Organometallic chemistry and catalysis on gold metal surfaces

Article abstract of DOI:10.1016/j.jorganchem.2007.11.023

As in transition metal complexes, C{triple bond, long}N-R ligands adsorbed on powdered gold undergo attack by amines to give putative diaminocarbene groups on the gold surface. This reaction forms the basis for the discovery of a gold metal-catalyzed reaction of C{triple bond, long}N-R, primary amines (R′NH₂) and O₂ to give carbodiimides (R′-N{double bond, long}C{double bond, long}N-R). An analogous reaction of C{triple bond, long}O, RNH₂, and O₂ gives isocyanates (R-N{double bond, long}C{double bond, long}O), which react with additional amine to give urea (RNH)₂C{double bond, long}O products. The gold-catalyzed reaction of C{triple bond, long}N-R with secondary amines (HNR′₂) and O₂ gives mixed ureas RNH(C{double bond, long}O)NR′₂. In another type of gold-catalyzed reaction, secondary amines HN(CH₂R)₂ react with O₂ to undergo dehydrogenation to the imine product, RCH{double bond, long}N(CH₂R). Of special interest is the high catalytic activity of gold powder, which is otherwise well-known for its poor catalytic properties.

Full text of DOI:10.1016/j.jorganchem.2007.11.023

Available online at www.sciencedirect.com
Journal of Organometallic Chemistry 693 (2008) 847–856
www.elsevier.com/locate/jorganchem
Organometallic chemistry and catalysis on gold metal surfaces
Robert J. Angelici *
Ames Laboratory and Department of Chemistry, Iowa State University, Ames, IA 50011-3111, USA
Received 13 August 2007; received in revised form 13 November 2007; accepted 13 November 2007
Available online 21 November 2007
Abstract
As in transition metal complexes, C„N–R ligands adsorbed on powdered gold undergo attack by amines to give putative diamino-
carbene groups on the gold surface. This reaction forms the basis for the discovery of a gold metal-catalyzed reaction of C„N–R, pri-
mary amines (R0NH2) and O2 to give carbodiimides (R0–N@C@N–R). An analogous reaction of C„O, RNH2, and O2 gives isocyanates
(R–N@C@O), which react with additional amine to give urea (RNH)2C@O products. The gold-catalyzed reaction of C„N–R with sec-
ondary amines (HNR02) and O2 gives mixed ureas RNH(C@O)NR02. In another type of gold-catalyzed reaction, secondary amines
HN(CH2R)2 react with O2 to undergo dehydrogenation to the imine product, RCH@N(CH2R). Of special interest is the high catalytic
activity of gold powder, which is otherwise well-known for its poor catalytic properties.
Ó 2007 Elsevier B.V. All rights reserved.
Keywords: Isocyanides; Carbon monoxide; Carbenes; Oxidation; Dehydrogenation; Gold
1. Introduction
means that the nature of organic reactant binding to a
metal surface is dependent on interpretations of a variety
of results from spectroscopic investigations.
Transformations of organic molecules are often cata-
lyzed by metals supported on high surface-area oxides such
as SiO2, Al2O3, or TiO2. Despite the widespread use of
these heterogeneous catalysts in industrial applications,
very little is known about the mechanistic details of the
reactions that occur on the metal surfaces. Proposed mech-
anisms for surface metal-catalyzed reactions [1,2] are typi-
cally based on reactions of organic molecules that occur in
solution-soluble transition metal complexes, where experi-
mental methods are able to yield more details of reaction
pathways. Often the coordination mode of the reacting
organic molecule can be established by X-ray diffraction
and NMR studies. Moreover, intermediates on the path
to products can sometimes be characterized. On the other
hand, in studies of supported metal-catalyzed reactions,
X-ray diffraction and NMR techniques are not generally
useful for characterizing either the initially-adsorbed reac-
tant nor intermediates that lead to the products. This
The purpose of the studies described in this paper was to
compare and contrast the binding and subsequent reac-
tions of specific organic molecules in metal complexes
and on metal surfaces. More specifically, we hoped to
understand whether or not principles that are useful in
understanding reactions of organic ligands in organometal-
lic complexes can be applied to the same organic molecules
adsorbed on a metal surface. Gold was chosen as the metal
surface because it does not react with air (O2) to form a
surface oxide at room temperature, which means that
organic molecules adsorb only to Au(0) atoms, and not
to gold in some positive oxidation state. While this simpli-
fies studies of organic molecule adsorption on a metal sur-
face, the gold surface is therefore also so unreactive that it
is a poor catalyst as is well-known [3]. However, as will
become evident, gold metal is not as catalytically inactive
as was previously assumed.
For the purpose of the proposed studies, it was neces-
sary to choose an organic reacting molecule that contained
a signature infrared absorption that would identify it when
*
Tel.: +1 515 294 2603; fax: +1 515 294 0105.
E-mail address: angelici@iastate.edu
0022-328X/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2007.11.023

848
R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
it was adsorbed on the gold surface and would shift to
another frequency when it reacted to give a new species
on the surface. Our initial studies, described below, began
with isocyanide molecules (C„N–R) which exhibit a
strong, characteristic m(N„C) infrared absorption, which
has been used extensively in characterizing metal com-
plexes with isocyanide ligands [4–10] and is also useful
for identifying isocyanides adsorbed on metal surfaces
[11]. Of course, the choice of an organic reactant for study
on a metal surface was also guided by a well-developed
understanding of its reactivity in metal complexes that
could be compared with its reactivity on the gold metal
surface.
Although gold metal has been traditionally known as a
poor catalyst [3], there has been tremendous recent interest
in nano-sized gold particles that have been shown to cata-
lyze a variety of reactions including the following [3,12–17]:
CO oxidation, epoxidation of propene, oxidation of vari-
ous hydrocarbons and alcohols, the water–gas shift reac-
tion, alkyne and alkene hydrogenation, and several
others. The catalytic activities of these nanogold catalysts
depend upon their stabilization as particles of less than
5 nm size. This is typically accomplished by forming them
on a metal oxide support, which is often TiO2 but may also
be many others (e.g., Fe2O3, CeO2, ZrO2, SiO2, Al2O3,
CoOx, and MnOx). The method used for preparing sup-
ported nanogold is very important for achieving an active
catalyst, and it is frequently not possible to reproduce cat-
alytic activity from one preparation to the next [3]. Because
of the surprising catalytic activity of nanogold, numerous
studies have sought to understand why nano-sized gold
particles are so much more active than large (bulk) gold
particles. This is a field of continuing research activity that
is regularly reviewed [3,12–17].
