Part A

General Principles,
Methods and Reaction Engineering

1 Introduction
1.1 Principles of Heterogeneous Catalysis
1.2 Development of the Science of Catalysis
1.3 Development of Industrial Catalysis
1.1 Principles of Heterogeneous
Catalysis
M. BOUDART
1.1.1 Introduction
Catalysis by solid materials has been observed quanti-
tatively at temperatures as low as 78 K and as high as
1500K; at pressures between and lo3 bar; with
reactants in the gas phase or in polar or nonpolar sol-
vents; with or without assistance of photons, radiation,
or electron transfer at electrodes; with pure metals as
unreactive as gold and as reactive as sodium; with
multicomponent and multiphase inorganic compounds
and acidic organic polymers; and at site time yields as
low as lop5s s l (one turnover per day) and as high as
lo9 s-l (gas kinetic collision rate at 10 bar).
With these facts in mind, the task of presenting the
principles of heterogeneous catalysis is a daunting one.
Indeed, the chemistry of catalysis is as varied as chem-
istry itself. Moreover, the related subjects cover many
disciplines: surface science; inorganic, organic, and or-
ganometallic chemistry; physical chemistry and spec-
troscopy; theoretical chemistry; materials science; mod-
eling and molecular graphics; and catalytic reaction
engineering. There exist catalytic materials that are
interesting but not applied. Then there are model cat-
alysts that serve as standards for comparison with
catalytic materials and commercial catalysts. Only the
latter are robust, economical and complex, but their
physical and chemical characteristics depend frequently
on hidden, unknown variables.
Principles of heterogeneous catalysis are more likely
to be formulated following the study of model catalysts
Handbook of Heterogeneous Catalysis
Edited by,G. Ertl, H. Knozinger, J. Weitkamc
Copyright 0VCH Verlagsgesellschaft mbH.1997
in ideal reactors than from performance data obtained
with commercial catalysts. The latter are obtained with
impure or mixed feedstreams reacting in complex re-
actors made of practical materials. Although principles
rarely solve the urgent problems of industry, they can
guide the researcher, inventor, and innovator of cata-
lysts and catalytic processes. This statement serves as
an operational definition of principles as discussed in
this chapter.
In this sense, principles are distinct from concepts.
The former are more general; the latter are more spe-
cialized, Principles are not tied to a class of reactants
or a type of solids. Concepts are to the mind what in-
strumentation or techniques are to the material world.
Examples of principles are the headings of the sections
of this chapter. Examples of concepts are discussed in
the beautiful book of Rideal [l]. The interplay between
concepts and techniques in catalysis has been illus-
trated by Thomas [2]. The editors of this Handbook
charged me to write on principles without telling me
what they are. Beyond the operational definition given
above, my interpretation of principle is based on re-
flections of Henri Poincare [3]. What I call a principle
in this chapter is a thought or law or rule of thumb that
is useful in as many situations as possible: its value is
directly related to its generality, as noted by Poincare.
1.1.2 Catalytic Cycle
The classical definition of a catalyst has been discussed
many times [4].In brief, a catalyst is a substance that
changes the kinetics but not the thermodynamics of
a chemical reaction. Another definition of a catalyst
is more informative: a catalyst is a substance that
transforms reactants into products, through an uninter-
rupted and repeated cycle of elementary steps in which
the catalyst participates while being regenerated in its
original form at the end of each cycle during the life of
the catalyst [5].
The catalytic cycle is the principle of catalytic action.
The actioity of the catalyst is defined by the number of
revolutions of the cycle per unit time, or turnover rate
or turnover frequency (TOF). The latter is often in the
References see page 22
2 1 Introduction
vicinity of one per second, in order of magnitude
(Section A.5.2.1). The life of the catalyst is defined by
the number of turnovers observed before the catalyst
dies. This number must be larger than unity; otherwise
the substance used is not a catalyst but a reagent. Cat-
alyst life can be very short, as in catalytic cracking of
oil, or very long, corresponding to as many as lo9 turn-
overs in ammonia synthesis.
In heterogeneous catalysis, the catalytic substance is
found as an active site or active center at the surface of
a solid. This solid needs to be activated, for example
reduced, before it works as a catalyst. For true catal-
ysis, the complete cycle must be carried out without
interruption in a single reactor. The fascination of cat-
alysis is that catalytic cycles exist in steadily increasing
numbers and complexity. Catalytic reactions are still
discovered and developed long before their mechanism
is understood at even a qualitative incompletle level. A
heuristic hypothesis accounting for the apparent luck
of the discoverers is discussed elsewhere [6]: it is based
on a powerful principle - that of the kinetic steady
state.
1.1.3 Kinetic Steady State
The principle is due to Bodenstein and has evolved
over time [6]. For simplicity, let us assume that each
step in the cycle must occur only once per cycle. Then,
according to the approximation of Bodenstein, the net
rate, i.e. the forward rate minus the reverse rate for
each step, is equal to the net rate at which the cycle
turns over. The validity of the Bodenstein approx-
imation relies on the high reactivity of all intermediates
in the cycle, as compared to the reactivity of reactants
and products. Because of this difference in reactivity,
the amount of these reactive Bodenstein intermediates
in the reacting system is very small as compared to the
amount of reactants and products. A colloquial way to
express this situation is to say that the catalyst is pres-
ent in the system in so-called catalytic amounts.
At the beginning of reaction, it takes some time,
called relaxation time, before the Bodenstein inter-
mediates reach their steady-state concentration. This
relaxation time is at least equal to the turnover time,
i.e. the inverse of the turnover frequency. However, for
the steady state approximation to have practical val-
idity, this relaxation time must be short as compared
to the time for reaction. At the end of reaction, if
the concentration of reactant becomes very small, the
Bodenstein approximation may not be valid either.
Thus, a possible breakdown of the Bodenstenn approx-
imation may occur for the catalytic removal of pollu-
tants down to the ppm level [7].
One reason why the Bodenstein principle of the
steady state or the quasisteady state is so important is
that it permits a definition of two other important
principles: the rate determining step and the kinetically
significant step, as discussed elsewhere in this Section
A.5.2.1. These principles are meaningless unless the
Bodenstein kinetic steady state prevails. This can be
appreciated by noting that the notion of a “slow” step
in a catalytic cycle at the steady state is a misnomer
since all steps proceed at the same net rate. The Boden-
stein principle remains the most important foundation
of our thinking about how catalytic cycles turn over.
Thus, the net rate of any elementary step in a cata-
lytic cycle is positive, i.e. the reaction proceeds from
left to right in the desired direction only if the chemical
affinity of that step is positive. The affinity is defined as
- ( d G / d ( ) , , where G is the Gibbs free energy and 5 is
the extent of reaction. But even if the standard affinity
of an elementary step is negative, this unfavorable
thermodynamic situation may be corrected so that the
affinity of the step becomes positive as it must if the
cycle is to turn over in the desired direction. The pos-
sibility of this correction is due to the kinetic coupling
between steps at the Bodenstein steady state. This cou-
pling shifts the unfavorable equilibrium of a step by
virtue of the principle of Le Chatelier if the concen-
tration of a Bodenstein intermediate either accumulates
if it is a reactant or is removed if it is a product. This
shift happens until the net rate of all steps in the cycle is
the same as it must be at the kinetic steady state. As
mentioned at the end of Section 1.1.2, this kinetic sit-
uation may explain the good luck of the discoverers of
catalytic cycles who do not know, and therefore cannot
control, the thermodynamics of individual steps in the
cycle.
Finally, it must be noted that on catalytic surfaces,
multiple steady states may exist as well as unsteady
phenomena in space and time. They are currently the
subject of many studies that have led to the discovery
of kinetic oscillations, chaos, and formation of surface
concentration patterns [8]. In spite of their fascinating
beauty, their discussion is not within the scope of this
chapter.
1.1.4 Microscopic Reversibility
This extrathermodynamic constraint says that at equi-
librium every physicochemical elementary process, and
its reverse, proceed at the same rate. In chemical ki-
netics, such a detailed balancing of individual reactions
in a cycle actually led Onsager to the formulation of his
reciprocal relations, as related by Denbigh [9]. Thus,
for any elementary step, ( k + / k - ) = K , where k+ and
k- are the forward and reverse rate constants, and K is
the equilibrium constant. This relation is strictly valid
only at equilibrium. Away from equilibrium, it remains
valid provided that transition state theory remains

applicable, a restriction that does not appear too severe
in heterogeneous catalysis, since transition state theory
remains applicable for desorption into a vacuum, an
extreme case of remoteness from equilibrium [lo]. Thus
the principle of microscopic reversibility appears to re-
main valid for any elementary step in a heterogeneous
catalytic reaction.
As to the situation at and near equilibrium for the
entire catalytic cycle, all steps in the cycle at equilib-
rium must be taken into account, in both directions
(Section A.5.2.1). However, at a remote distance from
equilibrium, the situation may change: only a limited
number of steps, or perhaps only one, remains kineti-
cally significant, perhaps only in the forward direction.
A typical situation in catalysis is that of catalyst satu-
ration with a reactant or a product, in which case the
rate of turnover may be independent of the extent of
reaction, a zero-order situation that obviously must
break down as equilibrium is approached.
As a result, the well-known imperative according to
which a catalyst is a substance that affects the rate of a
reaction in both directions, does not imply that the best
catalyst for a reaction from left to right, is also the best
catalyst for the reaction from right to left, when the
reaction is run far from equilibrium [7]. In this context,
by best catalyst, it is meant the one with the highest
turnover frequency, a simplification that does not hold
in practice when the figure of merit of a catalyst de-
pends on many attributes besides turnover frequency,
as will be discussed later.
As an example, consider ammonia decomposition.
At high pressures and low temperatures, the rate is zero
order and the rate constant k d is that of desorption
from a surface saturated with nitrogen. By contrast, at
low pressures and high temperatures, the rate is first
order in ammonia and the rate constant k, is that of
adsorption of ammonia [12]. Clearly there is no reason
to expect that the catalyst with the highest value of k,
times the concentration of ammonia would be found
on a catalyst that also exhibits the highest value of kd.
Here the comparison is made for a reaction taking
place in very different conditions of temperature and
pressure. For a reaction taking place under identical
conditions, see Ref. 11. Clearly the principle of micro-
scopic reversibility must be applied with caution when
it comes to a catalytic cycle, as opposed to a single
elementary step.
1.1.5 Principle of Sabatier
In many areas of chemical kinetics, semiempirical rules
have emerged under various names. These rules are re-
lated to linear free energy relationships between the
thermodynamic standard free energy of reaction and
the kinetic standard free energy of activation. Fre-
1.I Principles of Heterogeneous Catalysis 3
quently, if we neglect entropy values, these linear rela-
tionships deal only with the heat of reaction q and the
activation barrier E of an elementary step taking place
in the exothermic direction ( q > 0). If a is an empirical
quantity between zero and unity, the linear relationship
between kinetics and thermodynamics is of Brmsted-
Evans-Polanyi-Semenov type, or, in short, a Brernsted
relationship:
AE = aAq
where AE is a decrease in activation barrier corre-
sponding to an increase Aq in the heat of reaction. In
other words, if we neglect entropic effects, an elemen-
tary step will have a large rate constant in the exother-
mic direction when its heat of reaction q increases. In
the endothermic direction, since the activation energy
is then the sum of the activation barrier E and of the
heat of reaction q , an increase in the value of q will lead
to a smaller value of the rate constant in the endother-
mic direction.
Now, with rare exceptions [ 131, adsorption on a solid
surface is exothermic, and desorption from the sur-
face is endothermic. Thus, generally, an increase in
the binding energy of the adsorption complex will in-
crease the rate constant of activated adsorption. Since
activated adsorption, as first proposed by Taylor [14],
is the normal practical form of chemisorption, this in-
crease in the rate constant of adsorption with binding
energy is generally expected. By the same token, the
rate constant of desorption will generally become
smaller as the binding energy of the adsorption com-
plex increases.
What then is expected for the rate of adsorption and
the rate of desorption when the binding energy of the
adsorption complex increases and the concentration of
the adsorptive in the fluid phase remains the same at
a given temperature? If adsorption starts on a bare
surface and the desorbed species is identical to the
adsorbed species, the rate of adsorption will go down
as the surface fills up with adsorbate, and the rate of
desorption will increase correspondingly until both
rates are equal and equilibrium is reached. The surface
coverage at equilibrium will of course go up as the
binding energy of the adsorbate increases.
However, that situation is not catalysis. During
catalysis, it is possible that an adsorption-desorption
equilibrium may be reached at the steady state, but this
is not a necessary situation at all, because the surface
coverage for a given species at the steady state may be
very different from that at equilibrium. Moreover, in
catalysis, what is adsorbed is a reactant molecule and
what is desorbed is a product molecule. Overemphasis
on what controls only the activation of, say, methane,
or dihydrogen, or dioxygen in a catalytic cycle may be
a distraction from what controls the rate of the cata-
lytic cycle.
References see page 12
4 1 Introduction

400 1i
I cp\ I

i t
I

i
75

I I ,
-4 -3 -2 -1
500
t
(b’ Ii 600
60 80 100 120

Figure 2. Sabatier’s principle: volcano curves T, / K . The tem-

\ i
25 1 7
perature at which the -rate df decomposition of formic acid (ver-
tical axis) has a set value u, correlated with the heat of formation
of the bulk metallic formate, in kcal per equivalent (horizontal
axis) [17].

lo8l
4 3 2 1

Figure 1. The heats of adsorption of (a) ethene and (b) hydrogen

in kcalmol-I (vertical axis) correlated with the logarithm of the
relative rate constant for the rate of hydrogenation per unit surface
1o6
area of evaporated metallic film catalysts (horizontal axis) [16].

Yet, the bridge between thermodynamics and ki-

netics provided by the Br~nstedrelationship plus the
heuristic value of the Sabatier principle offer qual-
itative guidance in the search for catalytic materials.
Sabatier simply emphasized the unstable intermediate
compound formed with one of the reactants. This in- - - 52
termediate must be stable enough to be formed but not lo8
too stable, since it must decompose to yield ultimately lo6
- - 50
the final product or products. Sabatier’s principle is - - 48
to4
a qualitative expression of moderation and it can be
made semiquantitative by means of any quantity re- lo2- - 46
lated to the stability of the intermediate compound. - 44
1 -
The result is a volcano-shaped curve as first proposed
by Balandin [15]. These so-called volcano curves ap- 1 1 I I I I I
VllA Vllll V1112 V1113 IB
pear on a diagram with a measure of the rate of
a catalytic reaction along one axis and a measure of Figure 3. Activity pattern for the hydrogenolysis of ethane [20].
the stability of the putative intermediate compound on Relative rates per unit surface area of supported metal on left
the other axis. The latter can be a heat of adsorption of axis; percentage d character of the metallic bond according to
Pauling on right axis; volcano curves can be seen for the third
one of the reactants [16] (Fig. l), or the heat of for- transition series. A correlation between the percentage d character
mation of a bulk compound related to the surface and the rates of ethene hydrogenation as shown on Figs l(a) and
compound [17] (Fig. 2), or even the heat of formation I(b) was first pointed out by Boudart [20].
of any bulk compound that can be correlated with the
heat of adsorption [18], or simply the position of the
catalytic element along a horizontal series in the peri- (Fig. 1) can be observed, but this partial result is useful
odic table (Fig. 3) [19], or an electronic property of the in itself.
catalysts such as Pauling’s %d bond character of metals As an example, consider the decomposition of for-
1201. It is generally observed that the binding energy mic acid on transition metals (Fig. 2). A measure of q
of many species with a metal goes down monotoni- is taken to be the heat of formation of the bulk metal
cally as we move horizontally to the right in a tran- formate. At low values of q, the rate of reaction is that
sition series. Sometimes only one slope of the volcano of the entry step [6]. This is a rate of adsorption, and it

goes up with q as the rate constant for adsorption in-
creases in conformity with a Brernsted relationship. At
the other end of q values, if q is high enough, the rate of
reaction is that of the exit step. This is a rate of de-
sorption and it goes up as the value of q decreases. In
the case of increasing values of q starting from the left.
the steady state surface coverage goes up with q. With
decreasing values of q, the steady state surface cover-
age goes up as q decreases. As the top of the volcano is
approached from both sides, an intermediate value of
the steady state surface coverage is reached, and the
measured rate of the reaction is neither that of the en-
try step nor that of the exit step but depends on both of
them.
Even today, surfing the periodic table remains a
good guide in the search for catalytic materials. But the
principle of Sabatier, semiempirical as it may be, pro-
vides additional guidance in the ongoing search, and it
keeps improving as the methods of surface science give
us better values for the binding energy of pertinent
species at the surface of catalytic materials.
1.1.6 Active Sites and Catalyst Modifiers
1.1.6.1 Ideal Surfaces
The idea of active sites on metals goes back to Lang-
muir [21]. The surface exposes an array of sites that can
chemisorb an atom or molecule in a localized mode.
Surface mobility is possible but not during the course
of an experiment with localized chemisorption. Lo-
\ calized chemisorption is associated with a configura-
tional entropy as a result of the different ways to dis-
tribute m molecules among n sites. The configurational
entropy [22]
S = -R In ( e l l - 0)
where 0 is the fraction of sites covered is responsible for
the form of the Langmuir adsorption isotherm
0 = K[M1/(1 + K/lMl)
where K is the adsorption equilibrium constant and
[MI is the concentration of M in the fluid phase
~31.
Localized adsorption and configurational entropy
are key principles of the Langmuir model. A Langmuir
active site has been aptly called a coordinatively un-
saturated site (cus) by Burwell [24] in analogy with
homogeneous organometallic catalysts. A Langmuir
site is not only found on a metal surface but on any
surface susceptible of chemisorption: typical sites in-
clude metallic cus ions, anions, Lewis and Brernsted
acids or bases, and organometallic compounds or im-
mobilized enzymes. Sites may include more than one
species to form multiplets [15] or ensembles [25-271.
1.1 Principles of Heterogeneous Catalysis 5
The Langmuir model assumes that all available sites
are identical and noninteracting before or after chem-
isorption. Langmuir was well aware of the severe limi-
tations of this model. So is everyone else who uses it up to
this day. Why the model works so well is still debated.
Langmuir chemisorption is limited to a monolayer.
Adsorption by physical forces instead of chemical ones
occupies a central place in heterogeneous catalysis.
Physisorption in multilayers was cast into an adsorp-
tion isotherm by Brunauer, Emmett, and Teller (BET)
[28]. The BET isotherm was derived by assuming that
physisorption in the first layer takes place on a uniform
surface without lateral interactions between phys-
isorbed species and that all subsequent layers behave
like a liquid surface of the physisorbed species. Again,
in spite of the improbable restrictions behind the BET
isotherm, it is used universally, successfully and re-
producibly, to obtain the specific surface area of pow-
ders ranging from flour and moon dust to heteroge-
neous porous catalysts of all chemical compositions.
The general applicability of the BET model is truly as-
tonishing, but also gratifying, since the vast majority of
heterogeneous catalysts are made of nanoscopic ulti-
mate particles held in aggregates crisscrossed by meso-
pores (2-50nm in size). Many catalysts have specific
surface areas of the order of up to 200 m2 g-’ as mea-
sured by the BET method.
1.1.6.2 Real Surfaces
Crystalline solids exhibit crystalline anisotropy. Equili-
brated single crystals expose different faces with differ-
ent atomic structures so as to minimize total surface
energy. It would then be surprising if different crys-
tallographic planes exposing different active sites for
chemisorption and catalytic reactions did possess iden-
tical properties. Moreover, most catalytic solids are
polycrystalline. Each face may expose terraces, ledges,
kinks, and vacancies with different active sites. Nano-
scopic particles will expose edges and corners. If the
catalyst contains more than one component, as is most
generally the case, the surface composition may be dif-
ferent from that of the bulk and differently so for each
exposed crystallographic plane. Besides, solids nor-
mally contain defects of an electronic or atomic nature.
Next, solids contain impurities, known or unknown in
the bulk, but mostly unknown at the surface. Finally,
the surface atomic structure and composition may well
change with time as catalysis proceeds.
In short, it is normal to expect that a catalytic solid
surface exposes not one type of active site but many
types. It is so normal to expect such complexity that it
may be surprising today that when Taylor formulated
in 1925 his principle of active sites or active centers, his
paper created so much attention and remains one of
References see page 12
6 1 Introduction
the most often cited papers in heterogeneous catalysis,
even when the author of the citation has not read such
an antique paper [29].
What was most important in Taylor‘s seminal paper
is not the idea of active centers or the possibility of
different active sites. Taylor’s true innovation is con-
tained in two crisp sentences. The first one reads:
“There will be all extremes between the case in which all
the atoms in the suvface are acriue und thal in which
relatively few are so active.”
The second sentence reads:
“The amount of surface which is catalj~ticall~vactive is
determined by the reaction catalyzed.”
At the time, these statements were prophetic, as they
were not based on experimental evidence. However,
they inspired many workers in the field, including this
author and his co-workers who proposed, on the basis
of kinetic data [30, 311, that there were indeed struc-
ture-insensitive and structure-sensitive reactions [32] as
anticipated by Taylor much earlier in his prophetic
statements. The ultimate choice of the namles was in-
spired by earlier work of Kemball, on what he called
structure-sensitive reactions [33]. Possibly the most sur-
prising, well-documented fact is the existencle of many
structure-insensitive reactions in which all sites seem to
exhibit equal activity on several planes of a single
crystal: this is true structure insensitivity.
The idea of Taylor was also documented by Kral
[34], who proposed the Taylor ratio to express the
fraction of a given metal surface active in various re-
actions. The Taylor ratio was later used by Ponec and
Bond [35].
The most incisive comment of Taylor was the second
one as quoted above: it is as if the catalyst and the
catalyzed reaction adapt to each other. This adapta-
tion, if kinetically attainable, may lead to a steady state
surface that ultimately becomes uniform, although it
was not uniform at the start of the reaction. This pos-
sible adaptive behavior was emphasized by Boreskov
[36] on the basis of the extensive work of his labora-
tory. This adaptive behavior may be explained by what
is called surface reconstruction, which has been known
for sometime but has been clearly documented by the
recent surface science work of Somorjai [37]. This im-
portant aspect of heterogeneous catalysis will be dis-
cussed again in Section 1.1.7.
The relative importance of surface struciure as ex-
pressed by crystalline anisotropy and of surface defects
and composition underlines the difficulty of identify-
ing the active sites, simple or complex, responsible for
turning over the catalytic cycle. The identification and
counting of active sites in heterogeneous catalysis be-
came the Holy Grail of heterogeneous catalysis in
1925, and the situation remains the same today. Yet
progress has been made, in particular in counting sites
by various spectroscopic techniques or by selective
chemisorption.
The classic example of selective chemisorption is
the work of Emmett and co-workers dealing with the
analysis of the surface of an iron catalyst modified with
small amounts of aluminum oxide A1203 and potas-
sium oxide KzO [38]. By the BET method the total
surface area of a sample of a catalyst reduced in Hl can
be determined. Then the part of the total surface cov-
ered with metallic iron is measured by selective chem-
isorption of CO. The part of the surface of A1203 cov-
ered with K 2 0 is determined by chemisorption of COz.
The part of the surface covered with A1203 is de-
termined by difference. All details of the chemisorp-
tion measurements must be specified. At some time
they should be checked by spectroscopic means, as
they were in this case or any other case of titration by
chemisorption or selective chemisorption. The results
of Emmett et al. agree well with those obtained by
means of Auger electron spectroscopy (AES) on an-
other sample of the same batch of catalyst [39]. More
recently, Topsse et al. [40] have shown that high tem-
perature chemisorption of N2 on a variety of iron cat-
alysts, including a multiply promoted commercial syn-
thetic ammonia catalyst, counts only the most active
iron sites responsible for the catalytic reaction, whereas
chemisorption of CO counts all exposed iron sites.
As a result of reconstruction during ammonia syn-
thesis, the iron surface of a commercial catalyst ap-
pears to be reconstructed to expose mostly the more
active sites [41]; the identification and counting of ac-
tive sites in that case, and presumably in many other
cases, was an arduous task and took many years of
research. In general, active sites must be studied under
reaction conditions, as Tamaru has long insisted [42],
while the catalyst is working, that is, in situ. This is a
fine principle, but in situ studies are difficult, as very
few spectroscopic techniques lend themselves to the
characterization of a solid catalyst working in a gas-
eous environment at atmospheric pressure or above, or
in a liquid phase. In such situations, indirect evidence,
ingenuity, kinetic data, and spectroscopic data gath-
ered before and after reaction must make do in the
absence of in situ data.
1.1.6.3 Catalyst Modifiers
Commercial catalysts contain many species in more
than one phase. Some of these species have been added
over the years to correct certain defective features of
the original catalyst, or to improve its performance, or
to meet more stringent demands. As usual, adaptation
leads to complexity. For instance, a current three-way
automotive catalyst that oxidizes hydrocarbons and

