Catalytic reactions of carbene precursors on bulk gold metal

Article abstract of DOI:10.1021/ja900653s

Bulk gold metal powder, consisting of particles (5-50 μm) much larger than nanoparticles, catalyzes the coupling of carbenes generated from diazoalkanes (R₂C=N₂) and 3,3-diphenylcyclopropene (DPCP) to form olefins. It also catalyzes cyclopropanation reactions of these carbene precursors with styrenes. The catalytic activity of the gold powder depends on the nature of the gold particles, as determined by TEM and SEM studies. The reactions can be understood in terms of mechanisms that involve the generation of carbene R₂C: intermediates adsorbed on the gold surface.

Full text of DOI:10.1021/ja900653s

Published on Web 07/30/2009
Catalytic Reactions of Carbene Precursors on Bulk Gold
Metal
Yibo Zhou, Brian G. Trewyn, Robert J. Angelici,* and L. Keith Woo*
Ames Laboratory and Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011-3111
Received January 27, 2009; E-mail: angelici@iastate.edu; kwoo@iastate.edu
Abstract: Bulk gold metal powder, consisting of particles (5-50 µm) much larger than nanoparticles,
catalyzes the coupling of carbenes generated from diazoalkanes (R2CdN2) and 3,3-diphenylcyclopropene
(DPCP) to form olefins. It also catalyzes cyclopropanation reactions of these carbene precursors with
styrenes. The catalytic activity of the gold powder depends on the nature of the gold particles, as determined
by TEM and SEM studies. The reactions can be understood in terms of mechanisms that involve the
generation of carbene R2C: intermediates adsorbed on the gold surface.
Scheme 1
1. Introduction
Although nanosized gold particles (<5 nm) supported on metal
oxides catalyze a variety of reactions including CO, hydrocar-
bon, and alcohol oxidation,1 bulk gold metal is well-known for
its poor catalytic properties.1a We have shown recently2 that
bulk gold powder with particle sizes of ∼50 µm is capable of
catalyzing oxidative reactions of isocyanides,3,4 CO,5 and
amines.6,7 Gold powder also catalyzes the oxidation (O2) of CO
to CO2 and glycerol to glyceric and glycolic acids in basic (∼pH
14) aqueous solution.8 In the oxidative reactions of isocyanides
(CtNsR) with secondary and primary amines, there is good
evidence that adsorption of the isocyanide on gold activates it
toward attack by amines (Scheme 1) to give a diaminocarbene
intermediate (A) that reacts rapidly with O2 to give ureas (in
the case of secondary amines) and carbodiimides (in the case
of primary amines). In contrast to their high reactivities on gold
metal, diaminocarbene ligands in transition metal complexes
are usually unreactive,9 but their nonheteroatom analogues
(:CR2) are very reactive.10 Goals of the research described herein
were to explore the possibility that carbenes (:CR2) not
containing heteroatoms could be created on the surface of bulk
gold and then to compare their reactivities with those of
diaminocarbenes.
Transition metal carbene complexes have often been prepared
by treating diazoalkanes (R2CdNdN) with unsaturated metal
complexes.11,12 In the present studies, we sought to determine
whether or not carbene groups (:CR2) could be created on a
metal surface by treating diazoalkanes with gold metal powder.
We anticipated that such carbene groups would react with each
other or with other reactants that were present in the mixture.
By analogy with reactions of :CR2 ligands in metal complexes,10,12
one would expect carbene groups to couple to give olefins (2
R2CN2 f R2CdCR2 + 2 N2), in the absence of other reactants.
In the presence of olefins, carbenes on the gold might be
expected to give cyclopropanes, as is observed in transition
metal complex-catalyzed reactions of diazoalkanes and
olefins.10a
(10) (a) Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic Methods
for Organic Synthesis with Diazo Compounds: From Cyclopropanes
to Ylides; Wiley-VCH: New York, 1998 and references therein. (b)
Doyle, M. P.; Griffin, J. H.; Bagheri, V.; Dorow, R. L. Organometallics
1984, 3, 53, and references therein. (c) Brown, F. J. Prog. Inorg. Chem.
1980, 27, 1–122. (d) Do¨tz, K. H. In Reactions of Coordinated Ligands;
Braterman, P., Ed.; Plenum Press: New York, 1986; Vol. 1, pp 285-
370.
(11) (a) Gandelman, M.; Rybtchinski, B.; Ashkenazi, N.; Gauvin, R. M.;
Milstein, D. J. Am. Chem. Soc. 2001, 123, 5372. and references
therein. (b) Collman, J. P.; Brothers, P. J.; McElwee-White, L.; Rose,
E. J. Am. Chem. Soc. 1985, 107, 6110. (c) Woo, L. K.; Smith, D. A.
Organometallics 1992, 11, 2344. (d) Djukic, J.-P.; Smith, D. A.;
Young, V. G.; Woo, L. K. Organometallics 1994, 13, 3020. (e) Zhang,
J.; Gandelman, M.; Shimon, L. J. W.; Milstein, D. Organometallics
2008, 27, 3526.
(1) For reviews: (a) Bond, G. C.; Louis, C.; Thompson, D. T. Catalysi by
Gold; Imperial College Press: London, 2006. (b) Thompson, D. T.
Top. Catal. 2006, 38, 231. (c) Chen, M. S.; Goodman, D. W. Catal.
Today 2006, 111, 22. (d) Min, B. K.; Friend, C. M. Chem. ReV. 2007,
107, 2709.
(2) Angelici, R. J. J. Organomet. Chem. 2008, 693, 847.
(3) Lazar, M.; Angelici, R. J. J. Am. Chem. Soc. 2006, 128, 10613.
(4) Lazar, M.; Zhu, B.; Angelici, R. J. J. Phys. Chem. C 2007, 111, 4074.
(5) Zhu, B.; Angelici, R. J. J. Am. Chem. Soc. 2006, 128, 14460.
(6) Zhu, B.; Angelici, R. J. Chem. Commun. 2007, 2157.
(7) Zhu, B.; Lazar, M.; Trewyn, B.; Angelici, R. J. J. Catal. 2008,
260, 1.
(8) Ketchie, W. C.; Fang, Y.-L.; Wong, M. S.; Murayama, M.; Davis,
R. J. J. Catal. 2007, 250, 94.
(9) (a) N-Heterocyclic Carbenes in Transition Metal Catalysis; Glorius,
F. , Ed.; Topics in Organometallic Chemistry; Springer: Berlin/
Heidelberg, 2007; Vol. 21. (b) Strassner, T. Organometallic Oxidation
Catalysis; Topics in Organometallic Chemistry; Springer: Berlin/
Heidelberg, 2007; Vol. 22, p 125. (c) N-Heterocyclic Carbenes in
Synthesis; Nolan, S. P. , Ed.; Wiley-VCH: 2007.
(12) (a) Collman, J. P.; Rose, E.; Venbur; g, G. D. J. Chem. Soc., Chem.
Commun. 1993, 934. (b) Oshima, T.; Nagai, T. Tetrahedron Lett. 1980,
1251. (c) Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic
Methods for Organic Synthesis with Diazo Compounds: From
Cyclopropanes to Ylides; Wiley-VCH: New York, 1998; p 624.