None of the forms of gold metal that are used in the
studies described in this paper are nanogold. In order to
avoid complications that may arise from the presence of
a metal oxide support, we have used either gold powder
or a gold film. The gold powder was prepared by the simple
reduction of an aqueous solution of HAuCl4 with hydro-
quinone [18]. The resulting gold powder has a very low sur-
face area (0.35 m2/g), and its particle diameters are in the
range of 1000 nm, as determined by electron microscopy
[19]. This gold powder was re-used many times for both
the adsorption and catalysis studies. Typically, the gold
powder was cleaned before each use by treatment with
‘‘piranha” solution (concentrated H2SO4/30% aqueous
H2O2) [20,21], which restored the powder to its original
catalytic reactivity. If the gold powder were only washed
with a solvent, e.g., MeOH, between catalytic runs, its
activity would be slightly decreased (ꢀ10%). There was
no evidence for nano-sized gold particles in the gold pow-
der that was prepared and cleaned in this manner. The elec-
tron micrographs showed no particles that were less than
5 nm in size. A further test for nanogold particles involved
using the gold powder as a potential catalyst for the reac-
tion, CO + 1/2 O2 ? CO2. Nanogold particles are known
to catalyze this reaction even below room temperature
[3]; however, no reaction was observed between CO and
O2 (1.0 atm total) in the presence of 1.0 g of gold powder
at 45 °C in MeCN, which are conditions typically used
for our catalytic studies. Thus, it seems very unlikely that
nanogold particles are present in the gold powder. More-
over, gold powder had been previously shown to be a poor
catalyst for the oxidation of CO [22–24].
The gold films used in some of our studies were prepared
by first depositing a 15 nm layer of Cr on a glass slide.
Then, 300 nm of gold was vapor-deposited on the Cr.
These slides were also cleaned with piranha solution
between runs. Thus, all of the gold substrates used in the
studies described in this paper are much larger than the
supported nanogold particles that are so catalytically
active. In the following sections are discussed our studies
of the adsorption and catalytic reactions of isocyanides
(C„N–R), CO, and amines on bulk gold metal.
2. Isocyanides (CBN–R) on gold: adsorption, reactions, and
catalysis
Isocyanides in solution, e.g., 1,2-dichloroethane (DCE)
and hexane, readily adsorb to gold metal in all of its forms
(films, powders, and nanoparticles) [11]. Typically, a
m(N„C) frequency for an adsorbed isocyanide that is equal
to or higher than that of the free isocyanide indicates that
the isocyanide is adsorbed g1 through the carbon to one
Au atom. In our own studies, we observed for n-BuNC
adsorbed on a gold film a m(N„C) value (2224 cmꢁ1) that
is 76 cm ꢁ1 higher than that (2148 cmꢁ1) of free n-BuNC in
DCE solution [21]. Although the presence of only one
m(N„C) band for the adsorbed isocyanide suggests that
all isocyanide molecules are adsorbed on the same type
of gold site, this cannot be the case because n-BuNC gold
films that are immersed in a pure solvent, such as DCE,
desorb about 55% of the n-BuNC within 1 min while the
other 45% requires 4 h or longer. As the n-BuNC desorbs,
the m(N„C) of the n-BuNC remaining on the gold
increases slightly to 2232 cmꢁ1. Thus, there are two broad
types of adsorbed isocyanide: (1) one that is weakly
adsorbed with m(N„C) = 2224 cmꢁ1 and (2) one that is
strongly adsorbed with
(2232 cmꢁ1). There are also two types of n-BuNC sites in
powdered gold; one exhibits m(N„C) value of
a
higher m(N„C) value
a
2225 cmꢁ1 at saturation coverage, while the other shows
a m(N„C) band at 2233 cmꢁ1 for the strongly adsorbed
n-BuNC which remains on the surface after the weakly
adsorbed isocyanide is desorbed. While the m(N„C) values
for the weakly and strongly adsorbed n-BuNC are the same
on both gold film and gold powder, only 45% of n-BuNC is
strongly adsorbed on gold film, whereas 70% is strongly
adsorbed on gold powder. The much more irregular sur-
face of gold powder would be expected to have more
low-coordinate edge and corner gold atom sites than would
exist on gold films with more high-coordinate gold atoms
on the (111) terrace sites [25]. Gold powder would be

R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
849
expected to bond more isocyanides more strongly at the
larger number of low-coordinate sites.
surface area than the gold film, with n-BuNH2, it was
possible to identify the product of the reaction. We were
surprised to find that it was the carbodiimide Bun–N@
C@N–Bun, together with a significant amount of n-BuNC,
which presumably resulted from the rapid desorption of
the weakly adsorbed n-BuNC. Thus, it appears that most
of the reacting n-BuNC was the strongly adsorbed form,
which would be expected to be more reactive toward nucleo-
philes because it has a higher m(N„C) value than the weakly
adsorbed n-BuNC.
In transition metal complexes, isocyanide ligands whose
m(N„C) values are more than 40 cmꢁ1 higher than those of
the free isocyanides have been observed to be attacked by
amines (and other nucleophiles) [26–28] to give diamino-
carbene ligands (Eq. (1)). Typically, the rate laws for
NHR
LnM C N R
LnM C
ð1Þ
+ H2NR'
NHR'
During the studies of adsorbed n-BuNC with n-BuNH2,
we noticed that the desorbed n-BuNC product was also
converted to the carbodiimide product if the reaction was
allowed to stir for a longer time or was heated to 60 °C.
This meant that the reaction of n-BuNC with n-BuNH2
to form Bun–N@C@N–Bun was catalyzed by gold metal.
However, the fate of the two hydrogen atoms in the n-
BuNH2 reactant was not evident, but it seemed possible
that they were combining with O2 to give H2O, as the reac-
tions were conducted in air. The catalytic reaction of n-
BuNC with n-BuNH2 in the presence of gold powder was
then conducted under an argon atmosphere. Under these
conditions, no carbodiimide product was formed. This
meant that O2 is a reactant and that H2O is the likely prod-
uct. Therefore, the complete gold-catalyzed reaction is rep-
resented by Eq. (3). At 60° in hexane solvent, under an air
atmosphere and stirring with
these reactions are first order in the metal complex and
amine concentrations: Rate = k[LnM(CNR)][H2NR0]. This
rate law has been interpreted as supporting a mechanism
that involves nucleophilic attack on the isocyanide carbon
that is activated by a metal (M) that is in a positive oxida-
tion state and/or by ligands L that are weakly electron-
donating. In some reactions, the rate law exhibits an amine
concentration dependence that is higher than first order.
This is interpreted to indicate the involvement of a second
amine molecule, which promotes the rearrangement of the
initial isocyanide–amine adduct to the diaminocarbene
product [26–29].
Metal salts of Cu+, Ag+, Zn2+, Cd2+, and Au3+ are
known to catalyze reactions (Eq. (2)) of isocyanides with
primary and secondary amines to give formamidines
[28,30]. These reactions
NR
Au
ð2Þ
C N R
H C
+ HNR'2
n-BuNC þ n-BuNH2 þ ½O2 ! Bun–N@C@N-Bun þ H2O
NR'2
ð3Þ
are proposed to occur by initial metal coordination of the
isocyanide, which is attacked by the amine to give a diami-
nocarbene ligand that is displaced as the formamidine
product. The last step is modeled by reactions of diamino-
carbene complexes of Au(I) and Au(III) with PPh3 or CNꢁ
ligands that give the free formamidine product [28].
