Gold nanoparticle-initiated free radical oxidations and halogen abstractions

Article abstract of DOI:10.1039/b711573c

We report on the use of EPR spectroscopy and spin trapping technique to detect free radical intermediates formed in the presence of gold nanoparticles. Phosphine- and amine-protected gold nanoparticles were found to initiate air oxidation of organic subst

Full text of DOI:10.1039/b711573c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Gold nanoparticle-initiated free radical oxidations and halogen abstractions
Petre Ionita, Marco Conte, Bruce C. Gilbert and Victor Chechik*
Received 30th July 2007, Accepted 7th September 2007
First published as an Advance Article on the web 20th September 2007
DOI: 10.1039/b711573c
We report on the use of EPR spectroscopy and spin trapping technique to detect free radical
intermediates formed in the presence of gold nanoparticles. Phosphine- and amine-protected gold
nanoparticles were found to initiate air oxidation of organic substrates containing active hydrogen
atoms, such as amines and phosphine oxides. Nanoparticles protected by stronger bound ligands (e.g.,
thiols) were inactive in these reactions. We also found that gold nanoparticles are able to abstract a
halogen atom from the halogenated compounds, presumably due to the high affinity of gold metal for
halogens. Reaction of Au nanoparticles with chloroform showed an unusual inverse isotope effect. The
trichloromethyl spin adduct was observed when Au nanoparticles were mixed with CDCl3 but not with
CHCl3. This unexpected behaviour suggests that C–H bond breaking is not the rate-determining step in
Au-initiated hydrogen abstraction.
superoxide species. Other compounds, such as hydroperoxides or
Introduction
hydrides were also oxidised under the same conditions.8
Gold nanoparticles are an example of simple-to-make, well-
Similar catalytic properties of Au nanoparticles have been
observed by Meisel et al.9 Exposure of amino-TEMPO to
defined nanostructures, which can be used for a variety of
applications.1 One area where gold particles are attracting much
atmospheric oxygen in the presence of Au nanoparticles led to
interest is catalysis. Since the early reports on the remarkable
the formation of oxo-TEMPO (Fig. 1). The authors suggested a
activity of supported Au nanoparticles in CO oxidation,2 many
mechanism which involved electron transfer from the amino group
other reactions catalysed by Au nanoparticles showed unusual
to oxygen, with the formation of a hydroperoxide radical, leading
finally to the removal of the amino group and its replacement with
of homogeneous Au catalysts in such reactions as oxidations,
a ketone group.
activity and/or selectivity.3 Recent reports on the high activity
reductions, alkene additions, cycloisomerisations etc., further
expand the scope of potential applications of Au nanoparticle
catalysts.4 Au catalysis thus became one of the fastest developing
areas in homogeneous and heterogeneous catalysis alike.5
Despite the advances in Au nanoparticle catalysis, the mech-
anistic features of these reactions remain largely unexplored.
Some metal-catalysed reactions (e.g., oxidations and reductions)
Fig. 1 Oxidation of amino-TEMPO by gold nanoparticles.9
are likely to proceed via formation of free radical intermediates.
Such reactions are best studied by EPR spectroscopy and spin
In order to ascertain if formation of such free radical inter-
mediates is a general feature of the Au nanoparticle-catalysed
trapping methods. For instance, hydrogen spillover on supported
reactions, we used spin trapping and EPR spectroscopy to probe
Pd catalysts was investigated in late 1990s by Mile and Maher
other related systems. Here, we report the results of this study.