carbon monoxide, and reduces nitric oxide, is made of
the alumina support, platinum, palladium, rhodium,
and a cerium oxide modifier. A modifier is an additive
that fulfills one or more functions through mechanisms
that are rarely completely understood, an expression
that usually means that there exists no understanding
but just speculation. The addition of modifiers is often
based on empirical observations. Thus sulfur is usually
regarded as a poison of platinum naphtha-reforming
catalysts. Yet, during exploratory research at Chevron,
the bad performance of an experimental catalyst con-
taining rhenium turned around to enhanced perfor-
mance after many hours of time on feed. An explana-
tion of this behavior is that the feed contained very
small amounts of sulfur-containing molecules. In time
the sulfur became chemisorbed on rhenium atomically
dispersed on the platinum surface, thus breaking up
Pt-Re ensembles that led to excessive hydrogenolysis
of alkane feed molecules, thus enhancing performance
[431.
The paradox is that a modifier such as rhenium be-
comes a promoter, i.e. a good modifier, only after it
has reacted selectively with another modifier that is
normally considered as a poison, i.e. a bad modifier.
Modifiers can be good or bad depending on their
concentration. They can change the binding energy of
an active site or its structure, or disrupt an ensemble
of sites. They can affect differently the rates of two
parallel or consecutive reactions, thus modifying se-
lectivity. They can lead to increasing stability of the
catalytic phase by preventing crystal or particle growth
(sintering), thus controlling texture, i.e. pore size and
pore size distribution, and specific surface area of the
catalyst.
As an illustration, consider the two main modifiers
of an iron-based catalyst for ammonia synthesis. As
mentioned earlier, they are A1203 and K20. The first
promoter is definitely textural as it prevents the rapid
sintering of pure metallic iron. But it could also play a
role as a structural promoter, in stabilizing more active
sites on the iron surface [44]. As to KzO, it appears to
decrease the binding energy of nitrogen on adjacent
iron sites [45], thus making more free-iron sites avail-
able, at least at high pressure when the large amount
of ammonia made during the synthesis leads to an
excessively large coverage by nitrogen of an iron sur-
face that does not contain K20.
It can be said that the complexity of catalysts that
contain modifiers tends to repel academically oriented
researchers who are motivated more by what they
want to know rather than by what they need to know,
whereas the reverse normally applies to the researcher
fighting in the trenches of industrial conflicts. Similarly,
the study of poisoning of practically every catalyst
used in the conversion of organic molecules has largely
remained the province of industrial scientists and en-
1.1 Principles of Heterogeneous Catalysis 7
gineers. This type of poisoning is caused by organic
residues that can lead to the deposition of large
amounts of so-called coke, a polymer containing
mostly carbon and hydrogen. Deactivation of catalysts
by coke is a fact of life that the catalytic reaction en-
gineer must live with, if not by addressing the chem-
istry of the situation, at least by providing engineering
solutions such as periodic or continuous regeneration of
the poisoned catalyst by burning the coke in air. This
aspect of catalytic science and technology is touched
upon in the next section.
1.1.7 Catalyst Life Cycle
1.1.7.1 Preparation
“The Scientific Bases for the Preparation of Hetero-
genous Catalysts” is the title of a number of periodic
meetings organized by Bernard Delmon, with pro-
ceedings published by Elsevier, since 1975. The role of
solid state chemistry in the preparation of solid cata-
lysts is paramount [46]. The main hurdles in solid state
chemistry are the hidden variables that control nuclea-
tion and growth of the desired solid phase or phases.
The role of impurities and of minute changes in proto-
col militate against reproducibility of catalytic samples
even in a given batch.
Let us start with a common method to prepare
nanoscopic powders: a catalytic material with a large
specific surface area is often prepared by reducing an
oxide precursor in hydrogen. Here are some important
options and known variables.
First, the type of reactor must be considered. A fixed
shallow bed in a static atmosphere of reducing gas is the
least desirable. A better arrangement is a fixed bed of
precursor traversed by flowing hydrogen mixed with an
inert gas. Here are some of the details to be watched:
purity and pressure of the gas; size, diameter and
thickness of the fixed bed; control of the temperature,
whether it is fixed or programmed; flow rate or space
velocity; granulometry and chemical composition of
the precursor; and finally, post-treatment of the prod-
uct in a carefully controlled passivation prior to expo-
sure to the atmosphere.
Similar attention to detail must be paid to other op-
erations in the subsequent handling of the reduced
material as it is pressed into a pellet or an extrudate.
If a precursor is prepared by precipitation of water
solutions, the abundant science accumulated by gen-
erations of analytical chemists must be consulted. After
precipitates are obtained, their drying must be con-
trolled by numerous parameters that operate differently
with the size of the sample.
The difficulties of reproducibility and control are
compounded in the preparation of multicomponent
References see page 12
8 1 Introduction
and multiphase catalysts. No wonder that the prepara-
tion of a catalyst, as described in a patent, is the most
daunting task facing an industrial scientist or engineer.
The problem is not that details are conisciously with-
held from the patent but that the existence of hidden
variables is omitted by ignorance. Since the emergence
of industrial catalysis, in spite of the enormous ad-
vances of the science of heterogeneous catalysis, the
preparation of solid catalysts has remained the single
most important element that separates theory from
practice. The art of preparation accounts for the strict
secrecy surrounding the highly competitive manu-
facture of solid catalysts, as well as the costly litigations
alleging patent infringement. This regrettable situation
will ultimately be corrected with the further advance-
ment of scientific bases of manufacturing and of more
incisive fingerprinting instrumentation for patent pro-
tection.
1.1.7.2 Activation
With a few exceptions, the catalyst. as prepared or as
bought, is yet another precursor of the catalyst as used
in a reaction. Prior to use it must be activated in situ,
and this activation is in fact the last stage in the prep-
aration. Even if activation is often only the removal of
oxygen in the surface layers of a passivatetl prereduced
catalyst, that simpler operation is nevertheless affected
by the many variables affecting catalyst preparation
itself. For certain catalysts, a manufacturer will sell
ready-to-use batches at a premium to alleviate the
problems involved in controlled activation, especially
on a large scale.
Indeed, fine details of activation may determine the
surface morphology of the activated catalyst because of
surface reconstruction taking place during reduction or
activation. Morphology means exposure of preferred
crystalline faces or facets. In the open literature on
catalysis, it is not enough to specify the origin of the
catalyst by manufacturer and batch number, it is also
mandatory to list all details of the activation proce-
dure, including those listed above for the preparation
of the catalyst. Needless to say, these stern prescrip-
tions are rarely obeyed.
1.1.7.3 Reconstruction
After the start of a run with an activated catalyst, slow
changes in surface structure and composition are the
norm. Arbitrarily, we call these changes reconstruction
if they have a favorable effect on catalyst performance.
and deactivation if the change is unfavorable. An ex-
ample of reconstruction during use deals with surface
ensembles on platinum-rhenium catalysts, as discussed
in Section 1.1.6 with ensembles of sites broken up by
small amounts of sulfur-containing compounds in the
feed.
Another example is that of ammonia synthesis on
iron-based catalyst. Thus, the reconstruction of iron
single crystal tips by nitrogen was observed by field
electron microscopy [47]. Similarly, surfaces of iron
clusters are reconstructed by nitrogen or ammonia,
as inferred from Mossbauer effect spectroscopy [48].
This reconstruction leads to the growth of (1 11) planes
of iron that were later identified to be far more active
in ammonia synthesis than other iron planes [49].
During reconstruction, the catalyst evolves, according
to the principle enunciated by Boreskov [36], and in-
troduced in Section 1.1.6.2. According to this principle,
a steady-state minimizing surface free energy may be
approached during catalyst use. Thus, in ammonia
synthesis on iron, this evolution leads to the growth of
(1 11) facets that may well be the dominant ones ex-
posed at the surface of a commercial multipromoted
iron catalyst operating at its steady state [50].
Another case of beneficial reconstruction is the
roughening up of the surface platinum metals gauze
during the selective oxidation of ammonia to nitric ox-
ide by the Ostwald process. This phenomenon, also
called catalytic corrosion, was already well docu-
mented in the 1919 book of Taylor and Rideal[4]. This
book contains optical micrographs showing clearly the
so-called cauliflower morphology of corroded catalyst
wires. Later on, electron micrographs revealed the finer
texture of this reconstruction [51]. The principle of
Taylor and Boreskov is that a catalyst is an adaptive
system. for better or for worse. A final example of this
adaptivity is again in ammonia synthesis: after startup
of a converter, it can take months of operation before
performance reaches steady state, as small amounts of
dissolved oxygen continue to diffuse slowly from the
interior of the relatively large iron crystals of the cata-
lyst. This residual bulk oxygen moves to the surface
where it depresses activity until finally removed by feec
hydrogen.
1.1.7.4 Deactivation
Some catalysts do not show any measurable deactiva-
tion over periods of years, such as in ammonia synthe-
sis. Other catalysts lose an important fraction of their
activity after less than a minute of contact with feed,
as, in catalytic cracking of hydrocarbons. In the latter
case, this is due to coking (Section 1.1.6). With rapid
coking, continuous regeneration by coke burning be-
comes mandatory. This is one of the many examples in
catalysis where imperfect chemistry can be made ac-

ceptable by a workable engineering solution. Even
under laboratory conditions or on single crystal sur-
faces, coking during hydrocarbon catalysis is a great
obstacle to reproducibility. To extrapolate reaction rate
data to zero time of contact is empirical at best and all
too often unreliable. Lack of reproducibility leads in
turn to conflicting conclusions.
Thus, discussions concerning the structure sensitiv-
ity or insensitivity of a hydrocarbon reaction are con-
founded by the fact that the main reaction may be
structure insensitive while the deactivation or coking
reaction is structure sensitive. A clear example is given
by the hydrogenation of ethene on single crystals of
nickel: the reaction is structure insensitive as shown
by almost identical rates that are practically constant
with time on the (111) and (110) faces while the (100)
face is completely inactive, as a result of a rather ex-
treme case of structure sensitivity of the coking re-
action [52].
The next most common cause of deactivation of
solid catalysts is poisoning by impurities that diffuse
from the support in the case of supported metal cata-
lysts [53], or from the interior of a metal, or from the
feed, or from metallic components of the practical
reactor system. The latter type of contamination is
so unpredictable that expensive pilot plants are still a
prudent step on the way from the bench-scale reactor
to the commercial plant, this in spite of the steady
progress in catalytic reactor design.
Besides coking in the case of organic catalysis and
poisoning in general, the most common cause of deac-
tivation is loss of specific surface area. All high specific
surface area samples have metastable textures and,
given time, the thermodynamic drive to minimize sur-
face free energy will lead to losses of texture that may
become unacceptable. To combat this process, gen-
erally called sintering, textural promoters are added to
catalysts, such as alumina for iron catalysts, rare earth
ions for crystalline zeolites, or lattice oxygen-vacancy
blockers in oxide structures.
Another case of catalyst deactivation is through the
loss of an active ingredient in solution or in the gas
phase. For example, in the Ostwald process, the metal
in the gauze is vaporized slowly because of high tem-
perature in an oxidizing atmosphere. The gauze plates
are stacked up in the reactor, perpendicularly to the
gas flow; the reaction front moves toward the down-
stream plates as the upstream gauze disappears, until
the reactor is shut down for replacing the catalyst.
Fortunately, the vaporized noble metal oxides are re-
covered downstream.
In the laboratory or in the plant, catalyst deactiva-
tion is a fact of life not to be ignored. If it cannot be
circumvented by catalyst regeneration, the process may
be in jeopardy.
1.1 Principles of Heterogeneous Catalysis 9
1.1.7.5 Regeneration
The simplest case is when activity can be restored by
combustion of coke. Care must be taken not to sinter
the catalyst in such an operation, especially because of
the generation of steam, a common accelerator of sin-
tering. In the case of supported metal catalysts when
coke accumulates on both metal and support, re-
generation by combustion of coke agglomerates the
supported metal or destroys bimetallic clusters, as in
the case of platinum-iridium catalysts. Redispersion of
the metal into nanoscopic uni- or bimetallic clusters
can be achieved with the help of oxygen and chlorine
[55]. Successful regeneration is often the key to success
in practical heterogeneous catalysis. The mode of re-
generation has to be designed on a case to case basis.
Together with deactivation, regeneration must be con-
sidered from the beginning of a catalyst development
program. These challenges of heterogeneous catalysis
may be among the most difficult ones in the long march
from theory to practice.
1.1.7.6 Decommission
If a deactivated catalyst cannot be regenerated, Val-
uable components such as precious metals must be re-
covered economically. This form of decommissioning
is attractive. For instance, a catalyst consisting of pal-
ladium supported on active carbon and reaching the
end of its life, can be burned and the noble metal re-
covered for the preparation of a new catalyst [56].
If recovery is uneconomical, the catalyst must be
disposed of within the rules of environmental pro-
tection. This has become a serious problem in the case
of large amounts of spent catalysts containing heavy
metals, such as for the hydrodesulfurization of catalysts
used to treat petroleum fractions. Today it is necessary
to achieve safe decommission at the end of the life cycle
of any catalyst.
1.1.8 Tradeoffs
The four desirable attributes of a heterogeneous cata-
lyst are activity, selectivity, stability, and accessibility.
The first three also apply in homogeneous catalysis, but
accessibility is a problem proper to catalysis by solids.
We shall first briefly describe activity and selectivity,
since stability has already been discussed under the
name of its mirror image, namely, deactivation. Then
we will devote more attention to accessibility. Finally
we will examine the mutual conflicts between all four
attributes.
References see page 12
10 1 Introduction
1.1.8.1 Activity and Selectivity
To say that a catalyst is active means only that it is
alive, and that is not enough, either in theory or in
practice. Activity must be measured as a rate, and there
are many definitions or rates, some better than others.
As discussed in Section A.5.2.1, a readily available
measure of activity is a space-time yield, expressed in
units of amount of product made in the reactor per unit
time and per unit of reactor volume. As expressed by
Weisz [57], space-time yield should be within a broad
window of reality, the center of the window being
in order of magnitude equal to l,umolcmp' SKI,for
the catalyst to deserve consideration for large scab
industrial use. Note that conversion alone, or con-
version versus time, is useless as a measure of activity.
However, if we talk about selectivity, defined as the
amount of desired product obtained per amount of
consumed reactant, a value of selectivity without a
value of conversion is also useless. The product of
conversion and selectivity, commonly called yield for
short, is the simplest measure of selectivity. Since the
publication of the Handbuch der Katalyse [SS], enor-
mous advances have been made in our understanding
of the activity of heterogeneous catalysts. But quanti-
tative explanations of selectivity, either for parallel or
consecutive reactions, are still lacking. Selectivity, in-
cluding stereoselectivity and enantioselectivity, can be
enhanced by many means, including promoters, chiral
adsorbates, selective poisons, and solvents. The reader
is referred to a critical survey of these methods [59],
including the shape-selectivity of zeolites [60- 631.
Zeolites are microporous crystalline solids with cages
of molecular dimension that accept or reject in a se-
lective way certain reactants or products, on the basis
of their shape or the shape of a transition state between
reactants and products. The principle of shape se-
lectivity in heterogeneous catalysis may be considered
as the single most important principle in the control of
selectivity in heterogeneous catalysis. Catalysis by zeo-
lites exploits at a subtle molecular level the general
challenge of heterogeneous catalysis, namely the ac-
cessibility of the catalyst to mass and heat. It is not a
coincidence that the introduction of shape selectivity
was pioneered by a research team at Mobil led by P. B.
Weisz, as it was a team that developed over the years
many of the principles of catalyst accessibility.
1.1.8.2 Accessibility
Whether solid catalysts are porous or not, the diffusion
of heat and mass into and out of their surface is always
an important principle behind their chemical function.
Without gradients of concentration and temperature,
there can be no transport of reactants or products or
heat. This physicochemical topic is part of the broader
field of catalytic reaction engineering, and short in-
troductions to the physicochemical aspects of heat and
mass transfer are available. For such an introduction,
see Ref. 64.
We cannot do justice to the elegant principles and
dimensionless criteria that reflect the sophistication
with which accessibility can be tackled. We will only
discuss as examples two cases where internal or ex-
ternal accessibility can exert a profound influence on
catalytic selectivity.
Suppose we deal with the simplest consecutive net-
work of first-order catalytic reactions
A-+B-+C.
In the absence of complications, the concentrations of
A, B, and C will evolve with time. In the absence of
chemical complications, such as a direct bypass from A
to C, the curve depicting the concentration of C as a
function of time is tangent to the horizontal time axis
at zero time. But if the reactions take place in a porous
catalyst particle and if the rate of diffusion of B out of
the particle is not able to keep up with its rate of
chemical production, the C curve will have a nonzero
slope at the origin. In other words, the selectivity to B
at the beginning of reaction will be lower in the pres-
ence of diffusional limitation than in its absence. It
has already been noted above that diffusion in and out
of zeolite cages contributes, in part, to the shape se-
lectivity of zeolites.
Let us now turn to the diffusion of heat and mass to
the external surface of a nonporous metallic catalyst
such as the wires of noble metals used in the Ostwald
process (Section 1.7.3). The reaction rate is limited by
diffusion of ammonia through the very thin boundary
layer surrounding the metal to the metal surface. At the
same time the rate of heat generation due to the surface
reaction becomes higher than the rate of heat removal
away from the surface: the system is unstable and sur-
face ignition takes place so that the steady-state surface
temperature is much higher than the gas temperature
outside the boundary layer. The high selectivity of
the oxidation of ammonia to nitric oxide is due to
this limitation by diffusion of both mass to the surface
and heat from the surface. This situation is not unique.
For instance, methane is oxidized selectively to carbon
monoxide and hydrogen on nonporous platinum or
rhodium [65]: as in the Ostwald process, the surface
ignition temperature is very high ( % 1000 "C) and the
contact time very short (z1 ms).
Thus, accessibility is an attribute of a solid catalyst
that is not necessarily a nuisance to be taken care of. It
can also be an opportunity. But to ignore accessibility
can be disastrous. For instance, suppose that pore dif-
fusion is limiting the chemical rate of reaction. Then,
in the presence of a poison that penetrates progres-

sively into the porous texture of catalyst particles, pore
mouth poisoning can have catastrophic effects on the
catalytic yield.
Specialized tasks such as the design of the optimum
shape of porous catalyst particles or the spatial distri-
bution of the active material into a catalyst pellet in
various modes sometimes called egg shell, egg white,
and egg yolk, are better left to the catalytic reaction
engineer [66]. But awareness of accessibility in hetero-
geneous catalysis is absolutely essential whether the
catalyst be a large single crystal, a porous catalyst par-
ticle between 50pm or several mm in size, or a 100pm
washcoat of a supported metal catalyst on a porcelain
monolith or metallic foil.
1.1.8.3 Activity, Selectivity, Stability, and
Accessibility
It would be easy if these four desirable indices of cata-
lyst performance would move up in the same direction
upon an increase in a single process variable such
as temperature. This is not so, and choices have to
be made in every instance.
Start with activity. As a rate it generally goes up with
temperature. But stability usually goes down as tem-
perature goes up and so does accessibility. For se-
lectivity, that depends. Take a favorable example, the
catalytic decomposition of NO to N2 and 0 2 . The rate
of decomposition of NO on dehydroxylated alumina
goes down as temperature goes up [67], probably be-
cause it proceeds through an associative mechanism in
which the exothermic formation of a NO dimer pre-
cedes its decomposition into N2O and chemisorbed 0,
and the concentration of the surface dimer goes down
as temperature goes up [68]. However, the dissociation
of N20 into N2 and adsorbed oxygen proceeds through
an associative mechanism with a rate that increases
with temperature, as is normally the case. Hence the
selectivity of NO decomposition to N20 decreases with
temperature while the selectivity of decomposition of
NO to N2 increases with temperature. As to the rate
of catalytic decomposition of NO, it may decrease or
increase with temperature, depending on the relative
values of the heat of formation of the dimer and
the activation energy for the formation of the dimer
intermediate.
Thus no general rule can be given for the tradeoff
between activity and selectivity. When the gain in se-
lectivity is achieved by selective poisoning of the cata-
lyst, it will clearly be paid for by a decrease in activity.
An example is the change in selectivity for the parallel
hydrogenation on nickel of a polar molecule (acetone)
and a nonpolar one (cyclohexene). With adsorption of
a polar surface modifier, selectivity shifts in favor of
acetone and vice versa; it shifts in favor of cyclohexene
1.1 Principles of Heterogeneous Catalysis 11
with adsorption of a nonpolar modifier. In both cases,
however, increase in selectivity is achieved at the ex-
pense of activity because part of the surface is blocked
by the modifier acting as a polar or nonpolar surface
solvent [69].
The principle is clear: an increase in activity by tem-
perature rise is not necessarily accompanied by an in-
crease in the other indices of catalyst performance.
Activity as expressed by the principle of the window on
reality according to Weisz [57] appears to be a neces-
sary but not sufficient condition. Selectivity, stability,
and accessibility must then be considered, and tradeoffs
may have to be negotiated, perhaps by use of a multi-
functional reactor [70] with catalytic distillation or a
catalytic membrane. Again, success in heterogeneous
catalysis starts with chemistry but always ends with
catalytic reaction engineering. Disregarding this truism
may explain in part the vast amount of catalytic in-
dustrial research that ends without a paper, without a
patent, or without a plant.
1.1.9 Principles of Assisted Catalyst Design
A priori design of catalytic reactors, especially of clas-
sical ones of the h e d bed type, has made huge progress
in the past 50 years [71]. However, a priori design of
solid catalysts is not even in its infancy: it has not
been born. What can be discussed is what will be called
assisted catalyst design, a methodology based on the
traditional feedback loop: catalyst preparation, char-
acterization, testing, preparation of a modified catalyst,
and so on.
This has been the traditional method of catalysis
since it was first practised by Mittasch at BASF who,
according to Timm [72], issued in 1909 the following
directive to his team in ammonia synthesis:
“1. The search for a suitable catalyst necessitates car-
rying out experiments with a number of elements to-
gether with numerous additives;
2. The catalytic substances must be tested at high pres-
sures and temperatures just as in the case of Haber’s
experiments;
3. A very large number of tests will be required.”
Ten years later, the number of tests exceeded 10 000,
and more than 4000 catalysts had been studied. This
extraordinary effort was also extraordinarily successful.
What has changed since then is the way the system-
atic search is assisted today. Some of these modes of
assistance are surface chemistry of metals [49], oxides
[73], interfacial coordination chemistry [74], zeolitic
materials [75], kinetics [76], and catalytic reaction en-
gineering [66]. Various theoretical and computer meth-
ods [77] must be added to this list.
References see page 12
12 1 Introduction
With an arsenal of principles, concepts, instru-
mentation, and computers, the feedback loop of Mit-
tasch should lead to the improved or “designed” cata-
lyst in a much shorter time with a smaller number of
trial samples, especially with the possibility of charac-
terization in situ by X-ray absorption and X-ray dif-
fraction [78, 791. Clearly, the team reponsible for as-
sisted catalyst design must include surface chemists,
physical chemists, physical organic chemists, chemical
physicists, solid-state inorganic chemists, materials sci-
entists, and research chemical engineers. The practical
merit of this assisted Mittasch design is obvious. Its
scientific dividend is the possibility of learning as the
design proceeds, with the building up of a data bank of
rate constants and the formulation of more precise
models of active sites. Part of this approach is im-
bedded in the microkinetic methodology of Dumesic et
al. [80]. With new theoretical insights or principles,
scientific bases for catalyst preparation, and reproduci-
bility of catalyst behavior, the future of heterogeneous
catalysis during the useful lifetime of this Handbook
looks very bright indeed.
References
1. E. K. Rideal, Concepts in Catalysis, Academic Press, Lon-
don, 1968.
2. J. M. Thomas, Angew. Chem. Int. Ed. Engl. 1994, 33, 913.
3. H. Poincare, Science et Mkthode, Flammarion, Paris, 1908,
Chapter 1.
4. E. K. Rideal, H. S . Taylor, Catalysis in Theory and Practice,
Macmillan, London, 1919, Chapter 2.
5. M. Boudart in Perspectives in Catalysis (Eds: J. M. Thomas,
K. I. Zamaraev), Blackwell, Oxford, 1992, p. 183.
6. M. Boudart, Kinetics of Chemical Processes, Buttenvorth-
Heineman, Stoneham, MA, 1991, Chapter 3.
7. M. Boudart, Ind. Eng. Chem. Research 1994, 33, 2900.
8. G. Ertl, Top. Catal. 1994, I , 304.
9. K. G. Denbigh, The Thermodynamics of the Steady State,
Methuen, London, 1951.
10. J. C. Tully, Surface Science: The First Thirty Years (Ed.: C. B.
Duke), North Holland, Amsterdam, 1994, p. 667.
11. M. Boudart, Kinetics of Chemical Processes, Buttenvorth-
Heineman, Stoneham, MA, 1991, Chapter 9.
12. D. G. Loffler, L. D. Schmidt, J. Catal. 1976, 41, 440.
13. J. H. de Boer in The Dynamical Character of Adsorption,
Oxford University Press, Oxford, 1953, p. 234.
14. H. S. Taylor, J. Am. Chem. SOC.1931,53, 578.
15. A. A. Balandin, Adv. Catal. Rel. Subj. 1969, 19, 1.
16. 0. Beeck, Disc. Faraday SOC.1950, 8, 118.
17. W. J. M. Rootsaert, W. M. H. Sachtler, Z. Physik. Chem.
1960,26, 16.
18. K. Tanaka, K. Tamaru, J. Catal. 1963,2, 1963.
19. J. H. Sinfelt, Bimetallic Catalysts, Wiley, New York, 1983,
p. 14.
20. M. Boudart, J. Am. Chem. SOC.1950, 72, 1050.
21. I. Langmuir, J. Am. Chem. SOC.1915, 37, 1139.
22. D. H. Everett, Trans. Faraday SOC.1950, 46, 942.
23. K. Denbigh, The Principles of Chemical Equilibrium, Cam-
bridge University Press, 1955, p. 434.
24. R. L. Bunvell, Jr. in Catalysis: Science and Technology (Eds:
J. R. Anderson and M. Boudart), Springer-Verlag, Heidel-
berg, 1991, Vol. 9, Chapter 1.
25. N. I. Kobozev, Zhur. Fiz. Khim. 1939, 13, 1.
26. D. A. Dowden, Proceedings of the 5th International Congress
on Catalysis (Eds: G. C. Bond, P. B. Wells, F. C. Tompkins),
The Chemical Society, London, 1976, Vol. 1, p. 621.
27. V. Ponec, W. M. H. Sachtler, Proceedings of the 5th Intema-
tional Congress on Catalysis (Eds: G. C. Bond, P. B. Wells,
F. C. Tompkins), The Chemical Society, London, 1976, Vol.
1, p. 645.
28. S . Brunauer, P. H. Emmett, E. Teller, J. Am. Chem. SOC.
1938, 60, 309.
29. H. S . Taylor, Proc. Roy. SOC.(London) 1925, A108, 105.
30. M. Boudart, A. Aldag, J. E. Benson, N. A. Dougharty, C. G.
Harkins, J. Catal. 1966, 6, 92.
31. M. Boudart, A. Aldag, L. D. Ptak, J. E. Benson, J. Catal.
1968, 11, 35.
32. M. Boudart, Proceedings of the 5th International Congress on
Catalysis (Eds: G. C. Bond, P. B. Wells, F. C. Tompkins),
The Chemical Society, London, 1976, Vol. 1, p. 1.
33. C. Kemball, J. Chem. SOC.1956, 735.
34. H. Kral, Chemiker-Zeitung 1967, GI, 41.
35. V. Ponec, G. C. Bond, Catalysis by Metals and Alloys, Studies
in Surface Science and Catalysis, 1995, 95, p. 267.
36. G. K. Boreskov, Geterogennii Kataliz, Nauka, Moscow, 1988.
37. G. A. Somorjai, Ann. Rev. Phys. Chem. 1994, 45, 721.
38. P. H. Emmett in The Physical Basis for Heterogeneous Cat-
alysis (Eds: E. Drauglis, R. I. Jaffee), Plenum Press, New
York, 1975, pp. 3-34.
39. D. C. Silverman, M. Boudart, J. Catal. 1982, 77, 208.
40. H. Topsae, N. Topsere, H. Bohlbro, J. A. Dumesic, Proceed-
ings of the 7th International Congress on Catalysis, Part A (Eds:
T. Seiyama, K. Tanabe), Kodansha, Tokyo, 1981, p. 247.
41. M. Boudart, Top. Catal. 1994, I , 405.
42. K. Tamaru, in Dynamic Heterogeneous Catalysis, Academic
Press, 1978.
43. H. Kluksdahl, Chevron Research Company, US Patent
3415 737 (1968); W. C. Buss, R. L. Jacobson, R. H. Vorhis,
A. I. Salka, AM-78-51 Chevron Research Company, 1978
NPRA Annual Meeting; M. A. Freiburger, W. C. Buss, A.
G. Bridge, AM-80-52, Chevron Research Company, 1980
NPRA Annual Meeting.
44. D. R. Strongin, S . R. Bare, G. A. Somorjai, J. Catal. 1987,
103, 289.
45. B. Fastrup, Top. Catal. 1994, 1, 273.
46. B. Delmon, Introduction 2 la Cinktique Hktkrogdne, Technip,
Paris, 1979.
47. R. Brill, E. L. Richter, E. Ruch, Angew. Chem., Intern. Ed.
Engl. 1967, 6, 882.
48. M. Boudart, H. Topsere, J. A. Dumesic, The Physical Basis
for Heterogeneous Catalysis (Eds: E. Drauglis, R. I. Jaffee),
Plenum Press, New York, 1975.
49. G. A. Somorjai, Introduction to Surface Science and Catal-
ysis, Wiley, New York, 1994.
50. M. Boudart, D. G. Loffler, J. Phys. Chem. 1984,88,5763.
51. J. Turkevich, G. Garton, J. Chim. Phys. 1954,51, 516.
52. G. Dalmai-lmelik, J. Massardier, Proceedings of the 6th In-
ternational Congress on Catalysis London 1976 (Eds: G. C.
Bond, P. B. Wells, F. C. Tompkins), The Chemical Society,
London, 1977, Vol. 1, p. 90.
53. G. Leclercq, M. Boudart, J. Catal. 1981, 71, 127.
54. G. Kuczynski, J. J. Carbeny, J. Catal. 1973,28, 39.
55. S. C. Fung, CHEMTECH, January 1994,40.
56. G. Jaffe, personal communication.
57. P. B. Weisz, CHEMTECH, July 1992, 424.
58. G.-M. Schwab, Handbuch der Katalyse, Vol. 1-7, Springer-
Verlag, Wien, 1941-1943.