9
11734 J. AM. CHEM. SOC. 2009, 131, 11734–11743
10.1021/ja900653s CCC: $40.75  2009 American Chemical Society

Catalytic Reactions of Carbene Precursors on Bulk Au
A R T I C L E S
Figure 1. Electron micrographs of dull brown Au particles with sea urchin morphology. (a) Scanning electron microscopy (SEM) image measured at
50 000× magnification; (b) transmission electron microscopy (TEM) image measured at 64 000× magnification, with the inset showing an HRTEM image
focusing on the lattice fringes; (c) TEM image measured at 440 000×, with the inset showing an HRTEM image focusing on the FCC lattice of the gold
nanoparticles. The white arrows in (b) and (c) point to the areas that were magnified for the high resolution images (insets).
2. Results and Discussion
Another approach to the creation of carbene ligands in
transition metal complexes is the reaction (eq 1) of 3,3-
diphenylcyclopropene (DPCP) with unsaturated metal com-
plexes (LxM).13 Perhaps the most notable use of this reaction
2.1. Characterization of Bulk Gold Powder. Gold was chosen
as the catalytic metal for these reactions because it does not
react with O2 to form surface oxides.1a This means that the
surface gold atoms are in only the zero oxidation state, which
was expected to simplify interpretations of the role of the gold
metal in its catalytic reactions. The gold powder was prepared
by dissolving gold metal in aqua regia to form HAuCl4, which
was reduced with hydroquinone according to a published
procedure.22 The resulting dull brown powder was treated with
freshly prepared “piranha solution” (H2SO4/H2O2, 3:1), washed
with water and methanol, and then dried in an oven at 110 °C
[see Experimental Section]. After this treatment, the gold powder
had a dull brown appearance. Electron micrographs (Figure 1a)
of one sample of this dull brown Au powder showed large
particle sizes (∼500 nm) with attached spines (Figure 1b) that
had the overall appearance of a sea urchin. The spines consist
of Au nanoparticles (black in Figure 1b) (∼2 nm diameter)
embedded in a carbon-containing matrix. At a magnification of
440 000× (Figure 1c), fringes in the individual gold nanopar-
ticles are visible. They are parallel to each other in each
nanoparticle, but the direction of the fringes is different for each
nanoparticle. The fringe spacing (0.230 nm) corresponds well
with the spacing between [111] planes of FCC gold metal (0.235
nm).23-25
was in the synthesis of the first generation of Grubbs’ olefin
metathesis and ring-opening metathesis polymerization catalysts
14a
[(Ph3P)2RuCl2]dCHCHdCPh2
and [Cp*RuCl]2dCHCHd
CPh2.14c In addition, vinyl carbene complexes of Co,15 Ti,16
Zr,17 Ta,18 W,19 Ru,14 Os,20 and Re21 have been prepared using
DPCP. If DPCP were to form vinyl carbene groups on a gold
metal surface, one would expect them to couple with each other
to give triene products and/or to react with olefins to give vinyl
cyclopropanes. Through studies of these reactions of diazoal-
kanes and DPCP, we sought to gain an understanding of factors
that influence the formation of carbenes on metal surfaces and
also to explore the differences in reactivities of carbene groups
in metal complexes and on a gold metal surface. The observed
reactions also extend the range of reactions that are catalyzed
by bulk gold metal.
XRD analysis of the gold powder showed diffraction peaks
corresponding to the [111], [200], [220], [311], and [222] planes
of cubic gold crystallites (see Figure S2 in the Supporting
Information).23-25 An energy dispersive X-ray (EDX) spectrum
(see Figure S1 in the Supporting Information) of the body (∼500
nm) of a sea urchin particle shows that it consists of only gold.
An EDX spectrum of the spines (Figure S1) also shows only
Au to be present. While the EDX analysis detected only gold
in the spines, low atomic weight elements are not easily detected
by EDX. However, it is clear from Figure 1b,c that another
material is present between the gold nanoparticles in the spines.
The fringe spacing (0.682 nm) of this material does not
correspond to gold. To determine the chemical composition of
the material between the gold nanoparticles, the energy filtration
feature of the EELS was employed to image the spines.
Specifically, the spectrometer was tuned to capture the location
(13) For a review of cyclopropenes, see: Rubin, M.; Rubina, M.; Gevorgyan,
V. Chem. ReV. 2007, 107, 3117.
(14) (a) Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J. Am.
Chem. Soc. 1992, 114, 3974. (b) Nguyen, S. T.; Grubbs, R. H.; Ziller,
J. W. J. Am. Chem. Soc. 1993, 115, 9858. (c) Gagne, M. P.; Grubbs,
R. H. Organometallics 1992, 11, 3933. (d) Sanford, M. S.; Valdez,
M. R.; Grubbs, R. H. Organometallics 2001, 20, 5455. (e) Nishiyama,
H.; Park, S. B.; Itoh, K. Chem. Lett. 1995, 599.
(15) Fo¨rstner, J.; Kakoschke, A.; Stellfeldt, B.; Butenschon, H.; Wartchow,
R. Organometallics 1998, 17, 893.
(16) (a) Binger, P.; Mu¨ller, P.; Wenz, R.; Mynott, R. Angew. Chem. 1990,
102, 1070. (b) Binger, P.; Mu¨ller, P.; Benn, R.; Mynott, R. Angew.
Chem. 1989, 101, 647.
(17) Binger, P.; Mu¨ller, P.; Hermann, A. T.; Phillips, P.; Gabor, B.;
Langhau, F. Chem. Ber. 1991, 124, 2165.
(18) Huang, J.-H.; Lee, T.-Y.; Swenson, D. C.; Messerle, L. Inorg. Chim.
Acta 2003, 345, 209.
(19) (a) Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc.
1993, 115, 8130. (b) de la Mata, F. J.; Grubbs, R. H. Organometallics
1996, 15, 577. (c) Fujimura, O.; de la Mata, F. J.; Grubbs, R. H.
Organometallics 1996, 15, 1865. (d) de la Mata, F. J. J. Organomet.
Chem. 1996, 525, 183.
(22) Block, B. P. Inorg. Synth. 1953, 4, 14.
(23) Huang, C.-J.; Chiu, P.-H.; Wang, Y.-H.; Chen, W.-R.; Meen, T.-H.;
Yang, C.-F. Nanotechnol. 2006, 17, 5355.
(20) Green, M.; Orpen, A. G.; Schaverien, C. J. J. Chem. Soc., Dalton
Trans. 1989, 1333.
(24) Miller, J. T.; Kropf, A. J.; Zha, Y.; Regalbuto, J. R.; Delannoy, L.;
Louis, C.; Bus, E.; van Bokhoven, J. A. J. Catal. 2006, 240, 222.
(25) Kobayashi, Y.; Tadaki, Y.; Nagao, D.; Konno, M. J. Colloid Interface
Sci. 2005, 283, 601.
(21) Flatt, B. T.; Grubbs, R. H.; Blanski, R. L.; Calabrese, J. C.; Feldman,
J. Organometallics 1994, 13, 2728.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11735

A R T I C L E S
Zhou et al.