The spin adducts were obtained using DMPO and PBN spin traps
(Fig. 2).
et al.6 Although this work clearly demonstrated the potential
of EPR spectroscopy in uncovering the subtle details of the
mechanisms of the nanoparticle-catalysed reactions, there have
been virtually no further reports in this area, as highlighted in a
recent review.7
During our study of the ligand exchange reactions on the
surface of Au nanoparticles, we discovered that some of these
processes occur via a free radical mechanism.8 We detected
formation of sulfur-centred radicals in an exchange reaction
Fig. 2 Chemical structures of the PBN and DMPO spin-traps.
between triphenylphosphine-protected gold nanoparticles and an
alkanethiol. The formation of S-centred radicals was probably
due to the oxidation of thiol by molecular oxygen in the presence
of the gold nanoparticles.8,9 The proposed mechanism involved
Results and discussion
the adsorption of the oxygen on the nanoparticle surface as a
Oxidation of amines and phosphine oxides by gold nanoparticles
The properties of gold nanoparticles depend on their size and
on the nature or protecting ligands; the latter also determine
their solubility. We have screened a series of different types
Department of Chemistry, University of York, Heslington, York, UK YO10
5DD. E-mail: vc4@york.ac.uk; Fax: (+44)-1904-432516
3504 | Org. Biomol. Chem., 2007, 5, 3504–3509
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
of nanoparticles for the ability to generate free radicals in the
presence of air and organic substrates such as thiols or amines. The
protecting ligands tried included triphenylphosphine (1.4 nm),10
butanethiol (2.5 nm),11 tetraoctylammonium bromide (4.7 nm),12
citrate/chloride ions (13 nm),13 dodecylamine (3.0 nm).14
more amine after the formation of the spin-adduct also did not
increase the intensity of the EPR spectra of the spin adduct.
In the reaction of diphenylphosphine oxide with phosphine-
or amine-protected nanoparticles in the presence of DMPO or
PBN, a signal corresponding to the P-spin adduct was observed.
The radical was assigned as Ph2PO·.16 The spectrum assignment
was confirmed by comparison with a spectrum of an authentic
Ph2PO· radical obtained by oxidation of diphenylphosphine oxide
with either lead dioxide or hydrazyl radicals. Deoxygenation of
the reaction mixture prior to the addition of the spin traps led
to significant reduction of the EPR intensity, further confirming
that the oxidation requires the presence of molecular oxygen
(Scheme 1).
The substrates used in this work were n-butylamine [as we were
interested to see if Au-catalysed oxidation of amino-TEMPO9 (vide
supra) can be extended to other systems], and diphenylphosphine
oxide. Both substrates contain reactive hydrogen attached to a
heteroatom which could be abstracted by a peroxide-type species.
We found that only Au nanoparticles protected by triph-
enylphosphine and (to a lesser extent) alkylamines were able to
initiate oxidation reactions. This was also confirmed using other
organic substrates which were reported in the previous study (e.g.,
n-butanethiol, borohydride, t-butyl hydroperoxide).8 The lack of
activity of the thiol-protected particles can be partially explained
by poisoning of the reactive surface with thiolate ligands. The
inactivity of the relatively large citrate-based nanoparticles is
probably due to the small surface area and low nanoparticle
concentration. Low solubility of oxygen in water can also explain
the failure to detect any free radical intermediates in reactions in
water (which were carried out in the presence of citrate-stabilised
particles).
Triphenylphosphine-protected nanoparticles, in the reaction
with butylamine in the presence of air and DMPO as a spin-
trap, led to the corresponding N-centred spin-adduct of DMPO,
well evidenced in the EPR spectrum (Fig. 3). The N-centred
radical, assigned as C4H9NH·, was also observed with the PBN
spin trap. Reaction of dodecylamine-protected Au nanoparticles
with butylamine in the presence of air and a spin trap (DMPO or
PBN) also showed the presence of N-centred radicals; however,
the intensity of the EPR peaks of the spin adducts in this case was
lower. The assignment of EPR spectra was achieved by comparing
the spectra with the genuine spin-adduct formed by oxidation of
the corresponding amine with lead dioxide or hydrazyl radicals.8,15
Scheme 1 Au nanoparticle-initiated oxidation of organic compounds.