59. Selectivity in Heterogeneous Catalysis. Faraday Discuss.
Chem. SOC.1981, 72.
60. W. 0. Haag, R. M. Lego, P. B. Weisz, Faraday Discuss.
Chem. SOC.1981, 72, 317.
61. E. G. Derouane, P. Dejaifve, Z . Gabelica, Faraday Discuss.
Chem. SOC.1981, 72, 331.
62. J. M. Thomas, G. R . Millward, S. Ramdas, L. A. Bursil, M.
Audier, Faraday Discuss. Chem. SOC.1981, 72, 345.
63. P. A. Jacobs, J. A. Martens, J. Weitkamp, H. K. Beyer,
Faraday Discuss. Chem. SOC.1981, 72, 353.
64. J. J. Carberry in Catalysis: Science and Technology (Eds:
J. R. Anderson, M. Boudart), Springer-Verlag, Heidelberg,
1987, Vol. 8, Chapter 3.
65. D. A. Hickman, L. D. Schmidt, Science, 1993,259, 343.
66. R. Aris in Catalyst Design: Progress and Perspectives (Ed.: L.
L. Hegedus), Wiley Interscience, New York, 1987, Chapter 7.
67. A. Solbakken, L. H. Ryerson, J. Phys. Chem. 1960,64, 1903.
68. B. Turnham, Ph.D. Dissertation, Stanford University, 1974.
69. J. P. Wauquier, J. C. Jungers, Bull. SOC.Chim. 1957, 1280.
70. K. R. Westerterp, Chem. Eng. Sci. 1992, 47, 2195.
71. Chemical Reactor Theory, A Review (Eds: L. Lapidus, N. R.
Amundson), Prentice Hall, Englewood Cliffs, NJ, 1977.
72. B. Timm, Proceedings of the 8th International Congress on
Catalysis Berlin 1984, Verlag Chemie, Frankfurt, 1984, Vol.
I, p. 7.
73. J. Haber, Proceedings of the 8th International Congress on
Catalysis Berlin 1984, Verlag Chemie, Frankfurt, 1984, Vol.
I, p. 85.
74. M. Che, Proceedings of the 10th International Congress on
Catalysis Budapest 1993, Akademiai Kiado, Budapest, 1993,
Vol. A. p. 31.
75. W. 0. Haag, N. Y. Chen in Catalyst Design: Progress and
Perspectives (Ed.: L. L. Hegedus), Wiley Interscience, New
York, 1987, Chapter 6.
76. M. Boudart in Catalyst Design: Progress and Perspectives
(Ed.: L. L. Hegedus), Wiley Interscience, New York, 1987,
Chapter 5.
77. Metal-Surface Reaction Energetics (Ed.: E. Shusturovich),
VCH, New York, 1991.
78. B. S. Clausen, L. Gribaek, G. Steffensen, P. L. Hansen, H.
Topsare, Catal. Lett. 1993,20, 23.
79. J. M. Thomas, G. N. Greaves, Catal. Lett. 1993,20, 337.
80. J. A. Dumesic, D. F. Rudd, L. M. Aparicio, J. E. Rekoske,
A. A. Treviiio, The Microkinetics of Heterogeneous Catalysis,
American Chemical Society, Washington, DC, 1993.
1.2 Development of the Science of
Cata1y sis
B. H. DAVIS
1.2.1 Early Concepts: Berzelius, Liebig,
Faraday
Catalysis was an experimental fact and a subject of
debate prior to the introduction of the concept by Ber-
zelius during the 1835-1836 period. The first applica-
tion of catalysis, to produce ethanol by fermentation, is
1.2 Development of the Science of Catalysis 13
lost to antiquity. By the middle ages the use of sulfuric
acid to catalyze the production of diethyl ether was
widespread; written records to this synthesis date to
1552 [l].
The reception of the concept by the scientific com-
munity is illustrated by the work of Mitscherlich and
Berzelius. In 1834 Mitscherlich reported that when
alcohol was run into dilute sulfuric acid at 140 “C,ether
and water could be distilled from the mixture [2]. He
extended his observation by stating that decomposi-
tions and combinations of this kind were very frequent.
Mitscherlich introduced the term “contact” to describe
these actions, and summarized reactions that were
caused by contact: the formation of ether, the oxidation
of ethanol to acetic acid, the fermentation of sugar, the
production of sugar from starch by boiling sulfuric
acid, the hydrolysis of ethyl acetate by alkali, and the
formation of ethene from ethanol by heating with acid.
Berzelius, from 1821 on, summarized and reviewed
critically the scientific work done all over the world in
his “annual report” (Jahresberichte) [3]. The general-
izations in his reviews added as much or more to his
reputation as his own discoveries, and these were many
and important. In his annual review of 1835, he cov-
ered a number of reactions which take place in the
presence of a substance which remains unaffected.
Figure 1 provides some roots of catalysis, emphasizing
those considered by Berzelius; a number of additional
examples of reactions that predate Berzelius’ definition
can be found in Ref. 4.Trofast [3] presents an English
language version of Berzelius’ conclusions:
This is a new power to produce chemical activity be-
longing to both inorganic and organic nature, which is
surely more widespread than we have hitherto believed
and the nature of which is still concealed from us. When
I call it a new power, I do not mean to imply that it is a
capacity independent of the electrochemical properties of
the substance. On the contrary, I am unable to suppose
that this is anything other than a kind of special mani-
festation of these, but as long as we are unable to dis-
cover their mutual relationship, it will simplijy our re-
searches to regard it as a separate power for the time
being. It will also make it easier for us to refer to it if it
possesses a name of its own. I shall therefore, using a
derivation well-known in chemistry, call it the catalytic
power of the substances, and the decomposition by
means of this power catalysis, just as we use the word
analysis to denote the separation of the component parts
of bodies by means of ordinary chemical forces. Cata-
lytic power actually means that substances are able to
awaken afJinities which are asleep at this temperature by
their mere presence and not by their own afJinity.
Thus, the same concept was put forth from two per-
spectives within a 2 year period, and yet one had a
References see page 33
14 1 Introduction
CATALYSIS
THE ROOTS
Payen & Penoz (1833)
Isolate Ferment
Dobereiner (1 820’s)
E O H + HAc
Dobereiner (1808)
Starch+ EtOH
Starch -
Parrnentier (1 781)
H+
Sugar
Figure 1. Schematic representation of some of the studies that serve as a background for the formulation of the concept of catalysis by
Berzelius.
much more significant and lasting impact. The con-
cepts of Berzelius attracted much criticism that was
directed, for the most part, not at catalysis, but at the
concept of catalytic force. Berzelius has therefore been
credited with introducing the concept of catalysis even
though one could easily conclude that Mitscherlich
predated him by 2 years. Mitscherlich was not com-
pletely ignored since “contact catalysis” was utilized
into the middle of the 20th century.
The theories which were advanced during the 19th
century to explain the mechanism of catalysis may be
grouped into three classes: chemical. vibrational. and
physical. Proponents of the chemical theory view the
catalyst to operate through the continuous formation
and decomposition of unstable intermediate products.
The vibration theory was applied especially to fermen-
tation, and was at its zenith under the leadership of
Liebig. The ferment (enzyme) was then viewed to pos-
sess a particular internal motion that could be trans-
mitted to neighboring molecules, and thereby induce
reactions of them. The physical theory explains the
phenomena as being due to the condensation, and the
increase in concentration, of the reacting substances
at the surface of the catalyst, such increase in concen-
tration being brought about by capillary forces.
DuLong & Thenard (1 924)
Dobereiner (1923)
MeOH + Air
EtOH + EtOH
H. Davy (-1800)
CH, + Air (Pt)
It has long been recognized that a catalyst has the same
chemical composition at the end of the reaction that it
effected; however. it has been equally recognized that
its physical state may have undergone a profound al-
teration [ 5 ] . Thus, when ammonia is decomposed by
contact with heated metals there is nearly always an
alteration of the physical state of the metal, and this
has been attributed to the formation and decompo-
sition of intermediate compounds, metallic nitrides
in this particular case. While these views are obvious
today. it would be a demanding task to trace their his-
tory to “to the original statement of the concept.”
Whereas Berzelius defined “catalysis”, “catalytic
force”. and “catalytic action”, it remained for Arm-
strong [6] to define catalyst:
. . . so little has been done to ascertain the nature of
the influence of the contact-substance, or catalyst, as I
n,ould term it, the main object in view being the study of
the product of the reaction, that the importance of the
catalyst is not duly appreciated.
This is an outstanding example of an accurate his-
torical fact, and a presentation, on the one hand, of
an obvious definition, and on the other, an insightful
statement of needs for future research.

Adsorption is at the heart of any heterogeneously
catalyzed reaction. In 1800 adsorption was absorption,
and scientists were aware of it. The concepts of ad-
sorption are illustrated in an admirably lucid account
of its status at that time by de Saussure in a presen-
tation to the Geneva Society on April 16, 1812 [7]. The
questions de Saussure asked in the introduction of his
paper are equally valid for a manuscript that is being
written today. He begins with a statement that no
accurate experiments on the question of whether a
gas, when it penetrates into the pores of a solid body,
undergoes any diminution of bulk (volume) when no
chemical bonding takes place between the gas and the
solid. de Saussure, assuming that such contraction of
volume takes place, is led to a number of insightful
questions: What influence has the size of the pores on
this condensation? Are all gases equally condensed by
the same bodies? What influence has the density of the
gas on this condensation? When equal quantities of two
gases are in contact with a solid, do they adsorb in
equal quantities? Do the mixed gases, when condensed
in the solid, enter into combinations which they would
not form in the free state? He understood the need for
careful experimental results. Also, he clearly outlined
and defined many scientific aspects of adsorption.
de Saussure confirmed that adsorption in all cases,
except for oxygen, was complete at the end of 24-36 h.
He also demonstrated that adsorption of a gas is exo-
thermic, liberating a quantity of heat during the con-
densation that is often sensible to the feel, and suffi-
cient to raise a thermometer in contact with the solid
(charcoal) by several degrees. Aware that condensation
of a gas was exothermic and evaporation was endo-
thermic, he speculated that the solid should become
cooler during evaporation, and demonstrated experi-
mentally that this was indeed the case.
de Saussure speculated that charcoal should be acti-
vated by evacuation as well as by heating, and showed
this to be the case. He compared the amount adsorbed
on the whole and ground charcoal. The pulverized
sample adsorbed less gas, and he attributed this to the
amount of preadsorbed water. The rate of adsorption
of different gases was reported to be the same in all
solids of similar chemical properties even though the
amounts of gases condensed in the various solids may
vary greatly. He observed that exposure of a solid sa-
turated with one gas to a second gas would cause ad-
sorption of the second gas, resulting in the desorption
of some fraction of the first gas.
He stated that the affinity of the gas and the solid as
well as the condensibility of the gas must be considered
in addition to the porosity of the solid. Furthermore,
the forces of affinity and of condensibility act to oppose
each other. de Saussure considered the condensation
of gases in solid bodies to be analogous with the rise
of liquids in capillary tubes. Both are in the inverse
1.2 Development of the Science of Catalysis 15
ratio of the size of the interior diameters of the tubes
or pores.
Mellor [5] devotes considerable space to Faraday’s
“condensation theory” of catalysis. Taylor [8] wrote:
Faraday was theJirst to indicate the zone of adsorbed
material as the reaction space of a heterogeneous cata-
lytic action.
Robertson [9] and Schwab [lo], among others who
survey the concepts of catalysis, attribute the adsorp-
tion theory to Faraday. However, Faraday’s view of
catalysis was a limited extension of de Saussure’s views
that are described above.
Faraday, during about a 3 month period, investi-
gated the recombination of oxygen and hydrogen as
they remained in contact with platinum following their
liberation during electrolytic studies [l 11. He concluded
that this property belonged to platinum “. . . at all
times, and was altvays effective when the surface was
perfectly clean.” The platinum could be activated by
heating in acid or by employing mechanical clean-
ing; other approaches, such as treating or heating with
alkali, were effective only part of the time. Faraday
was convinced that the phenomenon which he observed
had a satisfactory explanation based on known princi-
ples, and did not require the assumption of any new
state or new property. Faraday clearly showed that
the presence of hydrogen sulfide, phosphine, or ethene.
among others, would prevent the platinum from exhi-
biting its properties of causing the combination of hy-
drogen and oxygen.
Faraday wrote that Dobereiner [ 121 considered the
effect to be an electric action in which hydrogen, being
very highly positive, represents the zinc of the usual
electric cell arrangement, and like it, attracts oxygen
and combines with it. Faraday stressed the need for a
perfectly clean and metallic surface. He considered the
effect to be produced by most, if not all, solid bodies,
weakly perhaps in many of them, but rising to a high
degree in platinum. He wrote that he was prepared to
admit that the sphere of action of particles extends
beyond those other particles with which they are im-
mediately and evidently in union, and in many cases
produces effects rising into considerable importance,
and he thought that this kind of attraction was a de-
termining cause of Dobereiner’s effect. Faraday com-
pared the spongy platinum to a hygroscopic body
which becomes moist due to the condensation of the
water vapor with which it is in contact. He then con-
siders that a gas, even when compressed at high pres-
sure, still has a low density and possesses sufficient
vacant volume that another gas can be added in the
void space. He wrote that the two gases, hydrogen and
oxygen, will not react even when compressed unless the
platinum surface is present to suppress, or remove,
their elasticity and/or the action of the metal in con-
References see page 33
 16 1 Introduction
densing them against its surface by an attractive force.
He compares the hydrogen-oxygen case to that where
Hall found COz and lime to remain combined under
pressure at temperatures at which they would not
have remained combined if the pressure had been re-
moved. The course of events may be stated, according
to Faraday by these principles:
(i> the deficiency of elastic poir>er and the uttraction of’
the metal j o r the gases are such that the gases are so
far condensed as to be brought within the action of
their mutual afinities at tlie existing temperature;
(ii) the deficiency of elastic p o ~ v r not
. merely subjecting
them more closely to the attractice influence o j t h e
metal, but also bringing thein into a more facormble
state for union, by abstracting I I part qf‘ that pott’er
(upon which depenh their elasticity) \t.hich else-
where in the mass of’ gases is opposing their cornbi-
nation.
The consequence of their combination is the produc-
tion of the vapor of water and an elevation of temper-
ature. The attraction of Pt for water is not as great as
for the gases so the water quickly diffuses through the
remaining gases. Fresh portions of the gases come into
juxtaposition with the metal, combine, and the vapor
formed also diffuses, allowing new portions of gas to be
acted upon. In this way the process advances and is
accelerated by the evolution of heat. The platinum is
not considered to cause the combination of any par-
ticles with itself, but only associating them closely
around it; the compressed particles are free to move
from the platinum, being replaced by other particles, as
a portion of dense air upon the surface of the globe, or
at the bottom of a deep mine, is free to rnove by the
slightest impulse into the upper and rarer parts of the
atmosphere. Thus, apart from introducing the need for
a clean surface, and what may be, in the most opti-
mistic view, a small step in the direction of stating the
concept of chemisorption, Faraday held views that
were essentially the same as those of de Sassure. This
is not meant to degrade Faraday’s work; in fact, he
applied extraordinary concepts and visionary thought
in an experimental program that extended over only 3
months.
During the period 1840-1880, Liebig’s view of cat-
alysis prevailed. Liebig matched Berzelius in scientific
stature, published his own journal, and had the advan-
tage that most chemists had been trained by him, either
directly or indirectly by his students. In the period
1836-46, Liebig was engaged in research on putrefac-
tion and in developing agriculture science where or-
ganic ferments were a subject for active investigation.
The isolation of an enzyme from barley malt in 1832
[ 131 ignited these intense research efforts.
Liebig was a proponent of the vibration theory of
fermentation and catalysis. Pasteur [ 141 proved that
alcoholic fermentation and putrefaction are caused by
the presence of certain low forms of life - bacteria.
Liebig attacked this view, and with vigor:
To suppose that putrefaction or fermentation is caused
by the physiological action of such creatures can only be
compared with the idea entertained by a child who would
explain the rapid current of a river through a mill-\vheel
by supposing that the mill-wheel, by its force, drives the
jtwter down the stream.
Liebig’s criticism did push Pasteur to more experi-
ments, and to firmly establish his views. Considering
the decomposition of hydrogen peroxide, Liebig [I 51
wrote.
Yet it is singular that the cause of the sudden separation
of the component parts oj peroxide of hydrogen has been
ciewed as diferent from those of common decomposi-
tion, and has been ascribed to a new polver termed the
catalytic force. Nobt., it has not been considered, that the
presence of the platinum and silver serces here only to
accelerate the decomposition; for without the contact of
these nietals, the peroxide of hydrogen decomposes
spontaneously, although very slowly.
Thus, Liebig presented a major concept that was to
be a part of the theory of catalysis advanced later by
Ostwald, although it did not have the framework of
kinetics that developed in the 40 years separating the
two.
After considering several examples of such decom-
positions, Liebig concludes [16] that:
No other explanation of these phenomena can be gicen,
than that a body in the act of combination or decom-
position enables another body, with which it is in con-
tact, to enter into the sume state. It is ecident that the
active state of the atoms of one body has an injuence
upon the atoms of a body in contact with it; and i f these
atoms are capable of the same change as the former,
they likewise undergo that change; and combinations
and decompositions are the consequence. But when the
atoms of the second body are not capable of such an ac-
tion, any further disposition to change ceases from the
moment at which the atoms of the first body assume the
state of rest, that is when the changes of transformations
of this bod!> are quite completed.
In that state in ivhiclz they exist within the pores or
upon the surjkce of solid bodies, their repulsion ceases,
and their Itthole chemical action is e.uerted. Thus combi-
nations Tt4ic.h oxygen cannot enter into, decompositions
bthich it cannot efect while in the state of gas, take place
with the greatest facility in the pores of platinum con-
taining condensed oxygen. When a jet of hydrogen gas,

