Electrochemistry and Optical Absorbance and Luminescence of Molecule-like Au38 Nanoparticles

Article abstract of DOI:10.1021/ja049605b

This paper describes electrochemical and spectroscopic properties of a well-characterized, synthetically accessible, 1.1 nm diam Au nanoparticle, Au₃₈(PhC₂S)₂₄, where PhC₂S is phenylethylthiolate. Properties of other Au₃₈ nanoparticles made by exchanging the monolayer ligands with different thiolate ligands are also described. Voltammetry of the Au₃₈ nanoparticles in CH ₂Cl₂ reveals a 1.62 V energy gap between the first one-electron oxidation and the first reduction. Based on a charging energy correction of ca. 0.29 V, the indicated HOMO-LUMO gap energy is ca. 1.33 eV. At low energies, the optical absorbance spectrum includes peaks at 675 nm (1.84 eV) and 770 nm (1.61 eV) and an absorbance edge at ca. 1.33 eV that gives an optical HOMO-LUMO gap energy that is consistent with the electrochemical estimate. The absorbance at lowest energy is bleached upon electrochemical depletion of the HOMO level. The complete voltammetry contains two separated doublets of oxidation waves, indicating two distinct molecular orbitals, and two reduction steps. The ligand-exchanged nanoparticle Au₃₈(PEG ₁₃₅S)₁₃(PhC₂S)₁₁, where PEG ₁₃₅S is -SCH₂CH₂OCH₂CH ₂OCH₃, exhibits a broad (1.77-0.89 eV) near-IR photoluminescence band resolvable into maxima at 902 nm (1.38 eV) and 1025 nm (1.2 eV). Much of the photoluminescence occurs at energies less than the HOMO-LUMO gap energy. A working model of the energy level structure of the Au₃₈ nanoparticle is presented.

Full text of DOI:10.1021/ja049605b

Published on Web 04/27/2004
Electrochemistry and Optical Absorbance and Luminescence
of Molecule-like Au38 Nanoparticles
†
‡
Dongil Lee, Robert L. Donkers, Gangli Wang, Amanda S. Harper, and
Royce W. Murray*
Contribution from the Kenan Laboratories of Chemistry, UniVersity of North Carolina,
Chapel Hill, North Carolina 27599
Received January 22, 2004; E-mail: rwm@email.unc.edu
Abstract: This paper describes electrochemical and spectroscopic properties of a well-characterized,
synthetically accessible, 1.1 nm diam Au nanoparticle, Au38(PhC
Properties of other Au38 nanoparticles made by exchanging the monolayer ligands with different thiolate
ligands are also described. Voltammetry of the Au38 nanoparticles in CH Cl reveals a 1.62 V energy gap
2 2
S)24, where PhC S is phenylethylthiolate.
2
2
between the first one-electron oxidation and the first reduction. Based on a charging energy correction of
ca. 0.29 V, the indicated HOMO-LUMO gap energy is ca. 1.33 eV. At low energies, the optical absorbance
spectrum includes peaks at 675 nm (1.84 eV) and 770 nm (1.61 eV) and an absorbance edge at ca. 1.33
eV that gives an optical HOMO-LUMO gap energy that is consistent with the electrochemical estimate.
The absorbance at lowest energy is bleached upon electrochemical depletion of the HOMO level. The
complete voltammetry contains two separated doublets of oxidation waves, indicating two distinct molecular
orbitals, and two reduction steps. The ligand-exchanged nanoparticle Au38(PEG135S)13(PhC
2
S)11, where
PEG135S is -SCH CH OCH CH OCH , exhibits a broad (1.77-0.89 eV) near-IR photoluminescence band
2
2
2
2
3
resolvable into maxima at 902 nm (1.38 eV) and 1025 nm (1.2 eV). Much of the photoluminescence occurs
at energies less than the HOMO-LUMO gap energy. A working model of the energy level structure of the
Au38 nanoparticle is presented.
