Dynamic Article LinksC<Nanoscale

Cite this: Nanoscale, 2012, 4, 5333

www.rsc.org/nanoscale PAPER

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online / Journal Homepage / Table of Contents for this issue

Electron transfer catalysis with monolayer protected Au25 clusters†

Sabrina Antonello,* Mahdi Hesari, Federico Polo and Flavio Maran*

Received 1st May 2012, Accepted 4th June 2012

DOI: 10.1039/c2nr31066j

Au25L18 (L ¼ S(CH2)2Ph) clusters were prepared and characterized. The resulting monodisperse

clusters were reacted with bis(pentafluorobenzoyl) peroxide in dichloromethane to form Au25L18+

quantitatively. The kinetics and thermodynamics of the corresponding electron transfer (ET) reactions

were characterized via electrochemistry and thermochemical calculations. Au25L18+ was used in

homogeneous redox catalysis experiments with a series of sym-substituted benzoyl peroxides, including

the above peroxide, bis(para-cyanobenzoyl) peroxide, dibenzoyl peroxide, and bis(para-

methoxybenzoyl) peroxide. Peroxide dissociative ET was catalyzed using both the Au25L18/Au25L18�

and the Au25L18+/Au25L18 redox couples as redox mediators. Simulation of the CV curves led to

determination of the ET rate constant (kET) values for concerted dissociative ET to the peroxides. The

ET free energy DG� could be estimated for all donor–acceptor combinations, leading to observation of

a nice activation–driving force (log kET vs. DG�) relationship. Comparison with the kET obtained using

a ferrocene-type donor with a formal potential similar to that of Au25L18/Au25L18� showed that the

presence of the capping monolayer affects the ET rate rather significantly, which is attributed to the

intrinsic nonadiabaticity of peroxide acceptors.

Introduction

When gold nanoparticles are sufficiently small, electronic-band

energetics leads to quantum confinement effects, which are conve-

niently studied by electrochemistry.1,2 Quantized double-layer

charging behavior of phenylethanethiolate or hexanethiolate coated

Au140 (core diameter, 1.6 nm)3–5 and Au225 clusters (2.0 nm)6 has

been evidenced by sensitive transient electrochemical techniques

such as differential pulse voltammetry (DPV), square wave vol-

tammetry, and cyclic voltammetry (CV). For these monolayer-

protected clusters (MPCs), stepwise charging of the metal core is

detected as an approximately regular sequence of redox peaks.

When the core size decreases, the separation between the formal

potential (E�) values of such redox couples increases. When the core

radius is made progressively smaller, a transition between quantized-

double layer charging behavior and molecule-like regime eventually

takes place,2 as detected through the observation of a sizeable

separation between the first oxidation and first reduction steps. This

E� separation, which is related to the energy gap between the highestoccupied and lowest unoccupied molecular orbitals (HOMO–

LUMO),7–11 becomes evident2 for clusters smaller than 1.5 nm, such

as Au75,12 Au38,

4,13 Au25,8,9,11,14 Au13,

15 Au11,16 and Au9.

17

Department of Chemistry, University of Padova, via Marzolo 1, 35131Padova, Italy. E-mail: flavio.maran@unipd.it; sabrina.antonello@unipd.it; Tel: +39 049 827 5147

† This article was submitted as part of a Themed Issue on metallicclusters. Other papers on this topic can be found in issue 14 of vol. 4(2012). This issue can be found from the Nanoscale homepage[http://www.rsc.org/nanoscale].

This journal is ª The Royal Society of Chemistry 2012

Within the few small gold nanoclusters so far characterized

and displaying distinct molecule-like redox and optical behav-

iors, Au25L18 (1 nm) occupies a special position.18 In its native

diamagnetic form, such as when synthesized in the presence of

tetra-n-octylammonium,19,20 Au25L18 bears a negative charge.21

Au25L18� tends to undergo aerobic oxidation to form the neutral

state Au25L180.22 The structures of both anionic and neutral

forms were unraveled in 2008.22–24 Although the structure of

[n-Oct4N+][Au25L18

�] shows some distortion, both redox states

share the same general features, i.e., the presence of a core

composed of a Au13 icosahedron capped by six Au2L3 staple-like

elements. The presence of these Au atoms, stellated on 12 faces of

the Au13 core, causes the ligands to be of two types: whereas in 6

ligands each sulfur atom is connected to two stellated Au atoms,

in the remaining 12 ligands the sulfur atom is connected to one

stellated Au atom and one core Au atom. Provided the sample is

truly monodisperse, 1H and 13C NMR spectroscopy quantita-

tively shows that the corresponding resonances of the two ligand

families are not equivalent.19

Whereas basic electrochemical information (such as E� values

and E� differences) has been reported for some molecule-like

clusters and under different environment conditions (solvents,

electrolytes, and temperature), much less is known about their

actual electron-transfer (ET) properties, such as intrinsic barriers,

activation parameters, and ET rate constants. The ET properties of

Au25L18 clusters (L ¼ phenylethanethiolate) have been studied in

more detail, albeit in former papers they were erroneously assigned

to Au38L240 instead of Au25L18

�.2ET characterization of the native

cluster, Au25L18�, showed that Au25 behaves in all regards as a

Nanoscale, 2012, 4, 5333–5342 | 5333

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

molecular redox species11 whose redox potentials can be varied by

proper substitution of the capping ligands.5,10 Marcus analysis of

the CV curves indicated that the intrinsic barrier for the oxidation

of Au25L18� (the intrinsic barrier,DG‡

0, is the value of the activation

free energy DG‡ of the self-exchange ET reaction, i.e., when the ET

free energy DG� is zero) is predominantly in the form of inner

reorganization,11 in agreement with the outcome of ET self-

exchange measurements (0 and�1 forms) in the solid state25 and in

solution.26 This evidence of a structural change on passing from

Au25L18� to Au25L18

0 is in keeping with the difference between the

corresponding structures22–24 and supported by Raman results

pointing to a Au–S bond length variation upon oxidation.26

In this paper we will describe the ET properties of stable

charge states of Au25L18, which were used as redox catalysis

mediators toward peroxide acceptors. Peroxides were chosen as

acceptors because they are usually reduced according to a

concerted dissociative ET mechanism. Concerted dissociative

ETs occur when direct formation of the fragmentation products

(eqn (1)) is favored over a reaction path involving formation of a

discrete intermediate followed by sigma bond cleavage (stepwise

mechanism, eqn (2) and (3)).27–30 Eqn (1)–(3) show the general

case of a neutral acceptor molecule, AB, forming a radical, A_,

and an anionic fragment, B�:

AB + e / A_+ B� (1)

AB + e % AB_� (2)

AB_� / A_+ B� (3)