for instance, is thrown upon spongy platinum, it com-
bines with the oxygen condensed in the interior of the
mass; at their point of contact water is formed, and as
the immediate consequence heat is evolved; the platinum
becomes red hot and the gas is inflamed. . . . In finely
pulvrrized platinum, and even in spongy platinum, we
therefore possess a perpetuum mobile - a mechanism like
a watch which runs out and winds itself up - a fotrce
which is never exhausted - competent to produce efSects
of the most powerful kind, and self-renewed ad infinitum.
Most phenomena, formerly inexplicable, are satisjhc-
torily explained by these recently discovered properties
of porous bodies. The metamorphosis of alcohol into
acetic acid, by the process known as the quick vine-
gar manufacture, depends upon principles, at a kno wl-
edge cfwhich we have arrived by a careful study of these
properties.
1.2.2 Wilhelm Ostwald
During the 50 years following the definition of catalysis
by Berzelius, there was little progress in understanding
it. Berzelius had organized enzyme catalysis, as well as
organic and inorganic catalysis, into a single discipline,
and it remained that way for the next half-century.
During this period, as in the period preceding its de-
finition, catalysis was a science based on trial and error
- Edisonian research - and reports of the observations.
However, during this period great advances in the un-
derstanding of chemistry were taking place, and these
provided the framework for an explosion of soundly
based concepts of catalysis. Organic chemistry was
gaining a structure through many developments that
included van? Hoff-Le Bel’s three-dimensional male-
cules, the recognition of four-valent carbon, the dekel-
opment of reliable chemical formulas based upon
Liebig’s method of elemental analysis of organic com-
pounds, the recognition of stereoisomers and the isola-
tion of enantiomers. Physical chemistry was introduced
and developed, primarily by Arrhenius, Ostwald, and
van? HoR its acceptance was documented with the
appearance of Ostwald’s two-volume book and the
Zeitschrift fur physikalische Chemie, edited by van?
Hoff and Ostwald. Kinetics became a subject for study
following the measurements by Wilhelmy in 1850 of
the acid-catalyzed inversion of cane sugar [I71 and
attained its first mathematical maturity in the work of
Harcourt and Esson [18]. Mayer introduced the con-
cept of the conservation of energy in 1842, and others
put the concept on a firm scientific basis [19]. Thernio-
dynamics was established, and Gibbs provided ad-
vanced concepts of statistical thermodynamics and the
phase rules. Boltzman provided his theories of the dis-
tribution of energy, the theory of approach to equilib-
rium, and entropy and probability, among others.
1.2 Development of the Science of Catalysis 17
Ostwald first came to catalysis through h s work on
the acceleration of homogeneous reactions by acids.
This work was widely accepted at the time although,
ultimately, it would be shown to be incorrect because he
believed that the acid, acting as a catalyst, did not enter
into the chemical change which it influences but rather
acted by its mere presence (contact catalysis).
Ostwald, reviewing a paper in which Stohmann [20]
utilized the vibrational theory of catalysis, wrote:
The abstractor has several objections to make to this
dejinition. First, the assumption of a “condition of
movement of the atoms in a molecule” is hypothetical
and therefore not suitable for purposes of dejinition. . . .
r f the abstractor were to formulate for himself the prob-
lem of characterizing the phenomena of catalysis in a
general way, he would consider the following expression
as probably most suitable: Catalysis is the acceleration
of a chemical reaction, which proceeds slowly, by the
presence of a foreign substance. It would then be neces-
sary to give the following explanations.. . . This accel-
eration occurs without alteration of the general energy
relations, since after the end of the reaction of foreign
body can again be separated from the field of the re-
action, so that the energy used by the addition can once
more be obtained by the separation, or the reverse.
However, these processes, like all natural ones, must
always occur in such a direction that the free energy of
the entire system is decreased.. . . The existence of cata-
lytic processes is to me therefore a positive proof that
chemical processes cannot have a kinetic nature.
With this review, the dominant concept of catalysis
passed from Liebig to Ostwald. In these comments
Ostwald showed that a catalyst could not change the
equilibrium. He, more than anyone, brought down the
vibrational theory as not being amenable to experi-
mental verification; at the same time his definition em-
phasized an equally unverifiable concept: a catalyst
cannot start a reaction which is not taking place with-
out it. In 1901, Ostwald [21] included as catalytic phe-
nomena:
(i) the release of supersaturation;
(ii) catalysis in homogeneous mixtures;
(iii) catalysis in heterogeneous systems;
(iv) enzyme action.
Ostwald had not accepted the kinetic theory at the time
of his definitions of catalysis, and incorrectly saw the
phenomena disproving the atomic hypothesis cham-
pioned by, among others, Boltzman. Like Liebig be-
fore him, Ostwald saw catalysis theory by analogy: a
catalyst acts like oil on a machine, or as with a whip on
a tired horse [22]. His work in catalysis was recognized
with the Nobel prize for chemistry in 1909.
References see page 33
18 1 Introduction
In Ostwald we have an example that demonstrates
that concepts that are shown, ultimately, to have little
validity, or even to be incorrect, may have great impact
in developing science. In many instances it is the drive
and forceful dominance by the individual that estab-
lishes the concept, and not the novelty of the concept.
Thus, most authors associate the concept of a catalyst
speeding up an existing reaction with Ostwald, not re-
alizing that Liebig advanced it 50 years earlier. With
the acceptance of physical chemistry, and the training
of a sufficient number of scientists to make use of the
advances in the field, catalysis was set to make many
advances beginning about 1900, and the Ostwald school
was in the driver’s seat.
1.2.3 The Concepts of Kinetics and
Intermediate Compounds
Dobereiner believed that platinum black carried oxy-
gen over to hydrogen. De la Rive suspected that a layer
of platinum oxide formed on the surface of platinum.
Fusinieri [23] went so far as to propose that waves of
oxidation and reduction were responsible for the re-
action of hydrogen and oxygen; he reported the visual
observation of such wave action. Faraday was unable,
however, to form a distinct idea of the power of Fusi-
nieri’s theory. However, Fusinieri defended his theory,
citing evidence to show that those with good eyesight
could see these waves. Berthollet suggested that plati-
num hydrides were formed in the presence of hydro-
gen and that these reacted with oxygen to produce
hydrogen.
Brodie [24] explained the decomposition of peroxides
on the basis of chemical affinity and of catalysis in
terms of coupled chemical reactions. For the reaction
(his uncorrected equations are used in the following):
12 + BazOz = 2BaI + 0 2
Brodie considered the reaction to take place by the de-
composition and reformation of water, according to
the two equations
12 + BazOz + H20 = 2HI + Ba2O + 0 2
2HI + BazO = 2BaI + H20
Viewing water as a catalyst provides the cyclic concept
of catalytic action.
Brodie also noted:
Viewsas to the polarization of oxygen, and the cause of
the decompositions effected by the alkaline peroxides,
which to a great extent are identical with the preceding
[24], and in which the same language and the same
notation are employed, have recently been put forward,
with considerable pretension, as new and originating
with himseK by Schonbein, Professor of Chemistry at
Basle. This chemist can scarcely be aware of the memoir
referred to [24], as in his numerous publications he
makes no allusion to it. A reclamation of priority of ten
years ought not to be required, but I am compelled to
call the attention of chemists to these circumstances in
order that I myself may not be considered to appro-
priate without acknouiledgment the ideas of discoveries
of another.
The intermediate compound theory has guided the
thoughts of many individuals whose contributions
moved catalysis to the forefront. For example, Sabatier
[25] wrote,
Having arrived at the end of m y career as a chemist, I
have thought that it might be of interest, and at the same
time of some utility, to relate how I was led to deuelop the
direct hydrogenation method by Jinely divided metals,
tvhich won for me the Nobel Prize in 1912. . . . It is the
faith in the theory of temporary compoundsfurnished by
the catalyst which has constantly guided me in these
mrious labors; it is these inductions to which I owe all
m y new results.
The concept of the intermediate compound remains
dominant today in the many “volcano plots,” and is
frequently associated with the names of Sabatier and/
or Balandin. Thus, the rate increases with the heat of
adsorption (or formation) until a sufficiently high value
is obtained so that the species is adsorbed so strongly
that the rate decreases; this produces the well-known
volcano plot [26] (Fig. 2). As with any principle or
correlation, it is frequently utilized in situations where
its use is questionable. Thus, without the data point for
Au, the volcano curve in Fig. 3 would be a straight
line.
Polanyi [27] utilized this concept in the development
of his transition state theory of catalysis. He assumed
that free homopolar valencies protrude from the surface
of a solid, and that a reaction AB + CD = AC + BD
would be catalyzed on the surface of such a solid
as represented in Fig. 4. For the reaction of H + H2,
Polyani provided the curve shown in Fig. 5. He
concluded that the catalytic effect will be strongest
when there is adsorption of moderate overall energy.
Surface valencies forming links of 50 000-5 1 000 gcal,
corresponding to Q’ = 0-2000, will be the most effi-
cient.
The highlights in chemical kinetics have been sum-
marized by Laidler in his outstanding history of physi-
cal chemistry (Table 1) [28]. The kinetic ideas intro-
duced by Wilhelmy and by Harcourt and Esson were
put forth by van? Hoff in a book, Etudes de dynamique
chimique, so that a compilation of the concepts of both
chemical thermodynamics and kinetics were available.
 1.2 DeveloDment of the Science of Catalysis 19

0 Pt (110)

0 Ni (110)

CuiNi (110)

88 Fe(100

I
600 I I I I 1

60 70 80 90 100 120
AHf (HCOOM), kcal/mol
Figure 2. Example of a typical volcano plot correlating the formate decomposition rate for both high-pressure and low-pressure single
crystal studies (reproduced from Ref. 26).

Final State

/
/
I -
Pdo
I
PA
' Aco,o,

B
-A
C
Initial State
initial State

Transition State
c",o

t
A
Final State

9731 1 I I I I 1 Figure 4. Polyani's representation of his transition state theory

0 200 400 600 (redrawn from Ref. 27).

-AH: (kJ/rnol 0) ki
A+C e Z
Figure 3. Example of a volcano plot with little data to sup- k3
port one side of the mountain (reproduced from an anonymous
reference). Zk'.B+C
Here the reaction that determines the velocity of the
van't Hoff introduced what is now considered to be process is the decomposition of Z, and the assumptions
the order of the reaction. Arrhenius also introduced of the low and steady-state concentration of Z permits
the concept of an intermediate, Z, in the reaction the rate of production of product B, catalyzed by C, to
scheme: be written as:

References see page 33

20 1 Introduction

science. The systematic studies by Hinshelwood ex-

I I I I I I I I
-6000 -4000 -2000 0 2000 4000 6000 8000 10,000
Heat of Adsorption of H, Molecule (a’)
Figure 5. Relationship between the rate of a catalyzed reaction
and the heat of adsorption of hydrogen - an early volcano-type
plot (redrawn from Ref. 27).
d[B]/dt= k3[Z] = k3Kl[A][C]; K1 = kl/k2
Langmuir [29] derived his isotherm from a consid-
eration of adsorption-desorption kinetics, and devel-
oped kinetic equations for a number of cases of catal-
ysis. The adsorption equation developed by Langmuir
was advanced independently by Michaelis and Menton
[30] to describe the kinetics of enzyme catalysis and
illustrates that by 1916 biocatalysis had again devel-
oped into a study separated from much of catalysis
tended the use of Langmuir isotherms to what is now
know as Langmuir-Hinshelwood kinetics. Hougen and
co-workers provided the basis for kinetic applications
in chemical engineering [3 11.
A development that provided a conceptual frame-
work for the kinetics of catalytic reactions was the in-
troduction of the activated reaction rate theory by
Eyring and co-workers [32, 331. Eyring’s theory pro-
vided a basis for the reaction coordinate shown for the
homogeneous and heterogeneous reaction in Fig. 6
[34]. In this figure, and even more so in many other in-
stances, the Ostwald view that a catalyst just speeds up
a reaction that is already occurring is evident since the
I activated complexes for the thermal and catalytic re-
action are depicted at a common point on the reaction
coordinate. Only recently has it become widely ac-
cepted that the role of the catalyst in nearly all cases is
to alter the reaction mechanism so that the thermal
pathway that requires a high activation energy is trans-
formed to two or more steps, each with a lower acti-
vation energy.
The concept of utilizing a stoichiometric number to
define, or at least limit the number of steps that define,
the rate was introduced by Horiuti [35], and many of
the mathematical derivations relating to this method
were developed by his group. The stoichiometric num-
ber is defined as the number of times that a step of the
reaction mechanism occurs. The first efforts to utilize
the technique to define the mechanism for ammonia
synthesis led to contradictory results; Enomoto and
Horiuti [36] reported a stoichiometric number of 2,
whereas the results of a combined kinetic and isotopic
study indicated that it was 1 [37]. Similar problems

Table 1. Highlights in chemical kintics (from Ref. 28)

Date Author Contribution

1850 Wilhelmy rate-concentration dependence

1865 Harcourt and Esson time course of reactions
1884 van’t Hoff differential method; temperature dependence
1889 Arrhenius Arrhenius equation
1891 Ostwald theory of catalysis
1899 Chapman theory of detonatin
1913 Chapman steady-state treatment
1914 Marcelin potential-energy surfaces
1917 Trautz; W. C. McC. Lewis collision theory
1918 Nernst atomic chain mechanism
1921 Langmuir surface reactions
1921- 1922 Lindemann; Christiansen unimolecular reactions
1927- 1928 Semenov; Hinshelwood branching chains
1931 Eyring and M. Polanyi potential-energy surface for H + H2
1934 Rice and Herzfeld organic chain mechanisms
1935 Eyring; Evans and M. Polanyi transition-state theory
1949 Porter and Norrish flash photolysis
1954 Eigen relaxation methods
1980 J. C. Polanyi spectroscopy of transition species
 1.2 DeveloDment of the Science of Catalysis 21

Kemball, however, wrote that the idea of an atomic

Activated
stot. chain based on the repeated reactions was proposed by
forgos rC0r)ion Nernst in 1916 [39]. In any event, catalysis and chain
reactions have features of a common concept.
An area of kinetics that has attracted much atten-
tion, the compensation effect where a high activation
energy is compensated for by an increase in the pre-
exponential term, derives from a one-page note by
a
I
I
Cremer and Schwab [40]. The initial experimental data
c
0
a
I
I
exhibiting the compensation effect was provided by
c
W
I
I Constable [41]. The debate continues as to the cause of
-
I
the linear relationship between 1nA and E; these in-

/
I
I
.-
0
+
c L clude views that it is based upon one of several theo-
a
c
i retical explanations, is merely an empirical observa-
‘,L ,!
GOS~OUS I
a
0 I
reactants I tion, or even to questions of its existence [42, 431.
I
Today, the progress of a heterogeneous catalytic re-
action can be resolved into at least five distinct steps:
products

(i) diffusion of the reactants to the catalyst;

A c o ction C o o r d i n a t e
Figure 6. Potential-energy curves for a reaction proceeding
homogeneously (full curve) and on a surface (dotted curve) that
illustrates the retention of the Ostwald concept that a catalyst only
speeds up a homogeneous reaction, and not one of changing the
mechanism (reproduced from Ref. 34).
(ii) the formation of the adsorption complex (reactant-
surface);
(iii) the chemical change on the surface;
(iv) the decomposition of the adsorption complex
(product-surface);
(v) diffusion of the reaction products from the catalyst.

Unless all five steps have the same activation energy,

have been encountered in applying the stoichiometric
number to other catalytic systems; thus, while it is an
attractive concept and provides assistance in eliminat-
ing some potential mechanisms from consideration, the
experimental data available today usually limit its use-
fulness.
The concept of turnover, the number of catalytic re-
actions per catalytic site per unit time came into com-
mon usage in enzyme catalysis. Only recently has this
concept been adapted to heterogeneous catalysis, par-
ticularly by Boudart and his students. Where the num-
ber of active sites can be defined precisely, e.g. H+ in a
sulfuric acid solution, the concept has great utility in a
definition of catalysis; however, it has been applied too
frequently in instances where the measure of the num-
ber of catalytic sites is uncertain, or worse.
Jost [38] attributes
the discovery of chain reactions to Bodenstein in 1913,
and the coining of the word to Christiansen . . . E.
Cremer, in her thesis with Bodenstein in 1927 on the
H2-Cl2 reaction, first noted the occurrence of chain
branching and instability caused by branching. The im-
plications of this concept, however, were not fully rec-
ognized either by her or by Bodenstein until Semenof’s
later publications.
the one step with the highest activation energy deter-
mines the rate.
Steps (i) and (v) received prominent attention during
the early application of kinetics to catalysis. Nernst,
among others, considered the rate-controlling step of
the dissolution of a solid to be due to the diffusion
of the molecule or ions through the nearly saturated
boundary layer that surrounds each particle. By anal-
ogy, in the late 1800s and early 1900s, Nernst, Boden-
stein, and others proposed that the rate of a catalytic
reaction was determined by the diffusion of reactants
and/or products through the layer(s) of adsorbed mol-
ecules. While this particular view of catalysis gradually
fell from favor, diffusion remains an important concept
in catalysis.
Thiele, working on the theory of catalysis at home in
his spare time, developed a connection between the rate
of reaction and the diffusional limitations imposed by
the porosity of a catalyst [44].This theory was am-
plified by many, particularly Wheeler [45]. Wheeler’s
work was very original and the results monumental; his
significant results included:
(i) emphasizing the apparent decrease of the activa-
tion energy, pressure and temperature gradients
expected in catalyst pellets;
(ii) influence of poisons on activity and selectivity;

References see page 33

22 1 Introduction
C Cat. Cracking (118” pills)
-Butane Dehydrogenation (1!8” pills)
-
0
40-
-=
0
L
20-
8 gums or waxes or by a deposit of carbon in organic
I I I I I I 1 1 L I
0 1 2 3 4 5
kexp.lkD
Figure 7. Wheeler’s estimate for the percentage of internal cata-
lyst surface available for effecting a number of important indus-
trial reactions (reproduced from Ref. 46).
(iii) role of diffusion in altering the observed kine-
tics from, for example, a series to parallel reac-
tions;
(iv) the estimate of the percentage of internal catalyst
surface utilized for many reactions (Fig. 7 ) .
1.2.4 Negative Catalysis - Autocatalysis
Negative catalysis was one of the concepts that added
to the confusion about catalysis. The work of Turner
[47] and Faraday [ l l ] clearly showed that some sub-
stances retarded the effectiveness of a catalyst. Mellor
[48] wrote,
The interesting feature is that these gases may be re-
garded as catalytic agents, which inhibit the action of
another catalytic agent. This phenonzenon will be called
negative catalysis.
Mellor explained that negative catalysis in ester hy-
drolysis had been explained, with more or less success,
by assuming:
(i) The degree of ionization of the ester or catalyzer
is diminished, or else the catalyzer combines with
the “foreign substance” so that the quantity of the
available catalytic agent is diminished.
(ii) The combination of the ester with the retarding salt
by which the active mass of the ester is diminished.
It was gradually recognized that negative catalysis
was merely the result of catalyst poisoning. Likewise,
the chemical concepts slowly developed to show that a
strong adsorptive bond between the poison and the
catalyst caused the effect to be highly specific. Due in a
large part to the extensive work by Maxted [49], it was
understood that the effect depended on definite types of
electronic configuration of both the catalyst and the
poison. A material would only be regarded as a poison
if it exerted an appreciable effect on catalysis even when
it was present in very small concentrations. Maxted
was careful to distinguish between catalyst activity loss
due to poisoning and that of “mechanically covering
up of a catalyst surface by less specifically held coat-
ings, such as the cloaking of a catalyst by a layer of
1 .
reactions at high temperatures.” Today, the study of
6 these topics have become so extensive that they are
important components of an International Symposia
on Catalyst Deactivation; the books based on these
symposia contain many papers that provide finer de-
tails and more complex examples of these two basic
concepts.
Autocatalysis was coined by Ostwald to identify a
class of catalytic reactions which are slow during an
induction period and then the rate accelerates rapidly
following this period. This effect will occur when one of
the products can act as a catalyst for the reaction.
Thus, the acid formed as a product of the hydrolysis of
an ester can act as a hydrolysis catalyst. Another in-
stance where autocatalysis will be observed is during
the reduction of a metal oxide with hydrogen, provided
the metal that forms catalyzes the reduction. Another
example occurs in the methanol-to-gasoline conversion
where a product, the C3+ alkenes in this case, combine
with methanol more rapidly than the hydrocarbons can
be formed from only methanol.
1.2.5 Adsorption
In the early 1900s, it was gradually recognized that
there are two types of adsorption. Langmuir’s work
[29] emphasized the type which has become known as
chemisorption. The usefulness of the other type, com-
monly known as physisorption, was emphasized by
the introduction of the Brunauer, Emmett, and Teller
(BET) method for measuring surface area. Taylor was
primarily responsible for the concept of activated ad-
sorption, and the need for an activation energy for
the transformation from the physisorbed state to the
chemisorbed state [50]. This proposal was made by
Taylor in an effort to more closely correlate the specific
activity of the catalysts with their adsorption charac-
teristics.
Wheeler [45] advanced the use of the physisorbed
state by introducing the concept that, by taking into
account the formation of multilayers of physically ad-
sorbed gas and capillary condensation in capillaries,

one could calculate the size distribution of pores pres-
ent in a catalyst. Barrett, Joyner, and Halenda [51]
provided a detailed procedure for quantitative calcu-
lations. Wheeler subsequently pioneered the develop-
ment of the theory of interpreting the influence of pore
size on many characteristics of catalytic activity and
selectivity. For example, the activation energy for a
reaction completely controlled by pore diffusion will
have only one-half the true activation energy; likewise,
a series reaction A + B + C may be made to appear
as parallel reactions by pore-diffusion control. More-
over, Wheeler pointed out that the catalyst preparation
conditions can be utilized to obtain a material with the
optimum surface area and porosity for the production
of a desired product. The isomerization of xylenes to
produce greater than 90% of the para isomer, in con-
trast to the 20-25% present at equilibrium, just by
controlling the pore opening of the HZSM-5 catalyst,
is one example of the application of the concept se-
lectivity control by diffusion. The concept of diffusion
in catalysis has now advanced to become a scientific
discipline that represents an independent field of study.
Washburn [52] developed a theory based upon mer-
cury penetration to provide a method to measure po-
rosity, especially for larger pores, and this concept was
converted into an experimental reality by the work of
Ritter and Drake [53].
The classification into five types of physical adsorp-
tion isotherms was accomplished by Brunauer et al.
(see Ref. 54). The desorption hysteresis associated with
each adsorption isotherm came to be related to the type
of pores that the material possessed. and while a single
source cannot be cited for this concept, de Boer and
co-workers had much to do with publicizing the im-
portance of the shapes of the hysteresis loops.
During 1914- 15, Langmuir put forth three concepts
important to chemisorption and its role in catalysis:
kinetics, his checkerboard model of the surface, and
chemical forces and bonding [55]. In developing his
concepts, Langmuir described heterogeneous catalysis
as a chemical drama that occurred in a single layer and
by the same chemical forces that held molecules and
solids together. Langmuir’s concepts dominated the
field for decades. and remain important today. The
kinetic nature of the adsorption and desorption steps at
equilibrium for a given pressure led to the Langmuir
Isotherm:
Q=kP/(l+kP)
where 0 is the surface coverage, k is the ratio of the
adsorption and desorption rate constants and P is the
pressure of the adsorbent. Langmuir’s notebook con-
tain his “fundamental laws of heterogeneous reactions”:
(i) The surface of a metal contains atoms spaced ac-
cording to a surface lattice (number per cm2).
1.2 Development of the Science of Catalysis 23
I
L
+Dist. from Metal
Figure 8. Schematic representation of the Leonard-Jones poten-
tial illustrating the activation energy needed to transfer from the
physically adsorbed to the chemisorbed state (redrawn from Ref.
56).
(ii) Adsorption films consist of atoms or molecules
held to the atoms forming the surface lattice by
chemical forces [ 5 5 ] .
Langmuir’s model implies that the amount of gas, say
CO, adsorbed on a metal, say Pt, permits a calculation
of the surface of the metal exposed to the gas; how-
ever, Langmuir did not emphasize this concept. He
did, however, introduce the checkerboard model for
the surface of the catalyst.
Brunauer and Emmett utilized in the 1930s the
combination of physical and chemical adsorption to
obtain a measure of the fraction of the surface of an
iron synthetic ammonia catalyst that consists of metal-
lic iron, alkali oxide promoter, and structural promoter
such as silica or alumina. Thus, low-temperature ni-
trogen adsorption provided a measure of the total sur-
face, chemisorbed CO a measure of free metallic iron,
and COz adsorption the surface concentration of alkali
oxide promoter, and by difference the fraction covered
by the structural promoter. Significant improvements
in the technique have occurred since the initial work,
but little has been altered concerning the concepts
involved.
Lennard-Jones [56] in 1935 provided a pictorial
scheme for the adsorption of, in this example, hydro-
gen (Fig. 8). This scheme illustrates that the transition
from van der Waals adsorption to that of the chem-
isorbed state needs an activation energy. This concept
allows for the activated adsorption championed by
Taylor; at the same time, in some special cases the ac-
tivation energy may be so low that adsorption may
occur with little or no activation. deBoer [57] consi-
dered adsorption from the viewpoint of an adsorption
time. The average stay of a molecule on the surface
should depend upon the interaction energy. Thus, as
the interaction energy increases, the average residence
References see page 33
24 1 Introduction

time increases; thus, as the energy increases the resi- place the concept of active sites on a firm foundation.
dence time increases from essentially no adsorption to About the same time that he advanced the active site,
physical adsorption to chemisorption. Taylor [63] made an observation that is equally in-
Surface mobility became an important concern fol- sightful, but it has received little attention:
lowing Langmuir’s work on chemisorption. Mobility,
as it applies to physical adsorption, means that the ad- An oxide catalyst surface is to be regarded as composed,
sorbate can move so freely on the surface that its state not of a single catalyst, but of two catalysts, metal ions
is that of a two-dimensional gas. For chemisorption, at and oxide ions and the nature of the changes induced in
least three definitions have been advanced [58]. Beeck the adsorbed reactant is determined by the charge of the
[59] suggested that adsorption be classified as mobile or ion on which the reactant molecule is adsorbed.
immobile according to whether the heat of adsorption
This statement contains the rudiments of an acid-base
fell with increasing coverage or remained constant.
theory of metal oxide catalysis.
Thus, if the adsorbate is mobile it should adsorb on
A liberal historian could trace the concept of active
high-energy sites first, resulting with a decrease in the
sites back at least to Loew [64]. He suggested that when
heat of adsorption with increasing surface coverage.
a molecule of the reactant contacts the catalyst, the
The field emission microscope [60] provided the first
“sharp comers” of the catalyst break up the molecule
direct method of studying surface mobility; it is sur-
into atoms, and these are more reactive. The finer the
prising that the developer of this technique, E. Muller,
particle, the greater the number of comers, and the
did not receive more recognition [61].
greater the activity of the catalyst.
Shortly following Taylor’s active site hypothesis,
Balandin proposed what has become known as the
1.2.6 Active Site - Geometric or Electronic? multiplet hypothesis. Balandin proposed that binding
between two atoms of an adsorbed molecule can be
Langmuir’s checkerboard model provided an initial
broken if they are attracted by two different catalyst
impetus to relating physical/chemical properties of the
atoms; a bond may be formed between atoms of an
solid and the catalytic characteristics. Taylor‘s name is
adsorbed molecule bound to the same catalyst atom
usually the one associated with the concept of an active
[65]. This hypothesis was applied by Balandin to many
site [62]. Taylor was not the first to propose a special
reactions, including the dehydrogenation of cyclohe-
site of unusually high activity; however, his simple
xane; here the spatial arrangement of the catalyst sur-
model and his stature focused much attention to the
face atoms must be such that metal-hydrogen bonding
active site. Thus, Taylor attributed special activity to
can result (Fig. 10) [66]. Those metals with little mis-
those atoms that, because of the uneven geometry of
the surface (Fig. 9) have many atoms whose coordina-
tion to other catalyst atoms is very low (unsaturation),
and these are the atoms that were attributed to provide
the seat of most of the catalytic conversions. This view
focused attention to the heterogeneity of the surface of
nearly all catalysts, and the fact that the total surface
would not be equally active in effecting chemical re-
actions. Constable assumed that these active sites fol-
lowed an exponential distribution [413. The application
of modern surface techniques have provided the data to