Introduction
This paper describes properties of a purified molecule-like,
monolayer-protected Au38 nanoparticle. Molecule-like nano-
particles have been previously isolated using fractional precipi-
Gold nanoparticles have received much attention owing to
their novel optical and electronic properties and potential optical
and chemical sensing applications. Au nanoparticles containing
4
,5
1
tation and chromatographic fractionation. A size-dependent
opening of a HOMO-LUMO (the highest occupied and lowest
unoccupied molecular orbitals) energy gap, representing the
<
200 down to a few tens of atoms are of special interest because
they encompass the transition between bulk and molecular
regimes, where electronic band energetics yield to quantum
confinement effects, and discrete electronic states emerge. Two
key issues in the study of such very small nanoparticles are
property dependence on nanoparticle size and on the capping
or protecting ligands bonded to the nanoparticle surface. Size
and ligand dependence are interrelated, since the fraction of
atoms of the nanoparticle that are ligand-bonded increases as
size decreases. The quantum size confinement is manifested as
property changes that bring out molecule-like forms of behav-
3
c
bulk-to-molecule transition, has been reported based on
voltammetric and near-IR studies of alkanethiolate-coated
monolayer-protected metal clusters (MPCs) with Au core masses
6
of 8 to 38 kDa. Recently, Hutchison and co-workers prepared
alkanethiolate-protected Au11 by replacing the triphenylphos-
(
2) (a) Alivisatos, A. P. Science, 1996, 271, 933. (b) Brus, L. E. J. Phys. Chem.
1986, 90, 2555. (c) Heath, J. R. Science 1995, 270, 1315. (d) Murray, C.
B.; Kagan, C. R.; Bawendi, M. G. Science 1995, 270, 1335.
3) (a) Link, S.; Beeby, A.; FitzGerald, S.; El-Sayed, M. A.; Schaaff, T. G.;
Whetten, R. L. J. Phys. Chem. B 2002, 106, 3410-3415. (b) Schaaff, T.
G.; Knight, G.; Shafigullin, M. N.; Borkman, R. F.; Whetten, R. L. J. Phys.
Chem. B 1998, 102, 10643-10646. (c) Chen, S.; Ingram, R. S.; Hostetler,
M. J.; Pietron, J. J.; Murray, R. W.; Schaaff, T. G.; Khoury, J. T.; Alvarez,
M. M.; Whetten, R. L. Science 1998, 280, 2098-2101.
(
2,3
ior.
†
Present address: Department of Chemistry, Western Michigan Uni-
versity, Kalamazoo, MI 49008-5413.
(
4) (a) Schaaff, T. G.; Shafigullin, M. N.; Khoury, J. T.; Vezmar, I.; Whetten,
R. L.; Cullen, W. G.; First, P. N.; Guti e´ rrez-Wing, C.; Ascensio, J.; Jose-
Yacam a´ n, M. J. J. Phys. Chem. B 1997, 101, 7885. (b) Whetten, R. L.;
Khoury, J. T.; Alvarez, M. M.; Murthy, S.; Vezmar, I.; Wang, Z. L.;
Stephens, P. W.; Cleveland, C. L.; Luedtke, W. D.; Landman, U. AdV.
Mater. 1996, 8, 428-433.
‡
Present address: Steacie Institute for Molecular Sciences, National
Research Council of Canada, Ottawa, Ontario, K1A 0R6, Canada.
(
1) (a) Hayat, M. A., Ed. Colloidal Gold: Principles, Methods, and Applica-
tions; Academic Press: San Diego, CA, 1991. (b) Wood, A.; Giersig, M.;
Mulvaney, P. J. Phys. Chem. B 2001, 105, 8810-8815. (c) Sampaio, J. F.;
Beverly, K. C.; Heath, J. R. J. Phys. Chem. B 2001, 105, 8797-8800. (d)
Vance, F. W.; Lemon, B. I.; Hupp, J. T. J. Phys. Chem. B 1998, 102,
(5) (a) Jimenez, V. L.; Leopold, M. C.; Mazzitelli, C.; Jorgenson, J. W.; Murray,
R. W. Anal. Chem. 2003, 75, 199. (b) Wei, G.; Liu, F.; Wang, C. R. C.
Anal. Chem. 1999, 71, 2085-2091. (c) Wilcoxon, J. P.; Martin, J. E.;
Provencio, P. Langmuir 2000, 16, 9912-9920. (d) Schnabel, U.; Fischer,
C.; Kenndler, E. J. Microcolumn Sep. 1997, 9, 529-534. (e) Templeton,
A. C.; Cliffel, D. E.; Murray, R. W. J. Am. Chem. Soc. 1999, 121, 7081-
7089.