In both cases, the electron can be provided either homoge-

neously (using soluble ET donors) or heterogeneously (on elec-

trodes), and the corresponding intrinsic barriers are strictly

correlated.31 The concerted mechanism is particularly favored by

a small bond energy of the breaking bond and a positive

oxidation potential of the leaving group, conditions that are met

by most peroxides.28,32,34 A transition between the concerted and

the stepwise mechanisms can be detected by progressively

varying the applied driving force, as observed with a series of

related perbenzoates.33 Borderline dissociative ET mechanisms

may also occur, depending on the nature of the cleaving bond

and interactions between caged fragments.31

Dissociative ETs to peroxide acceptors are generally nonadia-

batic and thus proceed with rates that are orders of magnitude

smaller than the adiabatic limit.33–39This behavior is most probably

associated with the failure of the Born–Oppenheimer approxima-

tion near the transition state, meaning that the reactant electronic

wavefunction does not instantaneously adjust along the reaction

coordinate near the transition state and, therefore, that the crossing

between the reactant and product curves is only narrowly avoided,

causing the reaction rate to drop significantly.37 We previously

analyzed this issue in detail and reached the conclusion that

dissociative ET reactions, such as to peroxides, are particularly

susceptible to proceed nonadiabatically.40 In this context, the

nonadiabatic dissociative ET theory by German and Kuznetsov41

has been of fundamental importance for obtaining further quan-

titative insights into these processes.34,37,38,40

The present study is based on the extensive knowledge

acquired on both the reduction of peroxides28,29,31–39 and our

5334 | Nanoscale, 2012, 4, 5333–5342

previous use of the homogeneous redox catalysis approach42 with

an unstable Au25 species.11 In particular, we previously described

results concerning homogeneous ETs in N,N-dimethylforma-

mide (DMF) in which we used an electrogenerated Au25species,11 Au25L18

2�, to reduce the dissociative-type acceptors

diphenyl disulfide and benzyl bromide. The ET rate constant

(kET) values nicely compared with those obtained using

conventional one-electron donors.43,44 However, owing to the

short lifetime of Au25L182� (�4 ms: on a longer time scale the

clusters collapse by releasing thiolate species), these studies could

be carried out only with acceptors undergoing relatively fast

reductions. Here we focus on stable redox states of Au25L18, i.e.,

Au25[S(CH2)2Ph]18z where z ¼ �1, 0, +1, and the use of sym-

substituted dibenzoyl peroxides. The latter undergo a particu-

larly facile reduction that nicely matches the E� values of the

Au25L18/Au25L18� and Au25L18

+/Au25L18 redox couples in

dichloromethane (DCM) containing 0.1 M Bu4NPF6 (TBAH),

i.e., �0.077 and 0.226 V, respectively.11

Precise control of the charge state of the clusters is crucial, and

details are provided for the ET reaction employed to form

Au25L18+ quantitatively. The redox catalysis experiments were

carried out with both Au25L18/Au25L18� and Au25L18

+/Au25L18

redox couples. Digital simulation of the CV curves led to deter-

mination of the corresponding kET values. A very good activation–

driving force relationship (log kET vs.DG�) could be obtained using

estimates of the corresponding ET free energyDG� values. By usinga different donor with a comparable E�, we found that for the

investigated acceptors the monolayer thickness affects the rate

rather significantly. Analysis of the results points to these MPCs as

acting in ET reactions as the 3D equivalents of self-assembled

monolayers (SAMs) formed on extended gold surfaces.

Experimental

Phenylethanethiol (Aldrich, 98%), hydrogen tetrachloroaurate

trihydrate (Aldrich, 99.9%), sodium borohydride (Aldrich, 99%),

dibenzoyl peroxide (Aldrich, 97%), decamethylferrocene (Alfa

Aesar, 99%), tetrahydrofuran (THF, Aldrich, 99.9%), methanol

(VWR, HPLC grade, 99.8%), and ethanol (Fluka, 99.8%) were

used as received. N,N-Dimethylformamide (Carlo Erba, 99.8%)

was treated with anhydrous Na2CO3 and doubly distilled at

reduced pressure under a nitrogen atmosphere. Dichloro-

methane (VWR, 99.8%) was freshly distilled over CaH2. Tetra-n-

butylammonium hexafluorophosphate (Fluka, 99%) was

recrystallized from ethanol. Tetra-n-butylammonium perchlo-

rate (TBAP, Fluka, 99%) was recrystallized from a 2 : 1 ethanol–

water solution and dried at 60 �C under vacuum. Low conduc-

tivity water was milli-Q water pro-analysis (Merck) or AnalaR

Normapur (NDH). Ferrocene (Carlo Erba, RPE) was recrys-

tallized by sublimation using a cold finger. The syntheses of

bis(pentafluorobenzoyl) peroxide, bis(para-cyanobenzoyl)

peroxide, and bis(para-methoxybenzoyl) peroxide were carried

out as described in the literature.45 No explosions were encoun-

tered with the peroxides thus prepared. Whereas the other

peroxides proved to be stable, bis(pentafluorobenzoyl) peroxide

showed appreciable decomposition on standing at room

temperature for some days. All peroxides were stored at 4 �C in

the dark and used with all precautions normally applied to

reactive peroxides.

This journal is ª The Royal Society of Chemistry 2012

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

A typical synthesis of Au25(SCH2CH2Ph)18 was carried out as

follows.19 After dissolving 1.01 g (2.56 mmol) of HAuCl4$3H2O

in 100 ml of THF, phenylethanethiol (2.1 ml, 15.6 mmol) was

added dropwise, under stirring, at room temperature. The

mixture was stirred for 45 min until the color changed from

yellow to milky white, and then NaBH4 (0.96 g, 25.4 mmol),

dissolved in 20 ml of water and cooled on ice, was added under

vigorous stirring to the mixture, kept at room temperature. A

black color, indicative of MPC formation, was immediately

produced in the solution mixture and a gas evolved. The reaction

progress was monitored by UV-vis absorption spectroscopy.

After 6–7 days the reaction mixture was filtered on paper and

THF was removed from the filtrate with a rotary evaporator,

leaving a solid covered by a colorless liquid (residual H2O from

aq. NaBH4). The water phase was removed and the solid was

washed several times with mixtures of 70/30 and then 80/20

methanol–water to remove excess of thiol. If the product is

mostly Au25L180, the final solid is a grayish powder. The powder

was collected on a Gooch G4 filter and washed with EtOH to

extract Au25L18� from the solid. The powder was collected off

the frit by dissolution in DCM and then the solution was rotary

evaporated to yield Au25(SCH2CH2Ph)18. The cluster soluble in

ethanol, Au25L18�, eventually undergoes aerobic oxidation, and

Au25L180 precipitates. To obtain pure samples of the�1 form, we

carried out the synthesis by adding 1.2 eq. tetra-n-octylammo-

nium bromide before addition of the thiol. The rest of the

procedure is as already described, but for the direct use of DCM

to collect the cluster from the Gooch.

UV-vis spectra of Au25 were recorded with either a Thermo

Scientific Evolution 60S spectrophotometer or, during syntheses,

an Ocean Optics QE65000 spectrophotometer equipped with a

DH-2000-BAL light source. The spectra resolution was 1 nm.

Matrix-assisted laser desorption ionization time-of-flight

(MALDI-TOF) mass spectrometry experiments were carried out

using an Applied Biosystems 4800 MALDI-TOF/TOF spec-

trometer equipped with a Nd:YAG laser operating at 355 nm.

The laser firing rate was 200 Hz and the accelerating voltage was

25 kV. The laser pulse intensity was kept as low as possible.

Trans-2-[3-(4-tert-Butylphenyl)-2-methyl-2-propenylidene]

malononitrile (DCTB) was used as the matrix. The nanoparticle

was dissolved in DCM containing DCTB to obtain a 1 mM

solution with a 1 : 400 nanoparticle/matrix ratio. 2 mL of solution

were dropcast onto the sample plate and air-dried. The spectra

were recorded in negative MS-Reflector mode.

The electrochemical experiments were carried out as follows.