GRANULE PROPER Figure 10. Schematic view of the location of surface atoms of
the catalyst (intersections of light lines) and the adsorbed cyclo-
Figure 9. Taylor’s representation of heterogeneity and the pres- hexane (heavy lines); the multiplet theory indicated that the
ence of various types of active sites on a nickel catalyst (repro- spacings had to match in order to form the appropriate bonding
duced from Ref. 62). (reproduced from Ref. 66).
 1.2 Development of the Science of Catalysis 25
~~

match between the spacing of the reactant hydrogen

atoms and the catalyst surface atoms were considered
to be active catalysts, whereas those with a large mis-
match would be inactive. Balandin expanded his scope
to include the activation enthalpy of the formation-
decomposition of the multiplet complex and made
extensive studies to obtain a measure of the sum of
energies of disrupted bonds in reactants and newly
formed ones in products. This view attracted much
attention and, because of Balandin’s position, probably
survived in its original form much longer that could
have been anticipated.
Gwathmey began studies in 1934 that were to show,
at least qualitatively, that different metal crystal faces
possess different catalytic activities. Gwathmey’s work
was based upon his captivation by the diffraction ex-
periments of Davisson and Germer in 1927 [67]. He
demonstrated the anisotropy of different crystal faces 36 6( 52
% of d-character
in chemical reactions as well as the view that ani-
sotropy of surface behavior was wide-spread, if not Figure 11. Correlation of the catalytic activity for ethene hydro-
universal. genation with the percentage d character of the catalyst (repro-
Kobozev [68] extended Balandin’s concept and in- duced from Ref. 72).
troduced the hypothesis of active ensembles. As origi-
nally proposed, the theory lacked detail and was not
readily susceptible to experimental verification. Van
Hardeveld and Hartog [69] placed the Kobozev con-
cept upon a more quantitative basis when they made
- 0
calculations for the dependence of the number of metal .-5
atoms of a specified coordination number upon the
L

crystal size. The work by the Dutch workers resulted

t ‘
0,

0
P
from their failure to reproduce initially the report by 2 2
Eischens and Jacknow [70] of infrared bands due the c

adsorption of nitrogen on nickel. Only when they re-
duced a fresh catalyst in situ could they obtain the
sought-after IR band. They combined quantitative IR
data and models of the surface of various sized metal
catalyst particles to develop their quantitative model.
The situation became even more complex with the
introduction of the concept of facile and demanding
reactions [7 11; furthermore, it was considered that a
single atom should be adequate for a facile reaction,
whereas two or many more atoms would be required
for a catalytic site adequate for demanding reactions.
Thus, some reactions should be dependent on surface
geometry but other may not.
The pioneering work of Beeck using evaporated
metal films led to the concept of catalytic activity being
related to the number of holes in the d band and/or the
geometry (Figs 11 and 12). The concept of relating
catalytic activity to the electronic theories of solid-state
properties was given impetus by the theoretical paper
by Dowden and the accompanying paper with experi-
mental data to show that the catalytic activity of nickel
decreased linearly as the nickel was alloyed with copper
to cause a parallel decrease in the number of d-band
holes. Furthermore, the activity approached zero for
m
L
- 3
5
-8 4
1 I J
.30 .35 .40 .45
Interatomic distance nm
Figure 12. Correlation of the catalytic activity for ethene hydro-
genation with the surface atom interatomic distance (reproduced
from Ref. 73).
the alloy composition where the number of d-band
holes approached zero [74]. Although this work pro-
vided impetus for many studies during the following
years, it was subsequently shown that surface enrich-
ment caused the composition of the outermost layer of
the Ni-Cu alloy used by Dowden and Reynolds to
differ dramatically from that of the bulk.
Schwab [75], using Hume-Rothery alloys, showed
that the addition of a second element to increase the
occupation ratios of the first Brillouin zone (conduction
band) provided a corresponding increase in the activa-
tion energies of donor reactions (Fig. 13). Schwab there-

References see page 33

26 1 Introduction
A B C a Y E
1 of the verification of the concept were numerous. These
I
I included deviations from the Arrhenius plots for cal-
I
1.5 2
EC - 2.5
Figure 13. Alteration energy for the conversion of formic acid as
electron levels are filled by varying the composition of binary
metal catalysts: (A - , hardness (kgmm-2); B -+, activation
energy (kcalmole-I); C 0 0 0 , resistance Rcm x lo4) (redrawn
from Ref. 7 5 ) .
fore concluded that electron holes must be considered
to be involved in the catalysis.
The pioneering work of Farnsworth [76], who strug-
gled from his first publication in 1929 until the 1950s to
develop adequate instrumentation to obtain data to
show the likelihood of differences in the atomic spacing
within the surface layer and the bulk as well as the re-
laxation of the surface layer inwards, and Germer [77]
showed that the surface layer in many instances un-
dergoes reorganization to a more stable configuration
that differs dramatically from the structure of the bulk.
Furthermore, the gas molecules may adsorb and ar-
range themselves into a regular two-dimensional lattice
of their own, and this may be dependent upon surface
coverage. In addition, chemisorption may cause recon-
struction of the surface arrangement [78]. The pioneer-
ing infrared studies by Eischens and Pliskin [79] pro-
vided definitive evidence that CO adsorbed on a Pt
surface in more than one form, contrary to the check-
erboard model of Langmuir, and that, by analogy with
carbonyl cluster molecules, these corresponded to the
bridged and linear forms of adsorbed CO.
The staggering number of developments in the theo-
retical physics of both metals and semiconductors in
the mid-1900s led their application in catalysis. For
example, most of volume 7 of Advances in Catalysis
was devoted to some aspect of this issue. Wagner [80]
showed that zinc oxide could accommodate excess zinc
atoms that were associated with interstitial positions of
the lattice and that the conductivity of the solid de-
pended upon the pressure of the oxygen in contact with
the solid, as predicted by the theory. Doping ZnO with
a metal of valence greater than Zn2+ requires an in-
crease in free electrons, whereas doping with a uni-
valent cation will decrease the electron concentration.
Parravano and Boudart [81] pointed out that while the
literature data and the activity for hydrogen-deuterium
exchange could be shown to be compatible with the
predictions of the semiconductor theory, the limitations
culating the activation energy and evidence for a com-
pensation effect in the low temperature range where
activation energies could be calculated. Likewise, the
data for the changes in activity for CO oxidation with
doping of NiO led to contradictory conclusions [81].
Even so, it was concluded that “catalytic behavior of
semiconducting oxides has been modified in opposite
directions by the addition of impurities which modify
their electrical characteristics in opposite directions”
1811.
Garner et al. [82] made an extensive study of the
catalytic oxidation of CO on cuprous oxide and, at
the same time, followed the conductivity of the solid,
which resembled that of the solid following saturation
with CO. They suggested that oxygen and CO react on
the surface giving a carbonate complex which then re-
acts with adsorbed CO to produce C02:
and
coi,d, + co& + 2c02
Vol’kenshtein [83] attempted to formulate adsorp-
tion on semiconductors in a manner that took into ac-
count not only the Fermi level but also the surface
states and the gas phase and adsorbed electronic struc-
ture of the adsorbate. This concept led Vol’kenshtein to
the conclusion that the act of adsorption could create
new sites for adsorption. Thus, there may be different
kinds of levels even for the same molecules adsorbed
on the same (from its chemical nature) surface. A
“homogeneous” surface could therefore become “het-
erogeneous” by the act of adsorption.
The studies to elucidate the electronic aspects of ca-
talysis dominated the 1950s, and continues even today.
Two models have emerged: one is the atomistic model
(a surface molecule of one or more atoms) and the
other is the band model. The atomistic model essen-
tially ignores the solid and concentrates attention to the
ensemble of one or a few surface atoms. The band
model describes the surface in terms of surface states
and localized energy levels available at the surface.
Vol’kenshtein serves as an example of an early propo-
nent of the valence band model. Knor [84] summarizes
the other approach. Morrison [85] surveyed the ap-
plication of solid-state theories to catalysis in 1977 and
in 1989; his views were more optimistic in 1977 than in
1989.
Today, the geometric effect is one of two dominant
themes of the catalysis by metals, the other being the
electronic effect. The geometric and electronic effects
are usually discussed as separate topics; however, it

appears that, just as with acid-base catalysis, the two
effects are interrelated so that, at best, one can only
document that one of the effects plays a dominant role
in the catalysis. In fact, it appears that it is necessary to
accept the concept that the initial catalytic material,
and especially with metallic catalysts, may interact with
the reactant(s) to redesign itself. This concept can be
illustrated by the observation of Beeck initially [86],by
the retention of 14C-labeled acetylene of ethylene [87],
and the surface science of, for example, Somorjai and
Ertl, that different metal catalysts will strongly, and
essentially irreversibly, bond different amounts of a re-
actant. Thus, one can view the metallic catalyst as re-
designing itself by doping its surface, and in some cases
the bulk, with sufficient reactant so as to attain an
electronic state where the bonding is sufficiently weak-
ened that chemical reactions may occur.
1.2.7 Selected Systems
1.2.7.1 Ammonia Synthesis
Le Chatelier’s principle indicates that ammonia for-
mation is favored at high pressure. Convinced of his
principle, Le Chatelier had his assistant compress a
mixture of hydrogen and nitrogen to a very high pres-
sure, and then to apply an electrical spark. The explo-
sive reaction killed the assistant. The explosion was
a result of faulty experimental procedure whereby air
was permitted to mix with the nitrogen-hydrogen
mixture, and it was in fact the oxygen-hydrogen re-
action, rather than ammonia formation, that caused
the explosion. This was another of hundreds of un-
successful early attempts to synthesize ammonia.
Larson [88] credits Perman with the initial synthesis
of ammonia using an iron catalyst. However, it was
the interactions of Nernst and Haber that led to the
successful demonstration of the catalytic synthesis of
ammonia and the results that were to generate the ex-
citement needed for rapid progress. Haber determined
experimentally the amount of ammonia present at
equilibrium. Nernst computed, using the heats of re-
action and other chemical constants, the equilibrium
amounts for many reactions. All values calculated by
Nernst were in agreement with the experimental values,
except for the experimental value for ammonia re-
ported by Haber. In diplomatic terms, a very spirited
debate arose over the correct value for ammonia [89].
It is sufficient here to state that Haber’s additional
measurements at high pressure, spurned by Nernst’s
comments, resulted in showing that the equilibrium
concentration of ammonia was sufficiently high to
indicate industrial significance. Haber’s most active
catalyst, osmium, was not of industrial significance.
One of Haber’s novel concepts was to use a process
1.2 DeveloDment of the Science of Catalysis 27
in which the unconverted hydrogen and nitrogen, after
removal by condensation of the ammonia produced,
were recycled to the reactor. In this connection, Haber
introduced the concept of space-time-yield [90]. Even
so, Haber’s work would only have resulted in a victory
for ego and academics if he had not been able to in-
terest BASF in the process.
Mittasch, a student of Ostwald, guided the research
at BASF that led to the iron synthetic ammonia cata-
lyst that is still utilized today in an improved form.
Mittasch was led, just as Sabatier, to a successful cata-
lyst following the intermediate compound concept; in
this case, the catalyst was viewed briefly to form a
nitride intermediate with subsequent reduction by hy-
drogen. While concepts were utilized by Mittasch, the
work was systematic and extensive. During about a 2
year period, 6500 experiments with about 2500 dif-
ferent catalysts were made [90], truly an example of
Edisonian research. A magnetite sample from Sweden
proved to be a surprisingly active catalyst, and, because
of its relative high density, the concept of compact
lvhile porous developed, subsequently to be proved
wrong. Haber considered temporarily that the high
density of osmium and similar metals was the explana-
tion for their effectiveness as an ammonia synthesis
catalyst. During the extensive study by Mittasch and
co-workers, he developed his concepts of promoter
action of catalysis, and building on the meager earlier
work, defined in the 1920s the options for the role of
promoters (Fig. 14) [91].
The development of the US synthetic ammonia in-
dustry provides an outstanding example of exceptional
accomplishment in the development of the technology
first and then the scientific understanding while at the
same time showing exceptionally disastrous political
actions. Within 6 years Larson and co-workers at the
Fixed Nitrogen Laboratory (FNL) developed a suit-
able catalyst. Taylor [92], considering the scientific
work conducted by Emmett, Brunauer, and co-workers
at the FNL following the development work, wrote:
These authors have given us the most detailed kinetic
study ever made by a single reaction, with all aspects of
the reaction studied, adsorptions, kinetics, influence of
reactant concentrations as a function of composition and
mode of preparations. The treatment is so comprehensive
that it is possible to present an almost complete account
of the phenomenon of surface catalysis by reference to
this one example alone.
The concept of the dissociation of nitrogen as the slow
step was demonstrated using both kinetic and isotopic
techniques. The catalysts were carefully characterized.
Kinetic concepts were developed and these, together
with the concept of virtual pressure, introduced by
Temkin and Pyzhev [93], provides a kinetic model that
References see page 33
28 1 Introduction

OA 50 A 100 A
100 6 50 B OB
Composition of the Catalyst
I. Simple activation m.and T. Additive effect
I ~ .support effect 19. Mutual activation
TI. Poisoning ZLI. Deactivation.
Figure 14. The five ways that a promoter may impact catalytic activity as the concentration of the promoter is varied (reproduced from
Ref. 91).

N+3H 1.2.7.2 Acid Catalysis

Figure 15. Energy profile of the progress of ammonia synthesis
on Fe (energies are in kJmol-I; reproduced from Ref. 94).
is still valid. However, the more detailed model that
has resulted from many surface science studies pro-
vides a much sounder basis with greater detail for
the concepts of how the catalyst operates (Fig. 15)
WI.
Acids have been a dominant theme in the study of re-
actions, and acid catalysis predates its definition. The
dissociation theories advanced by Arrhenius, and ap-
plied by Ostwald, did much to aid in developing the
understanding of acid catalysis. These authors domi-
nated in work to show that the catalytic power of a
solution of an acid is directly proportional to its elec-
trical conductivity and independent of the nature of the
anion. This indicated that the catalyst is the hydrogen
ion and that its effect is directly proportional to its
concentration in the solution. However, by 1900 acid
catalysis was in a state of confusion, primarily because
of the failure to recognize the role of the undissociated
acid and the salt effects. The concepts of acid-base
catalysis have been presented during the last 50 years in
a series of books by Bell [95]. Unfortunately, all too
frequently scientists forget that acid-base catalysis is
the title of Bell's first book, and treat acid catalysis in
isolation from the conjugate base.
Hydrocarbon conversions have been dominant in
the study of acid catalysis. Early work on this topic
involved acids such as the one formed by an alumi-
num chloride/hydrogen chloride mixture; however, the
combination of the corrosion and the difficulty of cat-
alyst recovery has limited its use to special situations,
and these two factors led to the failure of the cracking

process introduced by Gulf Oil that was based on this
catalyst. The applications of acid catalysis was first
realized in petroleum processing by Eugene Houdry
with his introduction of catalytic cracking. Actually,
the Houdry process provided much greater advances in
catalyst regeneration and process control than it did
in introducing new catalytic concepts, and the same
was also true of the introduction of fluidized catalytic
cracking. Acid-treated clays were the first successful
acidic heterogeneous cracking catalyst, but they were
subsequently replaced by a synthetic amorphous silica-
alumina catalyst. The silica-alumina catalyst is acidic
because charge-balance requires that a proton, or other
cation, be present for each aluminum ion that is sub-
stituted into the SiOz framework. The credit for devel-
oping this concept for the source of acidity in the
amorphous silica-alumina catalyst is usually attributed
to Thomas [96]. In time, the amorphous catalyst was
replaced by crystalline zeolite catalysts.
Ipatieff and Pines carried out experiments, using
Ale13 as the catalyst and HC1 as a promoter, in June
1932, to show that alkylation of an alkane with an
alkene is possible. This work, augmented by that of
Komarewsky, Grosse, and other UOP scientists, led
to commercial processes. Schmerling explained these
reactions in terms of the carbonium ion mechanism.
These workers showed that this catalyst was effective
for hydrocarbon isomerizations, and that side-reactions
were inhibited by the presence of hydrogen [97].
Acid catalysis is also involved in a more complicated
process, naphtha reforming. Here two functions are
involved, a metallic function for hydrogenation and an
acidic function to effect hydrocarbon isomerizations.
Thus, the concept of bifunctional catalysis was intro-
duced [98]. Here, an alkene is dehydrogenated to an
alkene at the metallic function, the alkene then mi-
grates, on the surface or through the gas phase, to an
acid site where it undergoes isomerization, then mi-
grates again to the metallic function where it is hydro-
genated to an isomerized alkane. The introduction of
bi- or polyfunctional catalysis led to the concept of
the need for the two kinds of sites to be located suffi-
ciently close to each other so that transport between
the sites would not be the rate-limiting step of the
overall process.
1.2.7.3 Zeolites
Flanigen [99] provides a historical perspective of the
introduction of zeolites into catalysis. The incubation
period of the concepts needed for the application of
zeolites in catalysis began in 1756 with the discovery
of the first natural zeolite mineral. Following the deter-
mination of the structure in 1930 [loo], the colloidal
scientist McBain coined the term molecular siece to
1.2 Development of the Science of Catalysis 29
indicate that the materials act on the molecular scale
[loll. The pioneering work by Barrer on the synthesis
and adsorption properties of zeolites attracted atten-
tion to the subject. He offered the first classification of
zeolites based on molecular size and accomplished the
first definitive synthesis of zeolites [102]. Inspired by
Barrer’s work, Milton, Breck and co-workers at Union
Carbide synthesized several commercially significant
zeolites (including A, X and Y) by 1954. Initially the
zeolites found application as superior drying materials
and for separations. The introduction of the ISOSIV
process was an application of the concept of molecular
sieving, and offered the possibility of separating normal
and isoalkanes.
By 1959, the concept of a zeolite as an acid catalyst,
specifically as an isomerization catalyst, had been taken
to the commercial level [99]. However, the acidic na-
ture and the possible potential of their use as cracking
catalysts predated this; for example, Houdry workers
included data for catalytic cracking for a zeolite cata-
lyst in a paper at the 1950 Faraday Society meeting.
While zeolites were recognized to have a high catalytic
activity, they could not be kept in this state for a suffi-
cient amount of time. Extensive studies at Mobil Oil
led to the utilization of materials exchanged by rare
earths. It was the practical concept of Plank and Ros-
inski to test both the fresh zeolite and the material
produced by high-temperature steaming to simulate the
regeneration process that led to the explosive intro-
duction of zeolite-based cracking catalysts [ 1031. The
steamed material exhibited a dramatic increase in the
production of gasoline together with a dramatic de-
crease in coke make, over that of the best amorphous
silica-alumina catalyst then in use. Since the develop-
ment of the pioneering concepts and the development
of many commercial processes based on zeolites, it has
grown into a specialization of its own with the for-
mation of the International Zeolite Association and the
introduction of scientific journals to record the dra-
matic growth in both concepts and commercial appli-
cations of zeolites.
This provided the intellectual and economic driving
force that has led to the production of at least 100
crystalline materials, and many of these with unique
catalytic properties. As one example, the ZSM-5 zeolite
was found to convert methanol to gasoline-range hy-
drocarbons, high in aromatics, with a very high selec-
tivity [104]. This property was considered to be related
to the unique structure of the catalyst, and has resulted
in the advancement of numerous additional concepts.
1.2.7.4 Ions in Catalysis
The introduction of the concept of ions, and the success
of Ostwald in relating homogeneous catalytic action to
References see page 33
30 1 Introduction
the hydrogen ion, made it likely that the concept would
be extended to heterogeneous catalysis at a early date.
However, it appears that this did not occur prior to
1905 when Kirkby proposed that the hydrogen-oxygen
reaction was connected with the corpuscular discharge
emitted by hot platinum, which he supposed would
produce ions [105]. In 1940, Emmett and Teller [lo61
considered four earlier attempts [ 1071 to explain cata-
lytic action of surfaces by the formation of ions which
then react with each other. Concentrating their atten-
tion on metallic catalysts, Emmett and Teller con-
cluded that it was not likely that surface catalytic re-
actions had an ionic nature. They pointed out that in
many cases it reduces to a question of definition as to
whether or not one is dealing with ions. For surface
phenomena it is much more difficult to prove an ion
than it is for the gaseous state or for solutions. Thus,
the use of the concept of ionic intermediates in catalysis
expanded for metal oxide catalysts, and probably as
reaction intermediates where they are not appropriate,
but the concept did not prosper for metal-catalyzed re-
actions. Couper and Eley [lo81 later calculated theo-
retically that all possible ionization processes for the
adsorption of hydrogen on the surfaces of transition
are prohibitively endothermic. This concept was con-
ceived on a train during travel to a scientific meeting.
The concept of alkyl cations provided a basis for the
understanding of hydrocarbon cracking. The landmark
publication on rearrangements in acyclic hydrocarbon
and other systems by Whitmore established the useful-
ness of carbocations in explaining such conversions.
Three groups independently adapted the concepts ad-
vanced by Whitmore and utilized carbocations to ex-
plain the differences between the products obtained
from thermal and acid-catalyzed cracking [I 091. With
time, pentavalent carbon cations (carbonium ions)
were also identified and these are utilized together with
trivalent carbocations (carbenium ions) in explaining
catalytic cracking. Olah was a leader in advancing this
area, particularly the use of the superacid HSbF6 and in
applying NMR techniques to precisely identify cation
structures; he received the Nobel Prize in 1994.
1.2.7.5 Hydrogenation
The study of the reactions of hydrogen have contri-
buted dramatically to the development of concepts of
catalysis. The early fascination by the conversion of
hydrogen and oxygen over platinum and other solids
was instrumental in laying the groundwork for the
definition of catalysis. The reaction between hydrogen
and oxygen was sufficiently advanced that a book was
written on this subject in 1934 [l lo].
The role of hydrogenation in advancing the concepts
of catalysis received major boosts from the work of
Sabatier and of Ipatieff. Sabatier and his co-workers
were instrumental in developing the field of hydro-
genation of organic materials. French scientists had
reported that the porous metals adsorb, with sponta-
neous destruction, acetylene. Due to scientific courtesy,
Sabatier therefore did not feel that he could work with
acetylene, so his initial studies were with ethylene. Sa-
batier was a corecipient of the Nobel Prize in 1912 for
his work on hydrogenating organic compounds in the
presence of finely divided metals.
Ipatieff was trained as an artillery officer in the Rus-
sian army. He began research after becoming a pro-
fessor of chemistry at the Artillery Academy. In 1905
he introduced his high pressure autoclave, taking ad-
vantage of his artillery training to develop a tight seal
for the reaction vessel [l 111. With his new high pressure
reactor, Ipatieff made significant advances in the hy-
drogenation of aromatics, high pressure polymeriza-
tion of alkenes, and the hydrogenation of sugars. In
spite of his involvement in administrative duties fol-
lowing the Russian revolution, he continued some
research, including destructive hydrogenation of hy-
drocarbons. In 1929, after many of his associates and
co-workers had been arrested, some being shot without
trial, he left Russia for Germany, never to return, and
prohibited from contact with his children. In 1930,
efforts were started that were to lead him to UOP and
Northwestern University in the US. Arriving in the US
at age 62, he subsequently accomplished more research
than most persons accomplish during a whole career.
Among these was the development of alkylation de-
scribed in the section on acids.
In 1929, serendipity led to the discovery of the cata-
lytic conversions of ortho-para hydrogen (0-p-Hz)
[112]. Convinced of the concept that metallic platinum
would catalyze at liquid air temperature the transfor-
mation of the room-temperature mixture of o-p-Hz to
one that was enriched in the para compound, the ex-
periment was run repeatedly without success. Since Pt
had to be a good catalyst, the investigators reasoned
that it must be poisoned by an impurity in the hydro-
gen. Thus, Harteck applied the classic gas clean-up
technique: pass the gas over charcoal at a low tem-
perature. Harteck and Bonhoffer were surprised to find
that charcoal was an effective catalyst but, having a
source of nearly pure para-H2, they immediately began
pioneering studies on the catalysis of the o-p-Hz con-
version. The intense academic interest caused the
number of publications to increase very rapidly, and to
decline just as rapidly as all of the curiosity-driven
questions were answered. In 1935 no research director
would have advocated funding this research area as
one with a commercial potential, and would have been
correct. However, in 1960, with space travel and the
use of liquid hydrogen both as a fuel and potential
coolant, the 0-p-H2 conversion assumed immense com-

