Available via license: CC BY-NC-ND 3.0
Content may be subject to copyright.
ISSN:1369 7021 © Elsevier Ltd 2011
APRIL 2011 | VOLUME 14 | NUMBER 4
134
Precious metal magic:
catalytic wizardry
New eras have dawned and the course of history has changed,
simply due to mastery of the elements. The Iron Age, Bronze Age,
and Nuclear Age are obvious examples, but the rise of industry
through synthetic manipulation of the elements and more recent
advances in medicine and nanotechnology bring us to yet another
source point in history, brimming with potential. Surprisingly, by
clever utilization of their chemical “inertness,” the chemistry of
precious metals has emerged as one of the exciting frontiers for
advanced understanding and applications, particularly in catalysis.
From then to now
Shiny, malleable, and resistant to corrosion, gold has been the most
coveted precious metal since the time of the Pharaohs. In addition to
its well-known uses in currency and jewelry, metalworkers found that
the properties of gold were more than meets the eye. For example, gold
alloy nanoparticles (defined as particles from 1 – 100nm in size) in the
glass of the famous Lycurgus cup, c.a. 4
th
century AD, cause the color to
change from green for reflected light (Fig. 1a) to red for transmitted light
(Fig. 1b)
1
. Nanoparticles of another precious metal, silver, are responsible
Magic: “Influencing the course of events by using mysterious or
supernatural forces.” Precious metals are alluring and magical because
of their inactivity toward chemical reactions; they are extremely stable
and hence are also termed “noble metals.” During the industrial revolution
mankind realized that noble metals have the power to influence the
course of chemical events – through catalysis. A catalyst is defined as
a substance that facilitates a chemical transformation without itself
being consumed in the process; this power has a mysterious, almost
magic-wand character. In this review we explore precious metal catalysis
through the wide-angle lens of historical development and the atomic scale
microscope of recent discoveries. Fundamental understanding of underlying
mechanisms for catalytic oxidation processes reveals the magic and
transforms the use of noble metals from instruments of adornment, trade
and Edisonian industrialization, to key players in a new era of catalysis by
design with potential for environmentally benign chemical processing.
Cassandra G. Freyschlag and Robert J. Madix*
School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA
*E-mail: rmadix@seas.harvard.edu
MT14_4p134-143.indd 134 21/03/2011 10:58:50
Precious metal magic: catalytic wizardry REVIEW
APRIL 2011 | VOLUME 14 | NUMBER 4
135
for the gold luster of medieval middle-eastern and renaissance Italian
pottery
2
. Gold nanoparticles lend red color to medieval stained glass, and
gold has even been consumed for its alleged healing effects, the earliest
occurrence being in ancient China, c.a. 2500 BC
3
.
First player: platinum
Despite being one of the last noble metals to be manipulated (due to
its higher melting point), platinum was nevertheless one of the first
to be used for catalysis. Platinum has been a common catalyst since
the 1820s, when Dobereiner invented a small “tinderbox”
4
. In this
pre-match device, hydrogen was generated in situ and reacted with
oxygen over a small supported platinum catalyst to yield a flame. In
1831, Peregrin Phillips patented the use of platinum as a catalyst for
production of sulfuric acid via the oxidation of SO
2
in the “contact
process”
5
. This innovation replaced the “lead chamber process” and
enabled economical production of more concentrated sulfuric acid,
which was critical for the growing chemical industry
6
. However, the
platinum catalysts suffered from poisoning and were replaced in the
early 1900s by vanadium oxides with alkali metal oxide promoters,
which are now the standard catalyst for the contact process
7
.
With the advent of the internal combustion engine and the need
for refining crude oil for fuel in massive quantities, catalytic cracking
(the breakdown of heavy hydrocarbons to readily combustible
molecules) using metal oxides became widespread. Platinum
nanoparticles supported on alumina became the catalytic material of
choice for transforming alkanes into branched alkenes and aromatics
– a process known as “catalytic reforming”
4
. This technological
advance provided cheap transportation that has reshaped cultures
throughout technologically advanced societies. The platinum enables
dehydrogenation of the larger organic species and the acidic alumina
on which it rests facilitates isomerization to form the higher octane
organic molecules. Adding either rhenium or iridium to the platinum
enhances activity, because the added element breaks the C-C bonds
in carbonaceous deposits, which would otherwise poison the platinum
surface
4
. As platinum was used in this fashion to increase the octane
rating of fuel entering the engine, it would also be used as the basis
of the catalytic converter to clean up the gases exiting the engine
– a veritable “jack of all trades.” This use represents a first step for
catalysts for environmental purposes.