The glassy carbon (GC) electrode was obtained by sharpening a

3 mm diameter Tokai GC-20 rod to obtain a disk with a radius of

0.55 mm. GC was connected to a copper wire with silver epoxy

and then sealed in glass tubing with epoxy (Torr Seal). The Pt

electrode was prepared from 99.9% Pt wire (Johnson Matthew)

and directly sealed in glass tubing. Its experimental radius was

0.52 mm. The disk electrodes were polished using silicon carbide

papers (500, 1000, 2400, and 4000), diamond pastes (Struers: 3, 1,

and 0.25 mm), alumina (BDH: 0.075 and 0.015 mm), and then

stored in ethanol. Before the experiments, the electrodes were

polished with 0.015 mm alumina, ultrasonically rinsed with

ethanol for 5 minutes, washed with acetone, and carefully dried

with a cold air stream. Both electrode materials were electro-

chemically activated in the background solution by means of

This journal is ª The Royal Society of Chemistry 2012

several voltammetric cycles at 0.5 V s�1 between the anodic and

cathodic potential limits of concern. The electrode areas of GC

and Pt were determined to be 9.64 � 10�3 and 8.51 � 10�3 cm2,

respectively, by measuring the voltammetric current for the

oxidation of ferrocene in DMF/0.1 M TBAP, in which ferrocene

has a diffusion coefficient of 1.13 � 10�5 cm2 s�1. A CHI 660c

electrochemical workstation was used for both DPV and most

CV experiments. Some redox catalysis experiments were also

carried out using an EG&G-PARC 173/179 potentiostat-digital

coulometer, an EG&G-PARC 175 universal programmer, and a

Nicolet 3091 12-bit resolution digital oscilloscope. For DPV, we

used a peak amplitude of 50 mV, a pulse width of 0.05 s, 2 mV

increment per cycle, and a pulse period of 0.1 s. For the CV

experiments, we employed the feedback correction to minimize

the ohmic drop between the working and the reference elec-

trodes. The corresponding CV curves were digitalized, corrected

for the background contribution by subtracting the curves

previously obtained in the absence of the substrate, and then

compared with digital simulations. The DigiSim 3.03 package

was used for all simulations, using a step size of 1 mV and an

exponential expansion factor of 0.5.

All electrochemical experiments were conducted under an Ar

atmosphere in an all-glass cell that was thermostatted at 25 �C.Except for the area determinations, the solvent was DCM con-

taining 0.1 M TBAH. A Pt wire was the counter electrode and an

Ag wire was used as a quasi-reference electrode; the Ag wire was

kept in a tube filled with the same electrolyte solution but sepa-

rated from the main compartment by a Vycor frit. The behavior

of the electrode and the quality of the solvent–electrolyte system

were first checked in the background solution by using CV and

DPV. At the end of each experiment, the reference electrode

potential was calibrated against the ferricenium/ferrocene (Fc+/

Fc) redox couple. E�Fc+/Fc was statistically calculated from an

extended series of independent experiments to be 0.460 V against

the KCl saturated calomel electrode (SCE). All potential values

are reported against SCE.

Results and discussion

Synthesis and characterization of Au25 clusters

The Au25(SCH2CH2Ph)18 clusters were synthesized by sodium

borohydride reduction of a solution obtained by reacting

HAuCl4 and phenylethanethiol in THF, as described in the

Experimental section. After purification, the purity and mono-

dispersity of the clusters were carefully checked by mass spec-

trometry, 1H and 13C NMR spectroscopy, UV-vis spectroscopy,

and electrochemistry. The negative mode MALDI-TOF mass

spectrometry spectrum shows an intense peak at MW 7394

corresponding to unfragmented Au25L18, which indicates that

the sample is truly monodisperse (Fig. 1, inset). The 1H and 13C

NMR spectroscopy behavior is as described in detail previ-

ously,19 i.e., each single proton and carbon atom of one ligand

family is not equivalent to that of the second family. Fig. 1 shows

the UV-vis spectra of Au25L180, [Au25L18

+][C6F5CO2�] (see

below), and [n-Oct4N+][Au25L18

�]. The spectra of the two redox

states obtained by direct syntheses, �1 and 0, display charac-

teristic differences particularly at ca. 400, 450, 600, and 800 nm.

The molar extinction coefficient of Au25L180 in DCM at 401 nm

Nanoscale, 2012, 4, 5333–5342 | 5335

Fig. 1 Comparison of the UV-vis absorption spectra of Au25L180 (red

curve), [n-Oct4N+][Au25L18

�] (black curve), and [Au25L18+][C6F5CO2

�](blue curve). The spectra were obtained in DCM, at room temperature,

and 0.020 mM concentration. The inset shows the negative-mode

MALDI-TOF spectrum of Au25L180.

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

is 3 ¼ 5.38 � 104 M�1 cm�1.19 The DPV and CV curves of

Au25L180 (Fig. 2, the potential range is limited to the region of

interest) are, qualitatively, as already described in detail for

Au25L18�.8,11 The main difference, however, is the equilibrium

potential, i.e., the potential spontaneously assumed by the elec-

trode when no current flows. Whereas for Au25L18� the potential

rests at a more negative value than that of the first oxidation

peak,8,46 for Au25L180 the equilibrium potential is at +0.088 V,

i.e., exactly in between the peaks of the Au25L18/Au25L18� and

Au25L18+/Au25L18 redox couples.

ET generation of Au25L18+

The redox catalysis experiments (see below) required preparation

of Au25L18+. This redox state was obtained in situ by oxidation of

Au25L180. The method relies on the use of an oxidant that is

suitable for use in organic solvents, does not bear protons, only

Fig. 2 DPV (black solid curve) and CV curves of 0.98 mM Au25L180

obtained on the GC electrode in DCM containing 0.1 M TBAH. Both

CV scans (0.1 V s�1) were initiated starting from the equilibrium

potential, 0.088 V, but the scan direction was either toward negative (blue

dashed curve) or positive (red solid curve) potential values. T ¼ 25 �C.

5336 | Nanoscale, 2012, 4, 5333–5342

generates diamagnetic species, and whose reaction with Au25L180

is quantitative.19 We exploited the properties of concerted

dissociative ETs to peroxides, in which the main reactions are

peroxide reduction (eqn (1), AB¼ROOR) followed by ET to the

ensuing radical (eqn (4), A ¼ RO).

A_+ e / A� (4)

E�AB/A_,B� ¼ E�

B_/B� � BDFE/F (5)

Since the standard potential for the concerted dissociative ET

to a neutral AB acceptor can be calculated through eqn (5),

where E�B_/B� is the standard potential of the leaving group and

BDFE is the dissociation free energy of the A–B bond,32 it

follows that because of their very small BDFE values peroxides

are ideal dissociative-type acceptors.28,32 Very facile reductions

are particularly expected for peroxides where the leaving group is

a benzoate-type anion, for which E�B_/B� is very positive.33

Another important property of most dissociative ETs is that

under usual AB reaction conditions the radical B_also undergoes

fast reduction by the very same donor species (eqn (4)), unless

competing radical reactions take place, such as H-atom

abstraction, rearrangements, bond cleavage, and coupling.

Hence, an overall two-electron uptake by the peroxide from the

same donor partner (electrode or two donor species in solution)

can often be expected.