Figure 2. Bright field (a), dark field (b), and energy-filtered (c) transmission electron micrographs (TEMs) of spines of the sea urchin gold. Micrograph (c)
was tuned for carbon. In the upper right (large arrow), lacey carbon from the Cu grid is visible.
of carbon. As is evident in Figure 2c, the areas (white or gray)
between the gold nanoparticles contain abundant carbon (small
arrow). Carbon is also visible in the bright field (Figure 2a)
and dark field (Figure 2b) micrographs. Also evident in Figure
2 is carbon on the surface of the large body of the sea urchin
gold. Thus, the sea urchin gold particles consist of a large gold
body that is coated with a carbon-containing material. Attached
to this body are spines consisting of nanogold particles
embedded in a carbon-containing matrix. The carbon-containing
material has a large lattice spacing (0.682 nm) which is not
consistent with graphite. Perhaps it is an organic material derived
from the hydroquinone used to prepare the gold powder.
To understand how the morphology and composition of the
gold metal affects its catalytic activity, 1.0 g of the dull brown
Au powder with the sea urchin morphology was used as the
catalyst for the reaction (eq 2) of 0.20 mmol of ethyl diazoac-
Figure 3. Effect of gold regeneration by piranha treatment on catalyst
etate (EDA) in 5 mL of CH3CN at 60 °C, while stirring with a
activity in the self-coupling of EDA (eq 2) using 0.20 mmol of EDA and
1.0 g of Au powder (with initial sea urchin morphology) in 5 mL of CH3CN
at 60 °C. The gold powder was washed with CH3CN (5 × 30 mL), CH2Cl2
(5 × 30 mL), and MeOH (5 × 30 mL) and then treated with “piranha
solution” before each cycle. ([) 1st cycle, (9) 2nd cycle, (2) 3rd cycle,
(0) 4th cycle, (O) 5th cycle, (b) 6th cycle.
magnetic stirbar in a round-bottom tube [see Experimental
Section]. After 6 h, the reaction generated a 47% yield of diethyl
maleate (cis-1) and diethyl fumarate (trans-1) in a ratio of 1.1:
1. When the reaction was allowed to continue for a longer time
(Figure 3), there was no increase in yield even though unreacted
EDA was still present. It is not clear why the sea urchin gold
becomes inactive during the reaction; perhaps an EDA-derived
material deposits on the gold surface. After the first reaction,
the gold, which now had a shiny gold appearance, was separated
by filtration and washed with large amounts of CH3CN, CH2Cl2,
and MeOH. After the gold was treated with piranha solution,
washed (H2O/MeOH), and dried, an SEM image (Figure 4) of
the powder showed much larger gold particles that had a very
different morphology. This shiny gold powder was then used
in a second reaction of EDA under the same conditions as the
first cycle; this reaction was much faster than the first reaction
Figure 4. Scanning electron microscopy (SEM) image at 15 000×
giving a quantitative yield of the products after only 7 h (Figure
3). The high activity of the gold in the second run must be due
to changes in the surface composition and morphology of the
gold during the first reaction and to the CH2Cl2/MeOH washing
prior to the second run. It is possible that a small amount of
carbon-containing material remains on the gold surface in the
second run. Throughout the course of the second run, the cis/
trans ratio of the products was always in the range 1:1-1.1:1.
When the gold was regenerated and reused in additional cycles,
the initial rates decreased (Figure 3), but the overall yield still
reached 90% even after 6 cycles. The decrease in activity may
be due to incomplete removal of organic material from the
magnification of the sea urchin gold after one cycle in the catalytic reaction
of EDA (eq 2).
surface, even after the piranha treatment. A parallel series of
recycling reactions was performed in which the catalyst was
only washed with CH2Cl2 and MeOH (without the piranha
solution) before each cycle. The results were similar to those
in Figure 3 (see Figure S4 in the Supporting Information),
suggesting that the piranha treatment did not have a large effect.
Typically, the gold powder was used in 10 catalytic reactions
before its activity had decreased to the point where it was
beneficial to recycle the gold by converting it back to HAuCl4,
9
11736 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Catalytic Reactions of Carbene Precursors on Bulk Au
A R T I C L E S
which was again reduced to Au powder. The activity of this
regenerated dull brown Au powder was often somewhat lower
than that of the originally prepared gold powder. It is not clear
why the activity of the Au powder varied even though it was
prepared in the same way. Perhaps even the piranha treatment
did not completely remove organic material that deposited on
the gold surface. It seems likely that the low catalytic activity
of the freshly prepared sea urchin gold was due to the carbon-
containing material that coated the gold.
Even the morphology of the fresh gold powder prepared by
the hydroquinone reduction of HAuCl4 varied. In one prepara-
tion, the gold particles (Figure S3) did not have the sea urchin
shape (Figure 1). Instead, it consisted of much larger particles
similar to those (Figure 4) of the sea urchin gold after it had
catalyzed one cycle of the EDA reaction (eq 2). This gold
(Figure S3) was more reactive in the first run than the sea urchin
gold. As for the sea urchin gold, its dull brown appearance
changed to shiny gold during the first reaction. The activity was
higher in the second run but then decreased in subsequent runs,
as was observed for the sea urchin gold (Figure 3).
Since a variety of reactions, including the formation of
cyclopropanes from EDA and olefins,26 are catalyzed by
homogeneous Au(I) and Au(III) complexes,27 we studied rea-
ction 2 under conditions that would reveal the presence of
soluble catalytic species, such as gold complexes or gold
(nano)particles, as the actual catalysts of reaction 2. A solution
of EDA (0.20 mmol) in CH3CN (5 mL) containing mesitylene
(0.06917 mmol) as a GC standard was stirred in air at 60 °C
with 1.0 g of gold powder; the gold used was in its most active
form, as it had been used in only one previous reaction (see
Figure 3). After 45 min of reaction, the yield of the olefin
products was 25% with a cis/trans ratio of 1.03:1.00. The
solution was then filtered from the gold and divided into two
equal parts. One part (not containing gold powder) was heated
to 60 °C for 5.5 h; no additional olefin product was formed.
This indicated that there were no soluble catalytic species in
the solution. To the other part of the reaction solution was added
the original 1.0 g of gold powder; after 5.5 h at 60 °C, the yield
of the olefin was now 100% with a 1.04/1.00 cis/trans ratio.
Thus, the solid gold powder is necessary for the catalysis.
Similar experiments performed in the gold powdered-catalyzed
reactions of CO, O2, and primary amines to give ureas5 and the
oxidative dehydrogenation of amines to imines6,7 showed that
these reactions are also not catalyzed by solution-soluble gold
species.