The catalytic converter is one of the best-known uses of catalysis
in modern society. Due to incomplete combustion of the fuel, its
sulfur content, and the equilibrium between oxygen and nitrogen
attained at the high temperature of the internal combustion engine,
the exhaust includes a toxic mixture of CO, hydrocarbons (C
x
H
x
), NO
x
,
and SO
x
, along with H
2
O and CO
2
. The sulfur oxides are eliminated
by desulfurization of the fuel in its manufacture. However, removal
of the remaining gases posed a seemingly insurmountable problem.
Fortunately, the CO and C
x
H
x
can be oxidized and NO
x
simultaneously
reduced using a so-called “three way” catalyst, yielding CO
2
, H
2
O, H
2
,
and N
2
8
. Bonds are broken on the noble metals, and then the atoms
recombine to form CO
2
, H
2
, N
2
, and H
2
O (eq. 1 – 4)
8
.
NO → N + O (Rh) (1)
O
2
→ O + O (Rh/Pd/Pt) (2)
CH → C + H (Pd/Pt) (3)
CO + O → CO
2
(Pd/Pt) (4)
The standard three-way catalyst consists of noble metals Rh and
Pt and/or Pd on an oxide or mixed-oxide monolith support. The Rh
acts to promote NO dissociation, leading to N
2
formation, while Pt
and Pd combust CO and C
x
H
x
to CO
2
and water
8
. To the layman
this technological advance is apparently magical. It is a triumph of
the science of catalysis. One remaining challenge is the “cold-start”
problem: controlling the unwanted emissions produced before the
catalytic converter reaches operational temperatures
8
.
Silver comes in second
Because we live in an oxygen-rich atmosphere, the conversion of natural
gas and other hydrocarbon sources to useful materials via chemical
intermediates produced by partial oxidation was a “natural” pursuit.
Partial oxidation and epoxidation catalysis using metallic silver emerged
in the early 1900s. Silver catalysts are used industrially in the oxidation
of methanol to formaldehyde and in the epoxidation of ethylene
4
.
Oxidation of ethylene is a good example of the importance of
catalyst selectivity, as a different catalyst yields a dramatically different
major product (Fig. 2)
4
. When ethylene is oxidized over platinum, since
platinum facilely cleaves C-H bonds, full combustion occurs, and the
products are CO
2
and H
2
O. Palladium salts afford the partial oxidation
product, acetaldehyde. Silver, however, uniquely yields ethylene oxide,
a precursor to polyester materials and ethylene glycol for antifreeze.
The origin of this singular capacity of metallic silver lies in the inability
of silver itself to break the C-H bond in ethylene, the relatively weak
binding of adsorbed atomic oxygen to the silver surface and the ability
of this surface-bound oxygen to attach itself to the C-C double bond in
ethylene – a unique confluence of circumstances. Moreover, this balance
is so delicate that propylene (adding one more carbon to ethylene)
Fig. 1 The Lycurgus cup is (a) green when viewed with reflected light and (b)
red when light is transmitted through it, due to precious metal (Au/Ag alloy)
nanoparticles within the glass. © Trustees of the British Museum.
(b)(a)
MT14_4p134-143.indd 135 21/03/2011 10:58:52
REVIEW Precious metal magic: catalytic wizardry
APRIL 2011 | VOLUME 14 | NUMBER 4
136
undergoes complete combustion
9
. Modern model studies under well
controlled conditions employing the methods of surface physics in
ultra high vacuum (UHV) studies have led to an understanding of the
kinetics and mechanisms of the reactions on silver surfaces
10,11
. Both
the methanol oxidation
12
and ethylene epoxidation
13
systems have
been microkinetically modeled, as well as that of the active oxygen
on silver
14
, aiding in understanding the molecular-level processes in
industrial conditions. As research has continued, the range of catalytic
reactions that noble metals facilitate has increased dramatically,
particularly with metallic gold, the relative newcomer to catalysis.
Gold is now green: a more current
environmental currency
The recent interest in gold as a catalyst was ignited by the discovery of
Haruta et al. that gold nanoparticles on reducible oxides were effective
catalysts for CO oxidation at low temperatures
15
. While gold, silver,
and platinum will presumably always be used to make jewelry, their
potential as green catalytic agents to redirect chemical synthesis so
as to significantly reduce harmful environmental impact is immense.
The chemical industry produces large amounts of waste and consumes
about 20 % of the energy in the U.S. industrial sector
16
, and the
average chemical company emits more CO
2
than companies in the
other six S&P sectors combined
17
. The demand for green catalysis –
utilizing catalysts that reduce the environmental impact of chemical
processes − requires selective, stable catalysts that function in a
benign medium and can be easily recovered
18
. It has been suggested
that low-temperature reactions on noble metals – gold in particular
– may be part of the solution
19,20
. Here, we will give an overview
of the catalytic applications and strengths of the noble metals, with
emphasis on heterogeneous catalysis. Then we will focus particularly
on fundamental understanding of the metals that have captured man’s
attention from the beginning: gold, and its fellow coinage metal, silver.
