Gold metal-catalyzed reactions of isocyanides with primary amines and oxygen: Analogies with reactions of isocyanides in transition metal complexes

Article abstract of DOI:10.1021/ja0618907

Despite its generally poor catalytic properties, bulk gold metal is observed to catalyze reactions of isocyanides (C≡N-R) with primary amines (H₂N-R′) and O₂ to give carbodiimides (R-N=C=N-R′) at room temperature and above. Detailed infrared reflection absorption spectroscopic (IRRAS) and kinetic studies show that the reaction occurs by initial η¹-adsorption of the isocyanide on the Au surface, which activates the isocyanide to attack by the amine. This attack is the rate-determining step in the catalytic cycle and has characteristics very similar to those of amine reactions with coordinated isocyanides in transition metal complexes. However, the metallic Au surface provides a pathway involving O₂ to give the carbodiimide product whereas homogeneous metal ion catalysts give formamidines [HC(=NR)(NHR′)].

Full text of DOI:10.1021/ja0618907

Published on Web 07/25/2006
Gold Metal-Catalyzed Reactions of Isocyanides with Primary
Amines and Oxygen: Analogies with Reactions of
Isocyanides in Transition Metal Complexes
Mihaela Lazar and Robert J. Angelici*
Contribution from the Ames Laboratory and Department of Chemistry, Iowa State UniVersity,
Ames, Iowa 50011-3111
Received March 20, 2006; E-mail: angelici@iastate.edu
Abstract: Despite its generally poor catalytic properties, bulk gold metal is observed to catalyze reactions
of isocyanides (CtN-R) with primary amines (H2N-R′) and O2 to give carbodiimides (R-NdCdN-R′) at
room temperature and above. Detailed infrared reflection absorption spectroscopic (IRRAS) and kinetic
studies show that the reaction occurs by initial η1-adsorption of the isocyanide on the Au surface, which
activates the isocyanide to attack by the amine. This attack is the rate-determining step in the catalytic
cycle and has characteristics very similar to those of amine reactions with coordinated isocyanides in
transition metal complexes. However, the metallic Au surface provides a pathway involving O2 to give the
carbodiimide product whereas homogeneous metal ion catalysts give formamidines [HC(dNR)(NHR′)].
Scheme 1. Modes of Isocyanide Adsorption on Metal Surfaces
Introduction
Organic isocyanides (R-NtC) adsorb on the surfaces of the
following metals: nickel,1 rhodium,2 palladium,3 platinum,4
silver,5 and gold in different forms (film,6 powder,7 or nano-
particles8). Some modes of isocyanide adsorption on these
different metals are shown in Scheme 1. On Au and Ag, η1 is
the observed CtN-R adsorption mode.9,10 The bonding
interaction between Au and the isocyanide primarily involves
σ donation of the lone pair of electrons from carbon to gold.
Bulk gold metal is a very poor π donor, as evidenced by its
very weak interaction with carbon monoxide,11 which suggests
that there is little if any π back-donation from the metal to the
isocyanide.
Diisocyanides, CtN-R-NtC, with a flexible alkyl carbon
chain (R), adsorb to the gold surface through both isocyano
groups.7d,12,13 On the other hand, when R is a rigid aromatic
unit and the isocyano groups are para to each other, only one
isocyano group adsorbs.7a,9,14 The free NC group has been used
to bond other molecules,15 metal clusters,16 nanoparticles,14,17
and to form multilayers on gold surfaces.13 The adsorption of
1,4 phenylenediisocyanide on gold nanoparticles is of special
interest because the diisocyanide is proposed to bind to two
different nanoparticles, with potential applications as a molecular
wire.8a, 8b
(1) (a) Friend, C. M.; Stein, J.; Muetterties, E. L. J. Am. Chem. Soc. 1981,
103, 767. (b) Friend, C. M.; Muetterties, E. L.; Gland, J. L. J. Phys. Chem.
1981, 85, 3256. (c) Hemminger, J. C.; Muetterties, E. L.; Somorjai, G. A.
J. Am. Chem. Soc. 1979, 101, 62.
(2) (a) Cavanagh, R. R.; Yates, J. T., Jr. J. Chem. Phys. 1981, 75, 1551. (b)
Semancik, S.; Haller, G. L.; Yates, J. T., Jr. J. Chem. Phys. 1983, 78, 6970.
(3) Murphy, K.; Azad, S.; Bennett, D. W.; Tysoe, W. T. Surf. Sci. 2000, 467,
1.
(4) (a) Kang, D. H.; Trenary, M. J. Am. Chem. Soc. 2001, 123, 8432. (b) Kang,
D. H.; Trenary, M. J. Phys. Chem. B 2002, 106, 5710. (c) Bong, D.; Tam,
I.; Breslow, R. J. Am. Chem. Soc. 2004, 126, 11796. (d) Avery, N. R.;
Matheson, T. W. Surf. Sci. 1984, 143, 110.
(5) (a) Han, H. S.; Han, S. W.; Joo, S.-W.; Kim, K. Langmuir 1999, 15, 6868.
(b) Ceyer, S. T.; Yates, J. T., Jr. J. Phys. Chem. 1985, 89, 3842. (c) Kim,
N. H.; Seung J. L.; Kim, K. Chem. Commun. 2003, 6, 724.
(6) (a) Kim, S.; Ihm, K.; Kang, T.-H.; Hwang, S.; Joo, S.-W. Surf. Interface.
Anal. 2005, 37, 294. (b) Dahleke, R.; Schollwock, U. Phys. ReV. 2004, 69,
085324/1. (c) Stapleton, J. J.; Daniel, T. A.; Uppili, S.; Cabarcos, O. M.;
Naciri, J.; Shashidhar, R.; Allara, D. L. Langmuir 2005, 21, 11061.
(7) (a) Robertson, M. J.; Angelici, R. J. Langmuir 1994, 10, 1488. (b) Shih,
K.-C.; Angelici, R. J. Langmuir 1995, 11, 2539. (c) Ontko, A. C.; Angelici,
R. J. Langmuir 1998, 14, 1684. (d) Ontko, A. C.; Angelici, R. J. Langmuir
1998, 14, 3071.