2.2. Self-Coupling of Carbene Precursors. To explore the
scope of the gold-catalyzed reactions of carbene precursors
beyond that of EDA (eq 1), reactions of several diazoalkanes
were investigated (Table 1). The gold catalyst that was used in
these reactions was not freshly prepared but had been used in
catalytic runs at least twice previously. Between runs, it was
treated with piranha solution, washed with CH2Cl2 and MeOH,
and dried at 110 °C before each use. As shown in Table 1, the
reactions of phenyl diazomethane (PhCHdN2), the diazoacetates
(EDA and Me-MPDA), and the diazoketone (PhCOCHdN2)
at 60 °C generated cis and trans olefins in high yields
(76-100%) when catalyzed by gold powder (entries 1-5 and
9). In the absence of gold, no reactions occurred. In contrast to
the reactions in Scheme 1, the presence of O2 did not influence
the reactions (entry 1 vs 2, entry 3 vs 4), which also means that
the putative carbene intermediate does not react with O2, as also
indicated by the absence of aldehyde or ketone products. When
the reaction of EDA was performed in the presence of 1 equiv
of n-butyl isocyanide, DABCO, pyridine, or triphenylphosphine,
no reaction occurred, presumably due to the competitive
adsorption of the isocyanide, amines, and phosphine28 on the
gold surface. The reaction of Me-MPDA gave only a 76% yield
of 3 under nitrogen due to the concomitant formation of the
azine byproduct R1R2CNdNCR1R2 in 15% yield. Under an O2
atmosphere, Me-MPDA gave an even lower yield of 3 due to
the formation of more azine as well as other uncharacterized
byproducts. Although trans olefins are thermodynamically more
stable than their cis isomers,29 the cis/trans ratios (1.1-3) favor
the cis isomer in the gold-catalyzed reactions of phenyl
diazomethane, EDA, and Me-MPDA (entries 1-6). These
results are similar to those obtained in self-coupling reactions
of PhCHdN2 catalyzed by CuX (cis/trans ) 1.3-2.5)12b and
those of EDA catalyzed by a ruthenium porphyrin complex
Ru(tmp), where tmp is 5,10,15,20-tetramesitylporphyrin (cis/
trans ) 15).12a The yield of the olefin product 4 in reactions of
R-diazoacetophenone (PhCOCHdN2) is affected significantly
by the solvent. In acetonitrile after 24 h, this reaction (eq 3)
gave a 58% yield of the olefin product 4 (consisting primarily
of the trans isomer, cis/trans ) 13:87) and 42% of oxazole
product 6 (entry 7, and eq 3) resulting from the reaction of the
carbene intermediate with the MeCN solvent; in the absence of
the gold powder, there is no reaction of PhCOCHdN2. In
toluene after 24 h, 4 was obtained in 70% yield with a cis/
trans ratio of 42:58 (entry 8); some starting material was still
present when the reaction was stopped. In 1,2-dichloroethane
after 24 h, the yield of 4 was the highest (86%), and the cis-
isomer was the favored product (cis/trans ) 52:48, entry 9).
The formation of oxazole 6 in 42% yield in the reaction (eq
3) of PhCOCHdN2 with MeCN solvent (entry 7) suggested that
other nitriles might give substituted oxazoles. Unfortunately,
the reaction of PhCOCHdN2 in benzonitrile solvent gave no
oxazole product and 4 in only 42% yield (cis/trans ) 2.2:1). In
propionitrile under the same conditions, 4 was formed in 65%
yield (cis/trans ) 1:4) but none of the oxazole.
A qualitative comparison of the self-coupling reactions (Table
1) of the diazoalkanes shows that the rates of reaction decrease
in the following order: PhCHdN2 > EDA > PhCOCHdN2 ≈
Me-MPDA . (MeO2C)2CdN2. At 60 °C, phenyl diazomethane
reacts completely within 15-30 min, while the diazo esters and
diazo ketone (EDA, Me-MPDA, PhCOCHdN2) require more
than 2 h, and dimethyl diazomalonate does not react at all (entry
10). At room temperature, the reaction of PhCHdN2 is complete
within 30 min, but EDA requires 48 h. The overall trend in
(26) Fructos, M. R.; Belderrain, T. R.; de Fre´mont, P.; Scott, N. M.; Nolan,
S. P.; D´ıaz-Requejo, M. M.; Pe´rez, P. J. Angew. Chem., Int. Ed. 2005,
44, 5284.
(28) (a) Uvdal, K.; Persson, I.; Liedberg, B. Langmuir 1995, 11, 1252. (b)
Steiner, U. B.; Neuenschwander, P.; Caseri, W. R.; Suter, U. W.;
Stucki, F. Langmuir 1992, 8, 90.
(27) (a) Li, Z.; Brouwer, C.; He, C. Chem. ReV. 2008, 108, 3239. (b)
Hutchings, G. J.; Brust, M.; Schmidbaur, H. Chem. Soc. ReV. 2008,
37, 1745–2140. (c) Hashmi, A. S. Chem. ReV. 2007, 107, 3180.
(29) McWilliams, K. M.; Angelici, R. J. Organometallics 2007, 26, 5111,
and references therein.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11737

A R T I C L E S
Zhou et al.
Table 1. Self-Coupling of Carbene Precursors Catalyzed by Gold Powdera
b
a 0.2 mmol of the diazo compound, 1.0 g of gold powder, 5 mL of solvent, 60 °C. Yields were determined by H NMR with Ph3CH as the internal
standard. c Cis/trans ratios were determined by GC and 1H NMR. d Low yield due to the formation of azine and some uncharacterized byproducts.
e Low yield due to oxazole product 6. f Low yield due to incomplete reaction.
1
these diazoalkane reactivities is similar to that in their transition
metal complex-catalyzed reactions.30
groups on the Cu surface or by nucleophilic attack of Ph-
COCHdN2 on the adsorbed :CHC(dO)R.31 Either of these
mechanisms could be involved in our gold powder-catalyzed
reactions. Studies32,33 of reactions of the dinuclear complexes
Cp2Ru2(µ-CH2)(CO)2(NCMe)30 and RhM(µ-CH2)(CO)3(dppm)2+,
where M ) Ru or Os33 and µ-CH2 is a bridging carbene ligand,
with diazoalkanes have led to the conclusion that they proceed
according to the mechanism in Scheme 2. Such a mechanism is
reasonable for the coupling of carbene units on the gold metal
surface. To our knowledge, the only previously reported studies
of gold metal-catalyzed reactions of diazoalkanes were those of
The generally accepted mechanism for transition metal
complex-catalyzed olefin formation from diazo compounds
involves initial formation of a metal carbene complex (MdCR2)
which undergoes nucleophilic attack by another diazo molecule;
this is followed by loss of N2 and double bond formation.12a,c
Mechanistic studies have ruled out a process involving olefin
formation from the reaction of two metal carbene complexes
(MdCR2). On the other hand, it was proposed that the self-
coupling of PhCOCHdN2 on Cu powder, which yielded olefin
4, could occur either by coupling of two adjacent :CHC(dO)R
(31) Yates, P. J. Am. Chem. Soc. 1952, 74, 5376.
(32) Akita, M.; Hua, R.; Knox, S. A. R.; Moro-oka, Y.; Nakanishi, S.;
Yates, M. I. J. Organomet. Chem. 1998, 569, 71.
(30) (a) Regitz, M.; Maas, G. Diazo Compounds: Properties and Synthesis;
Academic Press: London, 1996. (b) Aggarwal, V. K.; Vicente, J. D.;
Bonnert, R. V. Org. Lett. 2001, 3, 2785. (c) Uchida, T.; Katsuki, T.
Synthesis 2006, 10, 1715.
(33) (a) Samant, R. G.; Graham, T. W.; Rowsell, B. D.; McDonald, R.;
Cowie, M. Organometallics 2008, 27, 3070. (b) Cowie, M. Can.
J. Chem. 2005, 83, 1043.
9
11738 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Catalytic Reactions of Carbene Precursors on Bulk Au
A R T I C L E S
Scheme 2
because of the fast self-coupling of PhCHdN2, which was noted
in the self-coupling reactions in the previous section. On the
other hand, reactions of EDA with PhCOCHdN2 (entry 7), EDA
with Me-MPDA (entry 8), and PhCOCHdN2 with Me-MPDA
(entry 9) all afforded the corresponding cross-coupled products
10, 11, and 12, respectively. Since EDA, PhCOCHdN2, and
Me-MPDA had similar rates of self-coupling (see previous
section), they are able to cross-couple with each other.