Magicians with different tricks
The catalytic ‘bag of tricks’ is revealed by modern research to be
somewhat different for each precious metal, and they exhibit distinct
patterns that can be used to predict and design new catalytic systems.
Because such a myriad of transformations is desired in the field
of catalysis, this process requires varying levels of selectivity and
specificity of the catalyst to cleave bonds of one type and reform
others. One useful way to think about the differences between noble
metals is to classify them in terms of bond-breaking capability, as
demonstrated for the clean metals in Fig. 3. This list is meant to be
representative, not exhaustive.
The supposedly ‘inert’ gold catalyzes a surprising variety of reactions
21
,
including hydrogenation, selective partial oxidation reactions
22
and
Fig. 2 Ethylene oxidation over Ag, Pt, and with PdCl
2
has extremely different
selectivities, resulting in the major products being ethylene oxide, CO
2
and
H
2
O, and acetaldehyde, respectively. Based on figure appearing in
4
.
Fig.3 The precious metals used in catalysis can be "ranked" by their bond-breaking abilities. The capacity of each metal to break specific chemical bonds is
indicated. The ability of clean gold to activate O-O bonds is a subject of current debate, particularly for gold nanoparticles. Note the increasing "inertness" in
progressing from left to right in this series.
MT14_4p134-143.indd 136 21/03/2011 10:58:54
Precious metal magic: catalytic wizardry REVIEW
APRIL 2011 | VOLUME 14 | NUMBER 4
137
nucleophilic addition to pi systems. As depicted in Table 1, selective
oxidation of alcohols by gold leads to formation of aldehydes and ketones,
as well as esters
23,24
; and selective oxidation of amines yields nitriles and
aldehydes
25
. Gold nano-clusters catalyze the intra-molecular cyclization
of amines and alkenes
26
. On metallic gold, coupling reactions between
alcohols and aldehydes (either externally introduced or made in situ)
result in ester formation
23,27,28
. In addition to epoxidation of propylene
29
and styrene
30
, gold also facilitates the aziridination of olefins by the
direct addition of adsorbed NH
31
. The capabilities of gold for selective
oxidation processes using O
2
as the oxidant make it a green competitor
for production of commodity chemicals such as sodium glycolate, sodium
lactate, and sodium gluconate from cheap starting reagents, avoiding the
use of chlorinated reagents and hydrogen cyanide
22
. Gold outperforms
both palladium and platinum catalysts for diol partial oxidation, with
98 % selectivity and 1000 – 35 000 TOF (h
-1
) for the oxidation of
ethylene glycol to ethyl glycolate
22
.
De-mystifying the magic
Just as magic tricks captivate and intrigue audiences who desire to
understand how the illusions are performed, so catalysis begs the
questions ‘how?’ and ‘why?’ from the chemist. Since the mid-late
20
th
century, the development of new surface science and materials
research techniques has enabled researchers to provide fundamental
answers to those questions and to gain greater understanding not only
of noble metals as materials themselves and what they do, but how
they do it. This understanding can lead to the design of new tricks with
ever more impressive results.
One of the most powerful tools for dissecting this chemical magic is
that of temperature programmed reaction spectroscopy (TPRS). In this
technique, the desired reactants are placed on a metal crystal surface
of controlled structure and composition in ultra high vacuum (UHV)
at a temperature sufficiently low that initially no reactions take place.
Then, with the crystal in front of a multichannel mass spectrometer,
the crystal is heated linearly with time, leading to the activation of
various reaction channels available to the reactants. Products evolve
from the surface at temperatures characteristic of the activation
energies of their formation
32,33
. Multiple products can be monitored
at the same time, enabling the entire reaction to be deciphered.
In combination with isotopic labeling, this method reveals step by
step the processes of bond rupture and reformation of the catalytic
transformation. It can be complemented by spectroscopic techniques
such as vibrational
34
and photoelectron spectroscopies
35
, in order to
identify intermediates on the surface and follow the course of the
molecular rearrangements during the reaction. In some cases reactants,
products and reaction intermediates can be directly imaged using
scanning tunneling microscopy (STM)
35,36
.
Oxygen: the ‘breath’ of precious metal catalysts
for synthesis
Many important precious metal catalytic processes are oxidative, so
oxygen, both atomic and diatomic, has been the subject of much
fruitful academic debate and discovery in the surface science of
catalysis. On Pd, Pt, and Ag molecular oxygen readily dissociates to
form surface-bound atomic oxygen. When supplied from another
source, such as a metal oxide or a stronger oxidant, such as ozone,
atomic oxygen binds to Au surfaces. On specific single crystal surfaces
of these metals, this oxygen assumes different structures and has
different binding strengths (Fig. 4).