1
0091-10093. (e) Feldheim, D. L.; Keating, C. D. Chem. Soc. ReV. 1998,
2
7, 1-12. (f) Zanchet, D. M.; Micheel, C. M.; Parak, W. J.; Gerion, D.;
Alivisatos, A. P. NanoLett. 2001, 1, 32-35. (g) Nam, J.-M.; Thaxton, C.
S.; Mirkin, C. A. Science 2003, 301, 1884-1886. (h) Elghanian, R.;
Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; Mirkin, C. A. Science 1997,
2
77, 1078-1080. (i) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W.
(6) Woehrle, G. H.; Warner, M. G.; Hutchison, J. E. J. Phys. Chem. B. 2002,
106, 9979.
Acc. Chem. Res. 2000, 33, 27-36.
10.1021/ja049605b CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 6193-6199
9
6193

A R T I C L E S
Lee et al.
gold) was prepared using a literature procedure11 and stored in a freezer
at -20 °C. Water was purified with a Barnstead NANOpure system
phine ligands of a preformed Au11 core. In related work, Yang
and Chen reported that the Au11 nanoparticle’s HOMO-LUMO
7
(18 MΩ).
gap increases from 1.4 to 1.8 eV (from voltammetry and
absorbance spectra) upon replacement of the initial triphe-
nylphosphines with dodecanethiolate ligands. Quinn and co-
Synthesis of Au38(PhC
2
S)24. Au38(PhC
2
S)24 was synthesized as
9
described elsewhere. Briefly, in a standard two-phase Brust-style
synthesis,12 hydrogen chloroaurate (3.1 g, 11.1 mmol) was phase-
transferred into toluene with tetra-n-octylammonium bromide, followed
by addition of a 3-fold molar excess (relative to Au) of phenyl-
ethanethiol, forming a gold(I)-thiol polymer. This material was reduced
8
workers report a hexanethiolate-coated nanoparticle assigned
(without analytical evidence) as having a Au38 core and
exhibiting a 1.2 eV gap between the first voltammetric oxidation
and the first reduction.
by adding a 10-fold excess of aqueous NaBH at 0 °C, with vigorous
4
mixing, stirring the product solution at 0 °C for 24 h. After the aqueous
layer was removed, the toluene was rotary-evaporated at room
The molecule-like phenylethanethiolate-protected Au38 cluster
(
Au38 MPC) described here was isolated based on a strong
9
temperature. The Au38(PhC
with acetonitrile and was further purified from tetra-n-octylammonium
bromide by repeatedly dissolving the crude product in CH Cl and
2
S)24 was extracted from the crude product
solubility differentiation from larger nanoparticles. The focus
of this paper is its spectroscopic and electrochemical charging
properties; its synthesis and isolation and a thorough analytical
2
reprecipitating with ethanol until only traces of quaternary ammonium
9
confirmation have been presented earlier. Ligand-exchanged
1
cation could be detected ( H NMR). The product was characterized by
variants of this nanoparticle, Au38(PEG135S)13(PhC2S)11, Au38-
C6S)22(PhC2S)2, and Au38(C10S)19(PhC2S)5, where PEG135S-
1
H NMR, UV-vis spectra, and thermogravimetric analysis, as described
(
previously.
Ligand Place-Exchange Reactions. Thiolated poly(ethylene glycol)
(PEG135SH), decanethiol (C10SH), and hexanethiol (C SH) were
incorporated into the monolayer shell of Au38(PhC S)24 by ligand
exchange.1 PEG135SH was synthesized as described in the Supporting
Information. In a typical procedure, 45 mg of Au38(PhC S)24 in 5 mL
of CH Cl were stirred with an excess of ligand for 4 days, the solvent
9
is -SCH2CH2OCH2CH2OCH3, C6S- is hexanethiolate, and
C10S- is decanethiolate, respectively, are also described. The
electrochemical formal potentials for the first one-electron
reduction and one-electron oxidation of Au38(PhC2S)24 and its
ligand exchanged variants are separated by 1.65((0.03) V, with
minor dependencies on ligand, solvent, and temperature. The
large electrochemical potential spacing is consistent with a
molecule-like, discretized electronic energy level structure for
these Au38 MPCs. Correction of the electrochemical energy gap
for charging energy (also called10 “addition energy”) as
described later gives an HOMO-LUMO gap energy somewhat
above 1.3 eV, which corresponds very well to the observed ca.