Despite the large number of studies of adsorbed isocyanides
on gold surfaces, there are no reports of reactions of adsorbed
(11) (a) Grunwald, J. L.; Baiker, A. J. Phys. Chem. B 1999, 103, 1002. (b)
Ikezava, Y.; Saito, H.; Matsubayashi, H.; Toda, G. J. Electroanal. Chem.
1988, 252, 395.
(12) Henderson, J. I.; Feng, S.; Bein, T.; Kubiak, C. P. Langmuir 2000, 16,
6183.
(8) (a) Bae, S. J.; Lee, C.-R.; Choi, I. S.; Hwang, C.-S.; Gong, M.-S.; Kim,
K.; Joo, S.-W. J. Phys. Chem. B 2002, 106, 7076. (b) Lee, C.-R.; Kim, S.
I.; Yoon, C.-J.; Gong, M.-S.; Choi, B. K.; Kim, K.; Joo, S.-W. J. Colloid.
Interface Sci. 2004, 271, 41. (c) Joo, S.-W.; Kim, W.-J.; Yun, W. S.; Hwang,
S.; Choi, I. S. Appl. Spectrosc. 2004, 58, 218. (d) Joo, S.-W.; Kim, Y.-S.
Colloids Surf., A 2004, 234, 117.
(13) Lin, S.; McCarley, R. L. Langmuir 1999, 15, 151.
(14) Kim, H. S.; Lee, S. J.; Kim, N. H.; Yoon, J. K.; Park, H. K.; Kim, K.
Langmuir 2003, 19, 6701.
(15) Huc, V.; Bourgoin, J.-P.; Bureau, C.; Valin, F.; Zalczer, G.; Palacin, S. J.
Phys. Chem. B 1999, 103, 10489.
(16) Henderson, J. I.; Feng, S.; Ferrence, G. M.; Bein, T.; Kubiak, C. P. Inorg.
Chim. Acta 1996, 242, 115.
(17) Horswell, S. L.; O’Neil, I. A.; Schiffrin, D. J. J. Phys. Chem. B 2001, 105,
941.
(9) Swanson, S. A.; McClain, R.; Lovejoy, K. S.; Alamdari, N. B.; Hamilton,
J. S.; Campbell, S. J. Langmuir 2005, 21, 5034.
(10) Murphy, K. L.; Tysoe, W. T.; Bennett, D. W. Langmuir 2004, 20, 1732.
9
10.1021/ja0618907 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 10613-10620
10613

A R T I C L E S
Lazar and Angelici
isocyanides or attempts to determine whether these molecules
are activated to react when adsorbed to gold; however, the
addition of a hydrogen atom to the nitrogen atom of CtN-
CH3 adsorbed on Pt(111) has been reported.4a,4b It is well-known
that η1-coordinated isocyanides in transition metal complexes
are activated to attack by nucleophiles such as amines (eq 1)
and alcohols.18-20
with primary amines as is observed for isocyanide ligands in
metal complexes. Additional goals were to explore the pos-
sibility that diaminocarbene groups are detectable on the gold
surface, to determine whether formamidine or some other
products are formed in the reactions, and to determine whether
reactions of isocyanides and primary amines are catalyzed by
gold metal, even though bulk gold metal is known to be a poor
catalyst; however, it is noted that supported gold nanoparticles
are very active catalysts.27,28 Through these investigations, we
sought to understand the similarities and differences between
reactions of isocyanides in transition metal complexes and on
bulk gold metal surfaces.
An indicator of the susceptibility of such isocyanide ligands
to nucleophilic attack is the magnitude of the increase of the
NtC infrared stretching frequency, ν(NC), of the coordinated
isocyanide as compared to that of the free R-NtC molecule.
A value of ∆ν(NC) ) ν(NC)coord - ν(NC)free greater than 40
cm-1 in some systems indicates that a coordinated isocyanide
is capable of reacting with nucleophilic agents such as amines.19,20
Diaminocarbene complexes resulting from such a reaction have
been isolated for complexes of Pt, Pd, Au, Ag, and Fe.19,20
Isocyanide complexes of Au(I) and Au(III) that have been
reported to undergo such reactions are (C6F5)(p-CH3C6H4Nt
C)Au,21 (p-CH3C6H4NtC)AuCl,22 [(RNtC)2Au]+ (R ) C6H11,
p-CH3C6H4),23 and (C6F5)3(p-CH3C6H4NtC)Au.24 The isocya-
nide ligand in the gold clusters [Au8(PPh3)7(CNR)] (R ) tert-
butyl; i-propyl) also reacts with primary amines to form
diaminocarbene cluster complexes.25 The adsorption of isocya-
nides on metallic gold surfaces is also accompanied by an
increase in ν(NC) to give ∆ν(NC) values7a,9 between 55 and
90 cm-1, which suggests that adsorbed isocyanides should be
susceptible to attack by nucleophiles.
Experimental Section
Except where stated otherwise, all experiments were performed in
air, at room temperature. The isocyanides n-butyl isocyanide (n-BuNC),
1,1,3,3-tetramethylbutyl isocyanide (TMBNC), and cyclohexyl isocya-
nide (CyNC) were purchased from Sigma Aldrich and Strem Chemicals
and kept in a freezer at -20 °C. (Although isocyanides have a
penetrating odor, the quantities used in these studies were so small
that odors were avoided by performing all manipulations and reactions
in the hood. Glassware was washed with KOH/ethanol solutions. Gold
films containing adsorbed isocyanides have a slight odor when handled
outside the hood.) The amines n-dodecylamine (n-DoNH2), cyclohexyl-
amine (CyNH2), benzylamine (BzNH2), obtained from Sigma Aldrich,
and n-butylamine (n-BuNH2) obtained from Alfa Aesar, were used as
received. The solvents 1,2-dichloroethane (DCE), ethanol (EtOH) and
hexane are spectrophotometric grade and were obtained from Sigma
Aldrich.
