Gold-catalyzed cyanosilylation reaction: Homogeneous and heterogeneous pathways

Article abstract of DOI:10.1002/chem.200601791

Gold had been considered to be an extremely inert metal, but recently it was found that nanometersized gold particles on metal-oxide supports acted as catalysts for simple organic reactions, such as oxidation and hydrogenation, even at or below room tempe

Full text of DOI:10.1002/chem.200601791

FULL PAPER
DOI: 10.1002/chem.200601791
Gold-Catalyzed Cyanosilylation Reaction:
Homogeneous and Heterogeneous Pathways
Woo Kyung Cho, Jungkyu K. Lee, Sung Min Kang, Young Shik Chi, Hee-Seung Lee, and
InsungS. Choi* [a]
Abstract: Gold had been considered to
be an extremely inert metal, but re-
cently it was found that nanometer-
sized gold particles on metal-oxide sup-
ports acted as catalysts for simple or-
ganic reactions, such as oxidation and
hydrogenation, even at or below room
temperature. Herein, we report that
gold nanoparticles (AuNPs) of zero ox-
idation state (Au0) are catalytically
proceeded smoothly at room tempera-
ture with 0.2 wt% loading of AuNPs.
The reactions of aromatic aldehydes
were almost quantitative, except for
benzaldehyde derivatives containing
the electron-withdrawing NO2 group,
and a,b-unsaturated aromatic alde-
hydes were the most reactive sub-
strates. The reactions also went
smoothly for aliphatic aldehydes.
Mechanistic studies indicated that the
reactions proceeded both homogene-
ously and heterogeneously: homogene-
ous catalysis by leached gold species
and heterogeneous catalysis by the ad-
sorption of the reactants (aldehydes
and trimethylsilyl cyanide) onto
AuNPs. The ratio of homogeneous and
heterogeneous catalysis was estimated
to be ꢀ4:1.
Keywords: aldehydes · cyanosilyla-
À
active for a C C bond-forming reac-
tion
·
gold
·
heterogeneous
tion, the cyanosilylation of aldehydes.
The AuNP-catalyzed cyanosilylation
catalysis · homogeneous catalysis
Introduction
ture, AuNP-catalyzed CO oxidation.[2] Since the discovery
that metal-oxide-supported AuNPs were catalytically active
for CO oxidation, the scope of AuNP-catalyzed reactions
has been expanded to other reactions, such as selective oxi-
dation of alcohols,[3] epoxidation of propene,[4] NO reduction
with O2 and H2,[5] selective oxidation of cyclohexene,[6] hy-
drogenation of unsaturated hydrocarbons,[7] the water–gas
shift reaction,[8] and formation of carbodiimide between iso-
cyanide and primary amines.[9] Most of these reported reac-
tions could be characterized by 1) relatively simple chemical
transformations (that is, oxidation and hydrogenation),
2) use of metal-oxide (or active carbon or polymer)-support-
ed AuNPs as a catalyst, and 3) involvement of gases (for ex-
ample, O2, CO, H2, etc.) as a reactant. Albeit the reported
reactions are important themselves, there have been few re-
Gold is one of the least reactive metals, and has long been
considered as a poorly active catalyst. However, recently
there have been a growing number of reports that gold
nanoparticles (AuNPs), especially metal-oxide-supported
AuNPs, can serve as very efficient catalysts. Among the re-
ported reactions, the most famous example is the AuNP-cat-
alyzed CO oxidation[1] at very low temperature. This reac-
tion has attracted great attention experimentally and theo-
retically because it has commercial value. The Pt/Pd cata-
lysts that are currently used in cars for CO oxidation work
only at temperatures above 2008C. Therefore, CO pollution
is unavoidable during the first five minutes after starting the
engine. This problem can clearly be solved by low-tempera-
À
ports on the catalytic activities of AuNPs for C C bond-
forming reactions,[10] which are ubiquitous and one of the
most widely utilized organic transformations in synthetic or-
ganic chemistry.[11]
The catalytic activity of AuNPs has been attributed to the
adsorption of reactants, such as O2 and CO, onto AuNPs
(and/or Au/metal oxide interfaces).[1b,c] Of interest, the reac-
tion of CO with O2 has been known to take place on the
surfaces of AuNPs with an apparent activation energy of
almost 0 kJmolÀ1.[12] The selective hydrogenation of the C=
[a] W. K. Cho, J. K. Lee, S. M. Kang, Y. S. Chi, Prof. Dr. H.-S. Lee,
Prof. Dr. I. S. Choi
Department of Chemistry and School of Molecular Science (BK21)
Center for Molecular Design and Synthesis
KAIST
Daejeon 305-701 (Korea)
Fax: ( +82)42-869-2810
E-mail: ischoi@kaist.ac.kr
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2007, 13, 6351 – 6358
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6351

O versus the C=C group in a,b-unsaturated aldehydes has
been thought to be caused by the preferential adsorption of
the C=O groups onto AuNPs.[7a] In addition, the activation
of CO2 (presumably by CO2 adsorption onto AuNPs) has
been utilized to form cyclic carbonates and disubstituted
ureas.[13] Based on these findings, we reasoned that AuNPs
À
could be catalytically active for some C C bond-forming re-
actions, in which the activation of carbonyl groups (or possi-
bly any functional groups containing lone pairs) is required,
À
and herein we report the catalytic activity of AuNPs for C
C bond-forming reactions. We chose the cyanosilylation of
aldehydes (reaction of aldehydes with trimethylsilyl cyanide
(TMSCN))[14] as a model reaction for investigating the cata-
lytic activity of AuNPs, because the reactants had carbonyl
Scheme 1. AuNP-catalyzed cyanosilylation of aldehydes.
and lone-pair-containing cyano groups (Scheme 1). In addi-
tion, cyanohydrin trimethylsilyl ethers are industrially valua-
ble intermediates in the synthesis of cyanohydrins, b-amino
alcohols, a-hydroxy acids, and other biologically active com-
pounds.[15] It was reported that no reaction occurred be-
tween aldehydes and TMSCN in the absence of a catalyst.[16]
In this work, we synthesized AuNPs of Au0 oxidation
state[17] and studied the cyanosilylation of various aldehydes
at room temperature with AuNPs as catalyst. The reaction
mechanism was investigated in detail to determine whether
the AuNP-catalyzed cyanosilylation was heterogeneous and/
or homogeneous.