6
2
i,13
2
2
2
was rotary-evaporated, and the product was rinsed several times with
heptane (for PEG135SH) or ethanol (for alkanethiols), until the R-thiol
1
proton peak of the free ligand was eliminated from the H NMR
spectrum. The relative proportions of ligands in the resulting mixed
MPC monolayers were assessed by decomposing the MPC with iodine
1
and analyzing the liberated disulfides by H NMR.
Electrochemistry. Voltammetry was done with a CHI 660A
Electrochemical workstation, in 0.1 M Bu NPF solutions that were
4 6
1
.33 eV optical absorption onset of Au38(PhC2S)24. The
degassed and blanketed with high-purity Ar atmosphere during the
experimental procedure. The working electrode was a 0.4 mm Pt disk,
the counter electrode was a Pt wire, and the reference electrode was
consistency of electrochemical with optical HOMO-LUMO gap
energies is supported by spectroelectrochemistry showing a
partial bleach of the low energy optical absorbance upon
oxidation of the Au38 MPC and depletion of the HOMO level.
+
either a Ag/Ag electrode or a Ag wire quasi-reference electrode,
AgQRE. The working electrode was polished with 0.05 µm Al
slurries and cleaned electrochemically by potential-cycling in 0.1 M
SO solution. Reduced temperature voltammetry was done using cold
2 3
O
A number of electrochemical oxidation and reduction steps
are observable below and above, respectively, the initial
oxidation and reduction reactions of Au38(PhC2S)24. These
reactions appear as doublet steps of single-electron changes that
in the accessible potential range in CH2Cl2 lead ultimately to
H
2
4
baths of acetone/dry ice (-70 °C) and acetonitrile/dry ice (-40 °C).
+
Potentials in the tables in this paper are reported vs Ag/Ag which
was calibrated as -0.61, -0.64, -0.61, and -0.59 V vs the ferrocene/
ferrocenium couple in CH Cl , 2:1, 1:1, and 1:2 toluene/acetonitrile,
2 2
4
+
2-
Au38(PhC2S)24 and to Au38(PhC2S)24 . The potential spacing
respectively (AgQRE). (Irrespective of calibration, differences between
the potentials of peaks within a given voltammogram are more reliable
than actual potentials observed in different voltammograms taken in
varied solvents or temperatures. Potential differences are reported in
the paper; actual potentials are found in the Supporting Information.)
Background potential scans in electrolyte solutions were used to check
for any spurious peaks.
+
1/0
between the doublet of peaks for the Au38(PhC2S)24
and
+
2/+1
Au38(PhC2S)24
reactions is used to estimate the charging
energy for this nanoparticle’s reaction.1
0
The ligand-exchanged Au38(PEG135S)13(PhC2S)11 nanoparticle
is also photoluminescent, at energies spanning the HOMO-
LUMO gap energy.
Spectroscopy. UV-vis spectra were collected with a Shimadzu
UV-vis (model UV-1601) spectrometer. Photoluminescence (PL)
spectra were taken in a 90° geometry on a modified ISA Fluorolog
FL321 spectrometer. The spectrometer is equipped with a 450 W xenon
source and Hamamatsu R928 MPT (visible wavelengths) and InGaAs
(near-IR wavelengths, connected via a T channel) detectors. Sample
solutions were freshly prepared before each measurement.
Experimental Section
2
Chemicals. 2-Phenylethanethiol (PhC SH, 99%), decanethiol (98%),
hexanethiol (98%), tetra-n-octylammonium bromide (98%), and sodium
borohydride (99%) were used as received from Aldrich, as were toluene
(
(
Fisher, reagent grade), acetonitrile (Fisher, Optima), methylene chloride
Fisher, reagent grade), and ethanol (Aarper Alcohol and Chemical
Spectroelectrochemistry. The quasi-thin layer, demountable spec-
troelectrochemical cell14 consisted of two glass slides separated by a
Company). Hydrogen tetrachloroaurate trihydrate (from 99.999% pure
(
11) Handbook of PreparatiVe Inorganic Chemistry; Brauer, G., Ed.; Academic
(
7) Yang, Y.; Chen, S. Nano Lett. 2003, 3, 75.
Press: New York, 1965; p 1054.
(
8) Quinn, B. M.; Liljeroth, P.; Ruiz, V.; Laaksonen, T.; Kontturi, K. J. Am.
Chem. Soc. 2003, 125, 6644.