Gold Substrate Preparation. The gold films were prepared on glass
slides (25 mm/75 mm) by physical vapor deposition. First, a 15 nm Cr
layer was deposited and then 300 nm of gold.29 Evaporation and
deposition were performed in a cryopumped E306A Edwards Coating
System, maintaining the pressure during evaporation below 7 × 10-6
Torr and a deposition rate of 0.2 nm/s. Gold slides were stored in dry
air, in a desiccator. The gold powder was prepared from HAuCl4 by
reduction with hydroquinone according to a published procedure30 and
had a surface area of 0.35 m2/g and a particle diameter of approximately
1 µm.7a The gold powder was washed with methanol in a Soxhlet
extractor for 24 h and dried in air in the oven at 110 °C for 1 h.
Gold Surface Cleaning. Clean gold surfaces for isocyanide mono-
layer formation were obtained by treating previously used gold
substrates with “piranha” solution (a 3:1 mixture of concentrated H2SO4
and 30% aqueous H2O2; CAUTION: Add H2SO4 to H2O2, manipulate
carefully and neutralize everything with Na2CO3 immediately after
finishing the experiment). The gold slides were placed in 50 mL of
freshly prepared “piranha” solution for 5 min. Then the slides were
washed thoroughly with deionized water and methanol and dried in an
oven at 125 °C for 2 h. The gold powder was cleaned in a similar
manner using our previously published method.7c Both the gold film
and the gold powder, when treated in this way, gave clean surfaces
that exhibited no absorption peaks in their IR spectra.7c Such cleaned
Au surfaces adsorbed isocyanide like the freshly prepared Au substrates.
Monolayer Preparation. Monolayers were formed by immersing
the gold slides for 24 h into 50 mL of 5 mM isocyanide solution in
DCE. Upon removal from the isocyanide solution, the slides were rinsed
Reactions between isocyanides and primary amines to give
formamidines (eq 2) are sometimes catalyzed by metal ions.
Such is the case for group 11 and 12 metals (Cu+, Ag+, Zn2+
,
Cd2+, and Hg2+). It was proposed that these catalytic reactions
proceed by a mechanism that involves initial coordination (η1)
of the isocyanide to the metal ion followed by attack of the
amine on the isocyanide carbon atom, and then elimination of
the product formamidine.22 For Au(I) and Au(III) complexes,
the formamidine is liberated from the diaminocarbene complex
only upon displacement by other ligands such as PPh3 or CN-.26
In the present study, we sought to determine whether
isocyanides adsorbed on bulk gold metal are activated to react
(18) Vogler, A. Isonitrile Chemistry; Ugi, I., Ed.; Academic Press: New York,
1971; p 217.
(19) (a) Michelin, R. A.; Pombeiro, A. J. L.; Guedes da silva, M. F. Coord.
Chem. ReV 2001, 218, 75. (b) Chatt, J.; Richards, R. L.; Royston, G. H. D.
J. Chem. Soc., Dalton Trans. 1973, 1433.
(20) Crociani, B. Reactions of Coordinated Ligands, Vol. 1; Braterman, P., Ed.;
Plenum Press: New York, 1986; p 553.
(21) Uson, R.; Laguna, A.; Vicente, J.; Garcia, L.; Bergareche, B. J. Organomet.
Chem. 1979, 173, 349.
(22) Bonati, F.; Minghetti, T. J. Organomet. Chem. 1973, 59, 403.
(23) Minghetti, G.; Baratto, L. J. Organomet. Chem. 1975, 102, 397.
(24) Uson, R.; Laguna, A.; Laguna, M.; Fernandez, E.; Jones, P. G. J. Chem.
Soc., Dalton Trans.: Inorg. Chem. 1982, 10, 1971.
(27) Bond, G. C.; Thompson, D. T. Catal. ReV.-Sci. Eng. 1999, 41, 319.
(28) (a) Hutchings, G. J. Catal. Today 2005, 100, 55. (b) Bond, G. C. Catal.
Today 2002, 72, 5. (c) Grzybowska-Œwierkosz, B. Catal. Today 2006,
112, 3. (d) Bond, G. C.; Thompson, D. T. Appl. Catal. A: Gen. 2006, 302,
1. (e) 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.
(29) (a) Hampy, R. E.; Yost, F. G.; Ganyard, F. P. J. Vac. Sci. Technol. 1979,
16, 25. (b) Audino, R.; Destefanis, F.; Gorgellino, F.; Pollino, E.; Tamagno,
S. Thin Solid Films 1976, 36, 343.
(25) Bos, W.; Kanters, R. P. F.; Van Halen, C. J.; Bosman, W. P.; Behm, H.;
Smits, J. M. M.; Beurskens, P. T.; Bour, J. J.; Pignolet, L. H. J. Organomet.
Chem. 1986, 307, 385.
(26) (a) Parks, J. E. J. Organomet. Chem. 1974, 71 453. (b) Minghetti, G.; Bonati,
F.; Banditelli, G. Inorg. Chem. 1976, 15, 1718.
(30) Block, B. P. Inorg. Synth. 1953, 4, 14.
9
10614 J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006

Reaction of Isocyanides with Primary Amines and O2
A R T I C L E S
with ethanol for 5 s, to remove the unadsorbed isocyanide, and dried
in an air flow for about 10 s. The following procedure was used to
create an isocyanide monolayer on gold powder:7c 1 mL of 5 mM
isocyanide solution in DCE was added to a Pyrex test tube containing
1.00 g of gold powder. The test tube was capped with a Teflon-lined
screw cap, shaken for 30 s, and then allowed to settle for 24 h. The
isocyanide solution was decanted, and the gold powder was allowed
to dry in air for 24 h.
Infrared Reflection Absorption Spectroscopy (IRRAS). IRRAS
measurements were performed on a Nicolet 850 FT-IR spectrometer
with a liquid nitrogen cooled HgCdTe detector and p-polarized light
incident at 82°. During the analysis, the spectrometer was purged with
nitrogen. The spectra were collected by referencing 512 sample scans
to 512 reference scans at a resolution of 2 cm-1 with Happ-Genzel
apodization and reported as -log(R/R0), where R is the reflectivity of
the sample and R0 is the reflectivity of the reference. The reference
was a monolayer of n-octadecanethiol prepared on the same type of
gold slide by immersing the slide for 12 h in a 1 mM solution of
n-octadecanethiol in ethanol.31,32 The spectra were processed to remove
absorptions due to traces of water and CO2 and baseline corrected using
OMNIC software.