Figure 1. Characterization of synthesized AuNPs: a) TEM image, b) size
distribution (average diameter: 3.9 nm), and c) X-ray photoelectron spec-
trum.
The binding energies for Au 4f7/2 and Au 4f5/2 were observed
to be 83.7 and 87.4 eV, respectively; the peak-to-peak dis-
tance was 3.7 eV. These values are consistent with the Au0
oxidation state.[17] If AuI had been present, a peak (or at
least a shoulder on the Au0 4f5/2 peak) would have been ex-
pected near 84.9 eV. However, we did not observe any peak
or shoulder from AuI in the X-ray photoelectron spectrum.
The absence of Cl signals indicates that the surface of the
AuNPs was not contaminated by ionic species, such as
Results and Discussion
Synthesis and characterization of AuNPs: Figure 1a and b
show a typical transmission electron microscopy (TEM)
image and the size distribution of the synthesized AuNPs,
respectively. The average diameter of the AuNPs and its
standard deviation were 3.9 nm and 0.97, respectively, which
were obtained by counting at least 150 particles. The synthe-
sized AuNPs were also analyzed by X-ray photoelectron
spectroscopy (XPS). In the survey scan, peaks from gold
were shown at about 56.9 (5p3/2), 83.7 (4f7/2), 87.4 (4f5/2),
335.0 (4d5/2), 352.9 (4d3/2), and 546.8 eV (4p3/2) (Figure 1c).
Carbon 1s (285.0 eV) and oxygen 1s (531.9 eV) peaks were
also observed. The weak signal of bromine arose from N-
À
AuIIICl4 and AuICl2À.[10b,18] In addition, AuNPs were not dis-
solved in polar solvents, such as water and methanol. Taken
together, the characterizations confirmed that the synthe-
sized AuNPs had the Au0 oxidation state.
Cyanosilylation of aldehydes: First, we investigated the cata-
lytic activity of 1-octanethiol-protected AuNPs toward cya-
nosilylation of the a,b-unsaturated aldehyde, trans-cinnamal-
(C8H17)4Br, which was used as a phase-transfer reagent in
1
the synthesis of AuNPs. The silicon peaks were due to the
preparation method for sampling (see Experimental Sec-
tion).
dehyde. The H NMR spectrum showed almost 100% con-
version of the aldehyde to the corresponding cyanohydrin
trimethylsilyl ether after stirring for one hour. The peak at
À
The inset in Figure 1c represents the XPS signature of the
Au 4f doublet (4f7/2 and 4f5/2). The peak position, line shape,
and peak-to-peak distance of the Au 4f doublet are the im-
portant criteria to determine the oxidation state of gold.
9.57 ppm ( CHO) disappeared completely and a new peak
À
at 5.11 ppm ( CHCNOTMS) was observed. The product
was purified by column chromatography (hexane/ethyl ace-
tate 10:1) in the form of the trimethylsilyl ether (yield of
6352
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6351 – 6358

Mechanism of Gold-Catalyzed Cyanosilylation
FULL PAPER
isolated product 92%). In contrast, we could not observe
any conversion at all in the absence of AuNPs, even after
stirring for seven days at room temperature, which confirms
that AuNPs were required for cyanosilylation. The resulting
trimethylsilyl ethers of cyanohydrins were not stable during
silica-gel chromatography except for trans-cinnamaldehyde,
and thus the reaction yields were obtained by 1H NMR
spectroscopy with anisole as an internal standard. Table 1
Table 1. Cyanosilylation of aldehydes in the presence of AuNPs.
Entry
Substrate
t [h]
Yield [%][a]
1
4
93
2
1
100 (92)[b]
Figure 2. AFM image of bare gold thin films.
3
4
5
6
12
4
100
98
Interactions between reactants (aldehydes and TMSCN)
and AuNPs: Recent studies suggest that both CO and O2
are adsorbed onto the surface of AuNPs to near saturation
in the AuNP-catalyzed CO oxidation and that the rate-de-
termining step is the reaction between the two adsorbed re-
actants (CO and O2).[21] For the AuNP-catalyzed C=O hy-
drogenation of acrolein, the edges of AuNPs were identified
as active sites for favoring the adsorption of C=O groups of
acrolein and subsequent reaction to allyl alcohol.[7a] In addi-
tion, the cyclization of epoxides and the carbonylation of
amines have been known to occur through the adsorption-
driven activation of CO2.[13] Based on these reports, we
thought that aldehydes could be activated by adsorption
onto AuNPs in the cyanosilylation of aldehydes and
TMSCN.