(12) Brust, M.; Walker, M.; Bethell, D.; Schriffrin, D. J.; Whyman, R. J. Chem.
Soc., Chem. Commun. 1994, 801.
(13) (a) Hostetler, M. J.; Green, S. J.; Stokes, J. J.; Murray, R. W. J. Am. Chem.
Soc. 1996, 118, 4212. (b) Hostetler, M. J.; Templeton, A. C.; Murray, R.
W. Langmuir 1999, 15, 3782. (c) Song, Y.; Murray, R. W. J. Am. Chem.
Soc. 2002, 124, 7096.
(
9) Donkers, R. L.; Lee. D.; Murray, R. W. Langmuir 2004, 20, 1945-1952.
(
10) (a) Franceschetti, A.; Zunger, A. Phys. ReV. B 2000, 62, 2614. (b) Creutz,
C.; Brunschwig, B. S.; Sutin, N. ComprehensiVe Coordination Chemistry
II 2004, 7, 731-777.
6194 J. AM. CHEM. SOC.
9
VOL. 126, NO. 19, 2004

Molecule-like Au38 Nanoparticles
A R T I C L E S
a
Table 1. Formal Potential Spacings (V) for Au38(PhC2S)24 in
Nonaqueous Solvents Containing 0.1 M Bu4NClO4
solvent
T, °C ox4−ox3 ox3−ox2 ox2−ox1 ox1−re1 re1−re2 re2−re3 re3−re4
CH
2
Cl
2
25
0.73
0.64
0.65
0.64
0.72
0.63
0.67
0.64
0.29
0.44
0.32
0.25
0.20
0.38
0.30
0.22
1.62
1.68
1.65
1.60
1.67
1.67
1.68
1.62
2
1
1
:1 tol/MeCN 25
:1 tol/MeCN 25
:2 tol/MeCN 25
0.24
0.23
0.30
0.31
0.25
0.24
0.30
0.33
0.25
0.28
0.31
CH
2
Cl
2
-70 0.17
2
1
1
:1 tol/MeCN -40
:1 tol/MeCN -40
:2 tol/MeCN -40
0.24
0.29
0.27
0.33
0.32
0.25
a
Formal potentials are averages of reduction and oxidation peak
potentials in the DPV potential scans.
coatings. The electrochemical voltammetry of Au38(PhC2S)24
in Figure 1 is in contrast, qualitatively more like that of large
1
6
molecular species, such as the fullerenes.
The typical rest potential (Figure 1, arrow) of a solution of
0
the zerovalent (Au38(PhC2S)24 ) form of the nanoparticle is
typically about -0.1 V. The electrochemical formal potential
for the first one-electron oxidation of Au38(PhC2S)24 is only
slightly more positive than this. That the nanoparticle is readily
oxidized is consistent with a recent theoretical prediction17 for
Au38(SCH3)24 that filled levels lie just below the Fermi energy.
The first reduction step of Au38(PhC2S)24 lies, however, at a
rather negative potential that is separated from the potential of
the first oxidation by an appreciable energy gap of, in CH2Cl2
Figure 1. (a) 25 °C and (b) -70 °C differential pulse voltammograms
(DPVs) at 0.02 V/s, and (c) -70 °C cyclic voltammogram (0.1 V/s) of
Au38(PhC2S)24 in 0.1 M Bu4NPF6 in degassed CH2Cl2 at 0.4 mm diameter
Pt working, Ag wire quasireference (AgQRE), and Pt wire counter electrode.
Arrows indicate solution rest potentials, and * indicates wave for incom-
pletely removed O2, which varied from experiment to experiment.
(
Figure 1a), 1.62 V. This energy gap contains work terms
-
+
associated with charging the nanoparticle to Au38 and Au38 ,
the so-called charging or addition energy. This is estimated
1
0
0
.5 mm Teflon spacer, sealed by tightening screws, and enclosing Pt
using the separation between the Au38 and Au382+/1+ reaction
formal potentials, which from Table 1 (ox2 - ox1) is 0.29 V
in CH2Cl2. The difference gives a corrected energy of 1.33 eV,
which constitutes an electrochemical prediction of the HOMO-
LUMO gap energy.
+/0
mesh working and Pt wire counter electrodes, and a Ag wire AgQRE.