In the reactions of DPCP with PhCHdN2 (entry 3), EDA
(entry 4), and PhCOCHdN2 (entry 5), cross-coupled products
7, 8, and 9 are observed, which suggests that they have similar
self-coupling rates. On the other hand, the reaction of DPCP
and Me-MPDA (entry 6) gave only self-coupled products 3 and
5, which indicates that DPCP undergoes more rapid self-
coupling than Me-MPDA does. All of these results from the
cross-coupling studies may be summarized by assuming the
following order of self-coupling rates: PhCHdN2 . DPCP >
EDA > PhCOCHdN2 > Me-MPDA. Cross-coupled products
are observed only in reactions of reactants that have similar
self-coupling rates.
diazomethane and diazoethane which gave the polymers, polym-
ethylene and polyethylidene, respectively, and small amounts of
the olefins, ethylene and 2-butene;34 colloidal gold was the catalyst
in these reactions. More recently, polymethylene film formation
from CH2N2 has been investigated on gold surfaces; such films on
gold have been studied for their applications in the development
of sensors.35
Surface science studies of CH2 precursors (CH2N2 or CH2I2)
on metal surfaces (Pt(111), Ru(001), Pd(100), Rh(111), or
Cu(110)) provide evidence for CH2 on the metal surfaces, but
the nature of the attachment of the CH2 group to one or more
metal atoms has not been established.36 Oxygen atoms and NH
groups are isoelectronic with CH2. Their formation on Au(111)
is supported by spectroscopic studies and their chemical
reactions.37
This trend in self-coupling rates is also useful for understand-
ing the relative amounts of cross-coupled and self-coupled
products in the reactions. Thus, in the reaction of PhCHdN2
with DPCP (entry 3), the olefin products consist of relatively
little of the cross-coupled products 7 (24%) together with self-
coupled products 5 (18%) and 2 (58%) in the following ratios:
cis-2/trans-2/cis-7/trans-7/cis-5/trans-5 29:29:10:14:8:10. How-
ever, in the reaction of EDA with DPCP (entry 4), the higher
amount of cross-coupled product 8 (41%) as compared with
self-coupled products (1 (22%) and 5 (37%)) [cis-5/trans-5/
cis-8/trans-8/cis-1/trans-1, 16:21:21:20:8:14] indicates that DPCP
and EDA are closer in reactivity than DPCP and PhCHdN2.
The amounts of cross-coupled products formed in the other
reactions in Table 2 can also be explained by the trend in self-
coupling rates given in the previous paragraph.
In the reaction of PhCHdN2 and EDA (entry 1), where there
is no cross coupling, the presence of PhCHdN2 reduces the
rate of EDA self-coupling. This is indicated by the observation
that only 50% of the EDA undergoes self-coupling after 24 h,
while in the absence of PhCHdN2 (Table 1, entry 1), EDA
undergoes 100% self-coupling in only 2 h under the same
conditions. Thus, it appears that the more reactive PhCHdN2
and/or the stilbene product 2 adsorbs preferentially on the gold,
which reduces EDA adsorption and subsequent coupling.
In reactions where cross-coupling occurs, it is conceivable
that the cross-coupled products are formed by metathesis of the
self-coupled olefin products (eq 5), in which a surface carbene
acts as the catalyst. That this does not occur was shown by an
As described in the Introduction, 3,3-diphenylcyclopropene
(DPCP) is often used for the synthesis of vinyl carbene
complexes (eq 1). The goal of this aspect of the project was to
determine whether or not gold atoms on a gold metal surface
were also capable of promoting the rearrangement of DPCP to
vinyl carbene groups that would undergo self-coupling to give
the corresponding olefin. Indeed, when DPCP was stirred with
gold powder at 60 °C in CH3CN for 4 h, an 82% yield of the
triene product 5 was obtained in a 40:60 cis/trans ratio of
isomers (eq 4 and Table 1, entry 11). In the absence of gold,
no reaction occurred. In contrast, Grubbs reported that the
homogeneously catalyzed self-coupling of DPCP with [Cp*-
RuCl]2dCHCHdCPh2 gave only trans-5,14c while a tungsten
complex WCl2(O)[P(OMe)3]3 gave cis-5 as the major product.19c
The mechanism of the gold-catalyzed coupling reaction (eq 4)
is not known, but it seems likely to involve a surface-bound
vinylcarbene because it gives the same type of olefin product
that is obtained from the self-coupling of diazoalkanes. To our
knowledge, this is the first example of a reaction of DPCP that
is catalyzed by a metal surface.
2.3. Cross-Coupling of Carbene Precursors. If surface car-
bene intermediates are formed in the self-coupling reactions,
one might expect cross-coupled products when two different
carbene precursors are present. This is observed in many cases,
but not always as shown in Table 2. For example, the reactions
of PhCHdN2 with EDA (entry 1) and PhCHdN2 with
PhCOCHdN2 (entry 2) gave only self-coupled products. During
the course of the reaction of PhCHdN2 with EDA, we observed
that PhCHdN2 was consumed first and converted entirely to
cis and trans stilbene 2 within 2 h, but most of the EDA was
still unreacted. After 24 h, the yield and cis/trans ratio of
stilbenes did not change; however, more diethyl maleate (cis-
1) and diethyl fumarate (trans-1) were produced, but still only
50% of the EDA had been converted to 1. These results showed
that the self-coupling of PhCHdN2 is much faster than cross-
coupling of PhCHdN2 with EDA. Similarly, no cross-coupling
occurred when PhCHdN2 and PhCOCHdN2 were present,
examination of the reaction of PhCHdN2 and DPCP (Table 2,
entry 3), which gives 24% of the cross-coupled products 7.
When a mixture of DPCP/cis-stilbene/trans-stilbene (1:0.5:0.5)
and gold powder (1.0 g) was heated in CH3CN at 60 °C for
24 h, no cross-coupled product 7 was formed. Only the self-
coupled products cis- and trans-5 (81% total yield) and
unreacted cis- and trans-stilbene (in the same 1:1 cis/trans) ratio
were present in the product mixture. In a study of the reaction
(Table 2, entry 4) of DPCP and EDA, which gave a 41% yield
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11739

A R T I C L E S
Zhou et al.
Table 2. Cross Coupling of Carbene Precursors Catalyzed by Gold Powdera
a 0.20 mmol of one substrate mixed with 0.20 mmol of another substrate, 1.0 g of gold powder, 5 mL of solvent at 60 °C for 24 h. b Ratios were
c
determined by GC and H NMR. Yields were determined by H NMR with Ph3CH as the internal standard. d Toluene solvent. e ClCH2CH2Cl solvent.
1
1
f CH3CN solvent. g Low yields due to unidentified byproducts.
of cross-coupled product 8, DPCP was reacted with diethyl
maleate (cis-1) and diethyl fumarate (trans-1) under the above
conditions. Again, no cross-coupled products were detected
which rules out metathesis as the mechanism for the formation
of the cross-coupled products in the reactions of different
carbene precursors.
2.4. Cyclopropanation of Carbene Precursors with Olefins.
Another type of reaction that is characteristic of carbene ligands
in transition metal complexes is their reaction with olefins to
form cyclopropanes; there are many reactions of olefins with
diazo compounds that are catalyzed by transition metal
complexes.10,38 Therefore, reactions of a series of olefins and
diazo compounds were investigated with gold powder as the
catalyst (Table 3). The reactions were conducted at 80 °C for
24 h with a 100:1 olefin/diazo compound ratio because yields
of the cyclopropanes A were much lower at lower ratios. The
excess olefin made the cyclopropanation reaction competitive
with self-coupling of the diazo compound to give olefin B.