Reactions of Adsorbed Isocyanides on Au with Amines. A gold
slide prepared with adsorbed isocyanide on its surface, as described
above, was placed in 25 mL of a 5 mM amine solution in hexane for
a specific period of time in air. Then the slide was removed from the
solution, rinsed in an ethanol flow for 3 s, dried in a nitrogen flow for
30 s, and placed directly in the IRRAS spectrophotometer for the
spectral measurements.
Figure 1. Decrease of the ν(NC) peak intensity of n-BuNC adsorbed on
a Au film with time, upon immersion in solvents. On the y axis is the percent
of the peak intensity expressed as (I/I0) × 100, where I0 is the initial intensity
of the peak and I is the peak intensity at time t.
of the reaction solution and analyzing it by GC using the experimental
conditions described above and standard solutions of n-BuNC and But-
NdCdN-But for quantitation of the n-BuNC and Bun-NdCdN-
Bun present in the reaction solutions. The rate of the reaction of 0.5
mM n-BuNC with 50 mM n-BuNH2 under a pure O2 atmosphere at 60
°C was the same as the same reaction performed under an air
atmosphere, and the yield (68%) of Bun-NdCdN-Bun was nearly
the same as that (76%) achieved under an air atomosphere. When the
same reaction was performed under an argon atmosphere, the reaction
gave small amounts of the Bun-NdCdN-Bun product, but only 4%
of the n-BuNC had reacted after 2 h, whereas 100% of the isocyanide
reacted after the same time when the reaction was conducted in air or
O2. The small amount of reaction observed under argon was presumably
due to adventitious oxygen. These results show that O2 is a reactant in
these reactions.
To 1.00 g of gold powder with isocyanide adsorbed on the surface
was added 1 mL of 5 mM amine solution in hexane in air. The test
tube was capped, and the contents were stirred with a magnetic bar for
5 min in the case of n-BuNH2 and for 10 min for the rest of the amines.
A sample of this solution was analyzed by GC-MS to identify the
reaction products R-NdCdN-R′
Results and Discussion
GC-MS Analyses. These analyses were performed on a Micromass
GCT gas chromatograph mass spectrometer from Waters Corporation,
Beverly, MA. The instrument includes an Agilent 7683 series autosam-
pler, an Agilent 6890 gas chromatograph, and a Micromass time-of-
flight mass spectrometer. Instrument control, data acquisition, and data
processing used a Compaq Deskpro PC workstation with a Windows
XP operating system. The application software was MassLynx 4.0 from
Waters. The computerized libraries include the NIST 2002 from the
National Institute of Standards and Technology (147 198 entries), and
the Wiley Registry of Mass Spectral Data, 7th Edition, from John Wiley
& Sons, Inc. (329 171 entries). A Varian VS-5MS capillary column
(30 m length, 0.25 mm internal diameter, 0.25 µm film thickness, 5%
phenyl, 95% methyl silicone polymer) was used to determine product
yields. The GC analysis conditions were as follows: carrier gas (He);
mode (constant flow); initial column temperature (40 °C); temperature
program: time at initial temperature (2 min)/heating rate (15 °C/min)/
final temperature (280 °C).
Kinetic Studies of the Reaction between n-BuNC, n-BuNH2, and
O2 Catalyzed by Gold Powder. In a 10 mL round-bottom flask was
placed 1.00 g of clean gold powder. A mixture of n-BuNC, n-BuNH2,
and n-decane in hexane solution was added to the gold. n-Decane was
the internal standard for quantitative GC analysis. The flask was
thermostated at the desired temperature using an oil bath. The
experiments were performed in air at room temperature and 60 °C with
continuous stirring with a stirbar. One set of experiments was made
with 0.5 mM isocyanide, and the amine concentration was varied from
5 to 50 mM. Another experiment was made with 4.5 mM isocyanide
and 45 mM amine. For all experiments, the total solution volume was
3.0 mL. The reaction was monitored by periodically removing a sample
Adsorption/Desorption of Isocyanides on Gold Films.
IRRAS spectra of isocyanide monolayers on gold slides, formed
by placing the slides in 50 mL of 5 mM isocyanide in DCE for
24 h at room temperature, showed a strong ν(NC) absorption
at 2200-2225 cm-1. In this region of the spectra, no other peaks
of the analyzed molecules are present. On the gold film, the
ν(NC) absorptions shift to higher wavenumbers as compared
with those of the free isocyanides. For example, the ν(NC) value
for n-BuNC in DCE is 2148 cm-1, whereas that for n-BuNC
adsorbed on gold film is 2224 ( 1 cm-1. Values of ν(NC) for
the free and adsorbed isocyanide for the other two isocyanides
used in this study are: TMBNC, 2133 cm-1 (in DCE), 2205
cm-1 (adsorbed); CyNC, 2140 cm-1 (in DCE), 2216 cm-1
(adsorbed). The IRRAS spectra show that the isocyanide
monolayers on the Au film are stable for at least 2 or 3 weeks,
if the gold slides are kept in a desiccator (in air) at room
temperature.
To determine the stability of the n-BuNC isocyanide mono-
layer toward desorption in a solvent, the Au slides were placed
in 25 mL of hexane at room temperature. Periodically the slides
were removed from the hexane, rinsed with ethanol, dried in
an N2 flow, and the IRRAS spectra were recorded. The decrease
in intensity of the ν(NC) peak was used to determine the
approximate rate at which the isocyanide was removed from
the surface. By repeated immersions of the slide in hexane, it
was possible to determine an approximate rate for the desorption
process (Figure 1). During the desorption, the ν(NC) value of
the remaining adsorbed isocyanide shifts to higher values (for
example, 2225 cm-1 initially, but 2232 cm-1 after ∼40% of
(31) Walczak, M. M.; Alves, A. A.; Lamp B. D.; Porter, M. D. J. Electroanal.