Interestingly, the half-life time of the Ph2PO· spin adduct
prepared by the nanoparticle-initiated reaction was much shorter
(<1 min) than that of the same spin adduct prepared by oxidation
with lead dioxide (up to 1 h). This difference in the lifetime can only
be explained by the presence of gold nanoparticles. To confirm this,
we have added triphenylphosphine-protected Au nanoparticles to
the solution of a Ph2PO·–DMPO spin-adduct which was prepared
by oxidation of the corresponding substrate with lead dioxide. A
rapid disappearance of the EPR signal was indeed observed. In the
case of a PBN spin-adduct prepared in the same way, the influence
of the addition of the gold nanoparticles was almost unnoticeable.
Rapid disappearance of the DMPO spin-adduct in the presence
of Au nanoparticles is likely to be due to a further reaction on
the Au surface, or possibly a reaction with an Au(I) or Au(III)
species. Au(III) in particular is a strong oxidising agent and
is known to oxidise the spin-traps DMPO and PBN to form
persistent free radicals.17 Au(I) compounds are often considered
possible intermediates in the ligand exchange reaction on the
Au nanoparticle surface; hence we cannot exclude the presence
of small amounts of Au(I) either in solution or on the Au
surface. We must stress, however, that we observed no free radical
intermediates in control experiments when Au nanoparticles were
replaced by Au(I) or Au(III) compounds.
Halogen abstraction
We noticed that in some halogenated solvents, triphenyl-
phosphine-protected gold nanoparticles were particularly unsta-
ble, aggregating and precipitating within one or two days. In
order to check if this behaviour is accompanied by formation of
any radical intermediates, we added spin traps (e.g., DMPO and
PBN) to toluene containing halogenated solvents, such as chlo-
roform, dichloromethane, carbon tetrachloride, bromobutane,
iodomethane, and chlorobenzene. In all cases except chloroben-
zene and dichloromethane, very strong EPR signals of different
spin-adducts were observed. PBN and DMPO spin traps captured
the same radicals in all reaction mixtures. Dodecylamine-protected
Au nanoparticles also gave the same results. Unlike nanoparticle-
initiated oxidation reactions (vide supra), the halogen abstraction
reaction does not depend on the presence of oxygen in the reaction
mixture, as shown by carrying out spin trapping under a nitrogen
Fig. 3 EPR spectrum of the C4H9NH–DMPO spin adduct in toluene.
In order to optimise the spin trapping process and observe the
highest concentration of the spin adduct, different parameters
were varied, such as absolute and relative concentration of
reactants, and reaction time. It was found that in all cases the
formation of the spin adduct was very fast (<1 min), and increasing
the reaction time did not lead to any increase in the amount of
spin adducts (as judged by the EPR intensity). The addition of
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3504–3509 | 3505
©

View Article Online
atmosphere. Fig. 4 shows the EPR spectra of the DMPO spin
adducts with ·CHCl2 and ·CCl3.
the aH values showed large variation even for the PBN adducts
which enabled unambiguous assignment. The DMPO–CDCl2
spin adduct has not been reported in the literature, and we
have not been able to generate the ·CDCl2 radical by alternative
reactions; nonetheless, the hyperfine parameters strongly support
the assignment, and the PBN spin adduct of this radical gave a
perfect match with the literature parameters. Besides, formation
of the ·CDCl2 radical by halogen abstraction parallels similar
reactions for CCl4, BrCCl3, CH3I, C4H9Br.
The mechanism of halogen abstraction by the Au nanoparticles
could be explained by electron transfer from the nanoparticle to
the alkyl halide to create a radical anion which would then release
halide anion and form a carbon-centred radical. The Au nanopar-
ticle in this mechanism would become positively charged. Step-
wise charging of Au nanoparticles by electrochemical reduction–
oxidation is well known and thus support the feasibility of this
mechanism.22 Alternatively, the radicals can be formed by a direct
attack of Au nanoparticles on the alkyl halide leading to the
transfer of halogen atom (Scheme 2). This mechanism (driven
by the very high affinity of halogens for gold)23 would lead to
the formation of AuCl on the nanoparticle surface and possibly
displacement of the original ligand (e.g., triphenylphosphine).