1.0 g of gold powder, n-BuNC (1.7 ꢂ 10ꢁ2 mmol) and
n-BuNH2 (17 ꢂ 10ꢁ2 mmol) react for 90 h to give a 73%
yield of Bun–N@C@N–Bun. Since the n-BuNC had com-
pletely reacted, it meant that other products (27%) also
formed. These products were not identified, but there were
additional peaks at higher mass in the GC–MS, which
might be attributed to oligomers of n-BuNC. If the reac-
tions were conducted at higher concentrations of the isocya-
nide and amine, the percent yields decreased, presumably
due to the formation of more n-BuNC oligomers. These
lower yields mean that we have not found conditions that
make this reaction useful for the synthesis of larger amounts
of carbodiimides. Mixed carbodiimides R–N@C@N–R0
were obtained when relatively low concentrations of a vari-
ety of isocyanides (C„N–R) were reacted with a variety of
amines (R0NH2) [21,31].
As the m(N„C) value (2224 cmꢁ1) for n-BuNC adsorbed
on gold is much more than 40 cmꢁ1 higher than that
(2148 cmꢁ1
) of free n-BuNC, one expects n-BuNC
adsorbed on gold metal to undergo attack by amines to
give diaminocarbene groups on the surface, if one assumes
that the 40 cmꢁ1 guideline for isocyanides in metal com-
plexes also applies to isocyanides adsorbed on metal sur-
faces. In fact, we observe [21] that all of the n-BuNC
(both weakly and strongly adsorbed) is removed within
20 s from a gold film when a gold-plated slide containing
adsorbed n-BuNC is immersed into a hexane solution con-
taining n-BuNH2. The reaction of the adsorbed n-BuNC is
slower if the amine is less basic or more sterically bulky
than n-BuNH2. There was no evidence in the diffuse reflec-
tance infrared spectrum of the resulting gold film for diami-
nocarbene groups, which are expected to give a m(N–C@N)
absorption near 1550 cmꢁ1, which is characteristic of dia-
minocarbene ligands in metal complexes. Thus, the product
of the reaction of the adsorbed isocyanide with the amine
must have desorbed from the surface. By reacting n-BuNC
adsorbed on 1.0 g of gold powder, which has a much larger
Kinetic studies [21] of the gold powder-catalyzed reac-
tion (Eq. (3)) of n-BuNC and n-BuNH2 at 60 °C in hexane
show that the rate is independent of the O2 concentration/
pressure, as indicated by the observation that the rate is the
same in 1 atm of air or O2. However, the rate is first order
in the concentration of n-BuNH2. These results suggest a
mechanism (Scheme 1) in which the isocyanide first
adsorbs on the Au surface; as in metal complexes (Eq.
(1)), this makes the isocyanide susceptible to nucleophilic
attack by the amine in step 1. This is the rate-determining

850
R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
a key step in both the homogeneous and Au-catalyzed reac-
R'
R'
N C N
N C N
tions of isocyanides and amines, the Au surface provides a
pathway for O2 to react with the diaminocarbene intermedi-
ate that leads to a product that is very different from that
obtained with the homogeneous metal ion catalysts.
Of course, carbodiimide products are only possible
when the amine is a primary amine. With secondary
amines, it is still possible for the amine to attack an
adsorbed isocyanide, but the resulting intermediate a or b
(Scheme 1) cannot react with O2 to give a carbodiimide
product. However, isocyanides do react (Eq. (6)) with sec-
ondary amines in the presence of gold powder, to
R
R
(2)
(4)
– H2O
O2
O2
– H2O
R'
R
R
R
N
R'
N
H
N
C
N
N
H
(1)
(3)
C
H
C
H
NH2R'
Au Au
Au
Au Au
Au
Au
(b)
(a)
Scheme 1. Mechanism for the Au-catalyzed reaction of isocyanides
(R–N„C) with primary amines (H2NR0) and O2.
O
ð6Þ
Au
+ HNR'2 + ½ O2
C
R
N
C NR'2
RNH
step in the reaction. The resulting intermediate a either
rearranges to a diaminocarbene group b (as occurs in metal
complexes, Eq. (1)) in step 3 or reacts with O2 to give H2O
and the carbodiimide in step 2. Alternatively, the diamino-
carbene intermediate b reacts with O2 to give the products.
As the overall reaction rate does not depend on the O2 con-
centration, steps 2 and 4 must be fast as compared with
step 1. Although step 1 is readily understandable in terms
of a mechanism that is well-documented for reactions of
isocyanide ligands with amines (Eq. (1)), there is no compa-
rable metal complex model for the conversion of a diami-
nocarbene ligand to a carbodiimide. A reaction that
presumably involves the oxidation of a diaminocarbene
complex is that shown in Eq. (4) [32]. Water is suggested
to be another product of the reaction,
give ureas [36]. The reactions were performed under the
same conditions (60 °C, hexane or MeCN solvent, 1.0 g
of gold powder) that were used with the gold-catalyzed
reactions of isocyanides with primary amines (Eq. (3)).
The reaction occurs under 1 atm of air or O2, but the rate
was slightly faster in the presence of pure O2, although this
dependence was significantly less than first order in the O2
concentration/pressure. When the reaction was run under
an argon atmosphere, only a small amount (ꢀ1.5%) of
the urea product formed, presumably due to the presence
of adventitious O2. This formation of urea even under a
very low pressure of O2 means that the O2 is highly reactive
with intermediates on the gold surface. When n-BuNC was
reacted with different concentrations of HNPrn2 under
1 atm of O2 in MeCN solvent using 1.0 g of Au powder,
the rate of the reaction increases as the amine concentra-
tion increases, although it was somewhat less than first
order in the amine concentration. This amine dependence
suggested that a key step in the reaction was nucleophilic
attack of the secondary amine on the adsorbed isocyanide.
Then, one would expect the rate of the catalyzed reaction
to depend also on the steric and electronic properties of
the amines. This is indeed observed as shown by the urea
product yields (Table 1) obtained in reactions of n-BuNC
with different secondary amines under a standard set of
conditions (60 °C, MeCN solvent, 1 atm O2, 1.0 g gold
powder). The yields were determined after 24 h of reaction.
As compared with the bulky HNPrn2, the cyclic and
NHR
+ Ag2O
RN=C=NR'
ð4Þ
Cl2(RNC)Pd
C
NHR'
and it seems likely that Ag metal is the reduction product
especially because the stoichiometric reaction of C„N–R
with R0NH2 and Ag2O to give carbodiimides is catalyzed
by PdCl2 [32], as summarized in Eq. (5). Thus, it seems pos-
sible that O2 in Eq. (3) is an oxidizing agent like Ag+ in
Eqs. (4) and (5).