The oxygen/noble metal interaction is one of great complexity and
importance, as the oxidative catalysis of noble metals requires binding
of oxygen to the surface. For an in-depth discussion of the theory of
adsorption of oxygen on noble metals, see Hammer and Norskov
37
.
On Pt and Pd dioxygen dissociates and binds readily, forming well-
ordered structures
38-40
. Since these metals readily activate C-H bonds
in hydrocarbons, this oxygen can react with hydrocarbon fragments
on the surface, usually leading to complete combustion. Oxidation
occurs readily when the oxygen surface concentration is less than
one monolayer, but enhanced reactivity has also been reported for a
thin surface oxide of palladium
41
. On Au and Ag oxygen is bound less
strongly. Oxygen dissociation occurs on both Ag(110) and Ag(111) single
crystal surfaces, forming an ordered over-layer, less dense than on Pt and
Pd. On both silver surfaces, metal atoms are recruited by the oxygen in
order to form a 2D surface oxide with its own specific stoichiometry
42,43
.
Gold is more ‘inert’ to reaction because of the low oxygen
dissociation probability of molecular oxygen with the surface (<<10
-6
).
However, O
2
is used as the oxidant in reactions with gold nanoparticles
with
44
and without
45
an oxide support, and it has been shown that
identical reactions occur due to atomic oxygen on gold single crystal
surfaces. Hence, there is prima facie evidence that the reactive species
is adsorbed atomic oxygen. Using ozone exposure to achieve O
(a)
on gold in UHV
46
, the structure of O
(a)
on gold is disordered, with
Table 1 Gold facilitates a wide variety of reactions. The
examples below are representative of this variety, and
have been observed in solution and gas phase catalysis.
MT14_4p134-143.indd 137 21/03/2011 10:58:56
REVIEW Precious metal magic: catalytic wizardry
APRIL 2011 | VOLUME 14 | NUMBER 4
138
the oxygen adatoms resulting in a release of gold atoms and small
nanoparticle structures on the surface. When oxidized at a higher
temperature, less active, more highly structured “2D-oxide” structures
form
47
. In the presence of adsorbed surface oxygen on gold, a new
world of chemistry unfolds, as we will see below.
CO oxidation
Catalytic CO oxidation by noble metals is particularly important for
environmental protection, in ameliorating truck and auto exhaust and for
the prevention of poisoning of fuel cell electrodes. As one of the simplest
test reactions, it has also played a significant role in the development
of the fundamentals of reactions on noble metal surfaces. First, we will
address CO oxidation on Pt, which is the basis of the automotive catalytic
converter, and then we will discuss the remarkable development in the
oxidation of CO on gold nanoparticles supported on reducible oxides,
which opened the door to radically new directions in catalysis by gold.
CO oxidation on Pt proceeds by reaction of coadsorbed CO and
atomic oxygen. Below a critical steady state reaction temperature
the surface can become saturated with CO, so that dissociation of
dioxygen cannot proceed, and the adsorbed CO poisons the surface.
Above this minimum temperature, which depends on the metal, the
reaction proceeds readily
48
. By any measure this reaction is extremely
facile, because platinum readily dissociates dioxygen and binds CO in
molecular form, bringing the atomic oxygen and CO in close proximity
to facilitate reactions with a very low energy barrier (Fig.5a)
49
. Under
certain conditions of pressure and temperature dramatic periodic
oscillations in the rate occur, leading to time varying rates of formation
of CO
2
(Fig 5b)
50
. These oscillations are the result of two dimensional
concentration inhomogeneities across the surface. Their experimental
elucidation was a significant aspect of the research of Prof. Gerhard
Ertl, the Nobel Laureate in Chemistry in 2007.
CO can also be oxidized by metallic gold on oxide supports. Though
this fact was well documented in the 1970s, it was not until the mid
90s that it was shown by Haruta et al. that CO could be oxidized
at temperatures as low as -70 °C using reducible oxides as supports
for gold nanoparticles
15
. While it is clear that changing the support
changes the rate of this reaction, the mechanism for oxygen activation
and spillover is still under investigation. Gold nanoparticles were also
shown to be active for NO reduction by hydrocarbons.