Halogen adsorption on the nanoparticle surface then results in
particle aggregation (Scheme 2). This behaviour would explain the
poor stability of triphenylphosphine-stabilised Au nanoparticles
in some chlorinated solvents (e.g., CCl4). Although unambiguous
mechanistic conclusion is not possible, direct halogen abstraction
seems more likely.
Fig. 4 EPR spectrum of the DMPO–CHCl2 spin adduct prepared by
abstraction of a Cl atom from chloroform (a), DMPO–CCl3 spin adduct
prepared by abstraction of a Br atom from BrCCl3(b), and PBN–CCl3 spin
adduct formed by the same reaction (c).
The EPR parameters of the spin adducts formed are shown
in Table 1. The assignment (with the exception of DMPO–
CDCl2) was carried out by comparison with the literature values.
Although the aN values for all C-centred radicals are very similar,
Scheme 2 Abstraction of halogen (Hal) atoms by Au nanoparticles.
Although spin-trapping data unambiguously show that rad-
ical intermediates are formed in the Au nanoparticle-initiated
reactions, it is conceivable that these radicals are adsorbed on
the nanoparticle surface and do not exist as free species in
solution. For instance, spin adducts have been observed for some
transition metal-catalysed reactions, even though the reaction
stereochemistry ruled out formation of free radical intermediates.
This was tentatively explained by the group transfer between the
spin trap and the transition metal complex, which results in the
formation of the spin adduct.24 A similar process can be envisaged
for Au nanoparticle-initiated reaction (Scheme 3).
Table 1 Hyperfine splitting constants for the DMPO and PBN spin-
adducts obtained using triphenylphosphine-protected Au nanoparticles.
Values in brackets are literature data
Substrate
Radical
Spin trap aN/G
aH/G
aX/G
C4H9NH2 C4H9NH· DMPO
13.89
14.11
13.86
14.20
13.10
16.49
2.10
18.18
3.09
1.79 (aN)
PBN
—
Ph2PHO
Ph2PO·
·CCl3
DMPO
PBN
37.71 (ap)
18.73 (ap)
—
CCl4
DMPO
15.16
(13.17)18 (15.28)18
—
PBN
13.91
1.54
—
—
(13.60)19 (1.86)19
CHCl3
CDCl3
CH3I
·CHCl2
DMPO
PBN
13.66
14.41
(14.32)19
13.30
13.65
14.39
14.27
19.76
2.12
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
(3.03)19
15.62
19.77
2.08
·CCl3
DMPO
Scheme 3 Possible formation of spin adducts via formation of nanopar-
ticle-adsorbed radicals.
·CDCl2
·CDCl2
·CH3
PBN
DMPO
20.61
(14.20)20 (20.80)20
In order to test this possibility, we explored competitive trapping
of the C4H9S·, C4H9NH· and ·PPh2O radicals by a 1 : 1 mixture
of DMPO and PBN spin traps (Fig. 5). The same radicals were
generated using either of the two oxidants: diphenyl picryl hydrazyl
(DPPH), or Au nanoparticles in air. The radicals were then trapped
bythemixtureofspintraps, andtheratiosofthePBN–DMPOspin
adducts were calculated by simulating the experimental spectra as
PBN
14.80
(14.94)21
13.14
13.95
14.17
3.56
(3.63)21
15.17
1.52
CBrCl3
C4H9Br
·CCl3
·C4H9
DMPO
PBN
DMPO
PBN
22.34
3.62
14.83
(14.57)21
(3.22)21
3506 | Org. Biomol. Chem., 2007, 5, 3504–3509
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
different (but readily observed in the EPR spectra) DMPO–
PBN selectivity for this radical generated by reaction with Au
particles or DPPH (Table 2) therefore supports the interactions
of the free nitrogen-centred radicals with the nanoparticle surface
(Scheme 3).