The working and counter electrodes lie side by side, with the beam
path masked so that only products of the working electrode are optically
monitored. The (air) empty cell provided the reference spectrum, which
was subtracted from the results with the filled cell. Initial cyclic
voltammograms at 5 mV/s aided determining the appropriate electrolysis
potentials, chosen to produce 0, +1, or +2 charge states of the Au38
nanoparticles in the degassed, CH Cl cell solution. The electrolysis
2 2
typically required 3 to 5 min for completion (as monitored by current),
whereupon a spectrum was taken.
1+/0
and Au382+/1+ reactions are generally well-
The Au38
behaved in terms of chemical reversibility but the other redox
reactions display voltammetric symptoms of product instability.
The chemical reversibility is improved at lowered temperatures
3
+/2+
(
Figure 1b,c) where the Au38
wave becomes chemically
4+/3+
Results and Discussion
reversible and the Au38
reaction becomes nearly so. At room
temperature in Figure 1a (and later in Figures 4 and 5), the
Voltammetry, Absorbance Spectra, and Spectroelectro-
chemistry of CH2Cl2 Solutions of Au38(PhC2S)24 Nanopar-
ticles. Figure 1 shows cyclic and differential pulse voltammetry
3
+/2+
4+/3+
Au38
/Au38
oxidation doublet is not resolved owing to
apparent instability of the more highly oxidized forms. The
reduction steps are not as well behaved even at lowered
temperature, although it is obvious that there are two reduction
steps present (and more are seen in other solvents). The first
reduction step, at ca. -1.5 V, is partly reversible, and the second
reduction, much less so.
(DPV) of CH2Cl2 solutions of Au38(PhC2S)24 nanoparticles. The
current peaks in the DPV scans (Figure 1a) lie at the formal
potentials of the nanoparticle charge state couples. We see that
the formal potentials are unevenly spaced, in contrast to DPV
15
voltammograms of hexanethiolate-coated Au140 nanoparticles
that display roughly evenly spaced, serial one-electron changes
The pattern of electrochemical peaks in Figure 1 is two
separated doublets of oxidation steps and one of reduction (a
second reduction doublet is seen in Figure 5). The doublets are
not likely to be caused by the presence of isomers (as found
7
+
leading ultimately to core charge states as oxidized as Au140
and as reduced as Au1406-. The latter phenomenon is primarily
1
5
a quantized double layer (QDL) charging of the MPC cores
1
6
for fullerenes like C78), given the monodispersity evidence
that is observable at room temperature because of the sub-aF
capacitances (CCLU) of MPCs bearing low dielectric monolayer
1
8
from an HPLC experiment. The currents (*) between 0 and
1.5 V vary from experiment to experiment and are attributed
-
to residual oxygen. The potential spacing between current peaks
(
14) (a) Heineman, W. R.; O’Dom, G. W.; Murray, R. W. Anal. Chem. 1967,
3
9, 1666. (b) Anderson, C. W.; Halsall, H. B.; Heineman, W. R. Anal.
Biochem. 1979, 93, 366.
(16) (a) Yang, Y.; Arias, F.; Echegoyen, L.; Chibante, L. P. F.; Flanagan, S.;
Robertson, A.; Wilson, L. J. J. Am. Chem. Soc. 1995, 117, 7801. (b)
Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res. 1998, 31, 593.
(17) H a¨ kkinen, H.; Barnett, R. N.; Landman, U. Phys. ReV. Lett. 1999, 82, 3264-
3267.
(18) Lee, D.; Donkers, R. L.; DeSimone, J. M.; Murray, R. W. J. Am. Chem.
Soc. 2003, 125, 1182.
(
15) (a) The roughly even spacing of one-electron changes for Au140 on the
potential axis implies that CCLU, and the charging energy, is approximately
independent of potential. (b) Ingram, R. S.; Hostetler, M. J.; Pietron, J. J.;
Murray, R. W.; Schaaff, T. G.; Khoury, J.; Whetten, R. L.; Bigioni, T. P.;
Guthrie, D. K.; First, P. N. J. Am. Chem. Soc. 1997, 119, 9279. (c) Chen,
S.; Murray, R. W.; Feldberg, S. W. J. Phys. Chem. B. 1998, 102, 9898.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 19, 2004 6195

A R T I C L E S
Lee et al.
Figure 3. Schematic model energy level diagram for Au38(PhC2S)24, based
1-
on data taken in CH2Cl2 solution (except for reduction beyond Au38 where
data are from 2:1 toluene/acetonitrile).
of a continuum of electronic states. The energy separation
between the lower energy occupied level and the LUMO, 1.77
eV, is near that of the low energy absorbance peak in Figure 2.