48
94
72
22
7
8
3
4
97
100
[a] NMR spectroscopic yield. [b] Yield of isolated product.
shows the reactant scope of AuNP-catalyzed cyanosilyla-
tion.[19] The reactions of aromatic aldehydes were almost
quantitative (Table 1, entries 1–4), except for benzaldehyde
derivatives containing an electron-withdrawing NO2 group
(Table 1, entries 5,6), and the reactions went smoothly in the
case of aliphatic aldehydes (Table 1, entries 7,8). a,b-Unsa-
turated aldehydes seemed to be the most reactive: the con-
version was quantitative in a short period of time (Table 1,
entries 2,7).[20]
NMR spectroscopy was used to verify the interactions be-
tween TMSCN and AuNPs. The addition of TMSCN to a
CDCl3 solution of AuNPs shifted a peak of the TMS group
( Si
A
À
trum and another peak of the TMS group ( Si(CH3)3) from
À10.7 to 7.8 ppm in the 29Si NMR spectrum.[22] The observed
downfield shifts in the NMR spectra indicate that TMSCN
was adsorbed onto AuNPs.[16]
We found that a gold thin film was also effective in cata-
lyzing the cyanosilylation of aldehydes. The film was pre-
pared by thermal evaporation of titanium (5 nm) and gold
(100 nm) onto a silicon wafer (Si/SiO2). The atomic force
microscopy (AFM) image shows nanometer-sized features
of gold: the hemispherical features had a lateral dimension
of 50 nm on average (Figure 2). We immersed a freshly
cleaned piece of gold (ꢀ1”2 cm) in a CDCl3 solution con-
Organocyanides have been reported to be adsorbed onto
Au and other noble metals.[23] From studies of metal cyanide
complexes and the cyanide adsorbed onto metal surfaces by
using electron energy-loss spectroscopy (EELS), XPS, and
other techniques,[24] it has generally been accepted that the
linear coordination (s bonding) through the lone-pair elec-
ꢁ
taining
trans-cinnamaldehyde
(0.5 mmol),
TMSCN
trons of the nitrogen atom results in an increase of the C N
(0.6 mmol), and anisole (0.5 mmol), and then the mixture
was stirred at room temperature. The reaction went slower
than with AuNPs: after 9 and 24 hours the conversion was
47 and 78%, respectively. However, when we increased the
reaction temperature to 408C, the reaction was complete in
four hours. While we had difficulty in recycling AuNPs due
to the unavoidable loss of AuNPs at each recycling step, the
gold thin film made it easy to perform the recycling: we suc-
cessfully recycled the gold thin film at least five times.
stretching frequency compared with that of the free organo-
cyanide molecule. On the other hand, the coordination
ꢁ
through the C N p system has been known to result in a de-
ꢁ
crease of the C N stretching frequency. We investigated the
adsorption of TMSCN onto gold thin films by using IR spec-
ꢁ
troscopy (Figure 3). For free TMSCN molecules, the C N
stretching was observed at 2191 cmÀ1, which was consistent
with the reported value.[25] When a gold thin film immersed
in the TMSCN solution (CDCl3 phase) was taken out,
Chem. Eur. J. 2007, 13, 6351 – 6358
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6353

I. S. Choi et al.
Figure 3. IR spectra of the gold thin film a) before and b) after the ad-
sorption of TMSCN.
rinsed with chloroform, and analyzed by IR spectroscopy,
À1
ꢁ
the peak of the C N stretching appeared at 2156 cm . The
ꢁ
À1
C N stretching band was red-shifted by 35 cm . The red
shift indicated that the adsorption of TMSCN onto the gold
Figure 4. TEM images of AuNPs after the cyanosilylation.
thin film, composed of hemispherelike gold nanostructures,
ꢁ
occurred through the C N p system. In other words, the red
shift can be attributed to the combined effects of p(CN)-to-
metal electron donation and metal-to-p*(CN) back-dona-
tion.[23b] The red shift observed in the adsorption of TMSCN
onto gold is opposite to the case of isocyanides. It was re-
AuNPs and newly generated species might catalyze the reac-
tion. In particular, the TMSCN-induced dissolution of gold
would generate AuCN or AuCN2 . Therefore, we investigat-
À
ed the catalytic activity of AuCN in the reaction. When we
ported that the adsorption of organoisocyanides onto gold
stirred a mixture of trans-cinnamaldehyde (0.5 mmol),
[9]
ꢁ
led to an increase of the N C stretching frequency.
TMSCN (0.6 mmol), and AuCN (1 mol%) in CDCl3 for five
hours at room temperature, we observed about 35% conver-
sion. However, we did not observe any further increase in
the conversion even after three days. As a comparison,
trans-cinnamaldehyde was completely converted to trime-
thylsilyl ether in one hour with AuNPs as catalyst. In addi-
tion, KAuCN2 did not catalyze the cyanosilylation of trans-
cinnamaldehyde. Therefore, we thought that homogeneous
catalysis participated in the reaction to some extent, but the
exact species for the catalysis remains to be seen.
The TMSCN-induced dissolution of AuNPs would gener-
ate AuCN or its derivatives, but the observed catalytic effi-
ciency of AuNPs could not be fully explained by only the
catalytic activity of AuCN or its derivatives. Another possi-
ble species generated in the reaction mixture might be gold
chloride species (AuClx), and this hypothesis was scrutinized
experimentally. Recently, several researchers have reported
that AuClx can be used as catalysts for some reactions, such
as phenol synthesis,[26a,b] ring expansion of cycloalcohols,[26c]
and the formation of heterobicyclic alkenes.[26d] We exam-
ined whether AuNPs were dissolved by CDCl3 and whether
Dissolution of Au: The morphological change of AuNPs
before and after the reaction was characterized by TEM.