Interestingly, reactions of Au nanoparticles with chloroform
and deuteriated chloroform gave unexpected results. DMPO spin
trapping in a mixture of Au nanoparticles with chloroform (either
neat or in toluene solution) gave the ·CHCl2 adduct as the only
EPR-active product. However, the same reaction with CDCl3
showed simultaneous formation of ·CCl3 and ·CDCl2 spin adducts,
with a ratio close to 1, as estimated by the simulation of the
experimental spectra (Fig. 6).
Fig. 5 EPR spectrum obtained using an equimolecular mixture of DMPO
and PBN shows a mixture of DMPO–CCl3 (*) and PBN–CCl3 (#) spin
adducts.
a mixture of two components. The logic behind this experiment
was that DPPH-generated radicals definitely exist as free species in
solution. If nanoparticle-generated radicals are also free species,
one would expect to see exactly the same ratio of DMPO–PBN
spin adducts for DPPH and nanoparticle-generated radicals. If the
nanoparticle-generated radicals are adsorbed on the nanoparticle
surface, the ratio could be different.
The results of this competitive trapping are shown in Table 2.
It is worth noting that the selectivity of trapping depends strongly
on the nature of the radical. However, DMPO is a much better
substrate for the majority of radicals, and the DMPO spin adducts
were the predominant species in most cases.
For the N-, P- and S-centred radicals generated using Au
nanoparticles, we did not observe any PBN spin adduct. On the
other hand, PBN adducts were observed if the same radicals were
generated using DPPH (Table 2). Unfortunately, the stability of
the ·PPh2O spin adducts is very limited in the presence of Au
nanoparticles, and the kinetics of the spin adduct decay are quite
complex and poorly reproducible. Therefore, even the observed
large differences between the DMPO–PBN spin adduct ratios for
DPPH and Au nanoparticle-generated radicals could in principle
be explained by the instability of the spin adducts. On the other
hand, the DMPO and PBN adducts of the nitrogen-centred radical
were quite stable in the presence of Au nanoparticles. The slightly
Fig. 6 EPR spectrum obtained from reaction of triphenylphosphine-pro-
tected Au nanoparticles and CDCl3 in air. The observed spectra were
identified as a mixture of DMPO–CCl3 (*) and DMPO–CDCl2 (#) spin
adducts.
This result is difficult to explain. Formation of the ·CCl3
radical (which is unambiguously assigned by comparison with the
literature values and authentic species generated from CCl4 and
CBrCl3) can only be explained by the abstraction of the H atom
from CHCl3. This reaction (if it involves breaking of the C–H bond
in the rate-determining step) must exhibit a primary isotope effect,
e.g., it should be much faster for CHCl3 than for CDCl3. However,
we only observed ·CCl3 radical in the reaction of CDCl3, not with
CHCl3. If the formation of ·CCl3 radical does not involve breaking
of the C–H/C–D bond in the rate-determining step, it should
exhibit secondary isotope effects (which can be either normal or
inverse); however, secondary isotope effects are usually small (e.g.,
0.5 < kH/kD <2) and hence cannot explain complete absence of
the ·CCl3 radical in the CHCl3–Au nanoparticles reaction mixture.
We have carefully considered the possibility of artefacts. The
absence of the CCl3–DMPO adduct could be explained by a
contamination in CHCl3 (e.g., stabiliser). This was ruled out
by thoroughly purifying chloroform according to published
protocols.25 Additionally, a 1 : 1 mixture of CHCl3 and CDCl3
showed the presence of ·CCl3 radical in the concentration,
consistent with the amount of CDCl3 used, thus further confirming
the absence of reactive impurities in CHCl3.
