particularly selective
are faster and more highly time-resolved. Accurate coating of the surfaces of microfabricated devices. For
kinetic data are a necessary prelude to predictive models this we felt we needed kinetic data and a mechanistic
of the evolution of self-assembly processes. As the interpretation so that we could extrapolate to new
complexity and number of components in these situations such as deposition on textured surfaces, in
processes increase, predictive models will become more corners, etc. Phase-modulated ellipsometry has proven
important tools in process and product design. We to be a useful tool in kinetic studies (Parra, et al. 2003).
believe it is fair to make the analogy between self- Figure 6b shows data on the accumulation of
assembly and chemical reaction processes. Thorough monolayers (we have proven that the plateau deposition
understanding of chemical bonding, reaction is in fact a monolayer) fit to a model where we have
mechanisms, pathways and kinetics are crucial to considered both the diffusive rate of supply to the
reactor design. Analogs in self-assembly are surface and surface reorganization as important
intermolecular or interparticle potentials, trajectories potentially rate-controlling steps. Our unambiguous
and rates of assembly. conclusion is that surface reorganization dominates in
this case. These kinds of studies are very important to
As an example of work in this direction, we offer our the advancement of this field.
group's work on vesicle fusion. The essential steps of
this physico-chemical reaction are depicted in Figure 6a. Rates of self-assembly processes.
Vesicle fusion, just as LB deposition, is an increasingly
useful laboratory method to coat surfaces with Chemical engineers are seldom satisfied with the rates
monolayers of lipid molecules. We have been interested of uncatalyzed chemical reactions for technological
purposes. Biological self-assembly is, in some situations,
promoted or directed by protein or templates. The fact
that self-assembly proceeds down a free energy gradient
to a thermodynamically determined endpoint is
reassuring in some respects. However, metastable states
of varying duration may intervene. Equilibration is
unhurried; chemical manufacturing is not. It is
challenging to think of how one might design a catalyst
for a self-assembly process. Physical or chemical
templating, for example, a type of epitaxial growth, is
one possible route, which in self-assembly has been
explored more for its effect on structure than on rate
(Ball, 1997). External fields can profoundly affect the
rates of self-assembly processes (Koppi, et al. 1993) and
should be explored more extensively in this vein.
Perhaps more than in simpler chemical reaction
processes, the size and complexity of objects involved
Figure 6(a) Schematic diagram of vesicle fusion. in self-assembly are such that diffusion limitations are
Attractive forces between vesicle and surface cause more the rule than the exception, motivating a need for
distortion of vesicle shape, eventual rupture of the new research in transport phenomena applied to this
vesicle and ultimately deposition of a nearly field.
complete bilayer.

68
 Belcher, A. M., X. C. Wu, R. J. Christensen, P. K.
Characterization and process control. Hansma, G. D. Stucky, and D. E. Morse, "Control of
Crystal Phase Switching and Orientation by Soluble
Practical processes must be controlled to produce Mollusc-Shell Proteins," Nature, 381, 56 (1996).
valuable products; in the case of self-assembly, these
are products of intricate internal structure. Real-time Caruso, F., and H. Möhwald, "Protein Multilayer
structural analyses present instrumental challenges that Formation on Colloids through a Stepwise Self-
must be addressed. Since self-assembly often produces Assembly Technique," J. Am. Chem. Soc., 121, 6039
structured products over several length- and time-scales, (1999).
techniques capable of measuring, with resolution, over
multiple length- and time-scales are needed. Informative, Caruso, F., R. A. Caruso, and H. Möhwald,
on-line measurement is a prerequisite for process "Nanoengineering of Inorganic and Hybrid Hollow
control. Characterization of defects is particularly Spheres by Colloidal Templating," Science, 282, 1111
important for the applications envisioned. Furthermore, (1998).
self-assembly processing of nanostructured products is a
driver for new research in control of distributed Cesarano, J., and P. Calvert, "Freeforming Objects with
processes (Christofides, 2001). Control of distributed Low-Binder Slurry," U. S. Patent 6,027,326 (2000).
systems such as self-assembly will also rely heavily on
the development of tractable, predictive models. Cha, J. N., G. D. Stucky, D. E. Morse, and T. J. Deming,
"Block copolypeptide-mediated biomimetic synthesis of
Concluding remarks ordered silica structures," Nature, 403, 289 (2000).