Substituted dibenzoyl peroxides are soluble in organic solvents

and undergo dissociative reduction at particularly positive

potentials. For the oxidation of Au25L180 we used bis-(penta-

fluorodibenzoyl) peroxide. Fig. 3 shows the CV of this peroxide

in DCM on a GC electrode. The CV curve shows a broad peak

whose peak potential (Ep) is 0.50 V at 0.1 V s�1. From the peak

width DEp/2 (the difference between the potential at half-peak

height, Ep/2, and Ep), the value of the transfer coefficient a at

0.1 V s�1 is calculated to be 0.29 (DEp/2 ¼ 1.857 RT/Fa); a very

similar value of a, 0.27, is calculated from the scan-rate depen-

dence of Ep (|vEp/vlog v| ¼ 1.15 RT/Fa).47 These small a values

point to sluggish heterogeneous ET kinetics and are typical of

concerted dissociative ETs; for concerted dissociative ETs, in

fact, Ep can be smaller than E�AB/A_,B� by hundreds of

Fig. 3 Cyclic voltammetry for the reduction of 1.01 mM bis(penta-

fluorobenzoyl) peroxide at 0.5 V s�1 in DCM/0.1 M TBAH on the GC

electrode. T ¼ 25 �C. The inset shows the different potential regions at

which reduction occurs on GC (black) and Pt (red) electrodes.

This journal is ª The Royal Society of Chemistry 2012

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

millivolts.27,28 On the backward positive-going scan, the irre-

versible oxidation of the carboxylate is detectable at 1.83 V

(0.1 V s�1). The same peak can be observed by electrooxidation

of an authentic solution of the tetra-n-butylammonium salt of

pentafluorobenzoic acid.

The reaction of the peroxide with Au25L180 (eqn (6), Ar ¼

C6F5) was monitored by CV. Upon addition of 1 mM Au25L180

to a DCM solution already containing the peroxide at 0.5 mM

concentration, the peroxide CV curve disappears and only the

cluster CV pattern is visible, together with the carboxylate

oxidation peak. At the same time, the equilibrium potential

undergoes a positive shift to 0.67 V (Fig. 4). Both effects are in

keeping with quantitative formation of Au25L18+, in agreement

with eqn (6) and (7). The UV-vis spectrum of the resulting

Au25L18+ is shown in Fig. 1. Its distinct NMR behavior (as

opposed to Au25L180, which is a radical, Au25L18

+ is a diamag-

netic species) has been described in detail previously:19

(ArCO2)2 + Au25L18 / ArCO2_+ ArCO2� + Au25L18

+ (6)

ArCO2_+ Au25L18 / ArCO2� + Au25L18

+ (7)

To check whether reaction (6) was indeed a dissociative ET

process, we estimated the corresponding kET value. For perox-

ides,37 the calculation is usually best performed using the

German–Kuznetsov theory of non-adiabatic dissociative ET.41

This theory leads to a rate-constant expression in which kET is

proportional to HRP,2 where HRP is the electronic coupling

between the electronic wave functions describing the reactant

and product states. Estimating HRP, however, is a challenging

task and, in fact, it is the rate-constant expression that is

generally used to obtain HRP from experimental kET values

rather than the opposite. What we know from the analysis of

several dissociative ETs to peroxides is that theHRP value is most

often sufficiently small to make the preexponential factor of the

rate-constant expression significantly lower than the frequency

factor of adiabatic ET reactions.40 On the other hand, kET may

be simply described as follows (eqn (8)):

Fig. 4 Cyclic voltammetry for the reduction of 0.95 mM Au25L18+ at

0.1 V s�1 in DCM/0.1 M TBAH on the GC electrode. T ¼ 25 �C.

This journal is ª The Royal Society of Chemistry 2012

kET ¼ kZexp(�DG‡/RT) (8)

where Z is the preexponential factor and k is the electron

transmission coefficient. The dependence of DG‡ on DG� is best

expressed through the quadratic activation–driving force rela-

tionship proposed by Sav�eant for concerted dissociative ETs

(eqn (9)):27

DG‡ ¼ DG‡0

1þ DG

�

4DG‡0

!2(9)

Since we already know that Au25 behaves as a molecule-like

outer-sphere redox species,11 for the preexponential factor of eqn

(8) we use the average value obtained from Arrhenius plots of

redox reactions between benzoyl peroxides and outer-sphere

dimethoxybenzene donors.48 This factor, log A ¼ 8.7 (A ¼ 5 �108 M�1 s�1), is indeed comparable to that determined for the

dissociative ET to di-t-butyl peroxide by a series of aromatic

radical anions, log A ¼ 7.9.34 DG� is obtained from E� (eqn (5))

and E�AB/A_,B� of the redox couple Au25L18

+/Au25L18. By anal-

ysis of the irreversible voltammetric oxidation peak of the

benzoate formed upon reduction of the peroxide (cf. Fig. 3),

carried out along similar lines as already described for a series of

substituted benzoates,33 we estimate E�B_/B� ¼ 1.92 V. The O–O

BDFE is calculated by correcting the bond dissociation energy

(BDE) of the peroxide obtained from thermal decomposition

data, 30 kcal mol�1,49 for the entropy term, �8.5 kcal mol�1.34

The value estimated for the dissociative E� is 0.99 V. It is

rewarding to note that an independent estimate of the same

quantity, based on the potential-dependence of a (from the scan-

rate dependence of DEp/2),33,36 leads to a very similar value,

1.08 V. The DG� values of ET reactions (6) and (7) can thus be

estimated to be �0.76 and �1.69 eV, respectively.

For a homogeneous dissociative ET, the intrinsic barrier DG‡0

is the sum of inner and outer (DG‡0,out) reorganization free energy

contributions (eqn (10)). The inner reorganization energy

contains a term (DG‡0,in) associated with changes of bond lengths

and angles of donor and acceptor; DG‡0,in does not include the

contribution of the mode corresponding to the breaking O–O

bond, which is directly taken into account as BDE/4.27 In our

case, DG‡0,in is essentially related to the non-negligible inner

reorganization that Au25 undergoes upon ET: the latter is coin-

cident with the heterogeneous value, 4.2 kcal mol�1.11,31,50 For a

one-electron process and if the fluctuations in solvent polariza-

tion are treated by dielectric continuum theory, DG‡0,out can be

calculated according to eqn (11):51

DG‡0 ¼ BDE/4 + DG‡

0,in + DG‡0,out (10)

DG‡0;out ¼

NAe2

16p30

�1

3op� 1

3s

��1

2rDþ 1

2rA� 1

RDA

�(11)

where NA is Avogadro’s number, e is the charge of the electron,

30 is the permittivity of vacuum, 3op and 3s are the optical (square

of refractive index) and static dielectric constants of the solvent,

rD and rA are the radii of the donor and the acceptor (assumed to

be spherical), and RDA is the center-to-center separation distance

of the two reactants and is taken to be the sum rD + rA (van der

Waals contact in the solvent cage). We used rD¼ 12 �A23 and rA¼3.6 �A (from the diffusion coefficientD¼ 1.49� 10�5 cm2 s�1 and

Nanoscale, 2012, 4, 5333–5342 | 5337

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

the Stokes–Einstein equation) resulting in DG‡0,out ¼ 3.65 kcal

mol�1. From these values we estimated kET ¼ 9.2� 102 M�1 s�1.