Table 2 Relative percentage of DMPO spin adduct in competitive
binding of radicals with a mixture of DMPO–PBN spin traps. The data
in brackets refer to DPPH-generated radicals; all other data refer to the
radicals generated using triphenylphosphine-protected gold nanoparticles
Substrate
Percentage of DMPO spin adduct (%)
C4H9NH2
Ph2PHO
C4H9SH
CCl4
CHCl3
CDCl3
CH3I
100 (98.8)
100 (39.4)
100 (100)
98.5
87.7
39.1a/52.8b
81.9
Another artefact could be caused by the formation of ·CCl3
radical from an impurity in CDCl3 which cannot be efficiently
removed by purification (e.g., BrCCl3 or CCl4). For instance, we
found that the efficiency of the formation of ·CCl3 adduct from
CBrCl3
95.4
a For ·CCl3 spin-adduct. b For ·CDCl2 spin adduct.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3504–3509 | 3507
©

View Article Online
BrCCl3 is extremely high. In order to check if the formation of
radicals can be linked to the presence of impurities, we performed
the following control experiment. Two samples of the same batch
of CHCl3 were treated with H2O–NaOH and D2O–NaOH under
the same reaction conditions to obtain the samples of CHCl3
and CDCl3 with the same “handling history”.26 The formation
of CDCl3 was monitored by 13C NMR. However, reaction of Au
nanoparticles with these two solvents (in the presence of DMPO)
confirmed that CHCl3 gives only ·CHCl2 radical while CDCl3 gives
a mixture of ·CCl3 and ·CDCl2. This apparent inverse isotope effect
is hence a genuine phenomenon which cannot be explained by the
presence of impurities. Our results thus strongly suggest that C–
H bond breaking is not the rate-determining step in the overall
hydrogen atom abstraction mechanism.
Triphenylphosphine-protected Au nanoparticles10 were pre-
pared as follows. A 1% aqueous hydrogen tetrachloroaurate
trihydrate solution (10 mL) was added to a toluene solution
(10 mL) of tetraoctylammonium bromide (160 mg) and stirred
for 5 minutes. When the gold layer had transferred to the organic
phase, triphenylphosphine (230 mg) was added under stirring.
After 2 min, a freshly prepared aqueous solution (5 mL) of NaBH4
(140 mg) was rapidly added. The organic phase immediately turned
dark brown. The reaction mixture was stirred for 3 h. The toluene
layer was separated and the solvent was removed under vacuum
at 40 ◦C to give a dark brown solid. The solid was redissolved
in the minimum amount of dichloromethane and purified by gel
permeation chromatography using Bio-Beads SX-1 as a stationary
phase and dichloromethane as an eluent.
The spin trapping was performed as follows: the appropriate
spin trap (0.1 mL of 0.1 M solution in toluene) was added to the
substrate (0.1 mL of 0.1 M solution in toluene for non-halogenated
substrates, or 0.1 mL of neat halogenated solvent). The resultant
mixture was then added to a solution of gold nanoparticles in
toluene (0.2 mL 2 × 10−4 M). The mixture was transferred into a
glass tube and deoxygenated by bubbling nitrogen for ca. 1 min
prior to recording the EPR spectra. All experiments were carried
out at room temperature. The competitive spin trapping was
carried out in the same way using a 1 : 1 equimolar ratio of
DMPO–PBN. The spin trapping with DPPH and lead dioxide
was also used in the same way using a 0.1 mL solution of DPPH
in toluene (1 mM.) or ca. 50 mg of solid PbO2.
Conclusions
Triphenylphosphine-protected Au nanoparticles can activate
molecular oxygen, and the reactive species formed are capable
of abstracting a hydrogen atom from many organic molecules
including alkylamines and diarylphosphine oxides. The structure
of the free radicals thus formed was confirmed using EPR
spectroscopy and spin-trapping technique. Alkylamine-protected
Au nanoparticles are less active in the same reactions.