Figure 16. Plot of the number of papers on ortho-para hydrogen
catalytic conversion during the period 1910- 1988.
mercial importance, and the number of publications on
the topic showed a second rapid increase (Fig. 16). The
academic studies during the 1930s were directed toward
deciding which concept - dissociative adsorption or
magnetic effects acting on the undissociated molecule
were responsible for the transformation. In 1933, Bon-
hoeffer et al. [113] reviewed their work using metallic,
charcoal, and salt catalysts, and concluded that the two
mechanisms applied, but at different temperatures. At
low temperatures the conversion is monomolecular. It
occurs in the van der Waals adsorption layer, and
is most likely catalyzed by magnetic effects. At high
temperatures, the reaction involves atomic adsorption,
and this has become known as the Bonhoeffer-Farkas
mechanism.
The discovery in 1932 of the deuterium isotope and
methods to enrich its concentration ignited a period of
very fruitful research in catalysis. The Farkas brothers,
prepared by their work on o-p-H2 studies, were at the
forefront of this effort, but it was not easy for them to
do this. They were forced to leave Germany in 1933
because of the Nazi takeover of the Government,
and they settled in Sir Eric Rideal's laboratory in
Cambridge. In a 1 year period following their arrival in
England, they published 17 papers on the catalysis of
hydrogen. These publications illustrate a personality
trait of Rideal - he provided the brothers laboratory
space and facilities for their pioneering research - but
was a coauthor on only one of the papers, an unusual
gesture then or now. The first report by the Farkas
brothers demonstrated the occurrence of the thermal
and catalytic reaction
H2 + D2 = 2HD
1.2 Development of the Science of Catalysis 31
2-Butyne cis-2-ene
+
butane
Hz I 2-yne butane t +Hz
+H2
2-Bvt+yne* \ cis-2-ene II cis-2-ene
& +
'4
I Hz
Figure 17. Above: hydrogenation of 2-butyne to form 2-butene
and then the further hydrogen to produce only butane at the exit
of the reactor. Below: selective poisoning of the catalyst to pro-
duce nearly pure 2-butene (redrawn from Ref. 115).
They showed that the addition of D Z to ethene in the
presence of a nickel catalyst was accompanied above
60 "C by the exchange reaction
This mechanism differed from the one advanced earlier
that involve a half-hydrogenated intermediate during
D2-benzene exchange 11141. In 1939, A. Farkas sum-
marized the literature results and concluded that the
~
dissociative adsorption of hydrogen and the hydro-
carbon could explain the observed results. This paper
heightened interest in these reactions and initiated the
debate concerning the concepts of the dissociative or
the associative mechanism. During the subsequent
work, the Rideal-Eley mechanism, where interchange
between a chemisorbed atom and hydrogen adsorbed
in the van der Waals layer, was advanced. The ex-
perimental data for this later mechanism was shown
later to be in error, and by Rideal; however, the
Rideal-Eley concept survived and is applicable in
some cases even today. The measurements by Rideal
and his student Trapnell, the classic studies by Beeck,
and many others, showed that as chemisorption con-
tinued, the heat of adsorption decreased, and that this
altered the state from that of the irreversible adsorption
at low coverages to one of reversible adsorption at
higher coverages. The status today is that the concept
of dissociative adsorption describes the mechanism;
however, a few examples of associative adsorption
persist even today.
The hydrogenation of fats has become an important
industry. Studies to understand the process have led to
data that define the series nature of the reaction mech-
anism: the sequential hydrogenation of the triene to
the diene to the monoene and finally the saturated fat.
Burwell [115] presented a classic paper that showed
that hydrogenation of butenes was as or more rapid
than butyne; however, inhibition of butene hydrogena-
tion by the adsorbed butyne allowed nearly complete
References see page 33
32 1 Introduction
Figure 18. Mechanism of selective propene ammoxidation and oxidation over bismuth molybdate catalyst (redrawn from Ref. 133).
hydrogenation of the butyne prior to hydrogenation of
the butenes to form butane (Fig. 17). The classic book
by Bond on catalysis by metals includes many other
examples where hydrogenations have been utilized to
develop concepts of catalysis.
1.2.7.6 Oxidation
Catalytic oxidation dates at least from 1794 when Mrs.
Fulhame advanced the concept that the presence of
water was necessary to accelerate the oxidation of hy-
drocarbons [116]. Likewise, written documentation of
the oxidation of sulfur to sulfur trioxide in the presence
of saltpeter and then to sulfuric acid dates to the latter
half of the 15th century [117]. The patent for the man-
ufacture of sulfuric acid based upon the oxidation of
SO2 using air and a finely divided platinum metal pre-
dates the definition of catalysis by Berzelius. A pub-
lication by Knietsch [118] revealed concepts that had
been learned during intensive work at Badische: that
the Winkler concept that stoichiometric mixture of SO2
and oxygen, undiluted by other gases, had to be fed to
the catalyst was incorrect; the need to remove some of
the heat of reaction from the catalyst was recognized;
and it was claimed that the principles governing the
manufacture of the acid were now well understood.
I
25 O-
OXIDATION
1
n
The Hopcalite catalyst was unique since the dis-
covery and development work was conducted in uni-
versities, and under a tight time-frame required to de-
velop an effective gas mask for protection against a
poison gas, CO, during World War I [119]. The cata-
lyst that had sufticient activity to work at room tem-
perature was a combined effort by workers at The
Johns Hopkins University and the University of Cali-
fornia, Berkeley. After much study following World
War I, Frazer was able to explain the mechanism of
the catalysis by a two-point attachment of the 0 to the
metal atom and the C to an 0 of manganese dioxide.
The two-point attachment of the CO to the catalyst
surface woulg require stretching of the CO bond from
1.15 to 1.86A.
The Arabian alchemists are credited with a process
for the production of nitric acid as early as the 8th
century [120]. Rev. Milner reported in 1788 that nitric
oxide resulted from passing ammonia over heated
MnO2 [121]. The use of a platinum catalyst was ac-
complished by Henry in 1824 [122]. Kuhlmann re-
ported a number of catalysts, including F't, that cata-
lyzed the oxidation of ammonia to nitric acid [123]. A
heated platinized asbestos catalyst was used in 1891 by
Warren [124]. It remained for Ostwald [125] to make
a detailed study of ammonia oxidation, to settle on a
platinum catalyst, and for it to be a catalyst for build-

ing a commercial plant near Bochum, Germany. The
concept of using a platinum gauze catalyst was offered
by Kaiser [126], and has been considered to be the most
important innovation in ammonia oxidation to date
[127]. During 1920-30, extensive work was conducted
by Andrussow [128], Bodenstein [129], and Rasch-
ing [130] to understand the reaction mechanism and
three mechanisms were proposed: the imide mecha-
nism, the nitroxyl mechanism, and the hydroxylamine
mechanism (NH2OH intermediate). The hydroxyla-
mine mechanism, proposed by Bodenstein, became
widely accepted.
An excellent example of the admonition of Arm-
strong to study the catalyst is present in the under-
standing of oxidation with bismuth-molybdenum oxide
catalysts. Carbon- 13 and deuterium labeling conclu-
sively established that the formation of an allylic in-
termediate by the abstraction of an a-hydrogen was
the rate-limiting step [131]. Making use of "0 isotopic
tracer studies, Keulks and Matsuzaki [132] showed
that, for a prereduced catalyst, the low-temperature
reoxidation that incorporates oxygen is accompanied
with the Bio to Bi3, conversion, whereas the corre-
sponding high-temperature reoxidation is accompanied
with the Mo4' to Mo6+ conversion. The higher tempe-
rature reoxidation process is related to the rate-limit-
ing step for propene oxidation to acrolein at tempera-
tures below 400 'C. This step is identified as reoxidation
of the Mo centers by lattice oxygen diffusion. The
oxygen inlet site is identified as reduced Bi and the
oxygen outlet site is associated with Mo. Grasselli [133]
combined the results from his group and published
work to arrive at a detailed mechanistic cycle for the
selective ammoxidation or oxidation of propene over
bismuth molybdate (Fig. 18).
1.2.8 Summary
In catalysis, as in many areas of science, a concept will
be associated with one or a few individuals. Frequently
the concept will have been advanced in a similar form
and at an earlier time. In many instances, the concept
will be associated with the name of the forceful in-
dividual who provides a driving force to cause wide-
spread acceptance of the concept. One of the major
reasons for the more or less haphazard assignment of
credit for concepts is the scant attention paid to the
history of the subject. Thus, in catalysis, credit for
the development of a concept is likely to be assigned
on the basis of common usage.
References
I . S . J. Green, Industrial Catalysis, Ernest Benn, Ltd., 1928
2. E. Mitscherlich, Pagg. Ann. 1834, 31, 273.
1.2 Development of the Science of Catalysis 33
3. A. Trofast in Proceedings of Swedish Symposium of Cataly-
sis (Ed.: R. Larsson), Liber Laeromedel, Lund. Sweden,
1981, pp. 9-17.
4. J. R. Partington, A History of Chemistry, MacMillan, Lon-
don, 1964, pp. 261-264.
5 . J. W. Mellor, Chemical Statics and Dynamics, Longmans,
Green, and Co., London, 1904, p. 246.
6. H. E. Armstrong, Chemical News 1985. September 11, pp.
130- 156.
7. T. de Saussure, Ann. Philosophy 1815, 6 , 214; 331.
8. H. S. Taylor in Colloid Chemistry. Theoretical and Applied,
(Ed.: J. Alexander), The Chem. Catalog Co., Inc., New
York, 1931, Vol. 111.
9. A. J. B. Robertson, Platinum Metals Reu. 1975, 19, 64; M.
Farinelli, A. L. B. Gale, A. J. B. Robertson, Annals of Sci.
1974, 31, 19.
10. G.-M. Schwab in Catalysis. Science and Technologj,, (Eds.:
J. R. Anderson, M. Boudart), Springer-Verlag, Berlin, 1981,
Vol. 2, pp. 1-1 I).
11. M. Faraday, Phil. Trans. Roy. Soc. 1834, pt. I , p p . 55-
71.
12. J. W. Dobereiner. as described in: P. L. Dulong, L. J.
Thenard, Ann de Chim. et de phys. 1813, 23, 440; 1813. 24.
380.
13. J. Persoz. A. Payen, Ann. Chim. Phys. 1833, 53, 73.
14. L. Pasteur, Compt. Rend. 1858. 46, 179; 1859, 48, 640; 735;
1860, 50, 1083.
15. From Liebig's Complete Works on Chemistry, Justus Liebiy,
M.D., Ph. D., F. R. S., Professor of Chemistry in the Unicer-
sity of Giessen (Ed. by L. Playfair, from the last London
edition, much improved), T. B. Peterson and Brothers,
Philadelphia, no publication date, Chapter 11; p. 88.
16. J. Liebig, Familiar Letters on Chemistry, 3rd ed., Taylor,
Walton & Maberly, London. 1851, Letter X, pp. 125-138.
17. F. Wilhelmy, Ann. Phys. 1850,81, 413; 499.
18. A. F. Harcourt, W. Esson, Proc. Roy. Soc. 1865, 14, 470:
Phil. Trans., 1866, 156, 193; 1867, 157, 117.
19. B. H. Davis, Appl. Catal. 1992, 82, N-15.
20. W. Ostwald, Z. Phys. Chem. 1894, 15, 705-6; translation in
D. D . Runes (Ed.), A Treasury of World Sciences, Philo-
sophical Library, New York, 1962, pp. 782-784.
21. W. Ostwald, Nature 1902, 65, 522-6.
22 J. R. Partington, A History of Chemistry, MacMillan, Lon-
don, 1964, Volume 4, pp. 595-600.
23. G . Fusinieri, Giornale di Fisica 1825, 8, 259.
24 B. C. Brodie, Phil. Trans. Royal Soc., London 1862, 151,
837.
25 P. Sabatier, Ind. Eng. Chem. 1926, 18, 1005.
26 M. A. Barteau, Catal. Lett. 1991, 8, 175.
27 M. Polanyi, J. Sac. Chem. Ind. 1937, April 26, 123.
28. K. J. Laidler, The World of Physicai Chemistry, Oxford
University Press, Oxford, 1993, p. 233.
29. I. Langmuir, J. Am. Chem. SOC.1916, 38, 2221; 1918, 40,
1361.
30. L. Michaelis, M. L. Menton, Biochem. 2. 1913,49, 333.
31. S. W. Weller, Catal. Rev.-Sci. Eng. 1992, 34, 227-280.
32. S. Glastone, K. J. Laidler, H. Eyring, The Theory of Rate
Processes, McGraw-Hill, New York, 1941.
33. H. Eyring, M. Polanyi, 2. Physik. Chem. 1931, B12, 279.
34. K. Laidler, in Catalysis, (Ed.: P. H. Emmett), Reinhold,
New York, 1954, Vol. I , p. 236.
35. J. Horiuti, Advan. Catal. 1957, 9 , 339.
36. S. Enomoto, J. Horiuti, J. Research Inst. Catal., Hokkaido
Unit.. 1953, 2 , 87; 1955, 3, 185.
37. C. Bokhoven, M. J. Gorgels, P. Mars, Trans. Faraday Soc.
1959, 55, 339.
38. W. Jost in Ann. Rea. Pkys. Chem. (Ed.: H. Eyring), Annual
Reviews, Inc., Palo Alto, California 1966, Vol. 17, pp. 1-14.
34 1 Introduction
39. C. Kemball in Biographical Memoirs of Fellows of The
Royal Society, The Royal Society, London, 1975, Vol. 21,
pp. 517-547.
40. E. Cremer, G. M. Schwab, A. Physik. Chem. 1929, A144,
234.
41. F. H. Constable, Proc. Roy. SOC.(London) 1925, A108,
355.
42. E. Cremer, Advan. Catal. 1955, 7, 75.
43. M. P. Suarez, A. Palermo, C. M. Aldao. J. Thermal Anal.
1994,41, 807-816.
44. R. Randhava in Heterogeneous Catalysis, Selected American
Histories, (Eds.: B. H. Davis, W. P. Hettinger) ACS Symp.
Series, 1983,222, 173-177.
45. A. Wheeler, Advan. Catal. 1951, 3, 249; A. Wheeler Catal-
ysis (Ed.: P. H. Emmett) Van Nostrand, New York, 1955,
Vol. 2.
46. G. A. Mills in Heterogeneous Catalysis, Selected American
Histories, (Eds.: B. H. Davis, W. P. Hettinger) ACS Symp.
Series, 1983,222, 179-1 82.
47. E. Turner, Edin. Phil. J. 1824, 11, 99; 31 1 ; Pogg. Ann. 1824,
2, 210.
48. J. W. Mellor, Chemical Statics and Dynamics, Longmans,
Green, and Co., London, 1904, pp. 258; 285.
49. E. B. Maxted, Advan. Catal. 1951,3, 129-177.
50. H. S. Taylor, J. Am. Chem. SOC.1931, 53, 578.
51. E. P. Barrett, L. G. Joyner, P. P. Halenda, J. Am. Chem.
SOC.1951, 73, 373.
52. W. E. Washburn, Proc. Natl. Acad. Sci. 1921, 7, 115.
53. H. L. Ritter, L. C. Drake, Ind. Eng. Chem., Anal. Ed. 1945,
17, 787.
54. S . Brunauer, The Adsorption of Gases and Vapors, Princeton
U. Press, Princeton, 1942.
55. G. L. Gaines, G. Wise in Heterogeneous Catalysis. Selected
American Histories, (Eds.: B. H. Davis, W. P. Hettinger)
ASC Symp. Series, 1983,222, 13-22.
56. J. E. Lennard-Jones, Trans. Faraday Soc., 1932,28, 333.
57. J. H. deBoer, The Dynamic Character of Adsorption, Clar-
endon Press, Oxford, 1953.
58. A. W. Adamson, Physical Chemistry of Surfaces, 3rd ed.,
John Wiley & Sons, New York, 1976, pp. 652-654.
59. 0. Beeck, Adu. Catal. 1950,2, 151.
60. R. Germer, Adv. Catal., 1955, 7, 93.
61. E. W. Muller, T. T. Tsong, Field Ion Microscopy, Elsevier,
Amsterdam, 1969.
62. H. S. Taylor, Proc. Royal Soc., 1925, A108, 105; J. Phys.
Chem. 1926,30, 145.
63. H. S . Taylor in Colloid Symposium Monograph, (Ed.: H. B.
Weiser), The Chemical Catalog Co., Inc., New York, 1926,
pp. 19-28.
64. 0. Loew, J. Prakt. Chem. 1875, 11, 372.
65. A. A. Balandin, Advan. Catal. 1969, 19, 1; D. V. Sokolsky,
B. V. Erofeev, Chemtech 1973,3, 728.
66. B. M. W. Trapnell, Adv. Catal. 1951, 3, 4.
67. H. Leidheiser, Jr. in Heterogeneous Catalysis. Selected
American Histories, (Eds.: B. H. Davis, W. P. Hettinger, Jr.)
ACS Symp. Series., 1983,222, 121-130.
68. N. I. Kobozev, Acta. Physicochim USSR, 1938, 9, 805.
69. R. Van Hardeveld, F. Hartog, SurJ Sci. 1969, 15, 189.
70. R. P. Eischens, J. Jacknow, Proc. 3rd Int. Congr. Catal.
(Eds. W. M. H . Sachtler, G. C. A. Schuit, P. Zwietering),
North-Holland, Amsterdam, 1964, p. 627.
71. M. Boudart, Proc. 6th Intern. Congr. Catal., (Eds.: G. C.
Bond, P. B. Wells, F. C. Tompkins), The Chemical Society,
London, 1977, pp. 1-9.
72. 0. Beeck, A. W. Ritchie, Disc. Faraday SOC. 1950, 8,
159.
73. 0. Beeck, Modern Phys. 1945, 17, 61.
74. D. A. Dowden, J. Chem. Soc., 1950,242; D. A. Dowden, P.
W. Reynolds, Disc. Faraday SOC.1950,8, 184- 190.
75. G. M. Schwab, Disc. Faraday SOC.1950, 8, 166; G. M.
Schwab in Catalysis, (Eds.: J. R. Anderson, M. Boudart),
Springer-Verlag, Berlin, 1981, Vol. 2, pp. 1-11.
76. H. E. Farnsworth, Advan Catal. 1957, 9, 493.
77. L. H. Germer, Surj Sci. 1966,5, 147-151.
78. J. W. May, Ind. Eng. Chem. 1965, 57, 18.
79. R. P. Eischens, W. A. Pliskin, Advan. Catal. 1957, 9,
662.
80. C. Wagner, Z . Physik Chem. 1933, B22, 181.
81. G. Parravano, M. Boudart, Advan. Catal. 1955, 7,47.
82. W. E. Gamer, T. J. Gray, F. S. Stone, Disc. Faraday SOC.
1950, 8, 246.
83. F. F. Vol’kenshtein in Scientific Selection of Catalysts,
(Eds.: A. A. Balandin, et al.), English Translation, Israel
Program for Scientific Translations, Jerusalem, 1968, pp.
70-8 1.
84. Z . Knor, Advan. Catal. 1972, 22, 51.
85. S. R. Morrison, The Chemical Physics of Surfaces, Plenum
Press, New York, 1977; M. J. Madou, S . R. Morrison,
Chemical Sensing with Solid State Devices, Academic Press,
New York, 1989.
86. 0. Beeck, Disc. Faraday SOC.1950, 8, 118.
87. G. F. Taylor, S. J. Thomson, G. Webb, J. Catal. 1968, 12,
191.
88. A. T. Larson, J. Chem. Educ. 1926,3, 284-290.
89. K. Tamaru in Catalytic Ammonia Synthesis, (Ed.: J. R.
Jennings) Plenum Press, New York, 1991, pp. 118.
90. E. Farber, Chymia 1966, I l , 157.
91. A. Mittasch, Advan. Catal. 1950,2, 81.
92. H. S. Taylor, Am. Sci. 1946, October, 553.
93. M. Temkin, V. Pyzhev, Acta. Physicochem. USSR 1940,12,
327.
94. G. Ertl in Catalytic Ammonia Synthesis, (Ed.: J. R. Jen-
nings), Plenum Press, New York, 1991, p. 128.
95. R. P. Bell, The Proton in Chemistry, 3rd ed., Cornell Uni-
versity Press, 1973.
96. C. L. Thomas in Heterogeneous Catalysis, (Eds.: B. H.
Davis, W. P. Hettinger, Jr.) ACS Symp. Series, 1983, 222,
241-245.
97. V. N. Haensel, L. Schmerling, Advan. Catal. 1948, I , 27.
98. G. A. Mills, H. Heinmann, T. H. Milliken, A. G. Oblad,
Ind. Eng. Chem. 1953,45, 134.
99. E. M. Flanigen in Introduction to Zeolite Science and Prac-
tice, (Eds.: H. van Bekkum, E. M. Flanigen, J. C. Jander),
Elsevier, Amsterdam, 1991, pp. 13-34.
100. W. H. Taylor, Z. Kristallogr. 1930, 74, 1; L. Pauling, Proc.
Nut. Acad. Sci. 1930, 16, 453; Z. Kristallogr. 1930, 74,
213.
101. J. W. McBain, The Sorption of Gases and Vapors by Solids,
Rutledge and Sons. London, 1932, Chapter 5.
102. R. M. Barrer, J. SOC.Chem. Ind. 1945, 64, 130; R . M. Bar-
rer, J. Chem. SOC.1948, 2158.
103. C. J. Plank in Heterogeneous Catalysis. Selected American
Histories, (Eds.: B. H. Davis, W. P. Hettinger, Jr.) ACS
Symp. Series, 1983,222, 253-271.
104. G. T. Kerr, Scientific American 1989,261, 100.
105. P. J. Kirkby, Phil. Mag. 1905, 10, 467; A. J. B. Robertson,
Platinum Metals Rev. 1983,27, 31-39.
106. P. H. Emmett, E. Teller, in 12th Report on the Committee on
Catalysis, National Res. Council, Comm. on Contact Catal-
ysis, John Wiley & Sons, 1940, pp. 68-81.
107. A. K. Brewer, J. Phys. Chem. 1928, 32, 1006; J. E. Nyrop,
The Catalytic Action of Surfaces, Levin and Munkgaard,
Copenhagen, Denmark, 1937; L. V. Pisarzhevskii, Ukrain.
Khem. Zhur. 1925,1, Sci. part. 1-18; Acta Physiochim. URSS
1937,6, 555-574; 0. Schmidt, Chem. Rev. 1933,12, 363.
108. A. Couper, D. D. Eley, Disc. Faraday SOC.1950,8, 172.
109. C. L. Thomas in Heterogeneous Catalysis. Selected Ameri-
can Histories. (Eds.: B. H. Davis, W. P. Hettinger, Jr.) ACS

Symp. Series, 1983, 222, 241-245; H. H. Voge, ACS Symp.
Series 1983, 222, 235-240; R . C. Hansford, ACS Symp.
Series, 1983, 222, 241-252.
110. C. N. Hinshelwood, A. T. Williamson, The Reaction Between
Hydrogen and Oxygen, The Clarendon Press, Oxford, 1934.
111. H. Pines in Heterogeneous Catalysis. Selected American
Histories, (Eds.: B. H. Davis, W. P. Hettinger), ACS Svmp.
Ser. 1983,222, 23-32.
112. A. Farkas in Heterogeneous Catalysis. Selected American
Histories, (Eds.: B. H. Davis, W. P. Hettinger), ACS Symp.
Ser. 1983,222, 89-117.
113. K. F. Bonhoeffer. A. Farkas, K. W. Rummel, Z. Phj3sik
Chem. 1933, B21, 225.
114. J. Horiuti, M. Polyani, Nature 1933, 132, 931.
115. R. L. Burwell, Jr., C & E News, 1966, Aug. 22, 56-67.
116. Mrs. Fulhame, An Essay on Combustion, London, 1794; J.
W. Mellor, J. Phys. Chem. 1903, 7, 551.
117. A. M. Fairlie, Sulfuric Acid Manufacture, Reinhold Pub.
Corp., New York, 1936, pp. 23-26.
118. R. Knietsch. Ber. 1901, 34, 4069.
119. A. B. Lamb. W. C. Bray, J. C. W. Frazer, Ind. Eng. Chem.
1920, 12, 213.
120. Kirk-Olhmer Encyclopedia of Chemical Technology, 3rd ed.,
p. 853.
121. J. R. Partington, L. H. Parker, The Nitrogen Industry, Van
Nostrand Co., 1923, p. 210.
122. W. Henry, Phil. Trans. Roy. Soc. 1824, 114, 266.
123. F. Kuhlmann, Compt. Rendu 1838, 7, 1107.
124. H. N. Warren, Chem. News 1891, 63, 290.
125. W. Ostwald, Chem. Zeitung 1903, 27, 457.
126. K. Kiaser, US Patent 987 375, 1911.
127. B. S. Beshty, presentation at ACS Dic. ofHistory of Chem.,
Philadelphia, Fall Meeting, 1983.
128. L. Andrusow. Z. Angew. Chem. 1926, 39, 321; 1927, 40,
166; 1928, 41, 206.
129. M. Bodenstein. Angew. Chem. 1926, 40, 174.
130 F. Rasching, Angew. Chem. 1927, 40, 118.
131 W. H. M. Sachtler, N. H. deBoer, Proc. Int. Congr. Catul.
1965, 3. 252; C. R. Adams, T. Jennings, J. Cutal. 1964, 3,
549; 1962,2, 63.
I32 G. W. Keulks, T. Matsuzaki in Adsorption and Cataly.~ison
Oxide Surfaces, (Eds.: M. Che, G. C. Bond), Elsevier, Am-
sterdam, 1985, pp. 297-307.
133 R. K. Grasselli in Adsorption and Catalysis on Oxide Sur-
faces, (Eds.: M. Che, G. C. Bond), Elsevier, Amsterdam,
1985. pp. 275-283.
1.3 Development of Industrial
Catalysis
H. HEINEMANN
1.3.1 Introduction
Catalysis is involved in a very large percentage of pro-
cesses leading to the production of industrial chemicals,
fuels, pharmaceuticals, and to the avoidance, as well as
the clear-up, of environmental pollutants. The world-
1.3 Development of Industrial Catalysis 35
wide value of fuels and chemicals produced by catalytic
reactions is about $2.4 trillion per year, or more than
the gross national product of many nations. About
20% of the value of all commercial products manu-
factured in the USA is derived from processes involv-
ing catalysis. The vast majority of industrial catalytic
reactions involve heterogeneous catalysis but, since the
mid- 1980s, homogeneously catalyzed processes have
found increasing use. The value of the heterogeneous
catalyst market is estimated at $5 billion per year and
is expected to grow to $6.5 billion by the year 2000.
The cost of catalysts is normally a relatively small part
of processing costs and is estimated at 0.1% of the
value of fuels produced and about 0.22% of chemicals.
Industrial catalysis has developed slowly from an
art to a science and, even today, many industrial pro-
cesses are based more on enlightened empiricism than
on derivation from scientific knowledge. For centuries,
wine and soap makers have employed catalytic agents
on a know-how basis only. The lead chamber process
for the manufacture of sulfuric acid using nitric acid to
oxidize SO1 to SO3 in the presence of water was in-
troduced in the mid-18th century and is probably the
first conscious use of catalysis in industrial processing.
The lead chamber process was largely replaced, be-
ginning in about 1900, with a heterogeneous catalytic
process using platinum on acid-resisting material (as-
bestos or silica gel) as catalysts for the oxidation of SO2
with air. In subsequent years, catalysts comprising
VzOs and KzS04 on a silica support were substituted
for the more easily poisoned Pt.
Whereas other relatively small-scale uses of catalytic
reactions occurred in the 19th century, the first major
breakthrough of modern industrial catalysis was the
fixation of nitrogen by the Haber-Bosch process [l]
which was commercialized in 1913. From about 1910
on, industrial catalysis rapidly gained in importance
until it dominated chemical processing by the middle
of the 20th century [2]. Its growth was stimulated to a
considerable extent by the industrial needs of two
World Wars, and by the discovery of vast resources of
petroleum, which changed the energy balance of world
trade [3].
Table 1 presents a chronological summary of major
catalytic developments of great industrial importance.
It can be divided into three time periods of about 25-
30 years each, during each of which about 10 break-
throughs occurred. This first period from 1910 to 1940
involved the synthesis of base chemicals (e.g. ammonia,
methanol, nitric acid) and the conversion of natural
resources to chemical and fuel building blocks. The in-
creasing availability of inexpensive petroleum hydro-
carbons and of fuels produced from petroleum was the
main stimulus for new catalytic technologies in the pe-
troleum and petrochemical industries during the sec-
ond period between about 1935 and 1965 [4]. This
period was concentrated on fuels from petroleum and
References see page 48
36 1 Introduction