Reaction (7) is assumed to be an adiabatic (eqn (8), k ¼ 1)

marcusian outer-sphere ET. The ET rate constant was calculated52

using 1� 1010 M�1 s�1 for the rate constant for diffusion, Z¼ 3�1011 M�1 s�1,34 and DG‡

0,out¼ 7.8 kcal mol�1 (eqn (11)). Because of

the large driving force, the estimated value corresponds to the

diffusion limit, i.e., kET ¼ 1 � 1010 M�1 s�1. Decarboxylation of

the benzoyloxy radical, which occurs in ca. 1 ms,53 does not compete

with reaction (7) when the reactant concentrations are at least in the

millimolar range,19 a condition easily met in electrochemical

experiments. The occurrence of decarboxylation, however, was

taken into account for estimating E�B_/B�.

Direct reduction of dibenzoyl peroxides

The electroreduction of the other sym-substituted dibenzoyl

peroxides used as acceptors in the homogeneous ET study (see

below) also is electrochemically and chemically irreversible.

Besides bis(pentafluorobenzoyl) peroxide, we focused on bis-

(para-cyanobenzoyl) peroxide, dibenzoyl peroxide, and bis(para-

methoxybenzoyl) peroxide. The Ep values decrease as the

substituent is made less electron-withdrawing. On aGC electrode

at 0.2 V s�1, typical values are: in DCM/0.1 M TBAH, �0.03 V

(p-CN), �0.35 V (H), and �0.36 V (p-MeO) (Fig. 5). A similar

order is observed in DMF/0.1 M TBAP: 0.29 V (p-CN), �0.05 V

(H), and �0.13 V (p-MeO).54 Peroxide reduction forms the cor-

responding carboxylate anions, whose irreversible oxidation is

detected33,36 on the reverse positive-going scan at (DCM/0.1 M

TBAH, GC, 0.2 V s�1) 1.84 V (perfluoro), 1.61 V (p-CN), 1.43 V

(H), and 1.38 V (p-MeO).

Ep also depends quite dramatically on the electrode material: for

example, the values of bis(para-nitrobenzoyl) peroxide in DMF/0.1

M TBAP (0.2 V s�1) are 0.50 V (GC), 0.03 V (Au), �0.04 V (Hg),

and �0.50 V (Pt).54 The inset of Fig. 3 shows the difference

between the CV curves for the reduction of bis-(penta-

fluorodibenzoyl) peroxide on GC and Pt electrodes at 0.5 V s�1 in

DCM/0.1 M TBAH. On the other hand, it is well known that for

Fig. 5 Cyclic voltammetries for the reduction of the selected sym-

substituted dibenzoyl peroxides (from left to right: perfluoro, para-cyano,

unsubstituted, para- methoxy) at 0.1 V s:1 in DCM/0.1 M TBAH on the

GC electrode. T ¼ 25 �C.

5338 | Nanoscale, 2012, 4, 5333–5342

dissociative ETs GC provides the best approximation of an inert

electrode surface.27–30 The sluggishness of heterogeneous ET on

other electrode surfaces, however, can be exploited to ‘‘push’’ the

direct reduction wave toward more negative potentials, thereby

allowing better CV resolution (less overlapping of peaks) in the

homogeneous redox catalysis experiments. In the following section,

we will take advantage of this feature.

Au25 species as ET mediators

The kET values for the dissociative reduction of the selected

peroxides by Au25 donor species were determined using the redox

catalysis approach.42 In the absence of the peroxide acceptor, the

voltammetric reduction of the mediator is reversible (Fig. 6).

Upon addition of the acceptor, however, the reduction peak is

transformed into a chemically irreversible, catalytic peak. This is

because the electrogenerated active form of the mediator (the

donor) diffuses toward the bulk and reacts with the acceptor with

regeneration of the oxidized form of the mediator. The latter can

now be reduced again and this triggers the catalytic increase of

the CV peak current. The homogeneous ET reactions are as

shown in eqn (6) and (7), or their equivalent when the donor

redox couple is Au25L18/Au25L18� instead of Au25L18

+/Au25L18.

For the actual homogeneous redox catalysis experiments, we

used both Au25L18 and Au25L18� as the active donors. The use of

the direct reduction of Au25L18, however, is impractical because of

the proximity of the two redox waves of Au25. For all catalysis

experiments, we thus adopted a procedure in which Au25L18 is first

quantitatively oxidized by reaction with bis(pentafluorobenzoyl)

peroxide (in a molar ratio of 2 : 1, because the peroxide acts as a

two electron acceptor). CV curves of both reversible reduction

peaks were then acquired, usually in the potential scan rate (v)

range from 0.05 to 10 V s�1. The peroxide was added and a new

series of CV curves were recorded. The current of the catalytic peak

depends on v and the concentration of donor and acceptor.

Whereas the mediator concentration was kept constant at 1 mM,

we used three peroxide concentrations with the following molar

ratios ([acceptor]/[donor]) values: 0.5, 1 and 2 (perfluoro); 1, 2 and 5

Fig. 6 Cyclic voltammetries for the reduction of 1.12 mM Au25L18+ in

the absence (black line) and in the presence of 1 eq. (blue line) and 2 eq.

(red line) of bis(para-cyanobenzoyl) peroxide at 0.05 V s�1 in DCM/0.1

M TBAH on the Pt electrode. T ¼ 25 �C.

This journal is ª The Royal Society of Chemistry 2012

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

(p-CN); 5, 10 and 15 (H); 20, 40 and 60 (p-OMe). The molar ratios

were selected based on the actual driving force: the larger the latter

(i.e., the more DG� is made negative), the smaller the amount of

peroxide required to detect sufficient redox catalysis effects on the

CV time scale. Typical examples of catalytic CV curves obtained

with two peroxide concentrations are shown in Fig. 6. When v is

made smaller, the time scale of the experiment increases and so

does the catalytic activity, as evidenced by an increase of the peak

current of both peaks (Fig. 7). The increase is more significant for

the second peak because Au25L18� forms at a more negative

potential and is thus a more powerful reductant than Au25L18. As

already mentioned, for the redox catalysis measurements we used a

platinum electrode, which does not affect the reversible electro-

chemical behavior of Au25 but makes more sluggish the direct

reduction of the peroxide. Comparison of the experimental and

simulated curves obtained in the absence and in the presence of

peroxide led to the kET values shown in Table 1.55 We could obtain

reproducible kET values (i.e., for all v values and peroxide

concentrations) for both Au25L18 and Au25L18� donors. For bis-

(para-methoxybenzoyl) peroxide, redox catalysis could only be

observed at the second peak, a quite understandable result in view

of the trend of the values of Table 1 and considering that CV is the

electrochemical technique of choice only when kET > 1 M�1 s�1.

The kET values follow the order expected based on the

reducing power of the two Au25 redox states and the peroxide

reduction potentials. Noteworthy, the value log kET ¼ 3.30

obtained for the system Au25L18/bis(pentafluorobenzoyl)

peroxide is in remarkable agreement with the value of 2.96 esti-

mated using the dissociative ET theory and the various param-

eters described in the section about Au25L18+. This nicely shows

that despite the irreversibility of dissociative ETs, quite precise

estimates of ET rate-constant values are possible. This also

points to our estimate of the carboxylate E� value as quite

reasonable. For the other carboxylates, we carried out a similar

analysis33 of their irreversible oxidation peaks, using the same

rate constant value (k ¼ 1 � 106 s�1) for decarboxylation of

ArCO2_, and obtained the E�B_/B� values gathered in Table 1.

Fig. 7 Cyclic voltammetries for the reduction of 0.97 mM Au25L18+ in

the presence of 1 eq. of bis(pentafluorobenzoyl) peroxide at different scan

rates (top to bottom: 0.05, 0.1, 0.2, 0.5, 1, and 2 V s�1) in DCM/0.1 M

TBAH on the Pt electrode. T ¼ 25 �C. The current is normalized with

respect to the scan rate.