Chem. 1995, 396, 103.
(32) Zhong C.-J.; Zak, J.; Porter, M. D. J. Electroanal. Chem. 1997, 421, 9.
9
J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006 10615

A R T I C L E S
Lazar and Angelici
2the n-BuNC had been removed); this shift suggests that there
are different types of adsorption sites on the gold surface. Within
the first minute, approximately 55% of the adsorbed n-BuNC
is removed from the surface. After that, slow desorption removes
another 10% in approximately 2 h, whereas the remaining 35%
does not desorb even after 9 h (Figure 1). These IRRAS and
rate of desorption studies indicate that there are two categories
of adsorbed n-BuNC: (1) weakly adsorbed with ν(NC) ) 2225
cm-1 and (2) strongly adsorbed with ν(NC) ) 2232 cm-1
.
In a series of STM studies performed on gold films that were
prepared in the same way as those described in the present work,
Porter et al. demonstrated that the Au surface of evaporated
films, with a predominant Au(111) component, has a “rolling
hills” topography with a high step density.31,32 The surface can
be modeled as densely packed arrays of hemispheres on which
about 55% of the surface atoms are located on terraces and 45%
on step edges. For alkanethiol monolayers, the molecules
adsorbed on step sites are more strongly bonded than those on
the terrace sites. Our isocyanide desorption results are consistent
with approximately 55% of the sites (terrace sites) being weakly
adsorbing; the n-BuNC molecules adsorbed on these sites are
quickly desorbed in the first minutes of immersion in hexane.
Our previous DRIFT (diffuse reflectance IR Fourier transform)
spectroscopic investigations7c of the adsorption of isocyanides
on gold powder also showed two types of adsorption regimes,
as characterized by ν(NC) values of 2225 cm-1 at saturation
coverage and 2233 cm-1 for strongly adsorbed n-BuNC; these
values are the same as those for n-BuNC on gold film as
described above. For Au powder, approximately 30% was
weakly adsorbed, whereas 70% of the isocyanide (n-BuNC and
C18H37NC) was strongly adsorbed. The difference between 70%
strongly adsorbed isocyanide on gold powder and 45% on gold
film is likely due to differences in surface topography of these
two forms of gold.
Figure 2. Decrease of the ν(NC) peak intensity of isocyanides adsorbed
on gold slides immersed in 5 mM n-DoNH2/hexane solutions.
Table 1. Rates of Isocyanide Removal from RNC/Au Film in
Amine/Hexane Solutions
isocyanide
amine
t1 2 (s)
/
t′ 2 (s)
1/
amine pKa
n-BuNC
hexane
50
3
3.5
9
-
n-BuNH2
n-DoNH2
CyNH2
5
10.62
10.63
10.64
9.34
6.5
22
87
BzNH2
13
CyNC
hexane
70
5
4
11
16
-
17
13
28
60
n-BuNH2
n-DoNH2
CyNH2
10.62
10.63
10.64
9.34
BzNH2
TMBNC
hexane
100
7
7.5
12
8
-
10
17
27
20
n-BuNH2
n-DoNH2
CyNH2
10.62
10.63
10.64
9.34
BzNH2
intensity decreased drastically. Figure 2 presents the change in
ν(NC) intensity when a slide containing an adsorbed isocyanide
is immersed in 5 mM n-DoNH2 solution. For n-BuNC, after
only 20 s of immersion, the ν(NC) adsorption peak completely
disappeared. For CyNC and TMBNC, the ν(NC) intensity is
zero after 40 and 50 s, respectively. Similar results were obtained
with all amines (n-BuNH2, CyNH2, and BzNH2) used in this
study.
The time required for the disappearance of the ν(NC) peak
is very short compared with that in the desorption experiments
with the hexane solvent only (Figure 1). The faster rate of
isocyanide removal from the surface in the presence of the amine
may be due to a reaction of the adsorbed isocyanide with the
amine and/or an amine promoted desorption of the isocyanide;
these possibilities are discussed in detail in the next section. In
Figure 2, it can be seen that there are two different parts of the
process. In the first part, the ν(NC) peak intensity decreases
rapidly; this is followed by a slower decrease to zero. This
behavior also indicates the existence of two types of isocyanides
and adsorption sites on the gold surface, as was observed in
the solvent desorption studies (Figure 1).
Two other solvents, DCE and EtOH, were used in studies of
the desorption of n-BuNC from Au film. These polar solvents
remove all of the adsorbed isocyanide from the Au surface
within about 5 h (Figure 1). In these cases, an inflection in the
plot occurs at the point where approximately 60% of the
isocyanide is removed from the surface. The weakly adsorbed
molecules are desorbed in a few minutes, followed by slower
desorption of the remaining isocyanide within about 5 h. The
ν(NC) value of adsorbed molecules shifts from 2225 cm-1 in
the initial monolayer to 2234 cm-1 after 60% of the n-BuNC is
removed with DCE; this shift is from 2224 to 2229 cm-1 with
EtOH. Thus, in all of the isocyanide desorption experiments,
the strongly adsorbed isocyanide has a higher ν(NC) value,
which is consistent with the observation in transition metal
complexes in which a greater σ donation of the isocyanide to a
metal center is correlated with a higher ν(NC) value;19 this
greater σ-donation typically occurs in complexes where the
metal has a higher positive charge.
Reaction/Desorption of Isocyanides on Au Films in Amine
Solutions in Air. In a series of experiments, gold slides with
an isocyanide monolayer on the surface were immersed in 25
mL of 5 mM amine in hexane solution for a specific period of
time, in air, at room temperature. After being removed from
solution, the slide was rinsed with ethanol for 3 s, dried in a N2
flow, and placed in the IR spectrometer for analysis. After being
immersed in the amine solution for 10 s, the ν(NC) peak
Table 1 summarizes semiquantitative rate data for the removal
of the isocyanides from the gold film in the presence of amines.