‘Table 1. Major catalytic inventions.

~~~

Year of first Process Products

commercialization
~~ ~

= 1850 lead chamber process H2S04

x 1900 air oxidation of SO2 HzS04
1913 ammonia from N2 H2+ NH3
1915 ammonia oxidation HN03
x I920 water gas shift +
C02 H2
1923 methanol synthesis CH30H
1931 Bergius coal hydrogenation liquid fuels
1936 catalytic cracking gasoline
1938 Fischer-tropsch fuels
1941 fluid bed technology fuels and chemicals
1942 alkane alkylation gasoline
1950 naphtha reforming gasoline
1951 hydrocracking fuels
1955 Ziegler-natta polymerization plastics
1960 ethylene oxidation acetaldehyde
1962 steam reforming of CH4 syngas
1963 low-pressure ammonia synthesis NH3
1963 ammoxidation Nitriles
1964 zeolite catalysts fuels and chemicals
1964 oxychlorination chlorohydrocarbons plus chlorine
1966 alkene disproportionation higher alkenes
1967 bimetallic reforming catalysts Fuels
1968 Shape-selective catalysis chemicals
1970 low-pressure methanol synthesis CH3OH
1976 emission control catalysts environmental control
1978 methanol to gasoline synfuels
1982 crystalline aluminophosphates chemicals
x 1986 stereospecific synthesis pharmaceuticals
= 1986 NO, + NH3 ----t N2 + 2Hz0 environmental control
1990 combustion catalysts environmental control

on monomers and polymers derived from petroleum
processing.
The last period from 1965 to 1990 shows a gradual,
but accelerating, shift from catalytic processing of large
volumes of relatively inexpensive products to more
directed syntheses of smaller volume, more expensive
products (such as complex fibers, specialty and phar-
maceutical chemicals) and to major applications of
catalysis in environmental protection. Automotive
emission control [5], NO, removal from stack gases
and a wide range of chemical waste and byproducts
control were supplemented by new or modified pro-
cesses for manufacturing chemicals with higher selec-
tivity and fewer toxic byproducts.
Another trend in catalytic processing began about
1985 with the recognition that catalysis could assist
in the production of stereospecific and of chiral com-
pounds, particularly in the pharmaceutical and fine
chemicals industries. Whereas the bulk of catalytic pro-
cessing depends on solid catalysts, homogeneous catal-
ysis is playing an increasing role, mostly in the phar-
maceutical and fine chemical field. Zeolites, originally
introduced as petroleum refining catalysts have been
used in modified forms for a large variety of shape-
selective reactions, and have played a large role in
the successful attempts to substitute heterogeneous for
homogeneous catalysis, thus permitting an easier sepa-
ration of catalyst from substrate. These developments
reflect a shift from an industrial economy dominated
by price competitiveness alone to one in which envi-
ronmental regulations may supercede the cost factor.
In the following more detailed review of the devel-
opment of industrial catalysis, it is not always possible
to maintain a strict chronology, since progress was
often achieved by attaining a specific objective by im-
proving previous achievements through new and better
catalysts or by new engineering technology. The inter-
play of catalytic chemistry and process engineering is of
great importance, and the transition of a concept from
the laboratory to a competitive commercial product or
process involves at least as much effort and ingenuity
as the concept itself. Similarly, the scale-up of catalyst
manufacture requires both engineering and chemistry
F, 71.
1.3.2 The Period from 1910 to 1938
The Haber-Bosch synthesis of ammonia from the ele-
ments replaced dependence on guano, a rapidly dimin-
 1.3 Development of Industrial Catalysis 37

ishing resource, and on the very high temperature arc permitted the production of two moles of hydrogen per
process to produce cyanamide, which was thermally mole of carbon:
inefficient. The ammonia synthesis alleviated the fear
of world-wide starvation by providing fertilizers to in- C + H20 + CO + H2
CO + H7O C07 + H7
crease crop yields in line with population increases. It
-+
also provided nitrates for military explosives. It has
c + 2H20 c02 + 2H2
~~ ~

been said that World War I would have ended in 1916 --$

without Germany’s ability to produce gunpowder from In about 1962, a low temperature ( z 250 “C) shift
the ammonia synthesis. The military demand for ni- catalyst comprising copper, zinc oxide, and alumina
trates also expedited commercialization in 1915 of a was introduced to convert the CO remaining from the
process originally invented by Ostwald, which permit- high temperature shift, which is equilibrium limited.
ted conversion of ammonia to nitric acid, using iron-
An alternative route to synthetic fuels from coal
manganese-bismuth oxide catalysts.
via synthesis gas was invented by Fischer and Tropsch
The ammonia synthesis using promoted iron cata- and first commercialized in 1938. It is known as the
lysts owes its origin to the thermodynamic and kinetic
Fischer-Tropsch synthesis. Hydrogenation of carbon
studies of Fritz Haber, largely supported by grants
monoxide was carried out at 400-450 “C and about 7-
from the chemical industry. He determined the need for
30 bar pressure over alkalized iron turnings as catalyst.
high pressure and low temperature to assure reason-
Earlier catalysts had consisted of cobalt and of nickel
able yields of ammonia. His work faced serious ob-
or manganese oxides. Large scale plants in Germany
stacles in the development of a commercial process
produced about 800000t per annum of liquid fuels
because of the need to build large-scale high-pressure
during World War 11, using fixed bed reactors. The
equipment. Bosch and his co-workers achieved suc-
products were alkane hydrocarbons with an appreci-
cess in constructing and operating such equipment,
able amount of oxygenated compounds. After World
and Mittasch investigated numerous ammonia synthe-
Way 11, large-scale use of the Fischer-Tropsch synthe-
sis catalysts and scaled up their synthesis to commercial
sis was centered in South Africa where, since 1955,
scale. The availability of high pressure equipment con-
Sasol has operated first a fixed bed reactor plant with a
tributed greatly to the development of other catalytic
capacity of 250000t per annum and, later, two fluid
processes.
bed reactor plants with over 2 500 000 t per annum ca-
During experience with the high pressure catalytic
pacity. Since 1993, Shell Oil Company has operated a
ammonia synthesis, German scientists noticed in about
modified Fischer-Tropsch plant in Malaysia, which
1913 the presence of oxygen-containing compounds
uses syngas from natural gas to produce high-molecu-
in the product. Further research led to the catalytic
methanol synthesis from synthesis gas (CO H2) + lar-weight alkanes, which are then hydrocracked to
diesel fuel. This plant appears to use a cobalt catalyst.
which became commercial in 1923. The catalyst em-
There has been an abundance of publications since
ployed was ZnO-CrzO3, later replaced by CuO-ZnO
the 1950s to cover various scientific and/or techno-
catalysts. Manufacture of synthesis gas was originally
logical aspects of the Fischer-Tropsch synthesis, in-
based on gasification of coke from coal by means
cluding mechanisms as well as the effect of various
of air and steam. Only after World War I1 did pro-
catalyst metals, e.g. ruthenium, on product distribution.
duction of syngas by catalytic steam reforming of
In 1902, Sabatier demonstrated the catalytic hydro-
natural gas or of petroleum fractions become the
genation of ethylene using nickel catalysts. With the
dominant process, using nickel-based catalysts. This
availability of high-pressure equipment, Pier and co-
process was derived from work by Mond in the late
workers in the 1920s and 1930s extended catalytic
19th century.
hydrogenation to several important processes. such as
The oxidation of ammonia to nitric acid, first pro-
alkene saturation, hydrocracking, and aromatic satu-
posed by Ostwald in 1906, was commercialized in 1915,
ration, using metal sulfides as catalysts.
employing platinum catalysts. Just as the military de-
From the 1930s on, there has been a continuous
mand for gun powder had stimulated research on the
stream of process and catalysts improvements, in addi-
ammonia synthesis, the need for a domestic supply of
tion to novel processes.
liquid fuels resulted in research to convert coal to au-
tomotive and lubricant fuels. In 1931, Professor Ber-
gius in Germany demonstrated the liquefaction of coal
by hydrogenation in the liquid phase, with the coal and 1.3.3 The Period from 1938 to 1965
solid catalysts being suspended in a tar oil. The hydro-
gen required was again derived from synthesis gas. World War I1 stimulated the rapid development of
+
The water gas shift reaction (CO H 2 0 + C 0 2 + H2) three major processes which had been invented during
using a Fe304 catalyst containing some chromia at the period 1933-1945: catalytic cracking provided
z 350-450°C had been discovered before 1920 and much of the high-octane gasoline for the Allied air and

References see page 48

38 1 Introduction
1 7
+
Gas oil --_-_ _____
charge -+ c t
IC) “1
Air f o r --
regeneration
Purge ...... --
land forces; catalytic alkylation was a source of close
to 100 octane number gasoline which, blended with
catalytically cracked gasoline, provided a superior avi-
ation fuel; and catalytic dehydrogenation of methyl-
cyclohexane in petroleum naphtha resulted in a supply
of toluene for further nitration to trinitrotoluene
(TNT). Based in part on these developments and fur-
ther inventions, numerous other industrial applications
of catalysis occurred between 1945 and 1965.
1.3.3.1 Catalytic Cracking
Gasoline comprises only 15-25% of natural petroleum.
Its chemical components consist primarily of relatively
low octane number alkanes. Since motor and aviation
fuel gasoline was (and still is) the most desirable pe-
troleum fraction, methods to convert higher molecular
weight (C12+) hydrocarbons to large volumes of gaso-
line of high octane number were attractive. This re-
quired not only the breaking of carbon-carbon bonds,
but also the structural rearrangement of hydrocarbons
by isomerization, dehydrogenation, and aromatization.
All these reactions can be catalyzed by acidic catalysts.
Houdry based the first large-scale application of cata-
lytic cracking on the earlier recognition that acidic
clays were good cracking catalysts, and he overcame
several severe engineering problems, particularly heat
balancing the reactor design for an endothermic reac-
tion and exothermically regenerating a catalyst which
was continuously being deactivated by coke deposition.
The first Houdry units, which came into operation
shortly before World War 11, had fixed bed tubular re-
actors surrounded by a molten salt bed acting as a heat
transfer medium. A group of three reactor units was
operated in a cyclic mode with one reactor on-stream,
-Cracked product
t o fractionation
-Combustion and
-
purge gases
ICI ICI Catalyst
cases
0 Valves t o
permit having
one r e a c t o r each
on stream, on
one being purged by inert gas, and the third being re-
generated with air. A schematic drawing is shown in
Fig. 1 and indicates the complex valving required.
Numerous commercial units were built all over the
world during and after World War I1 and, at its peak,
catalytic cracking units in the USA alone produced
over 250 million tons of gasoline per year. Quite early
in the development of large-scale catalytic cracking,
major modifications of both unit design and catalyst
composition occurred. The cumbersome design of the
fixed bed units was replaced by units having separate
reactors and regenerators with the catalyst being
moved from one to the other and back. In one mode,
this was accomplished in a moving bed unit, in which
spent pelleted catalyst was continuously withdrawn
from the reactor and lifted either by elevators or pneu-
matically to the regenerator, from which regenerated
catalyst was, by gravity flow, reintroduced to the re-
actor. In the other, ultimately more successful mode,
both reactor and regenerator operated in a fluid bed
mode with small catalyst particles suspended in a gas
stream (Figs 2-4). Fluid bed operation also permits
good temperature control by heat transfer and has
found wide application in other exothermic or endo-
thermic reactions, e.g. in the partial oxidation of
hydrocarbons.
Catalyst improvements have progressed steadily from
the inception of catalytic cracking to the present day.
Acid-treated clays were replaced by synthetic silica-
alumina catalysts - first in pelleted, later in bead form.
Pore structure was adjusted for minimal diffusion lim-
itations. Resistance of the catalysts to abrasion and
erosion was important for moving bed operations. A
new set of parameters was required for fluid bed cata-
lysts (such as particle size distribution and burning
characteristics during regeneration).
 1.3 Development of Industrial Catalysis 39

Products ( t o fractionatorl

Cyclone vessel

Riser r e a c t o r

Cataly
Catalyst

'&-
Dispersion s t e a m
Feed injection

Figure 2. Schematic of FCC unit (UOP design).

D I L U T E PHASE

I I
I '
FLAPPER I I
VALVE I I
/ I

Figure 4. FCC regenerator.

STRIPPER Table 2. Catalytic cracking of gas oil.

RISER
/CONDUIT
BAFFLES Parameter SiOz- A1203 REHY

Conversion (%) 42.5 42.5

Cj, gasoline (vol %) 24.5 30.6
STRIPPED
SPENT
/ SiEAM F L U I DI 2 ED
CATALYST
Total Cq (VOI Yo)
Dry gas (wt YO)
9.4
6.5
8.2
4.1
CATALYST AND OIL Coke (wt %) 8.7 4.9
Gasoline octane No. F-I 85.6 90.1
Figure 3. FCC reactor

A major breakthrough occurred in 1964 at Mobil Oil
Co. with the finding that zeolites, and particularly zeo-
lite X and Y (faujasite) catalysts, showed tremendous
improvements in gasoline yield and octane number,
while coke deposition was reduced. An example is
given in Table 2, presenting a comparison of a silica-
alumina with rare earth containing Y zeolite (REHY)
for cracking a heavy gas oil at constant conversion. Re-
sults will, of course, vary with oil composition and con-
version. The zeolite is used with a silica-alumina matrix.
At the present time, almost all catalytic cracking
units in the world operate with zeolite catalysts. De-
pending on local refinery requirements, feedstock vari-
ation and government regulations, modifications of
zeolite containing catalysts are being used, which give
different yields and octane numbers. These differences
~

References see page 48

40 1 Introduction
Me\hyIcycloper’ane
Cyclohexene
41
Methylcyc’openfene
Jl
Cyclol77adiene Methylcyc openiodiene
”, Benzene
I
1
Figure 5. Reaction network for reforming of Ca hydrocarbons.
are relatively small, compared to the difference between
Si02-Al203 and REHY.
1.3.3.2 Catalytic Alkylation
Alkylation refers essentially to three reactions:
(i) The reaction of C3-C5 alkenes with isobutane to
produce C7-C9 isoalkanes of high octane number.
This process is catalyzed by strong acids, usually
H2SO4 or anhydrous HF.
(ii) The reaction of aromatic hydrocarbons with al-
kenes to produce alkyl aromatics. This process is
also acid-catalyzed.
(iii) The reaction of alkenes with alcohols, such as
methanol, to produce ethers, such as methyl-t-
butyl ether (MTBE), as high octane components
of gasoline. Relatively mild acids, such as ion ex-
change resins, are catalysts.
The alkene-isobutane reaction was invented by Ipatieff
and was rapidly commercialized during World War I1
in the USA to provide a very high octane number
blending stock with catalytically cracked gasoline for
aviation fuel. The process has been used to utilize
excess amounts of alkenes and has gained renewed
importance in the early 1990s as a component in “re-
formulated” gasoline, a hydrocarbon blend rich in iso-
meric alkanes that meets environmental restrictions.
The use of sulfuric or hydrofluoric acid, however, cre-
ates other environmental and toxic concerns. For al-
most 40 years, there has been a search for a solid non-
toxic catalyst for this reaction. No commercial process
using such a catalyst is in operation as of 1994.
Aromatics alkylation is important for ethylbenzene
production (a precursor of styrene) and for the de-
tergent industry. For many years, anhydrous alumi-
num chloride was the preferred catalyst for this re-
action until disposal problems made it undesirable. It
has been largely displaced by zeolites of the HZSM-5
type, which are strong enough solid acids and pose no
disposal problems.
4\
0-Hexane lsohexanes
$ n - Fexene
41
lsohexenes
1.3.3.3 Catalytic Dehydrogenation and Catalytic
Reforming
The demand for toluene as a precursor for TNT during
World War I1 was greater than could be supplied from
straight-run gasoline. A process for the catalytic de-
hydrogenation of methylcyclohexane contained in the
naphtha fraction of petroleum was instituted. This pro-
cess used molybdena catalysts, usually supported on
alumina or on bauxite. To avoid rapid deactivation of
the catalyst by coke deposits, the reaction was carried
out in the presence of external hydrogen (at about 15
bar). Reactors were mostly of the fixed bed type, but
one plant was constructed as a moving bed unit, the
first such unit to operate at elevated pressure.
It was soon recognized that it would be desirable
to complement the dehydrogenation reaction with an
acid-catalyzed reaction, which would permit isomer-
ization, cracking, and aromatization to proceed. Three
processes, Platforming, Catforming, and Houdryform-
ing, were almost simultaneously introduced in the
early 1950s, all of which employed platinum as a
metal dehydrogenation and alumina or silica-alumina
as acid catalysts. A network of hydrogenation-de-
hydrogenation and isomerization reactions (Fig. 5) al-
lowed conversion of Cs and C6 cycloalkanes to ar-
omatics and by product hydrogen. Catalysts doing this
job were called dual-functional. They formed the basis
for catalytic naphtha reforming, the second largest re-
finery process. Many catalyst and process improve-
ments and modifications have been introduced over the
years. The original processes operated at 30-40 bar
pressure in the presence of hydrogen and in a fixed
bed mode. Catalyst life was 0.5-3 years without re-
generation. It was obvious that, from a thermodynamic
point of view, lower pressure would improve the vol-
ume yield of reformate (Table 3). The introduction of
bimetallic catalysts by Chevron in 1967 accomplished
this, lowering the operating pressure by 14-17 bar. The
bimetallic catalysts contained rhenium, in addition to
platinum and acidic alumina. Alternatively, pressure
could be lowered in exchange for more frequent re-
generation. In a severe example of this, a moving cat-

Table 3. Yield-octane relationship for reforming (26-360 “C
naphtha at constant temperature.
+ O)a)
~~ ~ ~
Pressure (bar) Cs+ yield ( ~ 0 1 % ) Octane no. CS+(R
7 90 92
7 83 99
14 87 92
14 80 99
17 82 92
17 76 99
34 88 92
a) Research octane number in the absence of tetraethyl lead.
alyst bed operation was introduced. Since dehydro-
genation and reforming are strongly endothermic, a
reforming unit usually consists of at least three reactors
in series with interstage heaters. This complicates en-
gineering design for regenerative operation.
Reforming processes have been a major source of
hydrogen for other processes for many years. In the
early 1990s, with the introduction of “reformulated
gasoline”, there has been a reduction in the daily
volume of reformate, since benzene was considered
toxic and aromatics content of gasoline was re-
stricted. This in turn has led to a shortage of hydro-
gen in some refineries.
The alkylation of benzene with ethylene has been
mentioned earlier. Dehydrogenation of ethylbenzene to
styrene (a major monomer) has been, since 1951, a
catalytic process. Iron oxide catalysts containing Cr203
and KOH or K2C03 are used, and heat for the endo-
thermic reaction is supplied by superheated steam.
1.3.3.4 Hydrogenation and Hydrodesulfurization
Sulfide catalysts are probably the oldest class of cata-
lytic materials used in the production of liquid trans-
portation fuels. Developed after World War I for the
hydrogenation of coal tar liquids, the sulfide catalysts
are currently used in virtually every refinery in the
world (Fig. 6). They are employed to remove sulfur
and/or nitrogen compounds from hydrocarbons and to
hydrogenate petroleum, tar, or coal fuels. They may be
used before another catalytic process, such as reform-
ing or catalytic cracking, to hydrogenate aromatics to
cycloalkanes which are easier to crack, and to remove
sulfur compounds which may poison noble metal or
zeolite catalysts. They may also be used after a cata-
lytic reaction to finish the product and meet final
product specifications.
Modern sulfide catalysts are usually molybdenum-
or tungsten-based and supported on alumina, some
of which are promoted with cobalt and/or nickel.
This generation of sulfide catalysts has proved to
1.3 Development of Industrial Catalysis 41
be very successful in meeting past sulfur removal
from fuel goals. However, they may not be sufficient to
meet the constantly increasing specifications for sulfur
and nitrogen removal from hydrocarbon liquids and
for saturation of complex aromatics.
1.3.3.5 Hydrocracking
Hydrocracking of heavy petroleum fractions or residua
has gained increasing importance as more dependence
on heavier oils developed. Although tungsten and mo-
lybdenum oxides or sulfides supported on an acidic
compound, such as alumina or faujasite-type zeolites,
are the preferred catalysts for hydrocracking, precious
metals on these supports are also used. The catalyst
must have a proper balance of hydrogenation and
of acidic cracking function, which can be different for
different feedstocks.
Whereas most hydrocracking units operate in the
fixed- or trickle-bed mode, resid hydrocracking may
require a solid catalyst suspended in a heavy liquid,
using a slurry bed in several stages. Such a unit has
been built by Amoco.
Coal liquefaction by direct hydrogenation is de-
rived from the old Bergius process. Coal and heavy
tar hydrogenation differs from residual oil conversion
only in the fact that they are hydrogen poorer feed-
stocks and require more severe operating conditions.
The catalysts used are largely the same as in resid
hydrogenation.
Heavy oils and tars often contain organically bound
metals, particularly vanadium and nickel, which tend
to poison and also block the catalyst pores. Although
catalysts poisoned by coke deposits can be regenerated
by oxidation, conventional regeneration does not re-
move metals from the catalyst. Catalysts used in this
service must be physically robust and also relatively
inexpensive. Their ultimate disposal can create envi-
ronmental problems because of the toxic nature of
heavy metals.
There are several industrial catalytic hydrogenation
processes of small overall volume. Among them are:
the selective hydrogenation of alkenes and dialkenes
in pyrolysis gasoline to allow extraction of aromatics;
the selective hydrogenation of dialkenes in C3 and Cq
streams prior to polymerization, involving precious
metal catalysis in the liquid phase and the saturation of
fats.
1.3.3.6 Dehydrogenation
As mentioned earlier, the military and civilian require-
ments of World War I1 resulted in the rapid commer-
cialization of a number of catalytic processes. Among
References see page 48

42 1 Introduction
Oil f e e d
Hydrogen f e e d
1 Aqueous
solution
2-4 0
Dissolved
H, s
w
I
Recycle gas
High - p r e s s u r e
phase separator
Figure 6. Simplified diagram of a hydrodesulfurization process.
them were two dehydrogenation processes which, with
some modification, are still in use 50 years later. The
Houdry dehydrogenation process produced butenes
and butadiene for “Buna” rubber from n-butane, em-
ploying a chromia-alumina catalyst. There is appreci-
able coke deposition on the catalyst and swing reactors
are being used with alternate reactors being on stream
and on regeneration. Heat for the highly endothermic
reaction is provided by burning of the coke. Another
method using oxidative dehydrogenation over bismuth
molybdate or over perovskite catalysts was operated
for a while. It did not have to deal with temperature
swings. but had the disadvantage of not producing hy-
drogen, which instead was burned to water.
The Houdry dehydrogenation process has seen a re-
vival in the 1980s and 1990s, this time with isobutane
as feedstock and with the aim to provide isobutene for
alkylation with methanol in the production of MTBE,
used in as large as 5% by volume as gasoline additive.
British Petroleum and UOP have developed a propane
or butane dehydrogenation process using a supported
precious metal catalyst and a process design involving
a continuous moving bed. Commercial units have
come on stream in the early 1990s.
The need for a substitute for natural rubber during
World War I1 also stimulated a process to dehydro-
genate ethylbenzene to styrene. Alkali-promoted iron
catalysts were used, and this reaction, as well as the
butene to butadiene process, is carried out in the pres-
ence of steam to reduce polymerization of the desired
product. In recent years, a large polystyrene demand
has kept the process alive.
1.3.3.7 Isomerization
During World War 11, there developed a need to iso-
merize n-butane to isobutane for alkylation with C3
and Cq alkenes. This required a strong acid catalyst,
1.
1 Light p r o d u c t s
L
j
+
mu ?-
‘0
9W
.-vl
._
c a
4
‘0
W
.-
X
U
Oesulfurized
liquid product
and anhydrous aluminum chloride was used in the
presence of traces of water, either suspended in heavy
oils or supported on bauxite. Release of HC1 and high
corrosivity of the early process were reduced with in-
troduction of the UOP and Shell processes employing
Pt-alumina or Pt-mordenite catalysts, respectively. Iso-
merization of CS and Cg fractions has been practised,
although the equilibrium concentration of the desirable
highly branched isomers is relatively low at the kineti-
cally required temperatures.
Isomerization of alkyl aromatics and particularly of
xylenes became necessary to produce the para isomer,
a monomer for synthetic fibers. AlC13 was again the
catalyst in early operation to be replaced with ZSM-5
zeolite as a noncorrosive catalyst with no disposal
problems. Xylene isomerization gives an equilibrium
mixture of the isomers, which have to be separated to
recover the desired p-xylene. Recently, shape-selective
isomerization on ZSM-5 has permitted higher than
equilibrium yields of p-xylene.