C ꢃ N–R þ R0NH2 þ Ag2O P!dCl2 RN@C@NR0
þ H2O þ 2Ag
ð5Þ
Steps 2 and 4 in Scheme 1 are unlikely to involve the
generation of H2 from intermediate a or b, because H2
would not be converted to the H2O product by reaction
with O2 except at temperatures above 130 °C [13,33]. How-
ever, supported nanogold is reported to catalyze the reac-
tion of H2 and O2 to give H2O2 even below room
temperature [34,35].
Although there is little basis for proposing details of steps
involving O2 in Scheme 1, evidence for amine attack on
adsorbed isocyanide in step 1 is quite convincing. This step
is the same as that proposed for metal ion-catalyzed reac-
tions (Eq. (2)) of isocyanides and amines in solution. How-
ever, this homogeneous reaction gives formamidines
without the involvement of O2. Thus, while amine attack is
Table 1
Gold metal-catalyzed reactions of n-BuNC with secondary amines and O2
according to Eq. (6)a
NHR02
Urea yield (%)
HNPrn2
HNPrn2b
Piperidine
Morpholine
HNPrn2c
a
44
38
51
22
19
Reaction conditions: n-butyl isocyanide (2.0 mM), 80 mM amine, gold
powder (1.0 g), O2 (ꢀ1.0 L at ꢀ1 atm), at 60 °C in 5 mL MeCN for 24 h.
b
Air instead of O2.
1,1,3,3-Tetramethyl-n-butyl isocyanide (2.0 mM) instead of n-BuNC.
c

R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
851
therefore sterically-smaller piperidine ((CH2)5NH) gives a
higher yield (51%) of the corresponding urea. When mor-
pholine (O(CH2CH2)2NH), which is less nucleophilic but
sterically-similar to piperidine, is reacted with n-BuNC,
the yield (22%) of the urea is significantly lower. The higher
pKa of the protonated form of piperidine (11.2) as com-
pared with that for morpholine (8.36) is an indicator of
the higher nucleophilicity of piperidine. Therefore, both
the steric and electronic properties of the amine have pre-
dictable effects on the rates of reaction.
To test for the effect of the bulkiness of the isocyanide
on the rate of reaction, 1,1,3,3-tetramethyl-n-butyl isocya-
nide (CH3)3CCH2C(CH3)2NC was reacted with HNPr2n
and O2 under the standard conditions. As shown in the last
entry in Table 1, this bulky isocyanide gives a significantly
lower yield (19%) than n-BuNC (44%). The lower rate with
the bulky isocyanide could reflect a lower isocyanide cover-
age on the Au surface [37] or a reduced rate of amine attack
on the adsorbed isocyanide.
are known to be strongly electron-donating ligands in tran-
sition metal complexes [38]. So, it is not unreasonable to
propose that diaminocarbene groups add enough electron
density to the gold metal that it reacts with O2.
3. Carbon monoxide on gold: catalytic reactions with amines
As noted in the previous section, the susceptibility of
isocyanide ligands in metal complexes to attack by amines
(Eq. (1)) can be estimated from the m(N„C) value of the
isocyanide ligand. When this value is at least 40 cmꢁ1
higher than that of the free isocyanide, it is highly likely
that amines will attack it to give diaminocarbene ligands.
As the m(N„C) values of isocyanides adsorbed on Au
metal are approximately 80 cmꢁ1 higher than those of the
free isocyanides, these isocyanides react with amines. It
was these studies that led to the discovery of gold metal-
catalyzed reactions of isocyanides with amines and O2
(Eqs. (3) and (6)).
Results of the above studies can be reasonably under-
stood in terms of a mechanism (Scheme 2) that is related
to that proposed for the gold-catalyzed reactions (Eq.
(3)) of isocyanides with primary amines (Scheme 1). The
first step (1) in both of these mechanisms is amine attack
on an adsorbed isocyanide to give an isocyanide–amine
adduct a, which rearranges to the diaminocarbene interme-
diate b. The amine attack on adsorbed isocyanide is mod-
eled by analogous reactions of amines with isocyanide
ligands in metal complexes (Eq. (1)). The rate of this attack
in both mechanisms (Schemes 1 and 2) plays a major role in
influencing the overall rate of catalysis. In the reactions of
n-BuNC with secondary amines, the rate of this first step
(1) depends on the steric and electronic properties of the
reacting amine, but the rate is somewhat less than first
order in the amine concentration. The isocyanide–amine
adduct (a) resulting from this attack presumably rearranges
to the diaminocarbene intermediate b which reacts (step 3)
with O2 to give the urea product. The observed small
dependence of the rate of the gold-catalyzed reaction of
n-BuNC with HNPrn2 on the O2 concentration suggests that
step 3 (Scheme 2) is faster than step 1 but not by a large
amount. Details of step 3 are not known, and there are
no examples of reactions of diaminocarbene transition
metal complexes with O2 that lead to ureas. However, it
is possible that the diaminocarbene groups on the gold sur-
face add sufficient electron density to the gold metal that it
adsorbs O2, which reacts with the diaminocarbene interme-
diate b to give the urea product. Diaminocarbene ligands
Carbon monoxide (C„O), which is electronically-simi-
lar to isocyanides (C„N–R), might be expected to also
undergo attack by amines. In fact, reactions of amines with
CO ligands in metal complexes to form carbamoyl com-
plexes (Eq. (7)) occur when the m(C„O) value of the CO
O
+
+
ð7Þ
LxM C O
LxM C
+ 2 H2N–R
+ NH3–R
NHR
ligand is approximately 2000 cmꢁ1 or higher [39]. High
m(C„O) values are favored by metals (M) in positive oxi-
dation states and/or by ligands (L) in the complex that
are weakly electron-donating and/or strongly p-accepting
such as CO. Although detectable amounts of CO are
adsorbed on gold metal only at low temperatures (below
ꢀ150 K) [3,40], adsorbed CO has a relatively high
m(C„O) value of ꢀ2100 cmꢁ1, which suggests that it would
be susceptible to attack by amines. Despite the fact that lit-
tle CO would adsorb on gold at room temperature, that ad-
sorbed CO would be expected to undergo attack by
primary amines. If it were to react in the same way as iso-
cyanides with primary amines (Eq. (3)), one would expect
the reaction to give organic isocyanates (Eq. (8)). The final
products of the reaction would, however, be ureas
Au
O=C=N–R + H2O
H2NR
C O
+ H2NR + ½ O2
ð8Þ
O=C(NHR)2
R
R
N
R'
N
R
N
H
N
R
R'
N
N
C
R'
R'
(3)
(2)
(1)
N
C
O
H
R'
C
1
C
H
R'
O2
2
HNR'2
Au
Au
Au
(b)
(a)
Scheme 2. Mechanism for the Au-catalyzed reaction (Eq. (6)) of isocyanides (R–N„C) with secondary amines (HNR02) and O2.

852
R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
resulting from the subsequent reaction of the isocyanates
with additional amine.