The third column presents the time (t1/2) in which half of the
initial amount of adsorbed isocyanide is removed from the
surface, which is the time at which the ν(NC) peak intensity I
is half that of the initial I0. Considering that approximately 45%
of the total isocyanide is strongly bonded to the surface, column
9
10616 J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006

Reaction of Isocyanides with Primary Amines and O2
A R T I C L E S
Figure 3. GC chromatogram of the solution from the reaction of n-BuNC adsorbed on powdered gold with n-BuNH2. The peak at 3.72 min is n-BuNC,
whereas that at 9.44 min is n-Bu-NdCdN-n-Bu.
four presents the time (t′1/2) at which half of this strongly
adsorbed isocyanide is removed from the surface; thus, t′1/2 was
obtained by plotting I/I0 versus time considering an I/I0 value
of 0.45 as the starting point and measuring t′1/2 at I/I0 ) 0.225.
There are several general trends that are evident in the data in
Table 1. For the same isocyanide, the rate of the process (both
ability to use a smaller volume of solution, allowed us to
determine the product(s) formed when isocyanides adsorbed on
gold powder were reacted with primary amines. The isocyanide
was first adsorbed on 1.00 g of gold powder by immersing it in
a 5.00 mM isocyanide solution in DCE.7c After decanting the
solution and air-drying, the RNC/Au powder was immersed in
1 mL of a 5 mM amine solution in hexane, and the mixture
was stirred for 5 min in the case of n-BuNH2 and for 10 min
for the other amines. According to the Au film studies, all of
the isocyanide should be removed from the gold surface during
these times (Figure 2). The solution was analyzed by GC-MS
to identify the free isocyanide and/or reaction product. For the
reaction of adsorbed n-BuNC with n-BuNH2, the GC chromato-
gram (Figure 3) shows two new peaks (beside those of amine
and solvent) with retention times (tR) of 3.72 and 9.44 min.
These two peaks were identified as n-BuNC (at 3.72 min) and
N,N′-dibutylcarbodiimide (at 9.44 min) by comparing their mass
spectra with those of known compounds in the computer
libraries described in the Experimental Section. The n-BuNC
was also identified by comparing its GC-MS spectrum to that
of an authentic sample. On the basis of this product identifica-
tion, the reaction of the adsorbed n-BuNC with n-BuNH2 is
described by eq 3
t1/2 and t′1/2) is influenced by the nature of the amine. It is faster
for those amines with a linear carbon chain (n-BuNH2 and
n-DoNH2) and slower for the more bulky amines (CyNH2 and
BzNH2). The slower rates for BzNH2, as compared with CyNH2,
are probably due to its lower basicity/nucleophilicity. For
reactions using the same amine (e.g., n-BuNH2) but different
adsorbed isocyanides, the rates decrease (n-BuNC > CyNC >
TMBNC) as the bulkiness of the isocyanide increases, which
is consistent with a lower accessibility of bulky isocyanides to
attack by the amine.
In these studies, there is no IRRAS evidence in the ν(N-H)
region (3100-3500 cm-1) for adsorbed amine on the gold
surface, even in the absence of isocyanide. Also, there is no
evidence for a diaminocarbene group (see eq 1) on the surface;
such a species should give a ν(N-CdN) absorption near 1550
cm-1, which is characteristic of diaminocarbene groups in
transition metal complexes.19 Likewise, there is no IR band in
the region ∼2130 cm-1 that is characteristic of an adsorbed
carbodiimide (R-NdCdN-R) product. Because of their low
concentrations, it was not possible to detect desorbed R-NtC
or reaction products in the Au film immersion solutions.
n-BuNC + n-BuNH2 + O2 f
Bun-NdCdN-Bun + “H2O” (3)
Reaction/Desorption of Isocyanide on Au Powder in
Amine Solutions. Because the amount of desorbed isocyanide
and/or product was too small to be detected by GC-MS in Au
film experiments, we performed a series of experiments using
gold powder, which has a larger surface area than gold film: 3
× 10-1 m2/g for gold powder7a and 3 × 10-3 m2/slide
(calculated from refs 31 and 32). So, one gram of gold powder
is capable of adsorbing 100 times as much isocyanide as a gold
slide. This larger amount of isocyanide, combined with the
The observation of n-BuNC in the GC chromatogram (Figure
4) shows that some of the isocyanide simply desorbs from the
surface, without reacting with the amine. To determine whether
the desorbed isocyanide reacts with amine in solution, we
reacted a 0.5 mM solution of n-BuNC with 5 mM n-BuNH2 in
hexane, in the absence of gold. GC-MS studies of this solution
showed no evidence for N,N′-dibutylcarbodiimide after 5 days
of stirring at room temperature. This result is consistent with
an earlier report that stated that aliphatic isocyanides do not
9
J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006 10617

A R T I C L E S
Lazar and Angelici
Table 2. Isocyanide Conversion and Product Yield for the
Au-Catalyzed Reaction of n-BuNC with n-BuNH2
react with primary amines in the absence of a catalyst even at
temperatures as high as 120 °C.33 Thus, in the present studies,
the reaction between the isocyanide and the amine must occur
on the gold.
exp
n-BuNC (mmol)
T (°
C)
Ca (%)
η
b (%)
t1 c (h)
/2
1
2
3
4
1.6 × 10-3
20
20
60
60
97
98
100
100
42
47
70
73
72
120
5
17 × 10-3
Similar GC-MS analyses of solutions obtained from reactions
of other adsorbed isocyanides and amines were also performed
for the following reactants: (1) n-BuNC and n-DoNH2; (2)
n-BuNC and BzNH2; (3) CyNC and CyNH2; (4) CyNC and
n-BuNH2. In all cases, the GC-MS spectra showed two new
peaks, further identified to be the desorbed isocyanide and the
corresponding carbodiimide R-NdCdN-R′, where the R and
R′ groups originated from the reacting isocyanide and amine.
The carbodiimide reaction product is surprising because
isocyanide ligands in metal complexes react with amines to form
diaminocarbene complexes (eq 1), and in some cases the carbene
ligand can be displaced as the formamidine (eq 4).26a
1.6 × 10-3
17 × 10-3
40
a Conversion after: 8 days for (1), 19 days for (2), 12 h for (3), and 90
h for (4). b Yield of carbodiimide. c Time at which the conversion is 50%.