After the reaction, the TEM image showed the presence of
an amorphous layer on the AuNPs (Figure 4a). This layer
was thought to be the product that stuck on the surface of
AuNPs, which is in agreement with a previous report.[7a] Of
interest, we also observed AuNPs the diameter of which was
smaller than that of AuNPs before the reaction, indicative
of the dissolution of gold during the reaction (Figure 4b).
The dissolution of gold was also indirectly supported by in-
ductively coupled plasma mass spectrometry (ICP-MS). We
stirred a mixture of TMSCN and a gold thin film in CDCl3
for 72 hours, and analyzed the solution phase by ICP-MS
(we used a gold thin film instead of AuNPs, because it was
practically difficult to separate the CDCl3 phase from
AuNPs completely). The ICP-MS data showed a small but
detectable quantity of gold (<1 ppm) from the CDCl3
phase. Therefore, the TEM and ICP-MS analyses suggest
that TMSCN (and/or other species) dissolved gold from
6354
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6351 – 6358

Mechanism of Gold-Catalyzed Cyanosilylation
FULL PAPER
or not dissolution might generate AuClx. It was practically
difficult to entirely exclude the dispersed AuNPs from the
CDCl3 solution, and therefore the gold thin film was used
instead of AuNPs. The film was immersed in CDCl3 for 72
hours and then it was removed from the solution. We did
not detect any gold species from the solution phase in the
ICP-MS spectrum (<2 ppb), which indicates that CDCl3
itself did not dissolve AuNPs.
trans-cinnamaldehyde was complete in one hour. The slower
conversion would make it easier to monitor the reaction.
Figure 5 shows the results of mercury poisoning experi-
Solvent effect: The catalytic activity of AuNPs toward the
cyanosilylation of trans-cinnamaldehyde was also investigat-
ed with other solvents that did not contain chlorine. Table 2
Table 2. Cyanosilylation of trans-cinnamaldehyde with AuNPs in non-
chlorinated solvents.
Solvent
t [h]
Yield [%][a]
toluene
benzene
THF
DMF
DMSO
72
72
3
72
72
84
83
90
7–8
0
Figure 5. Mercury poisoning experiments with m-anisaldehyde. a) The re-
action was performed without mercury. b) Mercury was added to the
mixture after 1 h of reaction. c) Mercury was added at the beginning of
the reaction. d) A mixture of mercury and AuNPs was stirred for three
days and then the reactants were added.
[a] NMR spectroscopic yield.
shows that AuNPs were catalytically active for the cyanosi-
lylation in non-chlorinated solvents. AuNPs were most cata-
lytically active in THF and inactive in relatively polar sol-
vents, such as DMF and DMSO. We presume that the poor
solubility/suspensibility of AuNPs in DMF and DMSO
caused no (or little) catalytic activity in those solvents. The
ICP-MS spectrum confirmed that no gold species existed in
THF (<5 ppb). These results also indicate that AuCl (or its
derivatives) was not involved in the reaction.
ments. Without mercury, the conversion increased continu-
ously with time and reached 100% in 12 hours. When we
added mercury to the mixture after reaction for one hour,
we did not observe a complete suppression of the reaction,
but observed a decreased conversion. The reaction became
suppressed two hours after the addition of mercury and the
conversion was 80% in 12 hours (Figure 5b). The mercury
poisoning seemed to be slower than usual, therefore we
added mercury to the reaction mixture at the beginning of
the reaction. In this case, the suppression occurred a little
earlier, but the final conversion after reaction for 12 hours is
the same (80%; Figure 5c). Of interest, when we stirred a
mixture of mercury and AuNPs in CDCl3 for three days and
added the reactants (m-anisaldehyde and TMSCN), we ob-
served 20% conversion in 12 hours (Figure 5d).
Mercury poisoningexperiments : Based on all the results dis-
cussed above, we thought that the AuNP-catalyzed cyanosi-
lylation proceeded by both heterogeneous and homogene-
ous mechanisms. Homogeneous catalysis would be achieved
by leached gold species (presumably AuCNx). The adsorp-
tion of the reactants (aldehydes and/or TMSCN) onto
AuNPs of zero oxidation state would make heterogeneous
catalysis proceed. Therefore, we needed to investigate the
role of gold(0) in the reaction, and performed mercury poi-
soning experiments. Such experiments have been widely uti-
lized as one of the most decisive tests for distinguishing ho-
mogeneous catalysis from heterogeneous catalysis, although
careful precautions are required in the interpretation of ex-
perimental results.[27] This test is based on the ability of the
added mercury to poison metal(0) heterogeneous catalysts
by either the formation of amalgam or adsorption onto the
metal surface. Therefore, suppressed catalytic activity in the
presence of mercury(0) is generally regarded as strong evi-
dence for the presence of heterogeneous metal(0) catalysts.
We chose m-anisaldehyde as a model aldehyde, because
the AuNP-catalyzed cyanosilylation proceeded slower than
the reaction with trans-cinnamaldehyde. The conversion of
m-anisaldehyde was complete in 12 hours, while that of
The mercury poisoning experiments indicate that the
AuNP-catalyzed cyanosilylation was catalyzed heterogene-
ously to some extent by nanometer-sized AuNPs of zero oxi-
dation state and that mercury did not significantly affect the
homogeneous catalysis by soluble gold species. A control ex-
periment was also in accordance with the results. When we
stirred
a
CDCl3 solution of trans-cinnamaldehyde
(0.5 mmol), TMSCN (0.6 mmol), and AuCN (1 mol%) for
five hours at room temperature in the presence of mercury
(1 mmol), the conversion was 35% and did not increase
with longer reaction times. Without mercury, we obtained
the same conversion after five hours.