Epoxidation
As discussed above, ethylene epoxidation is not only an important
industrial process, but also a delicate one. It has been extensively
studied, but under UHV conditions, ethylene desorbs at temperatures
below which reaction proceeds, prohibiting the reaction from being
studied traditionally. This obstacle was circumvented by the use of
norbornene, which possesses a C-C double bond with the appropriate
structure to allow reaction with adsorbed atomic oxygen, confirming
that direct epoxidation of alkenes by adsorbed atomic oxygen is the
important reaction, even for ethylene
51
. It is possible to interrogate
the nature of the reactive intermediates formed in the ethylene
epoxidation by looking at the reaction in reverse. By adsorbing
ethylene oxide at 250 K, an intermediate forms on the surface and
redesorbs as ethylene oxide at 300 K
52
. This intermediate is proposed
Fig. 4 Representative atomically resolved structures of oxygen on (a) Pd,
38
(b) Pt,
40
(c) Ag, and (d) Au
47
that are active for oxidation catalysis. In most cases atoms
from the metal surface are enlisted to form a two-dimensional metal oxide. The energy for the chemisorption of oxygen is taken from
37
. Fig. (a) reprinted with
permission from
38
. © 2002 by the American Physical Society. Fig. (b) reprinted from
40
with permission from Elsevier. Fig. (d) reprinted with permission
47
. © 2006
American Chemical Society.
(b) Pt(111) p(2x1)-O (a) Pd(111) p(2x2)-O
(c) Ag(111) p(4x4)-O (d) Au(111) 0.2 ML O
MT14_4p134-143.indd 138 21/03/2011 10:58:57
Precious metal magic: catalytic wizardry REVIEW
APRIL 2011 | VOLUME 14 | NUMBER 4
139
to be an oxametallacycle intermediate, where the metal, oxygen,
and carbon atoms form a cyclic structure
52
. In order to further study
this intermediate as well as the oxidation of higher molecular weight
olefins on silver styrene oxidation has been studied on both the
Ag(110) and Ag(111) surfaces (Fig. 6)
53
. On Ag(111) it was found
that upon addition of styrene to an oxygen-covered silver surface,
an oxametallacycle intermediate forms, which leads to styrene oxide
formation
54
. On Ag(110) the dominant path is to form a combustion
intermediate which leads to other products. However, when Cs is
added as a promoter styrene oxide is the dominant product
55
.
Fig. 5 (a) STM shows CO oxidation on O/Pt(111) over time. The (2x2)-O (darker areas) and the c(4x2)-CO (lighter areas) reconstructions come into contact to
oxidize CO to CO
2
at the interface between structures
49
. (b) Photoemission electron microscopy (PEEM), shows reaction wave fronts, in which the oxygen-covered
areas are dark, and the CO covered areas are bright. These wave fronts can take both propagating (bottom) and standing (top) wave forms, depending on the
conditions
50
. Fig. (a) from
49
. Reprinted with permission from AAAS. Fig. (b) reprinted with permission from
50
. © 1990 by the American Physical Society.
(a)
(b)
Fig. 6 The dominant pathways for styrene oxidation on O/Ag(111) and O/Ag(110). The O/Ag (111) and (110) structures are shown to the left, after
42,43
,
respectively. STM images from
53
.
MT14_4p134-143.indd 139 21/03/2011 10:58:59
REVIEW Precious metal magic: catalytic wizardry
APRIL 2011 | VOLUME 14 | NUMBER 4
140
Selective epoxidation of propene is also very desirable for the
chemical and pharmaceutical industries, and cannot be performed by
silver because the C-H bonds of the methyl group are readily attacked
by surface-bound oxygen. Industrially, propene oxide is produced in a
wasteful two-stage process
29
. A direct catalytic oxidation route had
not been possible until a discovery by Hayashi et al.
56
. They found
that in the presence of both H
2
and O
2
, gold supported on TiO
2
can
facilitate a direct oxidation in the vapor phase, with selectivity of over
90 %. It was found that TiO
2
must be used as the support, and that H
2
enhanced the selectivity. It is essential that more fundamental research
be done to separate the mechanistic roles of the nanoparticles and the
supports, so that a better understanding of principles needed to design
operating conditions for the reaction can be realized.
Novel coupling reactions on Au and Ag
Though molecularly simple reactions such as CO oxidation and
epoxidation reactions are of technological importance to our society,
more complex synthesis reactions create the chemical intermediates
that are used to form a vast number of useful retail products. Again
the precious metals, particularly Au and Ag can play a substantial role
as catalysts for such reactions, because, perhaps paradoxically, they
are basically inert toward reactions with most molecules. However,
for oxidative reactions on Ag and Au, surface-bound atomic oxygen
activates the surface for selective partial oxidation with remarkable
precision
57
. Surface-bound oxygen atoms, O
(a)
, act as a Brønsted base
and selectively clip O-H, N-H, and S-H bonds in larger molecules,
forming a surface-bound molecular fragment and an adsorbed OH
group
58-61
. This reactivity pattern was originally established on metallic
silver for a wide variety of molecules
10
and extended in principle to
gold surfaces
62
. Recent studies have shown a much more extensive
pattern of oxygen activation on gold
30,31,63-65
, including most recently,
the self and cross-coupling of alcohols
24,66
and the coupling of alcohols
with aldehydes
28
to form esters (Fig. 7a).