N,N′-n-butylurea was identified as a minor product in the GC-
MS spectra of all of the reaction solutions. Unidentified products
with higher molecular weights (>300) were also observed.
The H2O product of this reaction (eq 3) could not be
positively identified in the reaction solutions by GC-MS.
Although it was possible to identify H2O in water-saturated
hexane solutions, the H2O peak in the GC chromatogram of
these solutions was only marginally distinguishable from that
of the background. The extremely low solubility of water in
hexane is well-documented.35 Although H2O could not be
positively identified as a product of the reaction, it seems to be
the most likely product of the reaction of O2. However, it is
possible that the two hydrogen atoms released from the amine
in the reaction (eq 3) combine with O2 and some of the
isocyanide or amine to give other products. The involvement
of O2 in the reaction is essential because Bun-NdCdN-Bun
is not formed in the reaction of n-BuNC and n-BuNH2 in the
absence of O2.
The presence of both the carbodiimide and water in the
product solution suggests the possibility that they could react
to form N,N′-di-n-butylurea. As noted above, small amounts of
N,N′-di-n-butylurea are observed in the reaction solutions.
However, this small amount is not sufficient to account for all
of the H2O that would be produced in the reaction. To determine
whether a carbodiimide and H2O could react under the condi-
tions (Table 2) of reaction 3, amounts of H2O (0.030 mmol,
0.54 µL) and But-NdCdN-But (0.030 mmol) comparable to
those produced in the Au-catalyzed reaction (eq 3) were stirred
in 20 mL of hexane. Even after 48 h at room temperature, there
was no decrease in the amount of But-NdCdN-But, which
indicates that it is unlikely that N,N′-di-n-butylurea would form
under the conditions in Table 2, assuming that Bun-NdCd
N-Bun and But-NdCdN-But have similar reactivities. This
lack of reaction between But-NdCdN-But and H2O may be
due to the very low solubility of water in hexane.
As noted in the Introduction (eq 2), some metal ions catalyze
the reaction of isocyanides with amines to give formamidines.
However, there are no catalytic reactions of isocyanides and
amines to give carbodiimides. Although a reaction (eq 5) of a
Pd diaminocarbene complex with Ag2O gives a carbodiimide,34
the Ag2O reactant makes this a reaction that is very different
than the reactions of isocyanides adsorbed on gold metal in the
current study.
Reactions of n-BuNC with n-BuNH2 and O2 Catalyzed
by Gold Metal. In the previous section we showed that a
reaction takes place on the Au surface between preadsorbed
isocyanide and amine in solution. To determine whether gold
catalyzes the reaction of isocyanides with amines, we performed
a series of experiments in which clean Au powder (1.00 g) was
stirred with a solution of n-BuNC (0.5 mM), n-BuNH2 (5 mM),
and 0.10 mg of decane (internal standard) in hexane at room
temperature (20 °C) in air. Periodically, the solution was
analyzed by GC to determine the quantities of isocyanide and
carbodiimide product. The initial amount of isocyanide was
selected to be similar to that used in previous experiments, which
would give saturation coverage of the gold powder at room
temperature7b (1.6 × 10-3 mmol n-BuNC). For other experi-
ments, the initial concentration of isocyanide was increased by
a factor of approximately 10. In every case, the amine
concentration was 10 times that of the isocyanide. As shown in
Table 2, the reactions were much faster at 60 than at 20 °C,
and the yield of Bun-NdCdN-Bun was higher (70-73%) at
60 °C. In addition to the Bun-NdCdN-Bun carbodiimide, di-
On the basis of the present studies and previous investigations
of reactions of isocyanide ligands in metal complexes with
amines (eq 1), we propose the mechanism in Scheme 2 for the
Au-catalyzed reactions of isocyanides with primary amines to
form carbodiimides. Initially, the isocyanide adsorbs to a Au
atom on the surface, which creates a more positive isocyanide
carbon atom,10 making it susceptible to attack by nucleophiles.
Attack of the amine on this carbon (step 1) leads to intermediate
a, which can transfer both hydrogen atoms to the gold surface
while releasing the carbodiimide product (step 2). Alternatively,
one of the NH2 hydrogen atoms in intermediate a can migrate
to the other nitrogen to give a diaminocarbene intermediate b
(step 3) comparable to diaminocarbene complexes formed in
(33) Saegusa, T.; Ito, Y. Isonitrile Chemistry, Ugi, I., Ed.; Academic Press:
New York, 1971; p 67.
(34) Ito, Y.; Hirao, T.; Saegusa, T. J. Org. Chem. 1975, 40, 2981.
(35) IUPAC-NIST Solubility Database. NIST Standard Reference Database
106: http://srdata.nist.gov/solubility/sol_detail.asp?sys_ID)37_252.
9
10618 J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006

Reaction of Isocyanides with Primary Amines and O2
A R T I C L E S
Scheme 2. Mechanism for the Au-Catalyzed Reaction of
Isocyanides (R-NtC) with Primary Amines (H2NR′) and O2
Table 3. Au-Catalyzed Reaction of n-BuNC with n-BuNH2 in
Hexane at 60 °C
amine concentration (mM)
24.8
5
12.4
37.5
50
kobs (h-1
1/2 (h)
η (%)a
)
0.13
5.25
70
0.34
2.01
78
0.72
0.96
74
1.07
0.64
76
1.39
0.49
76
t
a Yield at end of reaction.
reactions of isocyanide metal complexes with amines (eq 1).
This intermediate b could form the carbodiimide product by
transferring the two N-H hydrogen atoms to the gold surface
(step 4). Both carbodiimide-producing steps (2 and 4) are
proposed to involve transfer of the two hydrogen atoms to
oxygen to give H2O. Because the rate of reaction 3 is the same
in air and in pure O2, the participation of O2 in the overall
reaction must occur after the rate-determining step. This means
that we have no information about the details of steps 2 or 4.