R
N
R
N
O
C
H
O
O
C
H
C
H
H
(1)
A series of reactions (Table 2) of CO, a primary amine,
and O2 were conducted over stirred gold powder in MeCN
solvent at 45 °C [41]. The total gas pressure was approxi-
mately 1 atm with an approximately 2.5:1 ratio of CO
and O2, which were in stoichiometric excesses as compared
with the amine. Ureas were the major products after a 24 h
reaction time. The urea yields from reactions of three iso-
mers of butylamine decreased as the bulkiness of the butyl
group increased: n-BuNH2 (46%) > s-BuNH2 (12%) > t-
BuNH2 (2%). The less basic anilines, PhNH2 (21%) and
p-MeC6H4NH2 (24%), gave lower yields than n-BuNH2.
Thus, as in the reactions of isocyanides (Eqs. (3) and (6)),
the yields decrease as the size of the amine increases and
its basicity decreases.
The above trends suggest that the initial step in the reac-
tion (Scheme 3) is nucleophilic attack of the amine on
adsorbed CO to give a CO–amine adduct (a) or an
amino-oxocarbene (b) intermediate. It is also possible that
this intermediate is deprotonated to leave an anionic car-
bamoyl group (–C(@O)(NHR)ꢁ), as occurs in metal com-
plexes (Eq. (7)). Whatever is the nature of the
intermediate, it must transfer the two hydrogen atoms ini-
tially on the amine to an oxygen atom to give the putative
H2O product (which has not been specifically identified).
As for the analogous step (2 or 4 in Scheme 1) in reactions
of isocyanides with primary amines and O2 (Eq. (3)), there
is little basis for proposing details for step 2 in Scheme 3.
However, as noted for the isocyanide reactions, all of the
possible intermediates in the reaction (a, b, or the carbam-
oyl group) would donate electron density to the gold which
would make it more likely to activate O2.
or
H2NR
Au Au Au
Au Au Au
Au
(a)
(b)
(2)
O2
O
C
HNR'2
R–N=C=O + "H2O"
RNH
NR'2
(c)
Scheme 3.
CO and O2 under the conditions of the primary amine reac-
tions (Table 2). This would be expected because secondary
amines are unable to form isocyanates. (2) Small amounts
of the isocyanates are detected in some of the reactions
(entries 1, 4, 5, 6, 8 in Table 2). (3) When equimolar PhNH2
and HNPrn2 react with CO and O2 in the presence of gold
(entry 6 in Table 2), the product formed in highest yield
(19%) is the mixed urea PhNH(C@O)NPrn2. This result is
consistent with the initial formation of PhN@C@O, which
subsequently reacts with HNPrn2 to give the mixed urea
product. The higher known nucleophilicity [42] of HNPrn2,
as compared with PhNH2, would lead to preferential reac-
tion of HNPrn2 with PhNCO to give PhNH(C@O)NPrn2,
rather than PhNH(C@O)NHPh. Moreover, the reaction
of 0.10 mmol of PhN@C@O with 0.50 mmol of both
HNPrn2 and PhNH2 under the conditions in Table 2 (but
without gold) gives only the mixed urea PhNH(C@O)NPrn2.
(4) When equimolar n-BuNH2 and HNPrn2 are reacted
(entry 7), the ratio of products n-BuNH(C@O)NPrn2/(n-
BuNH)2C@O is 7.8. This is nearly the same ratio of these
products (7.2) as that obtained from the reaction of n-
BuNCO with equimolar n-BuNH2 and HNPrn2 in the
absence of gold; this result also supports the formation of
n-BuNCO as an intermediate in the catalytic reaction. (5)
In the reaction of equimolar n-BuNH2 and PhNH2 (entry
8), the major product is (n-BuNH)2C@O (41%) with the
A key assumption in the proposed mechanism (Scheme 3)
is that an isocyanate (R–N@C@O) is the initial product
in the reaction; it reacts with additional amine to give the
final urea product (Eq. (8)). There are several results that
support the proposal that an isocyanate is the initial prod-
uct: (1) The secondary amine HNPrn2 does not react with
Table 2
Gold metal-catalyzed reactions of primary amines with CO and O2 according to Eq. (8)a
Entry
Amine (mmol)
Products
Urea (%)
Isocyanate (%)
1
2
3
4
5
6
7
8
a
n-BuNH2 (0.5)
s-BuNH2 (0.5)
t-BuNH2 (0.5)
PhNH2 (0.5)
p-MeC6H4NH2 (0.5)
PhNH2/HNPrn2 (0.5/0.5)
n-BuNH2/HNPrn2 (0.5/0.5)
n-BuNH2/PhNH2 (0.5/0.5)
46
12
2
21
<1, trace
0
0
3
24
3
19b
39c, 5d
41d, 25e
3, PhNCO
0, n-BuNCO
2, PhNCO; trace, n-BuNCO
At 45 °C in 5 mL MeCN solvent for 24 h with 1.00 g of Au powder; CO and O2 total pressure = ꢀ1 atm.
b
c
PhNH(C@O)NPrn2.
(n-Bu)NH(C@O)NPrn2.
(n-BuNH)2C@O.
d
e
PhNH(C@O)NH(Bun).

R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
853
mixed urea PhNH(C@O)NHBun formed in 25% yield. This
is readily understood by considering n-BuNH2 as a stron-
ger nucleophile than PhNH2, which means that n-BuNH2
is more likely to attack adsorbed CO to give n-BuNCO,
which would react preferentially with n-BuNH2 to give
the major product (n-BuNH)2C@O. The smaller amount
of mixed urea product PhNH(C@O)NHBun could form
by competitive attack of PhNH2 on adsorbed CO and/or
on the n-BuNCO intermediate.
species were present in the solution, and the true catalyst
was the gold powder.
4. Oxidative-dehydrogenation of amines to imines
During the course of our studies of gold-catalyzed reac-
tions of isocyanides with amines and O2, we discovered that
secondary amines, in the absence of an isocyanide, were
oxidatively dehydrogenated to imines at rates that were
slower than their reactions with isocyanides. Thus, at
60 °C in MeCN solvent, the reaction of n-BuNC with piper-
idine ((CH2)5NH) and O2 over powdered gold (Table 1) was
nearly complete before significant amounts of the piperidine
oxidative-dehydrogenation product was detected. The sim-
plest examples of the oxidative-dehydrogenation of second-
ary amines occurred with di-i-propylamine, dibenzylamine,
and 1,2,3,4-tetrahydroisoquinoline (Table 3) to give the
corresponding imines (Eq. (9)) [44]. The presumed
Because nanoparticles of gold are so active in catalytic
reactions of CO [3,12–17], especially its oxidation with O2
to give CO2, it was important to establish that nanogold
was not the real catalyst of the reaction in Eq. (8). Of
course, it had been shown previously that bulk gold,
including gold powder, is a poor catalyst of CO oxidation
[14,24], which indicates that gold nanoparticles are not
present in bulk gold powder. So, it is unlikely that our gold
powder contains nanoparticles. However, as a further test,
we used our powdered gold to catalyze the reaction of CO
(0.7 atm) with O2 (0.3 atm) under the conditions in Table 2.