This journal is ª The Royal Society of Chemistry 2012

Analysis of the ET results

Usually, homogeneous redox catalysis plots are shown for the

dissociative reduction (the same holds true for oxidations) of a

single acceptor by a series of electron donors. Thus, log kET data

are plotted as a function of the donors’ E�s, which are easily

determined from their reversible CVs. In those cases where a

good estimate can be made of the acceptor E�, however, the datacan be shown as a function of the reaction free energy, DG�. Forthe investigated systems, in which different donors and acceptors

are used, we will follow the second approach. The E�AB/A_B�

values of peroxides are calculated as already described above for

bis(pentafluorobenzoyl) peroxide, i.e., by using eqn (5). Avail-

able thermal decomposition analysis of an extended series of

sym-substituted benzoyl peroxides shows that all BDEs can be

estimated to be 30 � 1 kcal mol�1;56 this introduces an error of

0.04 eV on DG�. BDFE was obtained by applying the �8.5 kcal

mol�1 correction.34 The resulting DG� values (Table 1) have a

maximum prudential error (including that stemming from the

E�B_/B� estimates) of � 0.2 eV. Nonetheless, Fig. 8 illustrates a

remarkably nice dependence of log kET on DG� spanning a large

free energy range, 1.05 eV. The slope is small and we calculate an

average transfer coefficient a (¼ vDG‡/vDG� ¼ �RTvln kET/

vDG�) of only 0.26, despite the proximity to DG� ¼ 0 where a

should be ca. 0.5.27 This is not an uncommon observation for

dissociative ETs,57 although in the present case the difference is

large. We attribute this outcome mostly as due to the error

associated with DG�, which also prevents detecting a sizeable

curvature of the plot (according to the dissociative ET theory, the

plot should be parabolic).27 In the framework of the present

paper, however, what needs to be stressed is the capabilities of

Au25 species to act as efficient electron donors. From an exper-

imental point of view, we also remark the usefulness of the

approach adopted to generate Au25L18+. Very recently,

the same species was obtained by oxidation of the native

cluster with 2,2,6,6-tetramethylpiperidin-1-oxoammonium tet-

rafluoroborate,58 but this approach would not be convenient for

ET catalysis experiments forming radicals or radical ions because

these reactions would be carried out in the presence of a well-

known spin trap (a nitroxide), with associated chemical

complications.59 We also note that the approach of using

different oxidation states of the same species to obtain more than

one ET rate constant, made possible by the stability of the

Au25L18/Au25L18� and Au25L18

+/Au25L18 redox couples, is rare

because for common mediators there are usually problems

caused by the intrinsic lability of the dianions or dications (e.g.,

basicity or acidity) formed from the corresponding radical ions.60

One important aspect to consider is that the electrons donated

to the peroxide and the radical ArCO2_ come from the MPC

HOMO. For Au25, the HOMOs do not spread significantly onto

the phenylethanethiolate ligands,61 although sulfur is most likely

involved as witnessed by the difference between the Raman Au–S

bands of Au25L18 and Au25L18�.26 This means that ET occurs

when the donor (gold core) and the acceptor are separated by a

distance roughly corresponding to the monolayer thickness. This

situation is thus the 3D analogue of what happens when a freely

diffusing species is reduced or oxidized by an electrode protected

by a self-assembled monolayer.62 According to the structure of

Au25,22–24 the average distance between the connecting Au atom

Nanoscale, 2012, 4, 5333–5342 | 5339

Table 1 Rate-driving force data for the dissociative ET to peroxides by Au25 donors in DCM at 25 �C

Donora Peroxide acceptor E�AB/A_,B�

b (V) E�B_/B�

c (V) DG�d (eV) log kETe (M�1 s�1)

0 Perfluoro 0.99 1.92 �0.76 3.30�1 Perfluoro 0.99 1.92 �1.06 4.600 4-CN 0.62 1.55 �0.39 1.78�1 4-CN 0.62 1.55 �0.69 3.300 H 0.22 1.15 �0.01 0.30�1 H 0.22 1.15 �0.29 1.60�1 4-OMe 0.17 1.10 �0.24 1.30Fc Perfluoro 0.99 1.92 �1.01 5.78

a The number indicates the redox state of the Au25 donor. Fc ¼ decamethylferrocene. b Dissociative ET formal potential. c Benzoate formal potential.d Uncertainty is 0.2 eV. e Uncertainty is 10%.

Fig. 8 Driving-force dependence of the dissociative ET rate constant for

the reduction of sym-substituted benzoyl peroxides by Au25 mediators

(black filled circles) in DCMat 25 �C. The plot also shows the log kET resultobtained with decamethylferricenium and bis(pentafluorobenzoyl) peroxide

(red filled square). The dashed line is the linear fit to the Au25 data.

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

and the outermost H atom of phenylethanethiolate ligands is ca.

7.5 �A. On the other hand, if one considers the structure of the

monolayer periphery of this small cluster it appears that, in fact,

most of the ligands are not at van der Waals distance from their

neighbors. This makes the actual average thickness of the

monolayer smaller. Based on the X-ray crystallographic struc-

tures, ET from the MPC gold core to comparatively smaller

molecules, such as the investigated peroxides, is thus occurring

when the actual donor and acceptor are separated by, say, 6–7 �A.

If the ET is adiabatic, this relatively small donor–acceptor

separation is not expected to affect the rate significantly. This is

probably the reason why we previously observed11 that

Au25L182� behaves toward adiabatic-type acceptors, such as

benzyl bromide and diphenyl disulfide,31,40 in good agreement

with the behavior displayed by common aromatic radical-anion

donors. On the other hand, we already stressed that peroxides are

intrinsically nonadiabatic acceptors, and it is thus conceivable

that even a slight increase of the donor–acceptor distance may be

relevant.

To address this issue we used another electron donor, deca-

methylferrocene, which has an E� value of �0.025 V, rather

similar to that of the Au25L18/Au25L18� redox couple, �0.077 V.

We applied the same strategy used for the catalysis involving

5340 | Nanoscale, 2012, 4, 5333–5342

Au25L18+/Au25L18, namely chemical reaction of neutral deca-

methylferrocene with bis(pentafluorobenzoyl) peroxide to form

quantitatively decamethylferricenium (which shifts the equilib-

rium potential to more positive potentials than its E�), followedby the usual set of CV runs in the absence and then in the

presence of peroxide. During the redox catalysis experiments, the

actual electron donor is thus decamethylferrocene. The corre-

sponding log kET value (Table 1 and Fig. 8), obtained by using

various v values and three bis(pentafluorobenzoyl) peroxide

concentrations (molar ratios ¼ 0.5, 1 and 1.5), is indeed 1.18 log

kET units larger than that obtained with the same peroxide and

the Au25L18/Au25L18� mediator couple. If the slight donor E�

difference is taken into account, Dlog kET becomes 1.37.