Chen, K., X.-P. Jiang, L. C. Kimerling, and P. T.

It has been remarked that new directions in science are Hammond, "Selective Self-Organization of Colloids on
launched by new tools more often than by new concepts Patterned Surfaces," Langmuir, 16, 7825 (2000).
(Dyson, 1997). Self-assembly is a blend of concepts and
tools; the field will advance more rapidly as we turn Chou, S.Y., C. Keimel, and J. Gu, “Ultrafast and Direct
concepts into techniques. Research in the chemical Imprint of Nanostructures in Silicon,” Nature, 417, 835
sciences and technology over the next few years, (2002).
pursued with a clear view to the barriers and
competition facing self-assembly, is likely to produce Christofides, P. D., "Control of Nonlinear Distributed
important new progress, not only for self-assembly Process Systems: Recent Developments and
processing but for the optimum processing routes to Challenges," AIChE Journal, 47, 514 (2001).
nanostructured chemical products.
Cremer, P. S., and S. G. Boxer, "Formation and
Acknowledgement Spreading of Lipid Bilayers on Planar Glass Supports,"
J. Phys. Chem. B., 103, 2554 (1999).,
This article is an updated and expanded adaptation of a
Perspectives articles that appeared in AIChE Journal 20 Davis, K. E., W. B. Russel, and W. J. Glantschnig,
months ago (Texter and Tirrell, 2001). During the "Disorder-to-Order Transition in Settling Suspensions
preparation of this article, MT received partial support of Colloidal Silica - X-ray Measurements," Science, 245,
from National Science Foundation: MRSEC Program 507 (1989).
Award No. DMR-0080034 and the Chemical and
Transport Systems Program Award No. CTS-0103516. Decher, G., "Fuzzy Nanoassemblies: Toward Layered
Polymeric Multicomposites," Science, 277, 1232 (1997).
Literature Cited
Dyson, F., Imaged Worlds, Harvard University Press,
Cambridge, MA, 1997, pgs. 50-51.
Asefa, T., M. J. MacLachan, N. Coombs, G. A. Ozin,
"Periodic Mesoporous Organosilicas with Organic Freemantle, M., "Novel Mesoporous Hybrids Prepared,"
Groups Inside the Channel Walls," Nature, 402, 867 Chemical and Engineering News, January 24, 2000, pg.
(1999). 33.

Ball, Philip, "Made to Measure: New Materials for the Gracias, D. H., J. Tien, T. L. Breen, C. Hsu, and G. M.
21st Century," Princeton University Press, Princeton, N. Whitesides, "Forming Electrical Networks in Three
J., 1997. Dimensions by Self-Assembly," Science, 289, 1170
(2000).
69
Griffith, M. L., and J. W. Halloran, "Freeform by Atomic Force Microscopy”, Langmuir, 16, 1806
Fabrication of Ceramics via Stereolithography, "J. Am. (2000).
Ceram. Soc., 79, 2601 (1996).
Sachs, E., M. Cima, P. Williams, D. Brancazio, and J.
Imhof, A., and D. J. Pine, "Ordered Macroporous Cornie, "3-Dimensional Printing - Rapid Tooling and
Materials by Emulsion Templating," Nature, 389, 948 Prototypes Directly from a CAD Model," J. Eng. Ind.,
(1997). 114, 481 (1992).

Jacoby, M., "Self-Assembling Microelectronics," Schreiber, F., "Structure and Growth of Self-
Chemical and Engineering News, 78, 7, August 21, Assembling Monolayers," Prog. Surf. Sci., 65, 151
2000. (2000).