Uncharacterized byproducts, perhaps oligomers of styrene, were
also detected in the GC of product mixtures.
the solvent, instead of toluene, improved the yield of cyclo-
propanes to 45% (entry 2). Without the gold catalyst, the
reaction of EDA with styrene (entry 3) gave a much lower yield
(10%) of the cyclopropane products. As compared with styrene,
4-methylstyrene (4-Mestyrene) reacted with EDA to give a
higher yield of cyclopropanes (54%, entry 4), while 4-meth-
oxystyrene (4-MeOstyrene) generated a mostly uncharacterized
product mixture and only a 6% yield of cyclopropanes (entry
5). 1-Hexene failed to react with EDA over gold powder either
in 1,2-dichloroethane solvent or in 1-hexene as the solvent.
Unreacted EDA was present in the final reaction mixture, which
indicates that self-coupling of EDA to olefin 1 (Table 1) was
inhibited by the 1-hexene. Cyclohexene and n-butyl vinyl ether
do not react with EDA under the gold-catalyzed conditions; only
the self-coupled olefins 1 are obtained in 52 and 65% yields,
respectively. The reaction of PhCOCHdN2 in neat cyclohexene
(34) (a) Nasini, A. G.; Trossarelli, L.; Saini, G. Makromol. Chem. 1961,
44-46, 550. (b) Nasini, A. G.; Saini, G.; Trossarelli, L. Pure Appl.
Chem. 1962, 4, 255, and references therein.
(35) (a) Seshadri, K.; Atre, S. V.; Tao, Y.-T.; Lee, M.-T.; Allara, D. L.
J. Am. Chem. Soc. 1997, 119, 4698. (b) Seshadri, K.; Wilson, A. M.;
Guiseppi-Elie, A.; Allara, D. L. Langmuir 1999, 15, 742. (c) Hu, W.-
S.; Weng, S.-Z.; Tao, Y.-T.; Liu, H.-J.; Lee, H.-Y.; Fan, L.-J.; Yang,
As shown in Table 3 (entry 1), the reaction of EDA with
styrene in toluene yielded 31% ethyl 2-phenylcyclopropanecar-
boxylate in a 27:73 cis/trans ratio. Using 1,2-dichloroethane as
9
11740 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Catalytic Reactions of Carbene Precursors on Bulk Au
A R T I C L E S
Table 3. Gold Metal-Catalyzed Cyclopropanation of Olefinsa
reacted in ClCH2CH2Cl solvent with the gold catalyst, no
cyclopropanation occurred; only 10% of the self-coupled olefin
4 was detected. The Au-catalyzed reaction of Me-MPDA with
styrene also gave no cyclopropanation product in toluene
solvent; even the yield of the self-coupled olefin product 3 was
much lower (<5%) than that (76%) in the reaction of Me-MPDA
in the absence of styrene (Table 1, entry 5) which suggests that
styrene inhibits the self-coupling of Me-MPDA. Attempts to
prepare cyclopropenes by reaction of alkynes (PhCCH,
4-MeOC6H4CCH, 1-hexyne, and 3-hexyne) with EDA and
PhCHdN2 with 100:1 RCCH/R1R2CN2 in ClCH2CH2Cl in the
presence of gold powder at 80 °C were unsuccessful.
The highly active PhCHdN2 reacts with styrene at 80 °C
over 10 h, even in the absence of gold, to give a 94% yield of
1,2-diphenylcyclopropane as a cis/trans (39:61) mixture (entry
8). In the presence of gold powder, the same reaction yielded
86% cyclopropanes with reversed diastereoselectivity (cis/trans
80:20, entry 9) and 12% stilbenes 2. Thus, the gold influences
the isomer selectivity of the reaction, although it does not
improve the overall cyclopropane yield. At room temperature,
the reaction of PhCHdN2 and styrene yielded 13% cyclopro-
panes and 63% stilbenes 2 using the gold catalyst (entry 10),
while 87% cyclopropanes and only 7% stilbene 2 were produced
in the absence of gold (entry 11). Thus, the yield of cyclopro-
panes is lower in the presence of the gold catalyst, because gold
is such an effective catalyst for PhCHdN2 self-coupling.
3,3-Diphenylcyclopropene (DPCP) also reacted with styrene
to afford a 19% yield of cyclopropanation product 13 as the
trans isomer14e (entry 7 and eq 6); in the absence of Au, no
cyclopropanation was observed. To our knowledge, the only
a 0.1 mmol of the diazo compound, alkene/diazo ) 100:1, 1.0 g of
gold powder, 30 mL of solvent, 80 °C, 24 h under N2. b Yields were
determined by 1H NMR with Ph3CH as the internal standard, based on
the diazo compounds. c Cis/trans ratios were determined by GC and 1H
NMR. d Not determined. e No catalyst. f Low yield due to
uncharacterized byproducts. g Room temperature, 24 h.
gave only the self-coupled olefins 4 in a 1.3:1.0 cis/trans ratio.
As for the gold powder catalyzed cyclopropanation reactions,
the homogeneous Fe(TTP) (TTP ) meso-tetra-p-tolylporphyrin)
catalyzes cyclopropanation reactions of styrenes with EDA much
faster than reactions of nonaryl alkenes; it also catalyzes more
electron-rich styrenes (e.g., 4-methylstyrene) faster than styrene
itself.38a On the other hand, the reaction of styrene with EDA
is only slightly faster than that of 1-decene when the catalyst is
Rh(TTP)(I).38k Possible reasons for the differences in olefin
reactivity using the iron and rhodium complexes have been
discussed.38a
In gold-catalyzed reactions of styrene with other diazo
compounds, PhCOCHdN2 reacts with neat styrene to give a
35% yield of the cyclopropane (entry 6), but much of the styrene
polymerized; less than 5% cyclopropanation was observed in
the absence of Au. When PhCOCHdN2 and styrene were
previous example of the cyclopropanation of styrene with DPCP
was catalyzed by the homogeneous complex (pybox)(Cl)2Ru-
(CHsCHdCPh2), where pybox is 2,6-bis(4(S)-isopropyloxazo-
lin-2-yl)pyridine.14e
In all of these gold-catalyzed cyclopropanation reactions, a
competing reaction is self-coupling of the diazo compound or
DPCP to give the corresponding olefin. To minimize the self-
coupling reactions, it is necessary to use a 100-fold excess of
the olefin. Yet even then, the only olefins that gave cyclopro-
panation products are styrene and its substituted derivatives.
As in homogeneous cyclopropanation reactions catalyzed by
Rh2(OAc)410b and metalloporphyrins,38a-c bulk gold favors the
trans product in reactions of EDA (entries 2, 4, 5, 6). Although
the mechanism of the gold-catalyzed cyclopropanation reactions
is not known, it seems likely that they proceed through carbene
intermediates adsorbed on the gold surface, which react with
the olefin (either adsorbed or in solution) to give the cyclopro-
pane products. To our knowledge, the only other metal that is
reported to catalyze cyclopropanation is copper powder, which
has been used in a variety of organic syntheses.39
Y.-W. Langmuir 2007, 23, 12901. (d) Bai, D.; Jennings, G. K. J. Am.