An unlikely pathway for steps 2 and 4 involves the formation
of free H2; bulk gold powder is reported27,36 to catalyze the
reaction of H2 and O2 at a significant rate only at temperatures
above 130 °C. It might be noted that supported gold nanopar-
ticles catalyze the reaction of H2 and O2 to give H2O2 even
below room temperature.37,38
IRRAS studies of the gold film did not reveal a ν(N-CdN)
absorption for intermediate (b) or (a); however, it is possible
that their absorptions are too weak to be observed. If it is
assumed that these intermediates are not present in significant
concentration, it means that the intermediate is rapidly converted
to product and that step 1, amine attack on the adsorbed
isocyanide, is the rate-controlling step in the catalysis. A kinetic
study20,39 of amine attack on the isocyanide ligand in Cl2(Ph3P)-
Pd(CtN-R) (eq 6) shows that the reaction follows a rate law,
Figure 4. Plot of kobs vs [n-BuNH2] for the Au-catalyzed reaction of
n-BuNC with n-BuNH2 at 60 °C in hexane.
presence of 1.00 g of Au powder at 60 °C in air. The large
excess of amine allows one to obtain pseudo first-order rate
constants (kobs) from plots of ln[n-BuNC] vs time, where kobs
) k[n-BuNH2] (Table 3). The plot of kobs vs [n-BuNH2] is linear
(Figure 4) demonstrating that the reaction is first order in the
amine concentration and that the overall rate law in the presence
of 1.00 g Au powder is Rate ) k[n-BuNC][n-BuNH2]. This
rate law, and especially the first order dependence on the amine
concentration, supports the proposal that step 1 in the mechanism
(Scheme 2) is rate-determining. In studies of the removal of
n-BuNC from a gold film upon reaction with amines in hexane
solution, it was observed (Table 1) that the less sterically bulky
and more basic amines removed the isocyanide more rapidly.
This is also consistent with amine attack (step 1) being the rate-
determining step. Because the steps beyond step 1 occur rapidly,
it is not possible to define the subsequent steps that lead to the
carbodiimide product. However, these results indicate that the
initial step in the Au-catalyzed reaction of isocyanide with
amines is very similar to that of isocyanide reactions in transition
metal complexes.
Rate ) k[Cl2(Ph3P)Pd(CNR)] [H2NR′], which is first order
in the amine concentration. It was also noted that the rates of
reaction increase as the basicity of the amine increases from
p-ClC6H4NH2 to p-MeOC6H4NH2. If the rate-determining step
in the Au-catalyzed reaction (eq 3) were amine attack on the
adsorbed isocyanide (step 1 in Scheme 2), the overall rate of
conversion of isocyanide to the carbodiimide product should
also be first order in the amine concentration.
To examine the amine concentration dependence of this
reaction (eq 3), a series of experiments was performed in which
1.6 × 10-3 mmol of n-BuNC in 3 mL of hexane was reacted
with excess amine (10-100 amine/isocyanide ratio) in the
Conclusions
Except for supported nanosized particles,27,28 gold metal is
generally a poor catalyst. However, we find that Au particles
of even 1 µm size catalyze reactions of isocyanides with primary
amines and O2 to form carbodiimides (eq 3) at room temper-
ature. IRRAS and DRIFT studies show that isocyanides adsorb
in an η1 fashion (Scheme 1) onto the surfaces of Au films and
powders and are thereby activated to react with amines. An
indication of this activation is the ∼75 cm-1 increase in the
isocyanide ν(NC) value upon adsorption. Previous studies
suggest that an increase in ν(NC) greater than 40 cm-1 upon
coordination in a metal complex is sufficient to activate an
isocyanide to attack by amines to give a diaminocarbene ligand
(eqs 1 and 6). As for reactions of metal complex-coordinated
isocyanides, the Au-catalyzed reaction of isocyanides is first
(36) Benton, A. F.; Elgin, J. C. J. Am. Chem. Soc. 1927, 49, 2426.
(37) Landon, P.; Collier, P. J.; Papworth, A. J.; Kiely, C. J.; Hutchings, G. J.
Chem. Commun. 2002, 2058.
(38) Edwards, J. K.; Solsona, B.; Landon, P.; Carley, A. F.; Herzing, A.;
Watanabe, M.; Kiely, C. J.; Hutchings, G. J. J. Mater. Chem. 2005, 15,
4595.
(39) Crociani, B.; Boschi T.; Nicolini M.; Belluco U. Inorg. Chem. 1972, 11,
1292.
9
J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006 10619

A R T I C L E S
Lazar and Angelici
order in the amine concentration, which suggests that amine
attack on an adsorbed isocyanide is the rate-determining step
in the overall mechanism (Scheme 2). Although the initial
activation of isocyanide to amine attack is common to both the
Au-catalyzed reaction and reactions of isocyanides in metal
complexes, the Au-catalyzed reaction (eq 3) gives a carbodi-
imide product whereas the homogeneous metal-ion-catalyzed
reaction (eq 2) gives a formamidine. It is apparently the ability
of hydrogen atoms to transfer to the Au surface and/or to the
O2 that leads to the carbodiimide product, a pathway that is not
available to the metal-ion-catalyzed reaction. These investiga-
tions highlight the commonality of key steps in reaction
mechanisms of organometallic complexes and heterogeneous
catalysts that nevertheless lead to different products.
Acknowledgment. This work was supported by the U.S.
Department of Energy, Office of Science, Office of Basic
Energy Sciences, Chemical Sciences Division, under contract
W-7405-Eng-82 with Iowa State University. We thank Professor
Marc D. Porter for use of his IRRAS equipment, Stephen
Veysey for the GC-MS measurements, and Dr. Bolin Zhu for
experimental assistance.
Supporting Information Available: IRRAS spectra of iso-
cyanides on Au-film. GC-MS spectra of carbodiimide products
(R-NdCdN-R′). Rate plots for reactions of n-BuNC with
n-BuNH2 under atmospheres of air, O2, and Ar. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA0618907
9
10620 J. AM. CHEM. SOC. VOL. 128, NO. 32, 2006