Koppi, K. A., M. Tirrell, and F. S. Bates, "Shear- Simon, P. F. W., R. Ulrich, H. W. Spiess, U. Wiesner,
Induced Isotropic-to-Lamellar Transition," Phys. Rev. “Block Copolymer-Ceramic Hybrid Materials from
Lett., 70, 1449 (1993). Organically Modified Ceramic Precursors”, Chem.
Mater., 13, 3464 (2001).
Kuhn, H., and A. Ulman, Chapter 1 in, Thin Films,
Volume 20, A. Ulman, ed., Academic Press, N.Y., 1995, Swalen, J. D., D. L. Allara, J. D. Andrade, E. A.
pg. 1. Chandross, S. Garoff, J. Israelachvili, T. J. McCarthy, R.
Murray, R. F. Pease, J. F. Rabolt, K. J. Wynne, and H.
Kumar, A., H. A. Biebuyck, and G. M. Whitesides, Yu, "Molecular Monolayers and Films," Langmuir, 3,
"Patterning Self-Assembled Monolayers - Applications 932 (1987).
in Materials Science," Langmuir, 10, 1498 (1994).
Texter, J. and M. Tirrell, "Chemical Processing by Self-
Lee, S-W., C. Mao, C. E. Flynn, A. M. Belcher, Assembly," AIChE Journal, 47, 1706 (2001).
“Ordering of Quantum Dots Using Genetically
Engineered Viruses,” Science, 296, 892 (2002) Tirrell, M., E. Kokkoli, and M. Biesalski, "The Role of
Surface Science in Bioengineered Materials," Surface
Leonenko, Z. V., A. Carnini, and D. T. Cramb, Science, 500, 61 (2003)
“Supported Planar Bilayer Formation by Vesicle
Fusion: The Interaction of Phospholipid Vesicles with Velev, O. D., A. M. Lehoff, and E. W. Kaler, "A Class
Surfaces and the Effect of Gramicidin on Bilayer of Microstructured Particles Through Colloidal
Properties using Atomic Force Microscopy”, Biochim. Crystallization," Science, 287, 2240 (2000).
Biophys. Acta, 1509, 131 (2000).
Whitesides, G.M., J. P. Mathias, and C. T. Seto,
Lewis, J. A., "Colloidal Processing of Ceramics," J. Am. "Molecular Self-Assembly and Nanochemistry: A
Ceram. Soc., 83, 2341 (2000). Chemical Strategy for the Synthesis of Nanostructures,"
Science, 254, 1312 (1991).
Manoharan, V. N., M. T. Elsesser, D. J. Pine, “Dense
Packing and Symmetry in Small Clusters of Whitesides, G.M., and B. Grzybowski, "Self-Assembly
Microspheres”, Science, 301, 483 (2003). at all Scales," Science, 295, 2418 (2002).

Melde, B. J., B. T. Holland, C. F. Blanford, and A. Stein, Yu, Y-C., V. Roontga, V.A. Daragan, K. H. Mayo, M.
"Mesoporous Sieves with Unified Hybrid Tirrell, and G. B. Fields, "Structure and Dynamics of
Inorganic/Organic Frameworks," Chem. Materials, 11, Peptide-Amphiphiles Incorporating Triple-Helical
3302 (1999). Protein-Like Molecular Architecture", Biochemistry, 38,
1659-1668 (1999).
Ottino, J.M., "Complex Systems," AIChE Journal, 49,
292 (2003). Yu, S., and D. Tirrell, "Nanometer-Scale Smectic
Ordering of Genetically Engineered Rodlike Polymers:
Reviakine, I. and A. Brisson, “Formation of Supported Synthesis and Characterization of Monodisperse
Lipid Bilayers from Unilamellar Vesicles Investigated Derivatives of Poly(-benzl, L-glutamate)," J. Am. Chem.
Soc., 28, 122 (2000).

70