Chem. Soc. 2005, 127, 3048. (e) Bai, D.; Ibrahim, Z.; Jennings, G. K.
J. Phys. Chem. C 2007, 111, 461.
(36) (a) Bent, B. E. Chem. ReV. 1996, 96, 1361. (b) Solymosi, F. Catal.
Today 1996, 28, 193.
(37) (a) Zielasek, V.; Xu, B.; Liu, X.; Ba¨umer, M.; Friend, C. M. J. Phys.
Chem. C 2009, 113, 8924. (b) Deng, X.; Friend, C. M. Surf. Sci. 2008,
602, 1066. (c) Ojifinni, R. A.; Gong, J.; Flaherty, D. W.; Kim, T. S.;
Mullins, C. B. J. Phys. Chem. C 2009, 113, 9820.
(38) (a) Wolf, J. R.; Hamaker, C. G.; Djukic, J.-P.; Kodadek, T.; Woo,
L. K. J. Am. Chem. Soc. 1995, 117, 9194. (b) Hamaker, C. G.;
Mirafzal, G. A.; Woo, L. K. Organometallics 2001, 20, 5171. (c)
Hamaker, C. G.; Djukic, J.-P.; Smith, D. A.; Woo, L. K. Organome-
tallics 2001, 20, 5189. (d) Evans, D. A.; Woeroel, K. A.; Hinman,
M. M.; Faul, M. M. J. Am. Chem. Soc. 1991, 113, 726. (e) Pfaltz, A.
Acc. Chem. Res. 1993, 26, 339. (f) Li, Y.; Huang, J.-S.; Zhou, Z.-Y.;
Che, C.-M. J. Am. Chem. Soc. 2001, 123, 4843–4844. (g) Nishiyama,
H.; Itoh, Y.; Matsumoto, H.; Park, S.-B.; Itoh, K. J. Am. Chem. Soc.
1994, 116, 2223. (h) Doyle, M. P.; Brandes, B. D.; Kazala, A. P.;
Pieters, R. J.; Jarstfer, M. B.; Watkins, L. M.; Eagle, C. T. Tetrahedron
Lett. 1990, 31, 6613. (i) Doyle, M. P.; Pieters, R. J.; Martin, S. F.;
Austin, R. E.; Oalmann, C. J.; Muller, P. J. J. Am. Chem. Soc. 1991,
113, 1423. (j) Yadav, J. S.; Subba Reddy, B. V.; Purnima, K. V.;
Nagaiah, K.; Lingaiah, N. J. Mol. Catal. A: Chem. 2008, 285, 36. (k)
Maxwell, J. L.; O’Malley, S.; Brown, K. C.; Kodadek, T. Organo-
metallics 1992, 11, 645.
3. Conclusions
Diazoalkanes (R2CdN2) and 3,3-diphenylcyclopropene (DPCP)
are well-established precursors for the synthesis of transition
metal carbene complexes. In the present studies, these precursors
(39) Liu, X.; Liu, Y.; Li, X.; Xiang, S.; Zhang, Y.; Ying, P.; Wei, Z.; Li,
C. Appl. Catal. A: General 2003, 239, 279, and references therein.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11741

A R T I C L E S
Zhou et al.
performed on a Tecnai G2 F20 microscope operated at 200 kV
accelerating voltage. The samples were sonicated in isopropyl
alcohol/water for 5 min, and then a drop of the suspension was
placed on a lacy carbon copper grid (from Ted Pella, Inc.) and
dried under ambient conditions. Electron Energy Loss Spectroscopy
(EELS) images were obtained on a sea urchin gold sample that
was sonicated in water for 5 min. A drop of the suspension was
placed on a lacey carbon copper grid (from Ted Pella, Inc.). The
Energy Filtered TEM (EFTEM) instrument was tuned for carbon
once the EELS images were optimized and was used to establish
the location of carbon on the sea urchin gold particles. Energy
dispersive X-ray analysis was used to establish the elemental
composition of the gold powders. The powder XRD diffraction data
were collected on a Scintag XRD 2000 X-ray diffractometer using
Cu KR radiation.
4.2. General Procedure for the Preparation of Gold Powder.
Gold powder was prepared from HAuCl4 by reduction with
hydroquinone using a procedure similar to that published previ-
ously.22 A solution of hydrogen tetrachloroaurate hydrate (10 g,
0.029 mol) in 30 mL of water was heated to 100 °C, and a solution
of hydroquinone (5.74 g, 0.052 mol) in 60 mL of hot water was
added slowly (during 30 min) through an addition funnel. The
mixture was maintained at ∼90 °C for 1 h, cooled to room
temperature, and filtered; the gold was washed with methanol to
remove the quinone product until the color of the filtrate changed
from dark red to colorless. The gold powder was then extracted
with methanol in a Soxhlet extractor for 16 h. After drying in an
oven at 110 °C for at least 1 h, the dull brown gold powder (5.63
g, 99% yield) was treated with 80 mL of freshly prepared “piranha”
solution (a 3:1 mixture of concentrated H2SO4 and 30% aqueous
H2O2. CAUTION: Add H2SO4 to H2O2 in an ice bath) at room
temperature in a large beaker with slow stirring. Vigorous foaming
and gas evolution occurred during the first 30 min.45 The mixture
was stirred for another 16 h and was then diluted with 400 mL of
distilled water. The gold was filtered on a coarse frit, washed 10
times with 100-mL aliquots of water followed by five 50-mL
methanol washings, and dried in air. The dull brown Au powder
was then heated overnight in air in a 110 °C oven and then cooled
to room temperature.
behave as if they form carbene intermediates on bulk gold metal
surfaces. The catalytic activity of the gold depends significantly
on the morphology and composition (including the presence of
carbon-containing material) of the metal surface; the freshly
prepared sea urchin morphology (with incorporated carbon) is
less active than the smoother gold powder that forms during its
catalysis of the self-coupling of EDA (eq 2). The reactivities
of the carbene intermediates generated from R2CdN2 and DPCP
are quite different than those of the diaminocarbene intermedi-
ates investigated previously2 (Scheme 1). Whereas the diami-
nocarbenes are readily oxidized (O2) to ureas or carbodiimides,
the nonheteroatom carbenes generated in the present studies do
not react with O2 but instead undergo self- or cross-coupling to
form olefins. The higher reactivity of diaminocarbenes with O2
may result from their stronger electron-donor properties which
make the gold surfaces sufficiently electron-rich to react with
O2.2 It is also possible that the differences in reactivity are due
to differences in their binding to the metal surface. Diaminocar-
benes are more likely to prefer coordination to one gold atom
while alkylidene-type carbenes (R2C:) may prefer to bridge two
gold atoms; these binding differences are observed in transition
metal complexes.10c,d The carbene coupling and cyclopropana-
tion reactions of R2CdN2 and DPCP extend the range of bulk
gold catalysis to reactions that are very different than the
oxidative reactions reported previously.2-8
4. Experimental Section
4.1. General Methods. Acetonitrile and 1,2-dichloroethane were
heated at reflux with CaH2 under N2 overnight and distilled before
use. Toluene was heated at reflux under N2 over sodium benzophe-
none and distilled before use. HAuCl4 was purchased from Strem
Chemicals, Inc.; triethylbenzylammonium chloride (TEBAC) and
ethyl diazoacetate (EDA) were purchased from Sigma Aldrich and
used as received. Styrene, 4-methylstyrene, 4-methoxystyrene,
acrylonitrile, 1-hexene, phenylacetylene, 1-hexyne, and 3-hexyne
were purchased from Sigma-Aldrich and purified by distillation
before use. It was necessary to distill these substrates from the
inhibitors as they prevented the cyclopropanation reactions from
occurring. Methyl(4-methylphenyl)diazoacetate (Me-MPDA),40 3,3-
diphenylcyclopropene (DPCP),18,41 R-diazoacetophenone,42 di-
methyl diazomalonate,43 and PhCHdN244 were prepared according
to literature procedures.