In general terms, after activation of the alcohol by surface-bound
oxygen to form adsorbed RO-, cleavage of the C-H bond by the surface
produces an aldehyde, in which the carbon bound to the oxygen has
a slight positive charge. This carbon is then readily attacked by the
oxygen of the RO- (Fig. 7a,2), and a C–O bond forms. The resulting
intermediate then loses a hydrogen to form the ester product. This
reaction also proceeds readily if the aldehyde is added directly to the
surface RO- species. Friend et al. find that when cross-coupling two
alcohols, the cross-coupled product is always the methyl-ester of the
longer chain alcohol. By balancing the β-H elimination abilities from the
alkoxy to form the aldehyde with the gas phase acidity of methanol,
ethanol, and butanol, they optimized the conditions under which the
cross-coupling product would be dominant
27
. Alcohol cross-coupling
has been achieved over supported gold nanoparticles as well, and the
mechanistic insight gained in UHV correlates well with solution phase
reactions. For example, Nielsen et al. report high selectivity for the
coupling of hexanol with methanol under similar conditions, ~90 %
methanol molar fraction
44
. Ethanol self-coupling to ethyl acetate has
also been observed
67
.
Using these fundamental studies of oxidative coupling of alcohols
on gold single crystals, it was envisioned that alcohols could self-
couple to selectively form esters in a steady state catalytic reactor
23,24
.
Using a new novel nanoporous gold material, which contains a very
small concentration of silver, high selectivity for oxygen-assisted ester
formation from methanol self-coupling was realized by flowing a mixture
of oxygen and methanol over the nanoporous gold at relatively low
temperatures (Fig. 8)
68
. This non-supported catalyst is intriguing for a
number of reasons, including the fact that oxygen dissociates and reacts
without the aid of an oxide support, the catalyst is not prone to sintering
at operational temperatures, and the exact amount of silver in the gold
changes the selectivity and optimal temperature of operation
68
.
The general similarity of the electron charge distribution surrounding
oxygen and nitrogen atoms in molecules suggests that similar reactions
Fig. 7 The mechanistic pathway for coupling of (a) alcohols and aldehydes and (b) amines and aldehydes on oxygen covered silver or gold.
(1) (1)
(2) (2)
(3) (3)
(a) (b)
(i) (i)(ii) (ii)
MT14_4p134-143.indd 140 21/03/2011 10:59:04
Precious metal magic: catalytic wizardry REVIEW
APRIL 2011 | VOLUME 14 | NUMBER 4
141
might occur with amines, where essentially the OH functionality in the
alcohol is replaced by an NH group in the amine (Fig. 7b). Indeed the
N-H bond in the amine is activated by surface oxygen, and the resulting
species adds facilely to the electron deficient center in aldehydes. These
similar reactions provide an excellent example of the use of general
patterns of reactivity to anticipate new synthesis pathways – enabling
catalysis by design, as opposed to trial and error.
As was predicted by the mechanism shown in Fig. 7b, silver and
gold surfaces have been shown to facilitate amide synthesis from
aldehydes and amines in UHV. Using TPRS and isotopic labeling,
the reaction mechanism has been clearly delineated (Fig. 9). At low
oxygen coverages, this reaction approaches ~100 % selectivity for the
formation of the coupling product, dimethylacetamide, on both silver
and gold surfaces; the route to methyl isocyanate and combustion
products is practically eliminated. Differences in selectivity and reaction
mechanism have been investigated by comparative study of acetlyation
of dimethlyamine with acetaldehyde on the both oxygen-covered Au
and Ag surfaces under UHV
69
. In this study and the previous studies
of dimethylamine acylation by formaldehyde on gold
70
and silver
71
,
it was observed that the oxygen-assisted reaction proceeds on the
surface via activation of N-H to form adsorbed amides. Coupling then
occurs via nucleophilic attack of the amide on the carbonyl carbon of
the aldehyde. This mechanism expands the known reactivity patterns
of adsorbed oxygen on silver and gold
57
, providing insight into similar
chemistry on gold nanoparticles
72,73
. For example, Christensen et al.
observe coupling of an alcohol and an amine to form an amide over
a supported gold catalyst
73
. The observed chemistry makes sense
when viewed in light of the mechanism proposed here. Oxygen on
gold activates both the O-H and N-H bond, the alkoxy then β-H
eliminates to form an aldehyde, which can be nucleophilically attacked
by the activated amide. The partial oxidation and coupling reactions
in solution phase catalysis using molecular oxygen as the oxidant
follow the same acid base reaction and coupling patterns as seen in
UHV using atomic oxygen on gold. Further study into these rich partial
Fig. 8 Unsupported nanoporous gold facilitates the same methanol cross
coupling to methylformate (HCO
2
CH
3
) as seen on gold single crystals in UHV.