This reaction gave no evidence for CO2 at least above our
detectability limit of a 0.1% yield for the reaction. Thus, it
seems unlikely that nanogold particles are present in our
powdered gold catalysts. It is interesting that a nanogold
catalyst consisting of gold (<10 nm) supported on an ion
exchange styrene–divinylbenzene polymer (Merck Ion
Exchange IV) catalyzes reactions of anilines with CO
(ꢀ40 atm) and O2 (ꢀ10 atm) to give diarylureas [43]. It also
catalyzes reactions of anilines with MeOH, CO, and O2 to
give carbamates ArNH(C@O)OMe. These reactions were
conducted under much more strenuous conditions
(175 °C, ꢀ50 atm gas pressure) than those used in the gold
powder-catalyzed reactions (Eq. (8)) of anilines and other
amines (Table 2), which suggests that the gold nanoparti-
cles are less active than gold powder in the reaction in
Eq. (8). On the other hand, the reaction of n-BuNH2 with
CO and O2 under conditions in Table 2, except for the use
of MeOH solvent instead of MeCN, did not give the carba-
mate n-BuNH(C@O)OMe; only (n-BuNH)2C@O was
formed.
H
C
H
H
Au
+ ½ O2
R1
R1
C
+ H2O
ð9Þ
N
R2
R2
N
H
H2O product of the reaction was not experimentally iden-
tified, but the reaction required O2 because dibenzylamine
and 1,2,3,4-tetrahydroisoquinoline did not react when the
O2 atmosphere was replaced with argon. The reactions
were performed in MeCN solvent at 60 °C for 40 h or in
toluene at 100 °C for 24 h. It is evident from yields of the
imine products in entries 1, 2 and 3 that benzyl amines un-
dergo oxidative-dehydrogenation more rapidly than the al-
kyl amine (i-Pr)2NH. The yields of the imine products in all
of the reactions were significantly higher when they were
run at the higher temperature (100 °C) in toluene. A small
amount of isoquinoline, resulting from dehydrogenation of
the CH2–CH2 bond in the initial oxidative-dehydrogena-
tion product 3,4-dihydroisoquinoline, was detected when
the reaction in entry 3 was run for longer times at
Another concern was that the gold powder might form
soluble gold species under the conditions of the reactions,
and these species might be the actual catalysts of the reac-
tion in Eq. (8). These might be either gold (I or III) com-
plexes resulting from oxidation of the gold metal and
coordination by amine ligands or nanoparticles of a
Au(0) colloid that is stabilized by adsorbed amine ligands.
If a soluble form of gold were the catalyst, the reaction
solution should be catalytically active after the gold pow-
der was removed. To test for this possibility, a reaction
of n-BuNH2, CO, and O2 with 1.0 g of gold powder, run
under the conditions in Table 2, was stopped after 6 h when
the (n-BuNH)2C@O urea yield was only 13%. The reaction
solution was then decanted from the gold powder, and the
solution was returned to a CO/O2 atmosphere. After
reheating the solution to 45 °C with stirring for 15 h, no
additional urea was formed, which shows that no catalytic
Table 3
Gold-catalyzed aerobic oxidation of secondary amines to imines according
to Eq. (9) and Scheme 4a
Entry
1
Substrate
Product
Yield [%]
15b
H
N
N
N
N
2
3
4
35b; 64c
40b; 87c
H
NH
NH
N
46b; 93c
41b; 75c
NH
NH
N
N
N
5
a
Reaction conditions: amine (0.20 mmol), 5 mL of solvent, gold powder
(1.0 g), O2 (ꢀ1.0 L at ꢀ1 atm).
b
At 60 °C for 40 h in acetonitrile.
At 100 °C for 24 h in toluene.
c

854
R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
N
N
100 °C. Although the oxidative-dehydrogenation of sec-
ondary amines with O2 has been reported to be catalyzed
by Ru/Al2O3 [45], catalysis of this reaction by bulk gold
metal is surprising because of its poor catalytic activity in
other reactions [3].
Au powder (1.0 g)/O2
+
N
HN
CH3CN, 60 °C, 40 h
4a
91% yield
2a
(0.2 mmol)
1a
(0.2 mmol)
ð10Þ
The cyclic secondary amines pyrrolidine ((CH2)4NH),
piperidine ((CH2)5NH), and hexamethylene imine
((CH2)6NH), undergo a more complicated oxidative-dehy-
drogenation reaction that results in the coupling of two of
the cyclic amines (Scheme 4). In MeCN solvent at 60 °C,
the yields of the products (Table 3) decrease as the ring size
increases from the 5-membered amine (46%) to the 7-mem-
bered amine (22%). The same products were obtained pre-
viously by the palladium-black catalyzed reaction of these
amines under an inert atmosphere (argon), rather than
O2, at 80 °C [46]. A proposed mechanism for the coupling
of the cyclic secondary amines is presented in Scheme 4. It
involves initial oxidative-dehydrogenation (step 1) of the
amine (1) to give the cyclic imine (2). This step is analogous
to the oxidative-dehydrogenation reactions (Eq. (9)) that
non-cyclic secondary amines undergo (entries 1, 2, and 3
in Table 3) under the same conditions. However, the less
sterically-crowded cyclic imines react further to undergo
attack by additional amine to give 3. There is no GC/MS
evidence for intermediate 2 or 3 in samples of the reaction
solution taken during the course of the reaction, which
means that 2 and 3 must react faster than 1, and step 1 is
the rate-determining step in the overall conversion of 1 to
4. To demonstrate that step 2 could occur under the condi-
tions of the reaction, D1-pyrroline (2a) was prepared [47]
and reacted with 1a in the presence of gold powder under
1 atm of O2 (Eq. (10)). This reaction gives product 4a with-
out any evidence for intermediate 3a. As
step 2 does not involve the reaction of O2, the reaction of
2a and 1a was conducted also under an argon atmosphere.
However, there was no reaction, which suggests that the
equilibrium in step 2 lies to the left, which means that the
oxidative-dehydrogenation of 3a to give 4a (step 3) pro-
vides the overall driving force for the conversion of 2a to
4a (Eq. (10)). (Another possible reason that the conversion
of 2a to 3a does not occur in the absence of O2 is that the
O2 activates the gold in some way.) It should also be noted
that the reaction of 2a with 1a (Eq. (10)), which is nearly
complete after only 20 h, is significantly faster than the
overall conversion of 1a to 4a.
If 3 is indeed an intermediate in the conversion of 1 to 4,
the oxidative-dehydrogenation of 3 must be fast as com-
pared with step 1 and with the overall conversion of 1 to
4. However, compounds 3a–c are unknown, perhaps
because the equilibrium in step 2 lies to the left, as sug-
gested by the lack of a reaction between 2a and 1a (Eq.