ET reactions, particularly at relatively short donor–acceptor

distances (rDA) and in the absence of hopping sites, proceed in a

single-step superexchange-mediated ET.63 In this mechanism, the

ET rate constant is described as decreasing exponentially with

the increase of rDA, i.e., kET f exp(�brDA) where b is the

exponential factor describing the falloff rate through the specific

bridge. While for saturated bridges b ¼ 0.8–1.0 �A�1, smaller

values are measured with oligophenylenes and benzylic deriva-

tives, 0.4–0.7 �A�1.63,64 If we apply the above exponential-decay

equation to the Au25 and decamethylferrocene results, the

observed b is 0.45–0.53 �A�1, in reasonable agreement with the

latter values although somehow on the lower side. Based on this

analysis, we suspect that the effective rDA could be smaller, by

probably ca. 1 �A (leading to b ¼ 0.63 �A�1), than the above lower

estimate.

Conclusions

Overall, our results show that, thanks to precise control of the

oxidation state, Au25 can be efficiently used for ET catalysis

when the stable redox couples Au25L18/Au25L18� and Au25L18

+/

Au25L18 are employed. Interestingly, the results also show that

due to the monolayer thickness the nature of the actual electron

acceptor (or donor, if the process is an Au25 mediated oxidation)

may affect the intermolecular ET rate rather significantly. In this

framework, Au25 MPCs thus act as the soluble 3D equivalents of

SAMs formed on extended gold surfaces (2D SAMs), where the

ET reaction between the electrode and the redox species takes

place through the SAM. This is unprecedented: although for

some radical-ion ET mediators part of the molecule may be

‘‘screened’’ by moieties where the singly occupied molecular

orbital does not spread, intermolecular ET rate constants are the

This journal is ª The Royal Society of Chemistry 2012

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

average of random D/A distance and orientation distributions in

the encounter complex and thus screening effects are undetect-

able, as opposed to what was found for intramolecular ETs.38,40

MPCs thus provide a unique opportunity to test the behavior of

fully screened ET mediators. It follows that if from one hand

Au25 can be efficiently used for ET catalysis, on the other hand

efficient prediction of the ET rate depends on the actual acceptor

type.

This work was financially supported by the Foundation

CARIPARO (Progetto di Eccellenza), the Italian Ministero

dell’Istruzione, dell’Universit�a e della Ricerca (PRIN grant

20098Z4M5E), and the University of Padova (PRAT grant

CPDA103389).

Notes and references

1 (a) A. C. Templeton, W. P. Wuelfing and R. W. Murray, Acc. Chem.Res., 2000, 33, 27–36; (b) J. Z. Zhang, Z. Wang, J. Liu, S. Chen andG. Liu, Self-Assembled Nanostructures, Kluwer Academic/PlenumPublisher, New York, 2003, pp. 271–307.

2 R. W. Murray, Chem. Rev., 2008, 108, 2688–2720.3 (a) J. F. Hicks, A. C. Templeton, S. Chen, K. M. Sheran, R. Jasti,R. W. Murray, J. Debord, T. G. Schaaff and R. L. Whetten, Anal.Chem., 1999, 71, 3703–3711; (b) J. F. Hicks, D. T. Miles andR. W. Murray, J. Am. Chem. Soc., 2002, 124, 13322–13328; (c)D. T. Miles, M. C. Leopold, J. F. Hicks and R. W. Murray,J. Electroanal. Chem., 2003, 554–555, 87–97; (d) R. Guo,D. Georganopoulou, S. W. Feldberg, R. L. Donkers andR. W. Murray, Anal. Chem., 2005, 77, 2662–2669.

4 B. M. Quinn, P. Liljeroth, V. Ruiz, T. Laaksonen and K. Kontturi,J. Am. Chem. Soc., 2003, 125, 6644–6645.

5 A. H. Holm, M. Ceccato, R. L. Donkers, L. Fabris, G. Pace andF. Maran, Langmuir, 2006, 22, 10584–10589.

6 R. L. Wolfe and R. W. Murray, Anal. Chem., 2006, 76, 1167–1173.7 D. Lee, R. L. Donkers, J. M. DeSimone and R. W. Murray, J. Am.Chem. Soc., 2003, 125, 1182–1183.

8 D. Lee, R. L. Donkers, G. Wang, A. S. Harper and R. W. Murray,J. Am. Chem. Soc., 2004, 126, 6193–6199.

9 V. L. Jimenez, D. G. Georganopoulou, R. J. White, A. S. Harper,A. J. Mills, D. Lee and R. W. Murray, Langmuir, 2004, 20, 6864–6870.

10 R. Guo and R. W. Murray, J. Am. Chem. Soc., 2005, 127, 12140–12143.

11 S. Antonello, A. H. Holm, E. Instuli and F. Maran, J. Am. Chem.Soc., 2007, 129, 9836–9837.

12 (a) S. Chen, R. S. Ingram, M. J. Hostetler, J. J. Pietron,R. W. Murray, T. G. Schaaff, J. T. Khoury, M. M. Alvarez andR. L. Whetten, Science, 1998, 280, 2098–2101; (b)R. Balasubramanian, R. Guo, A. J. Mills and R. W. Murray,J. Am. Chem. Soc., 2005, 127, 8126–8132.

13 O. Toikkanen, V. Ruiz, G. R€onnholm, N. Kalkkinen, P. Liljeroth andB. M. Quinn, J. Am. Chem. Soc., 2008, 130, 11049–11055.

14 D. Garc�ıa-Raya, R. Madue~no, M. Bl�azquez and T. Pineda, J. Phys.Chem. C, 2009, 113, 8756–8761.

15 L. D. Menard, S. Gao, H. Xu, R. D. Twesten, A. S. Harper, Y. Song,G. Wang, A. D. Douglas, J. C. Yang, A. I. Frenkel, R. G. Nuzzo andR. W. Murray, J.Phys. Chem. B, 2006, 110, 12874–12883.

16 Y. Yang and S. Chen, Nano Lett., 2003, 3, 75–79.17 F. Wen, U. Englert, B. Gutrath and U. Simon, Eur. J. Inorg. Chem.,

2008, 106–111.18 J. F. Parker, C. A. Fields-Zinna and R. W. Murray, Acc. Chem. Res.,

2010, 43, 1289–1296.19 A. Venzo, S. Antonello, J. Gascon, I. Guryanov, R. D. Leapman,

N. V. Perera, A. Sousa, M. Zamuner, A. Zanella and F. Maran,Anal. Chem., 2011, 83, 6355–6362.

20 J. P. Parker, J. E. F. Weaver, F. McCallum, C. A. Fields-Zinna andR. W. Murray, Langmuir, 2010, 26, 13650–13654.

21 (a) Y. Negishi, K. Nobusada and T. Tsukuda, J. Am. Chem. Soc.,2005, 127, 5261–5270; (b) J. B. Tracy, G. Kalyuzhny, M. C. Crowe,R. Balasubramanian, J.-P. Choi and R. W. Murray, J. Am. Chem.Soc., 2007, 129, 6706–6707.

This journal is ª The Royal Society of Chemistry 2012

22 M. Zhu, W. T. Eckenhoff, T. Pintauer and R. Jin, J. Phys. Chem. C,2008, 112, 14221–14224.

23 M. W. Heaven, A. Dass, P. S. White, K. M. Holt and R. W. Murray,J. Am. Chem. Soc., 2008, 130, 3754–3755.

24 M. Zhu, C. M. Aikens, F. J. Hollander, G. C. Schatz and R. Jin,J. Am. Chem. Soc., 2008, 130, 5883–5885.

25 J.-P. Choi and R. W. Murray, J. Am. Chem. Soc., 2006, 128, 10496–10502.

26 J. F. Parker, J.-P. Choi, W. Wang and R. W. Murray, J. Phys. Chem.C, 2008, 112, 13976–13981.

27 J.-M. Sav�eant, J. Am. Chem. Soc., 1997, 109, 6788–6795.28 F. Maran, D. D. M. Wayner and M. S. Workentin, Adv. Phys. Org.