Mercury would poison AuNPs by forming amalgam with
them and the formation of amalgam was confirmed by X-
ray energy-dispersive spectroscopy (EDS) by using scan-
ning-mode TEM. A mixture of mercury and AuNPs was
stirred in CDCl3 for three days and then the resulting mix-
Chem. Eur. J. 2007, 13, 6351 – 6358
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6355

I. S. Choi et al.
ture was analyzed. The TEM image showed black aggre-
gates and amorphous shells (Figure 6); the amorphous shells
existed around all black aggregates. EDS analysis along the
suppressed the heterogeneous catalysis of gold(0) and the
reaction proceeded both heterogeneously and homogene-
ously.
Proposed reaction mechanism: We propose that the ob-
served catalytic activity of AuNPs for cyanosilylation of al-
dehydes could be explained by dual catalysis (heterogeneous
and homogeneous catalysis; Scheme 2). In the heterogene-
Scheme 2. Proposed mechanism for AuNP-catalyzed cyanosilylation of
aldehydes.
ous catalysis, the reactants (aldehydes and TMSCN) are
preferentially adsorbed onto the surface of AuNPs, presum-
ably due to the nanoscopic property of AuNPs, and the acti-
vation energy of the reaction would become lowered by the
adsorption.[28] We assume that the contribution of heteroge-
neous catalysis to the reaction would be 20%, based on the
mercury poisoning experiments. Regarding the homogene-
ous catalysis, we suggest TMSCN-induced dissolution of
gold for the generation of catalytically active, soluble spe-
À
cies, such as AuCN, AuCN2 , and/or AuIII species. However,
we cannot completely exclude the possibility that AuNPs
solely acted as a reservoir of the catalytically active gold
species.
Conclusion
We performed a mechanistic investigation of the AuNP-cat-
alyzed cyanosilylation of aldehydes. The detailed experi-
ments indicated that the reaction proceeded via a combined
pathway of heterogeneous and homogeneous catalysis. Al-
though we could not identify the soluble species that were
catalytically active in the reaction, both AuNPs and gold
thin films were highly efficient for the cyanosilylation of var-
ious aldehydes. We believe that the coordination of both al-
dehydes and TMSCN onto gold would facilitate the hetero-
geneously catalyzed reaction by lowering the activation
energy. In addition to the coordination, the dissolution of
gold seemed to contribute significantly to the reaction
through homogeneous catalysis in the system investigated.
The newly discovered catalytic activity of nanometer-sized
gold structures for organic transformations has prompted us
to expand the usability of gold as a catalyst. The reported
results open the possibility of the use of Au0, in the form of
Figure 6. Line-scanning EDS analysis by using scanning-mode TEM after
the mercury poisoning experiment. A mixture of mercury and AuNPs
was stirred for three days. a) Line-scanning TEM image. The yellow line
indicates the scanned area. b) The red and cyan lines represent the rela-
tive amounts of gold and mercury, respectively.
yellow line definitely revealed that the amorphous shells
were composed of mercury and the black aggregates were
composed of gold. The existence of mercury is represented
as a cyan line and that of gold as a red line. Along the
yellow line, the constituent ratio between mercury and gold
was 10:1. In comparison, the EDS analysis of the synthe-
sized AuNPs showed only gold peaks. Two possible explana-
tions can be derived from the mercury poisoning and EDS
results. One scenario is that the formation of amalgam
would diminish the dissolution rate of gold and decrease the
conversion yield. In this case, the observed catalytic effect
of AuNPs would be derived solely from the leached gold
species. However, the 35% conversion with only AuCN, as
well as the observed decrease of the conversion upon addi-
tion of mercury, would support the explanation that mercury
À
either nanoparticles or thin films, for other C C bond-form-
ing reactions.
6356
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6351 – 6358

Mechanism of Gold-Catalyzed Cyanosilylation
FULL PAPER
(UK) with a monochromatized AlKa X-ray source (1486.6 eV). Emitted
photoelectrons were detected by a multichannel detector at a take-off
angle of 908 relative to the surface. During the measurements, the base
pressure was 10À9–10À10 Torr. Survey spectra were obtained at a resolu-
tion of 1 eV from one scan and high-resolution spectra were acquired at
a resolution of 0.05 eV from three scans.
Experimental Section
Materials: Hydrogen tetrachloroaurate
ACHTREUNG
ACS reagent), tetraoctylammonium bromide (N
AHCTREUNG
tanethiol (C8H17SH, 98.5%), benzaldehyde (redistilled, 99.5+ %), trans-
cinnamaldehyde (99%), m-anisaldehyde (97%), 3-nitrobenzaldehyde
(99%), 4-nitrobenzaldehyde (98%), crotonaldehyde (predominantly
trans, 99+ %), butyraldehyde (purified by distillation, ꢂ99.5%), anisole
(anhydrous, 99.7%), gold cyanide (AuCN, 99.99%), potassium dicya-
noaurate(I) (KAuCN2, 98%), and phosphotungstic acid (hydrate,
99.995%) were purchased from Aldrich and used as received. Sodium
borohydride (NaBH4, 98%, Junsei), p-anisaldehyde (99+ %, Acros),
TMSCN (96%, TCI), mercury (99.9995%, A.C.S. reagent, Sigma–Al-
drich), toluene (HPLC grade, 99.9%, Merck), and absolute ethanol
(99.8+ %, Merck) were used as received. CDCl3 (99.8%), [D8]toluene
(99.5%), [D6]benzene (99.5%), [D8]THF (99.5%), [D7]DMF (99.5%),
and [D6]DMSO were purchased from Cambridge Isotope Laboratories,
Ultrapure water (18.3 MWcmÀ1) from the Human Ultra Pure System
(Human Corp., Korea) was used.