High selectivity is favored at low surface oxygen coverages. From
68
. Reprinted
with permission from AAAS.
Fig. 9 TPRS allows the deciphering of the mechanism for catalytic production of amides by amine/aldehyde coupling on an oxygen covered gold surface. The
external cycle shows the sequence of reactions that transform dimethyl amine and acetaldehyde into dimethylacetamide (upper left). Competing partial oxidation
of dimethyl amine leads to methyl isocyanate (lower right) Lines from each product to the TPRS in the center indicate the mass spectrometric identification of the
products. Modified from
69
.
MT14_4p134-143.indd 141 21/03/2011 10:59:05
REVIEW Precious metal magic: catalytic wizardry
APRIL 2011 | VOLUME 14 | NUMBER 4
142
oxidation and coupling reactions will surely be expanded, and hopefully
industrial implementation will be realized in the near future.
Summary
Although catalysis has been increasingly well understood over the
ages, it was not until the last 60 years that clear and detailed reaction
mechanisms emerged to change the understanding and design of more
complex reactions. The last two examples of novel coupling chemistry
on oxygen-covered silver and gold surfaces may represent a turning
point in catalysis design – we have reached a confluence of catalyst
process design by reactivity principles derived in UHV advanced by
ambient and high pressure catalysis research to open new possibilities
for green catalysis by precious metals.
Acknowledgments
We gratefully acknowledge the support of the Department of
Energy Office of Science Graduate Fellowship Program (DOE SCGF),
made possible in part by the American Recovery and Reinvestment
Act of 2009, administered by ORISE-ORAU under contract no.
DE-AC05-06OR23100 (CGF) and the National Science Foundation,
Division of Chemistry, Analytical and Surface Science (RJM)
CHE- 0513936.
REFERENCES
1. Freestone, I., et al., Gold Bull (2007) 40, 270.
2. Padovani, S., et al., Appl Phys A (2006) 83, 521.
3. Wigley, R. A., Gold and Silver in Medicine. In Noble metals and biological systems:
their role in medicine, mineral exploration, and the environment, Brooks, R. R.,
(ed.), CRC Press, Boca Raton, (1992), 277.
4. Thomas, J. M., and Thomas, W. J., Principles and Practice of Heterogeneous
Catalysis, VCH Publishers, Inc., New York, 1997.
5. Golunski, S., JOM- J Min Met Mat S (2001) 53, 22.
6. Kneitsch, R., Pop Sci Monthly (1902) 61, 24.
7. Louie, D. K., Handbook of Sulphuric Acid Manufacturing, DKL Engineering, Inc.,
Thornhill, 1961.
8. Kaspar, J., et al., Catal Today (2003) 77, 419.
9. Barteau, M. A., and Madix, R. J., J Am Chem Soc (1983) 105, 344.
10. Wachs, I. E., and Madix, R. J., Surf Sci (1978) 76, 531.
11. Wachs, I. E., and Kelemen, S. R., J Catal (1981) 71, 78.
12. Andreasen, A., et al., Surf Sci (2003) 544, 5.
13. Stegelmann, C., et al., J Catal (2004) 221, 630.
14. Stegelmann, C., and Stoltze, P., Surf Sci (2004) 552, 260.
15. Haruta, M., et al., J Catal (1993) 144, 175.
16. Worrell, E., et al., Energy Use and Energy Intensity of the U.S. Chemical Industry,
LBNL-44314, Berkeley, 2000.
17. Reisch, M. S., C&EN (2009) 87, 10.
18. Astruc, D., Transition-metal Nanoparticles in Catalysis: From Historical
Background to State-of-the-Art. In Nanoparticles and Catalysis, Astruc, D., (ed.),
Wiley-VCH, Weinheim, (2008).
19. Min, B. K., and Friend, C. M., Chem Rev (2007) 107, 2709.
20. Christensen, C. H., and Norskov, J. K., Science (2010) 327, 278.
21. Hashmi, A. S. K., and Hutchings, G. J., Angew Chem Int Ed (2006) 45, 7896.
22. Pina, C. D., et al., Gold nanoparticles-catalyzed oxidations in organic chemistry. In
Nanoparticles and Catalysis, Astruc, D., (ed.), Wiley-VCH, Weinheim, (2008).
23.