Reaction of phosphine or amine-protected Au nanoparticles
with compounds possessing an active halogen atom (e.g., alkyl
bromides/iodides, chloroform, tetrachloromethane) led to the
abstraction of the halogen atom by the Au nanoparticles. The
radicals thus formed were also identified using spin trapping. The
results of the competitive trapping using a mixture of two spin
traps suggest interactions between the alkylamine-derived radical
intermediate with the Au surface.
CDCl3 was prepared by the following procedure.26 CHCl3
(5 mL) was mixed with a 1 M NaOH solution in D2O (15 mL)
at 35 ◦C under a N2 atmosphere and the mixture was stirred
vigorously for 1.5 h. In order to increase the yield of the
deuteration, the product was again treated with NaOH in D2O for
1.5 h at 35 ◦C to obtain 90% isotope purity. CHCl3 was prepared
by an identical procedure using H2O rather than D2O.
Interactions of chloroform with Au nanoparticles showed
unexpected inverse isotope effect. The ·CCl3 radical (presumably
formed by abstraction of a hydrogen atom from chloroform) was
observed in CDCl3 but not in CHCl3. This suggests that breaking
of the C–H bond is not the rate-determining step of the hydrogen
abstraction.
Acknowledgements
The authors would like to thank the EPSRC for funding (grants
GR/S45300/01 and EP/E001629/1).
Notes and references
Experimental section
1 M. C. Daniel and D. Astruc, Chem. Rev., 2004, 104, 293.
2 M. Haruta and M. Date, Appl. Catal., A, 2001, 222, 427.
3 X. Deng, B. K. Min, A. Guloy and C. M. Friend, J. Am. Chem. Soc.,
2005, 127, 9267; M. D. Hughes, Y.-J. Xu, P. Jenkins, P. McMorn, P.
Landon, D. I. Enache, A. F. Carley, G. A. Attard, G. J. Hutchings, F.
King, E. H. Stitt, P. Johnston, K. Griffin and C. J. Kiely, Nature, 2005,
437, 1132; J. Chou and E. W. McFarland, Chem. Commun., 2004, 1648.
4 D. Astruc, F. Lu and J. Aranzaez, Angew. Chem., Int. Ed., 2005, 44,
78; X. Deng and C. M. Friend, J. Am. Chem. Soc., 2005, 127, 17178;
A. Gupta, C. Rhim, C. Oh, R. Mane and S. Han, Green Chem., 2006,
8, 25; X. Zhang, H. Shi and B. Xu, Angew. Chem., Int. Ed., 2005, 44,
7132; A. S. K. Hashmi, Gold Bull., 2003, 36, 3; X. Yao and C.-J. Li,
J. Am. Chem. Soc., 2004, 126, 6884.
EPR spectra were recorded at room temperature in deoxygenated
toluene, using a Jeol JES-RE1X spectrometer. The typical settings
for the EPR spectra were: frequency 9.42 GHz, power 1 mW, sweep
width 100 G, centre field 3190 G, sweep time 60 s, time constant
30 ms, modulation frequency 100 kHz, modulation width 1 G,
gain 200. The spectra simulation was carried out using WinSim
software.27
DMPO, PBN and other chemicals were purchased from Aldrich
and used without further purification, except chloroform which
was purified as described in ref. 24. The butanethiol-,11 tetraocty-
lammonium bromide,12 citrate-,13 and dodecylamine-protected Au
nanoparticles14 were prepared following literature recipes. The
butanethiol-protected nanoparticles were additionally purified by
gel permeation chromatography using BioBeads SX-1 gel and
dichloromethane as an eluent. Other nanoparticles were purified
as described in the literature recipes.
5 A. S. K. Hashmi, Angew. Chem., Int. Ed., 2005, 44, 6990; M. Haruta,
Gold Bull., 2004, 37, 27.
6 A. Burt, M. Emery, J. Maher and B. Mile, Magn. Reson. Chem., 2001,
39, 85; A. F. Carley, H. A. Edwards, B. Mile, M. W. Roberts, C. C.