1H NMR and 13C NMR spectra were obtained with Bruker AVANCE 300
and AVANCE 400 spectrometers. 29Si NMR spectra were obtained with
Bruker AM-300 spectrometer. Chemical shifts in the 1H NMR spectra
were reported in ppm by using tetramethylsilane as an internal standard
and CDCl3 as a solvent. 13C NMR spectra were reported in ppm relative
to the center line of a triplet at 77.0 ppm of CDCl3. The IR spectrum of
TMSCN was obtained from a thin film prepared by evaporation of
CHCl3 solution on a KRS-5 disk. The IR spectrum of gold thin films was
obtained in single-reflection mode by using a dry N2-purged Thermo
Nicolet Nexus FT-IR spectrophotometer equipped with the smart aper-
tured grazing angle (SAGA) accessory. The p-polarized light was incident
at 808 relative to the normal surface of the gold thin film, and a narrow-
band mercury cadmium telluride (MCT) detector cooled with liquid ni-
trogen was used to detect the reflected light. We averaged 1000 scans to
yield the spectrum at a resolution of 4 cmÀ1. All IR spectra were reported
in the absorption mode. The amount of leached gold was measured by
ICP-MS (PQ3, VG Elemental).
Synthesis of 1-octanethiol-passivated AuNPs: AuNPs were synthesized
by following the two-phase method.[17]
A
yellow solution of
HAuCl4·3H2O (0.284 mmol) in deionized water (25 mL) was prepared.
(C8H17)4Br (0.667 mmol) as phase-transfer reagent was independently
N
ACHTREUNG
dissolved in toluene (25 mL) and added with rapid stirring to the aqueous
solution of the Au salt. An immediate two-layer separation resulted, with
an orange/red organic phase on the top and an orange-tinted aqueous
phase on the bottom. The mixture was vigorously stirred until all color
was removed from the aqueous phase, indicating quantitative transfer of
Acknowledgements
À
the AuCl4 moiety into the toluene phase. Next, 1-octanethiol (C8H17SH;
This work was supported by the Korea Science and Engineering Founda-
tion (KOSEF) through the Center for Molecular Design and Synthesis
(CMDS) and a Korea Research Foundation grant funded by the Korea
Government (MOEHRD, Basic Research Promotion Fund; KRF-2005-
070-C00078).
0.5 mmol) was added to the rapidly stirred two-phase mixture at an Au/1-
octanethiol molar ratio of 1:1.76. Upon the addition of 1-octanethiol, the
aqueous layer immediately became beige/murky white. Finally, NaBH4
(0.165 mmol) in deionized water (25 mL) was added to the rapidly stirred
mixture. The color of the organic phase changed from orange to black/
brown and then quickly to dark purple. The white precipitate and cloudi-
ness in the aqueous phase dissipated as the reaction proceeded. The reac-
tion was allowed to continue for 12 h with rapid stirring. After 12 h, the
organic phase was separated and reduced in volume to 5 mL by rotary
evaporation. The AuNPs were precipitated from the toluene solution by
adding ethanol (300 mL) and cooling at À208C for 12 h. The dark purple,
1-octanethiol-passivated AuNPs were then removed by centrifugation,
washed five times with ethanol to remove excess 1-octanethiol, and dried
under reduced pressure. 1-Octanethiol-passivated AuNPs were used for
all the reactions employing AuNPs in this study.
[1] a) M. Haruta, T. Kobayashi, H. Sano, N. Yamada, Chem. Lett. 1987,
405–408; b) M. S. Chen, D. W. Goodman, Science 2004, 306, 252–
255; c) N. Valden, X. Lai, D. W. Goodman, Science 1998, 281, 1647–
1650; d) M. M. Maye, Y. Lou, C.-J. Zhong, Langmuir 2000, 16,
7520–7523.
[2] D. Astruc, F. Lu, J. R. Aranzaes, Angew. Chem. 2005, 117, 8062–
8083; Angew. Chem. Int. Ed. 2005, 44, 7852–7872.
[3] a) D. I. Enache, J. K. Edwards, P. Landon, B. Solsona-Espriu, A. F.
Carley, A. A. Herzing, M. Watanabe, C. J. Kiely, D. W. Knight, G. J.
Hutchings, Science 2006, 311, 362–365; b) H. Tsunoyama, H. Sakur-
ai, Y. Negishi, T. Tsukuda, J. Am. Chem. Soc. 2005, 127, 9374–9375;
c) N. Lopez, J. K. Nørskov, J. Am. Chem. Soc. 2002, 124, 11262–
11263; d) R. J. H. Grisel, P. J. Kooyaman, B. E. Nieuwenhuys, J.
Catal. 2000, 191, 430–437; e) S. Carrettin, P. McMorn, P. Johnston,
K. Griffin, G. J. Hutchings, Chem. Commun. 2002, 696–697; f) C.
Bianchi, F. Porta, L. Prati, M. Rossi, Top. Catal. 2000, 13, 231–236.
[4] a) M. P. Kapoor, A. K. Sinha, S. Seelan, S. Inagaki, S. Tsubota, H.
Hoshida, M. Haruta, Chem. Commun. 2002, 2902–2903; b) T. Haya-
shi, K. Tanaka, M. Haruta, J. Catal. 1998, 178, 566–575.
[5] T. M. Salama, R. Ohnishi, M. Ichikawa, Chem. Commun. 1997, 105–
106.