4.3. Regeneration/Recycling of the Gold Powder Catalyst.
4.3.1. Using Piranha Solution. After use in a catalytic reaction, the
gold powder was filtered from the reaction mixture and washed
with CH3CN (5 × 30 mL), CH2Cl2 (5 × 30 mL), and then methanol
(5 × 30 mL). After filtration, the gold was stirred with 40 mL of
freshly made piranha solution for 16 h at room temperature. The
mixture was diluted with 200 mL of distilled water and filtered.
The gold was washed 10 times with 100-mL aliquots of water
followed by five 50-mL methanol washings and dried in air. The
Au powder was then heated overnight in air in a 110 °C oven,
after which it was cooled to room temperature for the next reaction
cycle.
1
Routine H and 13C NMR spectra were recorded on a Varian
VXR-400 or Bruker DRX-400 spectrometer. Mass spectra were
measured on a Finnigan TSQ 700 spectrometer. HRMS were
recorded on a Kratos MS 50 mass spectrometer. Capillary gas
chromatography was performed on an HP-6890 instrument equipped
with an HP-5 column and flame ionization detector. GC-MS
analyses were performed on a Finnigan TSQ 700 or Magnum ITD
instrument. The morphology and size of the particles in gold powder
samples were analyzed by scanning electron microscopy (SEM)
and transmission electron microscopy (TEM). The SEM measure-
ments were performed on a JEOL 840A microscope operated at
10 kV accelerating voltage and 0.005 nA beam current. The TEM
measurements with energy dispersive X-ray analysis (EDX) were
After many uses (∼10), the gold was dissolved in aqua regia to
form HAuCl4,22 which was converted back to gold powder
according to the above procedures.
4.3.2. By Washing with Solvent. Following use in a catalytic
reaction, the gold powder was filtered from the reaction mixture
and washed with CH3CN (5 × 30 mL), CH2Cl2 (5 × 30 mL), and
then methanol (5 × 30 mL). The gold powder was dried in an
oven overnight at 110 °C. After cooling to room temperature, it
was used in the next cycle. The use of gold that was recycled in
this way is shown in Figure S4 of the Supporting Information.
4.4. General Procedure for the Gold-Catalyzed Self-Coupling
Reactions under N2 and O2. A septum-capped glass tube (2.5 cm
× 18 cm, ∼65 mL volume) charged with gold powder (1.0 g) and
a stir bar was evacuated under vacuum and backfilled with nitrogen;
this procedure was repeated three times. A solution of the diazo
compound (0.20 mmol) or 3,3-diphenylcyclopropene (0.20 mmol)
(40) (a) Davies, H. M. L.; Cantrell, W. R.; Romines, K. R., Jr.; Baum,
J. S. Org. Synth. 1992, 70, 93. (b) Davies, H. M. L.; Hansen, T.;
Churchill, M. R. J. Am. Chem. Soc. 2000, 122, 3063.
(41) Bolesov, I. G.; Ignatchenko, A. V.; Bovin, N. V.; Prudchenko, I. A.;
Surmina, L. S.; Plemenkov, V. V.; Petrovskii, P. V.; Romanov, I. V.;
Mel’nik, I. I. Zh. Org. Khim. 1990, 26, 102.
(42) (a) Danheiser, R. L.; Miller, R. F.; Brisbois, R. G.; Park, S. Z. J. Org.
Chem. 1990, 55, 1959. (b) Danheiser, R. L.; Miller, R. F.; Brisbois,
R. G. Org. Synth. 1996, 73, 134.
(43) Baum, J. S.; Shook, D. A. Synth. Commun. 1987, 17, 1709.
(44) (a) Creary, X. Org. Synth. 1986, 64, 207. (b) Wulfman, D. S.;
Yousefian, S.; White, J. M. Synth. Commun. 1988, 18, 2349.
(45) Ontko, A. C.; Angelici, R. J. Langmuir 1998, 14, 1684.
9
11742 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Catalytic Reactions of Carbene Precursors on Bulk Au
A R T I C L E S
in a degassed solvent (5 mL) was added by syringe under nitrogen.
The mixture was stirred vigorously at 60 °C for the desired time
and then worked up by filtration to remove the gold powder. After
removing the solvent by rotary evaporation, triphenylmethane
(0.03-0.05 mmol, internal standard) and CDCl3 were added to the
reaction mixture residue to determine yields and cis/trans ratios of
4.6. General Procedure for the Catalytic Cyclopropanation
Reactions. A 100-mL three-necked round-bottomed flask charged
with gold powder (1.0 g) was evacuated under vacuum and
backfilled with N2; this procedure was repeated three times.
Degassed solvent (15 mL) and distilled alkene (10 mmol) were
added through a septum by syringe. A solution of the diazo
compound (0.1 mmol) in a solvent (15 mL) in an addition funnel
was added dropwise over a period of 6 h at 80 °C under a N2
atmosphere. After the addition was complete, the mixture was stirred
vigorously at 80 °C for an additional 18 h. The mixture was filtered
to remove the gold powder, and the solvent was removed by rotary
evaporation. Triphenylmethane (∼0.03 mmol) was added to the
reaction mixture residue as an internal standard to determine the
1
the products by H NMR spectroscopy.
The reactions under an O2 atmosphere were conducted in the
same way except all the reactants were combined in air, rather
than under N2. Then, a syringe needle attached to a rubber
balloon (∼1.0 L) containing O2 was inserted into the septum
covering the tube opening. The reactions were carried out and
worked up as described for the reactions under N2.
1
yields and cis/trans ratios of products by H NMR and GC.
4.5. General Procedure for the Catalytic Cross-Coupling Reac-
tions. A mixture of one diazo compound (0.20 mmol) and a different
diazo compound or 3,3-diphenylcyclopropene (0.20 mmol) and gold
powder (1.0 g) in acetonitrile (5 mL) was prepared in a glass tube
(2.5 cm × 18 cm, ∼65 mL volume). The tube was sealed with a
rubber septum allowing inclusion of air, and the mixture was stirred
vigorously (magnetic stir bar) at 60 °C for 24 h. After cooling to
room temperature and removing the gold powder by filtration, the
solvent was removed by rotary evaporation. Triphenylmethane
(∼0.05 mmol) was added to the reaction mixture residue as an
Acknowledgment. This research was supported by the U.S.
Department of Energy under Contract No. DE-AC02-07CH11358
with Iowa State University. We thank Matthew J. Kramer for useful
discussions.
Supporting Information Available: Reaction product spec-
troscopic characterizations, PhCHdN2 synthesis, figure of
recycled gold activity. This material is available free of charge
via the Internet at http://pubs.acs.org.
1
internal standard for H NMR analysis in CDCl3 solvent.
JA900653S
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11743