1.3.3.8 Oxidation
The oxidation [8] of aromatics, for example of naph-
thalene to phthalic anhydride, is an old art. Xylene has
been commercially converted to phthalic anhydride
since 1946. Vanadium oxide is the preferred catalyst,
although molybdena catalysts have also been used. The
process was greatly improved when fluid bed technol-
ogy was introduced and permitted better temperature
control of the highly exothermic reaction. The same
technology is used for oxidation of benzene to maleic
acid.
Direct partial oxidation of lower molecular weight
alkenes, particularly ethylene and propylene, is carried
out in a variety of catalytic processes to produce ethyl-
ene oxide, glycol, and propylene oxide. Silver is the
almost exclusive catalyst for ethylene oxide production

and it is promoted with alkali and with small amounts
of halides. Propylene oxide is today manufactured in a
homogeneous reaction by oxidizing propylene in the
presence of soluble molybdenum compounds with t-
butyl hydroperoxide, which, in turn, is obtained by air
oxidation of isobutane. t-Butyl alcohol is a byproduct.
Propylene is also the feed for oxidation to acrolein
over cuprous oxide and later over bismuth molybdate
catalysts. Acrylonitrile, animportant monomer, is manu-
factured from propylene, ammonia, and air in a fluid
bed over catalysts such as uranium antimonate or tin-
antimony mixed oxides.
Partial oxidation of methane to synthesis gas
(CO + H2) is an alternative route to steam reforming,
but is not a catalytic process. In the early 1990s, ex-
periments for catalytic partial oxidation of methane
were described in the literature, but no commercial
process is yet operating.
Oxidation of HC1 to chlorine is important since all
chlorination reactions produce stoichiometric amounts
of HC1. The Deacon process, reacting HCl with oxygen
over copper chloride, was discovered as early as 1860.
There have been several attempts to operate large scale
“Deacon” plants using cupric chloride catalysts, but all
failed because of corrosion problems. The ‘Kelchlor’
process developed by the M. W. Kellogg Co. operated
successfully in a DuPont plant for many years in the
1970s and 1980s. It involves a homogeneous reaction
of air or oxygen with HC1, using nitrogen oxide cata-
lysts. The reaction is carried out in sulfuric acid, which
absorbs the water produced and thus reduces corro-
sivity and also drives the equilibrium to higher yields.
In another partially homogeneous reaction, ethylene
is oxidized to acetaldehyde (which, in turn, can be oxi-
dized to acetic acid) with palladium chloride in the
presence of water. Reduced Pd metal is oxidized back
to PdC12 with cupric chloride and the reduced cuprous
chloride is reoxidized with HCl and oxygen. This
Wacker process has operated in numerous plants, and
a similar reaction is used to make vinyl acetate.
Acetic acid, a large scale chemical, is produced by
oxidation of acetaldehyde and and, since 1977, largely
by the Monsanto process, involving reaction of CO
with methanol in a homogeneous process promoted by
a rhodium catalyst.
Hydroformylation, the reaction of alkenes with CO
and hydrogen, was discovered by Roelen in the early
1940s, and the resulting 0 x 0 process still produces
propionaldelyde from ethylene and butyraldehyde from
propylene.
1.3.3.9 Polymerization
The need for high octane gasoline during World War
I1 led to efforts to dimerize and trimerize ethylene
and propylene, and the development of the solid phos-
1.3 Development of Industrial Catalysis 43
phoric acid process (phosphoric acid on kieselguhr).
The previously mentioned monomers, butadiene, iso-
prene, and styrene, were copolymerized to synthetic
rubbers.
In the 1950s polyethylene and nylon and its de-
rivatives became the first products of the plastics and
synthetic fiber industry. The early processes were of the
high-pressure radical type. In 1955 Ziegler disclosed
a low pressure polyethylene process employing an
aluminum alkyl and titanium tetrachloride catalyst.
Natta, at about the same time, produced polypropylene
by catalytic polymerization. While the original Ziegler
catalyst was used in solution, later catalysts employed
were solids, such as supported chromia or molybdena,
used with promoters. More recently, metallocene cata-
lysts permit highly selective polymerizations.
The polymer industry uses a wide spectrum of cata-
lysts, many of which are proprietary. An early and very
successful catalyst was Dabco (triethylene diamine),
introduced in 1959 for the production of urethane
foams from isocyanates and alcohols. Dabco is still in
use today. A more recent example of catalytically pro-
duced polymer chains are rigid polymers of great
strength and light weight. Kevlar has a tensile strength
higher than that of a steel fiber of the same dimensions,
weighing only one-fifth as much. It is produced em-
ploying eight catalytic steps, using a combination of
acid, hydrogenation, and oxidation catalysts. Polymers
such as Kevlar are likely to replace steel as a structural
material in many applications.
1.3.3.10 Zeolites
Although the use of zeolites as catalysts has been dis-
cussed in several of the preceding sections, the im-
portance of zeolite chemistry to catalytic processing
and engineering deserves special attention as one of the
major breakthroughs of the 1960s. The first announce-
ment of the ability of faujasites to catalyze an acid-
dependent reaction (isomerization) was made by Rabo
of Union Carbide Corp. in 1960. This was followed the
introduction of zeolite cracking catalysts by Mobil
Oil Co. Prior to about 1960 there had been a general
belief that crystalline materials possessed poor catalytic
properties and there also were relatively few natural
zeolites available. The synthesis and the blending of
zeolites with clays and aluminas provided the stepping
stones required for large-scale applications.
Early processes used synthetic X- and Y-zeolites and,
in a few cases, natural erionite. In the mid-l960s, the
synthesis of ZSM-5, a medium pore-size zeolite, was
described by Mobil Oil Co. workers. The pore size of
ZSM-5 permitted controlled molecular diffusion or
shape selectivity to supplement acidic properties. This
material then found broad applications, for example, in
methanol conversion to gasoline, aromatics processing,
References see page 48
44 I Introduction

and in cracking. Numerous other zeolites have since Table 4. Average emissions from gasoline cars [12] as grams per
then been synthesized and found use in a wide variety mile over Environmental Protection Agency (EPA) federal test
of reactions. Incorporation of metal functions into the procedure.
pores of ZSM-5 resulted in new hydrocracking and in
Hydrocarbons
new isomerization catalysts. Changes in the silica: alu-
mina ratio of zeolites permitted control of the acid Year of model CO NO, Exhaust Evaporative Total
function.
Zero miles
1966 78 3.4 7.2 4.5 11.7
1986 I 0.5 0.2 0.6 0.8
1.3.4 The Period from 1965 to 1990 Reduction ((Xu) 98 84 97 88 93
50 000 mi1e.r
Although some major specific catalytic process accom- 1966 89 3.4 8.1 4.5 12.6
plishments are discussed below, this time period is 1986 13 1.3 1.0 0.6 1.6
Reduction (%I) 85 62 88 87 87
characterized by a large number of new processes and
process modifications relating to smaller-size chemical
processes and to the production of fine chemicals.
Three recent review papers [l 11 have analyzed this The concept of shape selectivity, while originating in
trend in the USA, Japan, and Europe. In Europe alone, the petroleum refining industry, is finding increasing
there have been over 100 new developments between applications in the synthesis of polymers and of phar-
1980 and 1990, which obviously cannot be discussed in maceutical, agricultural, and enzymatic syntheses.
detail within the present framework, but which are of
undoubted value for obtaining specific products. Lists
of these developments are presented in the reference 1.3.4.2 Environmental Catalysis
papers.
The first impact of environmental concerns [9] occurred
in the early 1960s with the recognition that automo-
1.3.4.1 Shape Selectivity bile emissions provided a serious health hazard as did
the use of tetraethyl lead as a gasoline additive. While
The shape selectivity of zeolites was discovered in the there had been scientific work on the conversion of
mid-1960s and has since found many applications. Its carbon monoxide and of nitrogen oxides, large-scale
importance continues to grow. It is based on the con- application was expected to be expensive and not com-
cept that the pores of certain zeolites in which chemical mercially attractive. It was due to government regu-
reactions are catalyzed are of the same order of size as lations in the United States, which set standards for
the molecular dimensions of the reacting molecules. CO and NO, emissions from automobiles, that a ma-
Depending on the size and shape of reacting molecules, jor effort was made by various automobile and petro-
they may or may not diffuse into the zeolite, and some leum companies operating individually and in com-
potential reaction products can, for steric reasons, not bination to come up with a catalytic technology to
form in the zeolite pores. Early applications of shape- meet the required standards. The success of this effort
selective reactions were Selectoforming and M-form- is shown in Table 4. The history of the development of
ing, which permitted the selective cracking of normal automotive emission control catalysts is an interesting
or of slightly branched alkanes of low octane number, one. It is also interesting to note that early work under
but not of more highly branched, high octane number the pressure of compliance with regulation deadlines
alkanes. The shape-selective isomerization of xylenes was concentrated on nonprecious metal catalysts which
has already been mentioned. Distillate dewaxing with could give about 90% reduction of CO content but
either ZSM-5 or mordenite catalysts became commer- not the > 98% reduction required by law. Similarly,
cial in the early 1970s. Toluene disproportionation to the requirement that the catalysts were to perform for
benzene and xylene is being practised on an appreci- at least 50000 miles of a car’s operation shifted the
able scale. In 1982, the methanol to gasoline (MTG) commercial development to precious metal catalysts
process was commercialized in New Zealand with the even though they were appreciably more expensive
construction of a 750 000 t per annum plant to convert than nonprecious metals or metal oxides.
natural gas to gasoline via methanol. The process uses Increased consciousness by the public of the health
the shape-selective properties of ZSM-5. hazards of many toxic materials emitted into the air
In 1982, crystalline aluminophosphates were in- and into the water during chemical manufacturing and
troduced, which exhibit shape selectivity similar to other processing techniques, as well as the problem of
zeolites, but at lower acidity, and are used in a number disposal of nonbiodegradable products, has led to a
of fine-chemicals syntheses. rapid increase in efforts to ameliorate toxic effects and

+ 1m-
u
W
g
L
80-
A
c
.d
.- 60-
c
c
W
.- 40-
2
5
g 20-
u
environmental damage. Economics will shift to enviro-
economics as a driving force. Most of the processes
developed thus far involve catalysis. There are three
driving forces:
(i) society’s increasing environmental awareness,
concern and expectations;
(ii) corporate responsiveness and action;
(iii) legislation and regulation.
Consumer demands focus on clean air, water, and land
to safeguard the biosphere, and these concerns were
clearly expressed in the International Conference in
Rio de Janeiro in June 1992. Among the items that
seem to predominate in consumers’ minds are the de-
struction of the ozone layer, recycling programs, envi-
ronmentally friendly packaging, and product labeling
that reflects environmental considerations. The politi-
cal pressure exercised by these concerns is clearly dem-
onstrated by the Green Parties in Europe and the
Sierra Club in the United States. The cost-risk factor
has to be considered but will not be easy to explain to
the public. Environmental costs will increasingly be-
come an integral part of process economics.
Automobile exhaust emissions are determined by
fuel composition, engine design, and operation, and
by catalytic converters located in the exhaust system.
Exhaust gases from spark-ignition automobiles contain
harmful CO, NO, and hydrocarbons (incompletely
conibusted fuel constituents). Automotive converters
using oxidation catalysts were introduced in the US in
1975, converting CO and unreacted hydrocarbons to
C02 and H20. Three-way automotive converters cata-
lyze these oxidative reactions and also simultaneously
reduce nitrogen oxides (Fig. 7). They have been widely
adopted to meet federal NO, standards, beginning in
1981.
1.3 Development of Industrial Catalysis 45
v
k
Catalysts for automotive emissions are now well de-
veloped and in general are capable of meeting presently
mandated US standards for CO, NO, and HC emis-
sions from auto exhaust gases. New cars today emit
only a few percent of the exhaust pollutants in cars
built 25 years ago, before emissions control. Modem
three-way converter catalysts control CO and hydro-
carbons by oxidation. Simultaneously, NO, is removed
by reduction, according to:
2CO+ 2 N 0 + 2C02 + N2
HC + 2 N 0 C02 + iH2 + N2
-+
The generic converter catalyst structure was invented
in the early 1950s. Catalytic materials in automobile
converters are generally supported on a ceramic hon-
eycomb monolith. The honeycomb, made of cordierite
(2Mg0.2A1203 . 5Si02), contains 30-60 square chan-
nels per square centimeter, is coated with an activated
alumina layer called the washcoat. Catalytic Pt, Pd,
and Rh are highly dispersed on the high-area washcoat.
Rhodium metal displays high activity for the removal
of CO, HCs, and NO,, and is particularly important
for the reduction of NO,. Fresh Pd is more active than
Pt in most cases for CO, NO,, and hydrocarbon con-
version, but is more susceptible to poisons such as lead
and sulfur, and sinters rather easily in a reducing at-
mosphere. Ceria may be added at a loading of 2-
30 wt % to the washcoat of three-way catalysts to store
oxygen, to promote the water-gas shift reaction and to
stabilize the noble metals against thermal damage.
Nickel can be added to scavenge H2S.
Automobile pollution control necessarily interrelates
fuels composition, engine combustion performance and
catalytic converter remediation systems. Figure 7 de-
scribes the effectiveness of NO,, CO, and HC control
by the Pt/Rh catalyst as a function of the optimal
References see page 48
46 1 Introduction
Elimination of toxic reagents and intermediates
- methylene diphenyl diisocyanate (MDI)
- gasoline alkylation
- methyl methacrylate
- elimination of chlorofluorohydrocarbons
Elimination of toxic waste streams and b y - p r o d u c t s
- resorcinol
- acetaldehyde
- bioactive molecules
Elimination of toxic emissions
- catalytic combustion (NO1
Figure 8. Some specific examples of primary prevention
stoichiometric air: fuel ratio (equivalence ratio A). In
addition, fuel economy and maximum power vary
considerably with air: fuel ratio. Major steps have
been taken by auto manufacturers to accommodate
the requirements for efficient operation of converter
catalysts.
Although it is estimated that 96% of the exhaust H C
and CO and 76% of the NO, can be eliminated by the
use of catalytic converters, there are needs and oppor-
tunities for improvements. Three-way catalysts operate
near the stoichiometric air: fuel combustion ratio and
exhaust temperatures are normally 400-600 "C. How-
ever, today's catalytic converters do not reach mini-
mum operating temperatures until about 100 s after
the cold start of an automobile. During this warm-up
period, more than 50% of hydrocarbons and C O emis-
sions pass unreacted. The problem of cold-start pollu-
tion must be addressed by seeking to provide special
procedures and catalytic apparatus.
There are two major modes of pollution control:
(i) primary prevention, such as preemptive pollution
or avoidance;
(ii) secondary prevention, such as waste or pollution
cleanup.
Primary prevention eliminates toxic byproducts by
increased selectivity. An example of this is the Dow
Chemical cumene process, which permits the highly
selective alkylation of benzene with propylene using a
three-dimensional dealuminated mordenite (3-DDM)
catalyst. Cumene is preferentially formed and the pro-
duction of disubstituted benzene is inhibited.
Some other examples are given in Figs 8 and 9.
Other typical cases are removal of phosgene from the
production of methylenediphenyldiisocyanate;the re-
moval of hydrofluoric acid from alkene-alkane alky-
lation; avoidance of sulfuric acid use in the production
of methyl methacrylate; finding substitutes for chloro-
fluorohydrocarbons as propellants and refrigerants;
avoiding production of AlC13 sludges by using ZSM-5
zeolite for the production of ethylbenzene from ben-
Current technology
O C N G Z H 2 G NCO
Emerging technology
Figure 9. Elimination of toxic reagents and intermediates: MDI.
zene and ethylene. In all these cases, catalytic steps
have been found or are being investigated to play a
major role. A particularly interesting case is catalytic
alkylation. The requirement for clean-burning gasoline
of relatively little volatility has resulted in the need to
reduce the aromatics and butane content of gasoline.
To maintain octane requirement, additional alkylation
capacity is required. Replacement of hydrofluoric and
sulfuric acid, with solid acid catalysts would reduce
hazard and operating costs. Clean-burning gasoline
also will increasingly demand inclusions of oxygenated
hydrocarbons in the gasoline such as MTBE. MTBE in
turn demands both more methanol production and
more isobutene than are currently available. This has
led to emphasis on improved or new routes of catalytic
dehydrogenation of butanes, as mentioned earlier.
A Elimination of Toxic Waste Streams and Byproducts
A few examples here are elimination of sulfonating
agents in the production of resorcinol and substituting
aromatic alkylation and oxidation; the Wacker process
for the production of acetaldehyde gives aqueous sol-
utions of, as well as volatile, chloroorganics. Other
catalysts and processes can be substituted (often at
some cost) to minimize or eliminate the undesirable
byproducts.
B Elimination of Toxic Emissions
Catalytic combustion appears to be one way of avoid-
ing the production of large quantities of NO,y. NO,y is
formed in a kinetically controlled reaction in the high
temperature of the flame (Fig. lo). At temperatures
below 1600 "C, relatively little NO, is formed, whereas
at 1700 "C, almost three times as much and, at 1800 "C,
eight times as much NO, is formed. Gas turbine power
generation is becoming increasingly attractive, particu-
larly in areas where natural gas is available, which, of
course, is an environmentally more attractive material
than coal. Turbine efficiency increases with combus-
tion temperature. By catalytic combustion, high tem-

80
100 ms reaction time
60 -
40 -
20 -
0 --
1300 1400 1500 1600 1700 1800
Temperature ("C)
Figure 10. NO, formation.
perature homogeneous flames are eliminated without
losing efficiency. In this connection, it is interesting
to note a combination of catalytic processes which
has been proposed to increase turbine efficiency. The
chemically recuperated gas turbine involves both cata-
lytic combustion and catalytic steam reforming.
Examples of secondary prevention are (i) automotive
emission control. (ii) NO, control from stationary ex-
hausts; and (iii) waste stream purification. The contri-
bution of catalysis to the decrease in air pollution can
be shown for the period 1966-1986 by a 90% reduction
in carbon monoxide emissions, 30% reduction in SO2
emissions, 70% reduction in NO emissions, and 10%
reduction in surface ozone production. However, there
are novel ways of NO, decomposition or reduction,
such as the large-scale use of the reaction of ammonia
with NO, over titania/vanadia catalysts, and proposed
novel methods of NO decomposition, using, for in-
stance, copper/ZSM-5 catalysts.
1.3.4.3 Other Industrial Applications of Catalysis
Since the 1950s, there have been numerous cases of
catalytic processing in the fine chemicals area, which
involved high-price, relatively low-volume products
[lo]. Since the mid-l970s, the emphasis on catalytic
steps in the synthesis of pharmaceuticals and of agro-
chemicals, as well as food additives, has sharply in-
creased. Major progress in novel catalytic reactions
has, to some extent, shifted from hydrocarbon con-
versions and fuels to environmental control and to or-
ganic (and, to a lesser extent, inorganic) chemical syn-
thesis. Important among biochemicals is the chirality,
i.e. the handedness or optical rotation of the molecule,
which determines its bioactivity. One of the chiral iso-
mers can be very effective, while the other may be in-
1.3 Development of Industrial Catalysis 47
0 CH:Ph CHzPh 0
Figure 11. Physiological effects of chirality: aspartame.
Current Technology to R,S-Ibuprofen
A+-<E-EL* w~
Emerging Technology to S-Ibuprofen
3
Pd ~ C O ~ H
R,S-Ibuprofen
h
S-Ibuprofen
CHJOH
\\
/ Enzyme R,S- Ster
R-E,ster 4
Figure 12. Technology to ( R .S)- and to @)-ibuprofen.
effective or even harmful. A typical example is the case
of thalidomide, which is a sedative for pregnant women
in its ( R )form, but is a potent teratogen in its ( S )form,
causing children to be born with deformed limbs. An-
other example is aspartame, which, in one isomeric
form is sweet, in the other is bitter (Fig. 11). It is ob-
vious that isomer separation and, if possible, selective
synthesis of the desired isomer are essential. Ibuprofen,
a widely-used pain-relieving drug, is more effective in
its ( S ) form than in ( R , S )form. Figure 12 shows the
common synthesis to the ( R , S ) form involving two
catalytic steps and the improved technology, also in-
volving catalysis. In the latter case, it is shown how
( R ,S)-ibuprofen can be converted to (S)-ibuprofen by
means of enzyme catalysis. The US market for chiral
compounds is expected to grow from $0.5 billion in
1990 to $3 billion in 2000.
All life processes depend on the biological catalysts
called enzymes. An enzyme is essentially a protein
molecule, although it may have other atoms or mole-
cules attached. A protein molecule is a long chain,
made up of hundreds or thousands of amino acid mol-
ecules, assembled in a very specific order. When dis-
solved in water, this chain naturally kinks and knots
up. The sequence of amino acids making up the protein
determines the shape that the protein knots itself into,
and it is this shape that allows the protein to catalyze
reactions. The molecules that participate in the re-
action fit into crevices in the protein like a key in a
lock, causing shape selectivity. Once inside the crevice,
References see page 48
48 1 Introduction
the molecules are held in just the right relative ori-
entation for the reaction to proceed.
Enzymes have been industrially used for centuries, as
in wine-making. While this use was originally without
any scientific understanding, a large industry has de-
veloped since the 1970s which uses these enzymes on
the basis of scientific knowledge. The conversion of
cornstarch into corn syrup by the enzymes alpha- and
gluco-amylase generates corn sweeteners for soft drinks
and other uses. The enzymatic conversion of acryloni-
trile to acrylamide has been commercialized in Japan.
The stereospecificity of enzyme-catalyzed reactions has
been used to advantage in polymer synthesis as well.
Workers at ICI have developed a combined enzymatic
chemical process for the synthesis of polyphenylene
from benzene.
A major difference between enzymatic and hetero-
geneous catalytic approaches is the fact that enzymes
work well in dilute aqueous solutions and at moderate
temperature, pressure and pH. The reactions catalyzed
by these systems are typically environmentally friendly
in that few byproducts or waste products are generated.
The catalysts and the materials that they synthesize are,
as a rule, biodegradable and therefore do not persist in
the environment. The reactions are typically selective
with extremely high yields, and enzymes can be used to
catalyze a whole sequence of reactions in a single re-
actor, resulting in vastly improved overall yields with
high potential specificity.
Considerable work is being done and has been done
on immobilized enzymes. This refers to a technology in
which the active part of the protein molecule is chemi-
cally attached to an inert, inorganic substance such as a
silica, alumina or a zeolite. This then permits these im-
mobilized enzymes to be used like heterogeneous cata-
lysts in reactors in which the liquid or gaseous substrate
is passed over a solid material. Advantages of such
techniques lie in easier separation of the catalysts from
the reactants.
References
1. S. A. Topham in Catalysis, Science and Technology, (Eds: J.
R. Anderson, M. Boudart), Springer-Verlag, Berlin, 1985,
Vol. 7, pp. 1-50.
2. F. Aftalion, A History of the International Chemical Industry,
University of Pennsylvania Press, Philadelphia, 1991.
3. H. Heinemann in Catalysis, Science and Technology, (Eds:
J. R. Anderson, M. Bondart), Springer-Verlag, Berlin, 1981,
Vol. 1, pp. 1-41,
4. J. A. Rabo in New Frontiers in Catalysis: Proceedings of the
10th International Congress on Catalysis, (Eds: L. Guczi,
F. Solymosi, P. Tetenyi), Elsevier, Amsterdam, 1993, Vol. A,
pp. 1-30.
5 . K. C. Taylor in Catalysis, Science and Technology, (Eds: J.
R. Anderson, M. Boudart), Springer-Verlag, Berlin, 1984,
Vol. 5. pp. 119-170.
6. A. B. Stiles, Catal-vst Manufacture, Marcel Dekker, New
York, USA, 1983.
7. M. V. Twigg (Ed.), Catalyst Handbook, 2nd edn, Wolfe
Publishing, London, 1989.
8. G. Chinchen, P. Davies, R. J. Sampson in Catalysis, Sci-
ence and Technology, (Eds: J. R. Anderson, M. Boudart),
Springer-Verlag, Berlin, 1987, Vol. 8, pp. 1-67.
9. J. A. Cusumano, Catalysis - The Key to Environmentally
Compatible Chemical Technologies, Am. Assoc. Ach. of Sci-
ence, 1992.
10. J. Haggin, Chem. Eng. News 1994, 72, 22-25.
11. J. N. Armor, Appl. Catal. 1991, 78, 141; E. Min, P. Zhou,
Appl. Catal. A 1993, 95, 1; A. Chauvel, B. Delmon, W. F.
Hoelderich, Appl. Catal. A 1994, 115, 173.
12. R. Ecklund, G. A. Mills, Chem. Tech. 1989, 19, 552.