[6] M. D. Hughes, Y.-J. Xu, P. Jenkins, P. McMorn, P. Landon, D. I.
Enache, A. F. Carley, G. A. Attard, G. J. Hutchings, F. King, E. H.
Stitt, P. Johnston, K. Griffin, C. J. Kiely, Nature 2005, 437, 1132–
1135.
[7] a) C. Mohr, H. Hofmeister, J. Radnik, P. Claus, J. Am. Chem. Soc.
2003, 125, 1905–1911; b) P. Claus, A. Brückner, C. Mohr, H. Hof-
meister, J. Am. Chem. Soc. 2000, 122, 11430–11439; c) J. Jia, K.
Haraki, J. N. Kondo, K. Domen, K. Tamaru, J. Phys. Chem. B 2000,
104, 11153–11156.
Typical procedure for cyanosilylation of aldehydes: A CDCl3 solution of
aldehyde (0.5 mmol), TMSCN (0.6 mmol), and anisole (0.5 mmol) was
prepared in a round-bottomed flask. AuNPs (0.2 wt% with respect to al-
dehyde) were added to the solution, and the resulting mixture was stirred
at room temperature. The progress of the reaction was monitored by
thin-layer chromatography (TLC) and the NMR yield (conversion%)
was calculated by using anisole as an internal standard.
Mercury poisoningexperiments : A CDCl3 solution of m-anisaldehyde
(0.5 mmol), TMSCN (0.6 mmol), and anisole (0.5 mmol) was prepared in
a round-bottomed flask. AuNPs (0.2 wt%) were added, followed by mer-
cury (1.0 mmol) after a predetermined time, and the resulting mixture
was stirred at room temperature. The conversion of the reaction was cal-
culated by 1H NMR spectroscopy by using anisole as an internal stan-
dard.
Instrumentation: The synthesized AuNPs were characterized by TEM
and XPS. For the TEM analysis, AuNPs dispersed in chloroform were
dropped by using a micropipette onto a carbon-coated copper grid and
the solvent was evaporated. All the TEM images were obtained with
JEM-2010 and JEM-2100F instruments (JEOL, Japan). To obtain the
TEM images of AuNPs after cyanosilylation of aldehydes, AuNPs were
stained with phosphotungstic acid. For the XPS study, AuNPs dispersed
in chloroform were dropped by using a micropipette onto a small piece
of silicon wafer and the solvent was evaporated. X-ray photoelectron
spectra were obtained with a VG-Scientific ESCALAB 250 spectrometer
[8] a) H. Sakurai, A. Ueda, T. Kobayashi, M. Haruta, Chem. Commun.
1997, 271–272; b) T. Tabakova, V. Idakiev, D. Andreeva, I. Mitov,
Appl. Catal. A 2000, 202, 91–97.
Chem. Eur. J. 2007, 13, 6351 – 6358
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6357

I. S. Choi et al.
[9] M. Lazar, R. J. Angelici, J. Am. Chem. Soc. 2006, 128, 10613–
10620.
216, 41–52; e) C.-r. Lee, S. J. Bae, M.-s. Gong, K. Kim, S.-W. Joo, J.
Raman Spectrosc. 2002, 33, 429–433; f) E. Lee, S. S. Yi, M. S. Kim,
K. Kim, J. Mol. Struct. 1993, 298, 47–54; g) D. E. Richardson, H.
Taube, J. Am. Chem. Soc. 1983, 105, 40–51.
[10] a) S. Carrettin, J. Guzman, A. Corma, Angew. Chem. 2005, 117,
2282–2285; Angew. Chem. Int. Ed. 2005, 44, 2242–2245; b) A. K.
Gupta, C. Y. Rhim, C. H. Oh, R. S. Mane, S.-H. Han, Green Chem.
2006, 8, 25–28. In these cases, a cationic gold species, AuIII, was the
active species; c) H. Tsunoyama, H. Sakurai, N. Ichikuni, Y. Negishi,
T. Tsukuda, Langmuir 2004, 20, 11293–11296.
[11] a) B. M. Trost, F. D. Toste, A. B. Pinkerton, Chem. Rev. 2001, 101,
2067–2096; b) V. Ritleng, C. Sirlin, M. Pfeffer, Chem. Rev. 2002,
102, 1731–1770; c) C.-H. Jun, Chem. Soc. Rev. 2004, 33, 610–618.
[12] M. Haruta, CATTECH 2002, 6, 102–115.
[24] a) H. Ibach, D. L. Mills, Electron Energy Loss Spectroscopy andSur-
face Vibrations, Academic, New York, 1982; b) B. A. Sexton, N. R.
Avery, Surf. Sci. 1983, 129, 21–36; c) C. M. Friend, J. Stein, E. L.
Muetterties, J. Am. Chem. Soc. 1981, 103, 767–772; d) K. Bange, R.
McIntyre, J. K. Sass, N. V. Richardson, J. Electroanal. Chem. 1984,
178, 351–355; e) K. Kishi, S. Ikeda, Surf. Sci. 1981, 107, 405–416.
[25] J. A. Seckar, J. S. Thayer, Inorg. Chem. 1976, 15, 501–504.