(10)) in the absence of O2, as discussed above. So, com-
pound 5 (in Scheme 5), which contains secondary and ter-
tiary amine groups in the same relative positions as in 3b,
was chosen as a model for the conversion of 3b to 4b. When
5 was reacted with O2 under the standard conditions (Table
3), the formation of the imine 6 was nearly complete in only
10 h which is much faster than the oxidative-dehydrogena-
tion of any of the secondary amines in Table 3. Therefore,
step 3 in Scheme 4 is indeed likely to be much faster than
the overall conversion of 1 to 4 and is consistent with step
1 being the rate-determining step in the overall conversion.
As a side note, one can understand why the oxidative-
dehydrogenation of 5 to 6 is faster than that of 1b to 2b
by considering the effect of amine groups on the dehydro-
genation of hydrocarbons. The dehydrogenation (–H2) of
nitrogen-containing hydrocarbons, e.g., R2N–CH2–
CH3 ? R2 N–CH@CH2 + H2, is thermodynamically more
favorable than the dehydrogenation of simple hydrocarbon
analogs, e.g., H3C–CH3 ? H2C@CH2 + H2 [48]. It has
H
N
N
Au/O2
N
(CH2)n
–2 H2O
n(H2C)
n(H2C)
(1)
(4)
–H2O
Au/O2
–H2O
(3)
(1)
Au/O2
NH
N
N
N
H
N
N
4b
3b
1/Au/O2
N
(CH2)n
(2)
n(H2C)
n(H2C)
(2)
(3)
Au powder (1.0 g)/O2
CH3CN, 60 oC, 40 h
HN
N
N
N
n
0
1
2
1, 2, 3, 4
a
b
c
5
6
(97% yield)
(0.2 mmol)
Scheme 4.
Scheme 5.

R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
855
also been found that the dehydrogenation of nitrogen-con-
taining hydrocarbons is kinetically faster than the dehydro-
genation of simple hydrocarbons over Pd/C and Rh/C
heterogeneous catalysts at 110 °C [48]. It is therefore rea-
sonable to assume that it is the tertiary amine group in 5
(and presumably in 3a–c) that makes the oxidative-dehy-
drogenation of 5 faster than that of 1a–c and the other
amines in Table 3.
because the metal–diaminocarbene bond is so resistant to
oxidative degradation [49]. Clearly, the gold metal surface
must play a role in cleaving the metal–diaminocarbene
bond. However, O2 does not adsorb on bulk gold in detect-
able amounts at room temperature. So, it is presumably the
strongly electron-donating ability of the diaminocarbene
group [50] that makes the gold sufficiently electron-rich
to react with O2.
To determine whether or not the powdered gold was
converted to catalytically-active soluble species such as
gold colloids or gold complexes, the reaction of piperidine
(1b) with O2 (1 atm) in the presence of gold powder (1.0 g)
at 60 °C in MeCN solvent (same conditions as in Table 3)
was stopped after 6 h when the imine (4b) yield was only
9%. After the reaction solution was decanted from the gold
powder, the solution was returned to an O2 atmosphere
and reheated to 60 °C for 15 h. The yield of 4b remained
at 9% indicating that there were no soluble gold species
that were the actual catalysts of these reactions. Thus, as
in the reactions of primary amines with CO and O2, it is
the bulk gold powder that is the actual catalyst of the oxi-
dative-dehydrogenation of secondary amines.
In contrast to the reactions (path 1) of diaminocarbene
intermediates (A in Scheme 6) that lead to urea products,
carbene intermediates derived from primary amines react
(path 2) with O2 by transferring two of its hydrogen atoms
to O2 to give H2O and carbodiimide products. It is not
clear why this intermediate A prefers to give a carbodiimide
(path 2) rather than a urea (path 1). It is evident that the
reactivity of diaminocarbene groups on gold metal surfaces
requires further study.
By analogy with reactions of adsorbed isocyanides with
amines (Scheme 6), carbon monoxide reacts (Eq. (8))
readily with primary amines and O2 to give isocyanates
(R–N@C@O). The first step in this reaction is presumably
amine attack on an adsorbed CO, whose m(C„O) value is
sufficiently high that such an attack is expected on the basis
of reactions of CO ligands in transition metal complexes.
The nature of the resulting intermediate (Scheme 3) is less
well-defined than that in the isocyanide reactions (Scheme
6), but it reacts with O2 to give the isocyanate product
while transferring two hydrogen atoms to O2. It is presum-
ably the strong electron-donating ability of the intermedi-
ate that activates the gold to react with O2.
The final topic of this report is the oxidative-dehydroge-
nation of secondary amines. This is a very different type of
reaction than those of isocyanides and CO. While it must
involve cleavage of N–H and C–H bonds, together with a
combining of hydrogen atoms with O2, any mechanistic
proposal for this reaction would be highly speculative at
this stage. Attempts to catalyze the oxidative-dehydrogena-
tion of secondary amines (Eq. (9)) [44], as well as the oxi-
dative-amination of CO (Eq. (8)) [41], using copper and
silver metal powders were unsuccessful.
One of the important results of these studies is that
they demonstrate that bulk gold metal is capable of high
catalytic reactivity in certain reactions. Although nano-
gold catalysis is now a subject of intense current study,
it is worth taking another look at the catalytic properties
of bulk gold metal, the type of gold that is normally cho-
sen for its inertness in decorative and electronic
applications.
5. Concluding comments
An initial goal of these investigations was to determine
whether or not the reactivity of isocyanides adsorbed on
gold metal could be predicted from our understanding of
isocyanide reactivity in transition metal complexes. On
the basis of the m(N„C) values of isocyanides adsorbed
on gold metal, it was expected that amines would attack
them to give diaminocarbene groups. The observed reac-
tions (Scheme 6) of adsorbed isocyanides with primary
and secondary amines can be readily understood by assum-
ing the amines attack the adsorbed isocyanides. High
m(N„C) values for isocyanides adsorbed on a variety of
other metals [11] suggest that these metals too will activate
isocyanides to nucleophilic attack by amines and other
nucleophiles. After the initial attack, the resulting diamino-
carbene reacts with O2 by two different pathways depend-
ing on whether the amine is a secondary (path 1) or
primary amine (path 2). Path (1) leads to cleavage of the
Au–C bond with the formation of the urea. Although this
appears to be a logical reaction, it is not a reaction that has
been reported previously for diaminocarbene metal com-
plexes. In fact, metal complexes with diaminocarbene
ligands are often used as catalysts for oxidative reactions
R
N
R1
N
Acknowledgments
R2 = alkyl
(1)
R
N
C
R
N
R1
N
R2
C
O
H
O2
+ HNR1R2
H
I thank Diana Mihaela Lazar and Bolin Zhu for their
key contributions and personal qualities that they brought
to these investigations. This research was supported by the
US Department of Energy under Contract No. DE-AC02-
07CH11358 with Iowa State University.
R2
C
Au
Au
R
2 = H
(2)
R–N=C=N–R1 + H2O
A
Scheme 6.