Chem., 2001, 36, 85–166.29 S. Antonello and F. Maran, Chem. Soc. Rev., 2005, 34, 418–428.30 A. Houmam, Chem. Rev., 2008, 108, 2180–2237.31 A. B. Meneses, S. Antonello, M. C. Ar�evalo, C. C. Gonz�alez,

J. Sharma, A. N. Wallette, M. S. Workentin and F. Maran, Chem.–Eur. J., 2007, 13, 7983–7995.

32 S. Antonello, M. Musumeci, D. D. M. Wayner and F. Maran, J. Am.Chem. Soc., 1997, 119, 9541–9549.

33 S. Antonello and F. Maran, J. Am. Chem. Soc., 1999, 121, 9668–9676.34 R. L. Donkers, F. Maran, D. D. M. Wayner and M. S. Workentin,

J. Am. Chem. Soc., 1999, 121, 7239–7248.35 (a) M. S. Workentin, F. Maran and D. D. M. Wayner, J. Am. Chem.

Soc., 1995, 117, 2120–2121; (b) S. Antonello, F. Formaggio,A. Moretto, C. Toniolo and F. Maran, J. Am. Chem. Soc., 2003,125, 2874–2875; (c) F. Polo, S. Antonello, F. Formaggio,C. Toniolo and F. Maran, J. Am. Chem. Soc., 2005, 127, 492–493.

36 S. Antonello and F. Maran, J. Am. Chem. Soc., 1997, 119, 12595–12600.

37 S. Antonello, F. Formaggio, A. Moretto, C. Toniolo and F. Maran,J. Am. Chem. Soc., 2001, 123, 9577–9584.

38 S. Antonello, M. Crisma, F. Formaggio, A. Moretto, F. Taddei,C. Toniolo and F. Maran, J. Am. Chem. Soc., 2002, 124, 11503–11513.

39 (a) R. L. Donkers and M. S. Workentin, Chem.–Eur. J., 2001, 7,4012–4020; (b) D. Magri and M. S. Workentin, Org. Biomol.Chem., 2003, 1, 3418–3429.

40 S. Antonello, A. Venzo and F. Maran, J. Electroanal. Chem., 2011,660, 234–242.

41 (a) E. D. German and A. M. Kuznetsov, J. Phys. Chem., 1994, 98,6120–6127; (b) E. D. German, A. M. Kuznetsov andV. A. Tikhomirov, J. Phys. Chem., 1995, 99, 9095–9101.

42 (a) C. P. Andrieux, C. Blocman, J. M. Dumas-Bouchiat, F. M’Hallaand J.-M. Sav�eant, J. Electroanal. Chem., 1980, 113, 19–40; (b)C. P. Andrieux and J.-M. Sav�eant, J. Electroanal. Chem., 1986, 205,43–58.

43 (a) T. B. Christensen and K. Daasbjerg, Acta Chem. Scand., 1997, 51,307–317; (b) S. Antonello, K. Daasbjerg, H. Jensen, F. Taddei andF. Maran, J. Am. Chem. Soc., 2003, 125, 14905–14916.

44 (a) H. Lund and K. Daasbjerg, Acta Chem. Scand., 1993, 47, 597–604;(b) Y. Huang and D. D. M. Wayner, J. Am. Chem. Soc., 1994, 116,2157–2158.

45 (a) C. G. Swain, W. H. Stockmayer and J. T. Clarke, J. Am. Chem.Soc., 1950, 72, 5426–5434; (b) C. A. Barson and R. A. Wisdom,Eur. Polym. J., 1972, 8, 1139–1144.

46 Y. Song, A. S. Harper and R. W. Murray, Langmuir, 2005, 21, 5492–5500.

47 R. S. Nicholson and I. Shain, Anal. Chem., 1964, 36, 706–723.48 C. Walling and C. X. Zhao, Tetrahedron, 1982, 38, 1105–1112.49 The BDE value was taken as the average between pertinent literature

values: (a) Ref. 45b; (b) C. X. Zhao, X. K. Jiang and J. Y. Zhang,J. Fluorine Chem., 1985, 27, 401–410.

50 Not explicitly reported in ref. 11 is that the inner reorganizationenergy corresponding to the Au25L18

+/Au25L180 redox couple is

slightly larger than that of the Au25L180/Au25L18

� couple, 3.7 kcalmol�1

51 R. A. Marcus, Faraday Discuss. Chem. Soc., 1982, 74, 7–15.52 R. A. Marcus and N. Sutin, Biochim. Biophys. Acta, 1985, 811, 265–

322.53 J. Chateauneuf, J. Lusztyk and K. U. Ingold, J. Am. Chem. Soc.,

1988, 110, 2886–2893.54 G. M. Rigodanza, Laurea in Chimica, M.Sc. thesis, Padova, 1995.55 Simulations took into account the E� values of the two redox steps of

Au25, eqn (1) (E� for the direct reduction of AB, with AB¼ peroxide),

Nanoscale, 2012, 4, 5333–5342 | 5341

Publ

ishe

d on

07

June

201

2. D

ownl

oade

d by

Uni

vers

ity o

f L

eeds

on

13/0

8/20

13 0

2:55

:05.

View Article Online

eqn (4) (E� for the ArCO2_/ArCO2� redox couple), rate constant for

the decarboxylation of ArCO2_, reactions 6 and 7 and theirequivalent for Au25L18/Au25L18

� as the donor redox couple.56 A. T. Blomquist and A. J. Buselli, J. Am. Chem. Soc., 1951, 73, 3883–

3888.57 H. Lund, K. Daasbjerg, T. Lund and S. U. Pedersen,Acc. Chem. Res.,

1995, 28, 313–319.58 Z. Liu, M. Zhu, X. Meng, G. Xu and R. Jin, J. Phys. Chem. Lett.,

2011, 2, 2104–2109.59 L. Eberson, Adv. Phys. Org. Chem., 1998, 31, 91–141.60 Organic Electrochemistry, ed. H. Lund and M. M. Baizer, Marcel

Dekker, New York, 3rd edn, 1991.

5342 | Nanoscale, 2012, 4, 5333–5342

61 C. M. Aikens, J. Phys. Chem. Lett., 2010, 1, 2594–2599.62 (a) J. F. Smalley, H. O. Finklea, C. E. D. Chidsey, M. R. Linford,

S. E. Creager, J. P. Ferraris, K. Chalfant, T. Zawodzinsk,S. W. Feldberg and M. D. Newton, J. Am. Chem. Soc., 2003, 125,2004–2013; (b) J. J. Gooding, F. Mearns, W. Yang and J. Liu,Electroanalysis, 2003, 15, 81–96.

63 (a) M. D. Newton, Chem. Rev., 1991, 91, 767–792; (b) A. Nitzan,Annu. Rev. Phys. Chem., 2001, 52, 681–750; (c) M. N. Paddon-Row,Aust. J. Chem., 2003, 56, 729–748.

64 R. E. Holmlin, R. F. Ismagilov, R. Haag, V. Mujica, M. A. Ratner,M. A. Rampi and G. M. Whitesides, Angew. Chem., Int. Ed., 2001,40, 2316–2320.

This journal is ª The Royal Society of Chemistry 2012