Rowlands, F. E. Hancock and S. D. Jackson, J. Chem. Soc., Faraday
Trans., 1994, 90, 3341; A. F. Carley, H. A. Edwards, B. Mile, M. W.
Roberts, C. C. Rowlands, S. D. Jackson and F. E. Hancock, J. Chem.
Soc., 1994, 1407.
3508 | Org. Biomol. Chem., 2007, 5, 3504–3509
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
7 B. Mile, Curr. Org. Chem., 2000, 4, 55.
8 P. Ionita, B. C. Gilbert and V. Chechik, Angew. Chem., Int. Ed., 2005,
44, 3720.
9 Z. Y. Zhang, A. Berg, H. Levanon, R. W. Fessenden and D. Meisel,
J. Am. Chem. Soc., 2003, 125, 7959.
produced the same results as PPh2H. 31P NMR showed that solutions
of PPh2H rapidly oxidise toform PPh2HO under ambient conditions.
We thank the reviewer for correcting our initial error.
17 A. Nakajima, Y. Ueda, N. Endoh, K. Tajima and K. Makino,
Can. J. Chem., 2005, 83, 1178.
10 W. W. Weare, S. M. Reed, M. G. Warner and J. E. Hutchinson, J. Am.
Chem. Soc., 2000, 122, 12890.
11 M. Brust, M. Walker, D. Bethell, D. J. Schiffrin and R. J. Whyman,
J. Chem. Soc., Chem. Commun., 1994, 801.
12 J. Fink, C. J. Kiely, D. Bethell and D. J. Schiffrin, Chem. Mater., 1998,
10, 922.
13 J. Turkevich, P. C. Stevenson and J. Hiller, Discuss. Faraday Soc., 1951,
11, 55.
18 D. L. Haire, U. M. Oehler, P. H. Krygsman and E. G. Janzen, J. Org.
Chem., 1988, 53, 4535.
19 M. J. Davies and T. F. Slater, Chem.-Biol. Interact., 1986, 58, 137.
20 P. J. Barker, S. R. Stobart and P. R. West, J. Chem. Soc., Perkin Trans.
2, 1986, 127.
21 E. G. Janzen, U. M. Oehler, D. L. Haire and Y. Kotake, J. Am. Chem.
Soc., 1986, 108, 6858.
22 R. Guo, D. Georganopoulou, S. W. Feldberg, R. Donkers and R. W.
Murray, Anal. Chem., 2005, 77, 2662.
14 D. V. Leff, L. Brandt and J. R. Heath, Langmuir, 1996, 12, 4723.
15 P. Ionita, B. C. Gilbert and A. C. Whitwood, Lett. Org. Chem., 2004,
1, 70.
23 H.-S. Oh, J. H. Yang, C. K. Costello, Y. M. Wang, S. R. Bare, H. H.
Kung and M. C. Kung, J. Catal., 2002, 210, 375.
16 We originally used diphenyl phosphine PPh2H as a substrate and
assigned the trapped radical as ·PPh2. The reviewer suggested that
due to fast oxidation of secondary phosphines, the trapped radical may
be ·PPh2O. Control experiments confirmed this suggestion, as PbO2
oxidation of both PPh2H and PPh2HO gave the same spin adduct,
and use of PPh2O as a substrate in experiments with Au nanoparticles
24 J. F. Peyronel and H. B. Kagan, New J. Chem., 1978, 2, 211; J. K. Stille
and K. S. Y. Lau, Acc. Chem. Res., 1977, 10, 434.
25 W. L. Armarego, D. D. Perrin, Purification of laboratory chemicals,
Butterworth-Heinemann, Oxford, 4th edn, 1997.
26 E. A. Symons and J. D. Bonnett, J. Catal., 1985, 93, 209.
27 http://epr.niehs.nih.gov/pest.html.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3504–3509 | 3509
©