856
R.J. Angelici / Journal of Organometallic Chemistry 693 (2008) 847–856
[28] B. Crociani, in: P. Braterman (Ed.), Reactions of Coordinated
References
Ligands, vol. 1, Plenum Press, New York, 1986, pp. 553–638.
[29] U. Belluco, R.A. Michelin, P. Uguagliati, J. Organomet. Chem. 250
(1983) 565.
[30] T. Saegusa, Y. Ito, S. Kobayashi, K. Hirota, H. Yoshioka, Bull.
Chem. Soc. Jpn. 42 (1969) 3310.
[1] F. Zaera, Top. Catal. 34 (2005) 129.
[2] M.R. Albert, J.T. Yates Jr., The Surface Chemist’s Guide to
Organometallic Chemistry, American Chemical Society, Washington,
DC, 1987.
[3] G.C. Bond, C. Louis, D.T. Thompson, Catalysis by Gold, Imperial
College Press, London, 2006.
[31] B. Zhu, R.J. Angelici, unpublished results.
[32] Y. Ito, T. Hirao, T. Saegusa, J. Org. Chem. 40 (1975) 2981.
[33] A.F. Benton, J.C. Elgin, J. Am. Chem. Soc. 49 (1927) 2426.
[34] P. Landon, P.J. Collier, A.J. Papworth, C.J. Kiely, G.J. Hutchings,
Chem. Commun. (2002) 2058.
[35] J.K. Edwards, B. Solsona, P. Landon, A.F. Carley, A. Herzing, M.
Watanabe, C.J. Kiely, G.J. Hutchings, J. Mater. Chem. 15 (2005)
4595.
[36] M. Lazar, B. Zhu, R.J. Angelici, J. Phys. Chem. C 111 (2007) 4074.
[37] K.-C. Shih, R.J. Angelici, Langmuir 11 (1995) 2539.
[38] (a) M.-T. Lee, C.-H. Lu, Organometallics 23 (2004) 976;
(b) R.A. Diggle, S.A. Macgregor, M.K. Whittlesey, Organometallics
23 (2004) 1857;
[4] M.L. Kuznetsov, Russ. Chem. Rev. 71 (2002) 265;
M.L. Kuznetsov, Uspekhi Khim. 71 (2002) 307.
[5] L. Weber, Angew. Chem., Int. Ed. 37 (1998) 1515.
[6] Y. Yamamoto, Coord. Chem. Rev. 32 (1980) 193.
[7] D. Lentz, Angew. Chem., Int. Ed. Engl. 33 (1994) 1315.
[8] E. Singleton, H. Oosthuizen, Adv. Organomet. Chem. 22 (1983) 209.
[9] P.M. Treichel, Adv. Organomet. Chem. 11 (1973) 21.
[10] (a) L. Malatesta, F. Bonati, Isocyanide Complexes of Metals, Wiley,
New York, 1969;
(b) F.E. Hahn, Angew. Chem., Int. Ed. Engl. 32 (1993) 650.
[11] D. Lazar, R.J. Angelici, in: J.-M. Basset, R. Psaro, D. Roberto, R.
Ugo (Eds.), Modern Surface Organometallic Chemistry, Wiley-VCH
Verlag, Weinheim, in press.
[12] G.C. Bond, D.T. Thompson, Appl. Catal. A: Gen. 302 (2006) 1.
[13] D.T. Thompson, Top. Catal. 38 (2006) 231.
[14] G.C. Bond, D.T. Thompson, Catal. Rev. Sci. Eng. 41 (1999) 319.
[15] M.S. Chen, D.W. Goodman, Catal. Today 111 (2006) 22.
[16] (a) G.J. Hutchings, M. Haruta, Appl. Catal. A: Gen. 291 (2005) 2;
(b) G.J. Hutchings, Catal. Today 100 (2005) 55.
[17] B.K. Min, C.M. Friend, Chem. Rev. 107 (2007) 2709.
[18] B.P. Block, Inorg. Synth. 4 (1953) 14.
[19] M.J. Robertson, R.J. Angelici, Langmuir 10 (1994) 1488.
[20] A.C. Ontko, R.J. Angelici, Langmuir 14 (1998) 1684.
[21] M. Lazar, R.J. Angelici, J. Am. Chem. Soc. 128 (2006) 10613.
[22] Ref. 14, page 350, Fig. 5.
`
(c) G. Frenking, M. Sola, S.F. Vyboishchikov, J. Organomet. Chem.
690 (2005) 6178.
[39] R.J. Angelici, Acc. Chem. Res. 5 (1972) 335.
[40] M. Chen, Y. Cai, Z. Yan, D.W. Goodman, J. Am. Chem. Soc. 128
(2006) 6341.
[41] B. Zhu, R.J. Angelici, J. Am. Chem. Soc. 128 (2006) 14460.
[42] S.G. Entelis, O.V. Nesterov, Russ. Chem. Rev. 35 (1966) 917.
[43] F. Shi, Y. Deng, J. Catal. 211 (2002) 548.
[44] B. Zhu, R.J. Angelici, Chem. Commun. (2007) 2157.
[45] K. Yamaguchi, N. Mizuno, Angew. Chem., Int. Ed. 42 (2003)
1480.
[46] (a) S.I. Murahashi, N. Yoshimura, T. Tsumiyama, T. Kojima, J.
Am. Chem. Soc. 105 (1983) 5002;
(b) N. Yoshimura, I. Moritani, T. Shimamura, S.-I. Murahashi, J.
Chem. Soc., Chem. Commun. (1973) 307.
[47] M. Ochiai, M. Inenaga, Y. Nagao, R.M. Moriarty, R.K. Vaid, M.P.
Duncan, Tetrahedron Lett. 29 (1988) 6917.
[48] A. Moores, M. Poyatos, Y. Luo, R.H. Crabtree, New J. Chem. 30
(2006) 1675.
[49] M.M. Rogers, S.S. StahlTopics in Organometallic Chemistry, vol. 21,
Springer, Berlin/Heidelberg, 2007, p. 21.
[50] F. Glorius, N-Heterocyclic carbenes in transition metal catalysisTop-
ics in Organometallic Chemistry, vol. 21, Springer, Berlin/Heidelberg,
2007, pp. 21.
[23] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal. 115 (1989)
301.
[24] W.C. Ketchie, Y.-L. Fang, M.S. Wong, M. Murayama, R.J. Davis,
J. Catal. 250 (2007) 94.
[25] J.C. Love, L.A. Estroff, J.K. Kriebel, R.G. Nuzzo, G.M. Whitesides,
Chem. Rev. 105 (2005) 1103.
[26] R.A. Michelin, A.J.L. Pombeiro, M.F.C. Guedes da Silva, Coord.
Chem. Rev. 218 (2001) 75.
[27] J. Chatt, R.L. Richards, G.H.D. Royston, J. Chem. Soc., Dalton
Trans. (1973) 1433.