Xu, B. J., et al., Angew Chem Int Ed (2009) 48, 4206.
24. Liu, X. Y., et al., J Am Chem Soc (2009) 131, 5757.
25. Gong, J., et al., Chem Commun (2009) 761.
26. Kitahara, H., and Sakurai, H., Chem Lett (2010) 39, 46.
27. Xu, B., et al., J Am Chem Soc (2010) 132, 16571.
28. Xu, B. J., et al., Nat Chem (2010) 2, 61.
29. Qi, C., Gold Bull. (2008) 41, 224.
30. Deng, X., and Friend, C. M., J Am Chem Soc (2005) 127, 17178.
31. Deng, X., et al., Angew Chem Int Ed (2006) 45, 7075.
32. Redhead, P. A., Vacuum (1962) 12, 203.
33. King, D. A., Surf Sci (1975) 47, 384.
34. Weldon, M. K., and Friend, C. M., Chem Rev (1996) 96, 1391.
35. Niemantsverdriet, J. W., Spectroscopy in Catalysis: An Introduction, Wiley-VCH,
Weinheim, 2000.
36. Chen, C. J., Introduction to scanning tunneling microscopy, Oxford University
Press, New York, 1993.
37. Hammer, B., and Norskov, J. K., Adv in Catal (2000) 45, 71.
38. Lundgren, E., et al., Phys Rev Lett (2002) 88, 246103.
39. Zheng, G., and Altman, E. I., Surf Sci (2000) 462, 151.
40. Devarajan, S. P., et al., Surf Sci (2008) 602, 3116.
41. Hinojosa, J. A., et al., J Phys Chem C (2008) 112, 8324.
42. Klust, A., and Madix, R. J., J Chem Phys (2007) 126, 084707.
43. Canepa, M., et al., Phys Rev B (1993) 47, 15823.
44. Nielsen, I. S., et al., Catal Lett (2007) 116, 35.
45. Angelici,
R. J., J Organomet Chem (2008) 693, 847.
46. Saliba, N., et al., Surf Sci (1998) 410, 270.
47. Min, B. K., et al., J Phys Chem B (2006) 110, 19833.
48. McCrea, K., et al., High-pressure CO dissociation and CO oxidation studies
on platinum single crystal surfaces using sum frequency generation surface
vibrational spectroscopy. In Surface Chemistry and Catalysis, Carley, A. F., et al.,
(ed.), Kluwer Academic/Plenum, New York, (2002).
49. Wintterlin, J., et al., Science (1997) 278, 1931.
50. Jakubith, S., et al., Phys Rev Lett (1990) 65, 3013.
51. Roberts, J. T., and Madix, R. J., J Am Chem Soc (1988) 110, 8540.
52. Lukaski, A. C., and Barteau, M. A., Catal Lett (2009) 128, 9.
53. Zhou, L., and Madix, R. J., unpublished results.
54. Zhou, L., and Madix, R. J., J. Phys. Chem. C (2008) 112, 4725.
55. Zhou, L., and Madix, R. J., Surf. Sci. (2009) 603, 1751.
56. Hayashi, T., et al., J. Catal. (1998) 178, 566.
57. Madix, R. J., and Roberts, J. T., The Problem of Heterogeneously Catalyzed Partial
Oxidation: Model Studies on Single Crystal Surfaces. In Surface Reactions, Madix,
R. J., (ed.), Springer-Verlag, Berlin, (1994).
58. Stuve, E. M., et al., Surf Sci (1981) 111, 11.
59. Thornburg, D. M., and Madix, R. J., Surf Sci (1990) 226, 61.
60. Thornburg, D. M., and Madix, R. J., Surf Sci (1989) 220, 268.
61. Quiller, R. G., et al., J Chem Phys (2008) 129, 064702.
62. Outka, D. A., and Madix, R. J., J Am Chem Soc (1987) 109, 1708.
63. Deng, X., et al., J Am Chem Soc (2005) 127, 9267.
64. Deng, X., et al., J Phys Chem B (2006) 110, 15982.
65. Liu, X., and Friend, C. M., Langmuir (2010).
66. Xu, B., et al., Chem Sci (2010) 1, 310.
67. Jorgensen, B., et al., J Catal (2007)
251, 332.
68. Wittstock,
A., et al., Science (2010) 327, 319.
69. Freyschlag, C. G., et al., submitted.
70. Xu, B. J., et al., Angew Chem Int Ed (2010) 49, 394.
71. Zhou, L., et al., Chem Commun (2010) 46, 704.
72. Ishida, T., and Haruta, M., Chemsuschem (2009) 2, 538.
73. Klitgaard, S. K., et al., Green Chem (2008) 10, 419.
MT14_4p134-143.indd 142 21/03/2011 10:59:07