[26] a) A. S. K. Hashmi, T. M. Frost, J. W. Bats, Org. Lett. 2001, 3, 3769–
3771; b) A. S. K. Hashmi, M. Rudolph, J. P. Weyrauch, M. Wçlfle,
W. Frey, J. W. Bats, Angew. Chem. 2005, 117, 2858–2861; Angew.
Chem. Int. Ed. 2005, 44, 2798–2801; c) J. P. Markham, S. T. Staben,
F. D. Toste, J. Am. Chem. Soc. 2005, 127, 9708–9709; d) L. Zhang,
S. A. Kozmin, J. Am. Chem. Soc. 2005, 127, 6962–6963.
[13] F. Shi, Q. Zhang, Y. Ma, Y. He, Y. Deng, J. Am. Chem. Soc. 2005,
127, 4182–4183.
[14] a) D. E. Fuerst, E. N. Jacobsen, J. Am. Chem. Soc. 2005, 127, 8964–
8965; b) D. H. Ryu, E. J. Corey, J. Am. Chem. Soc. 2004, 126, 8106–
8107; c) J.-M. Brunel, I. P. Holmes, Angew. Chem. 2004, 116, 2810–
2837; Angew. Chem. Int. Ed. 2004, 43, 2752–2778; d) H. Grçger,
Chem. Eur. J. 2001, 7, 5246–5251; e) B. Karimi, L. MaꢁMani, Org.
Lett. 2004, 6, 4813–4815; f) R. J. H. Gregory, Chem. Rev. 1999, 99,
3649–3682.
[27] a) Y. Na, S. Park, S. B. Han, H. Han, S. Ko, S. Chang, J. Am. Chem.
Soc. 2003, 126, 250–258; b) B. P. S. Chauhan, J. S. Rathore, M. Chau-
han, A. Krawicz, J. Am. Chem. Soc. 2003, 125, 2876–2877; c) Y. Lin,
R. G. Finke, Inorg. Chem. 1994, 33, 4891–4910; d) D. R. Anton,
R. H. Crabtree, Organometallics 1983, 2, 855–859.
[15] C. G. Kruse in Chirality in Industry (Eds.: A. N. Collins, G. N. Shel-
drake, J. Crosby), Wiley, Chichester, 1992, Chapter 14.
[28] For heterogeneous catalysis, there was a report showing that chemi-
cal reactions can occur at the core surface of metal nanoparticles
passivated with thiols. The hydrogenation reactions of allyl alcohol
and maleic acid can be catalyzed by alkylthiol-passivated platinum
nanoparticles and charged-thiol-passivated platinum nanoparticles,
respectively (S. E. Eklund, D. E. Cliffel, Langmuir 2004, 20, 6012–
6018). In our reaction system, the adsorption of the reactants onto
1-octanethiol-passivated AuNPs may not be achieved with ease.
However, when the self-assembled monolayers (SAMs) of alkylth-
iols are formed on gold, the alkylthiol-coated gold surfaces are not
perfect and contain various defects, such as pinholes, loosely packed
SAMs, and large uncovered sites (X.-M. Zhao, J. L. Wilbur, G. M.
Whitesides, Langmuir 1996, 12, 3257–3264; D. Losic, J. G. Shapter,
J. J. Gooding, Electrochem. Commun. 2002, 4, 953–958). Therefore,
we believe that the reactants can be adsorbed onto AuNPs at these
defect sites. In addition, the defect sites also would be a starting site
for the dissolution of gold (A. C. Templeton, M. J. Hostetler, C. T.
Kraft, R. W. Murray, J. Am. Chem. Soc. 1998, 120, 1906–1911; R.
Paulini, B. L. Frankamp, V. M. Rotello, Langmuir 2002, 18, 2368–
2373; F. P. Zamborini, R. M. Crooks, Langmuir 1997, 13, 122–126).
[16] M. Bandini, P. G. Cozzi, A. Garelli, P. Melchiorre, A. Umani-
Ronchi, Eur. J. Org. Chem. 2002, 3243–3249.
[17] a) M. Brust, M. Walker, D. Bethell, D. J. Schiffrin, R. Whyman, J.
Chem. Soc. Chem. Commun. 1994, 801–802; b) D. V. Leff, L.
Brandt, J. R. Heath, Langmuir 1996, 12, 4723–4730.
[18] A. Kumar, S. Mandal, P. R. Selvakannan, R. Pasricha, A. B. Man-
dale, M. Sastry, Langmuir 2003, 19, 6277–6282.
[19] For NMR spectra of the reactions, see the Supporting Information.
[20] AuNPs were not effective in catalyzing ketones and imines com-
pared with aldehydes. We observed 61% conversion of 2-pentanone
after 72 h stirring at room temperature, but in other cases of ketones
and imines the conversion was less than 10%. The increased loading
of AuNPs did not improve the conversion.
[21] a) M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. J. Genet,
B. Delmon, J. Catal. 1993, 144, 175–192; b) Y. Iizuka, T. Tode, T.
Takao, K.-i. Yatsu, T. Takeuchi, S. Tsubota, M. Haruta, J. Catal.
1999, 187, 50–58; c) M. Haruta, GoldBull. 2004, 37, 27–36.
[22] For NMR spectra of the AuNP–TMSCN complex, see the Support-
ing Information.
[23] a) M. F. Mrozek, S. A. Wasileski, M. J. Weaver, J. Am. Chem. Soc.
2001, 123, 12817–12825; b) D. H. Son, S. J. Ahn, Y. J. Lee, K. Kim,
J. Phys. Chem. 1994, 98, 8488–8493; c) S.-W. Joo, T. D. Chung, W. C.
Jang, M.-s. Gong, N. Geum, K. Kim, Langmuir 2002, 18, 8813–8816;
d) H. A. Chun, D. W. Boo, K. Kim, M. S. Kim, J. Mol. Struct. 1990,
Received: December 13, 2006
Revised: March 13, 2007
Published online: May 11, 2007
6358
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6351 – 6358