Finally, the electrochemical changes shown above the LUMO,
which have been augmented with data taken in toluene/
acetonitrile solution (Table 1), are all roughly equally spaced,
by about the same energy as the estimated charging energy.
Consequently we cannot identify higher energy orbitals other
than the one shown.
Figure 2. UV-vis spectra (25 °C) of (a) Au38(PhC2S)24 (black line), (b)
1+
2+
Au38(PhC2S)24 (red line), and (c) Au38(PhC2S)24 (green line) in degassed
CH2Cl2 solution. The three spectra are of the same solution; the 1+ and
2
+ charge states were generated by electrolysis in the spectroelectrochemical
cell. These data are shown in a wavelength axis plot in Figure S-2.
for Figure 1 and other experiments discussed later are listed in
Table 1; the actual potentials are given in Table S-1.
The absorbance spectrum (black line) of a CH2Cl2 solution
The reversible bleaching effect of electrochemical charging
in spectra of films of semiconductor quantum dots has been
1
8
of Au38(PhC2S)24 displays (Figure 2), as described before,
a
steplike structure that is typical for gold MPCs with very small
2
1
recently reported. For the Au38 “quantum dots”, if the low
energy spectral absorbance (Figure 2) and the electrochemical
results (Figure 1) indeed reflect a common energy gap electronic
structure, then electrochemically depopulating the HOMO level
should attenuate the optical absorbance. The results in Figure
3
core dimensions. It is instructive to consider the low energy
part of this spectrum. The absorbance around 700 nm appears
(
from fitting to a sum of Gaussians,19 Figure S-1) to be two
overlapping peaks, at 675 nm (1.84 eV) and 770 nm (1.61 eV).
Following the absorbance spectrum into the near-infrared, the
absorbance appears to be extinguished as an absorbance edge
at 1.33 eV, as shown in the Figure 2 inset. This optical band
gap energy is identical to the preceding electrochemical estimate
of the HOMO-LUMO gap energy for the Au38(PhC2S)24
nanoparticle.
Based on the observations in Figures 1 and 2, we propose,
as a rough working model, that the Au38(PhC2S)24 electronic
structure contains three well-defined molecular orbitals, two
2
were obtained by electrochemical oxidation of Au38 nano-
14
particle solutions in an optically transparent thin layer cell.
The Figure 2 inset of the band edge region shows that the
0
spectrum (black line) of Au38(PhC2S)24 MPCs changes sub-
1+
2+
stantially upon oxidation to the Au38 (red) and Au38 (green)
states. There is a loss in absorbance intensity near the absorbance
band edge with an apparent shift in the edge toward higher
energy. There are other changes at higher energies in the
spectrum that we do not attempt to interpret at this time.
Importantly, the zero charge state spectrum (Figure 2, inset) is
quantitatively restored (90 to 98%, data not shown) upon
(electron pair occupied) levels below the Fermi level and an
empty one above. Each orbital spawns a serial pair of
electrochemical reaction steps as electron donors and acceptors.
This preliminary orbital diagram model, shown in Figure 3,
while very simple, is not unique; Banin et al.,20 for example,
described an analogous state-occupancy diagram based on STS
observations of InAs quantum dots.
Besides portraying the HOMO-LUMO gap, Figure 3 also
shows a second occupied orbital 0.44 eV below the HOMO.
This orbital is predicted by the second oxidation doublet seen
in the low-temperature voltammetry in Figure 1b,c. The 0.44
eV energy spacing between the first and second oxidation
doublets is based on applying a 0.29 V charging energy
correction to the 0.73 V spacing between the two doublets (Table
1
+
2+
0
reduction of the Au38 and Au38 states back to Au38 . This
ensures that the observed optical bleach is due to core-charging,
not core degradation, and re-enforces the rationality of the model
of Figure 3.
We must at this point remark on the obvious difference
between the HOMO-LUMO gap energy (>1.3 eV) determined
here for Au38(PhC2S)24 and the 0.9 eV value previously reported
3c
by us and Whetten for alkanethiolate-coated 8 kDa MPCs and
8
by Quinn. The present Au38(PhC2S)24 and previous 8 kDa
samples are both well-characterized as to core size, at least as
to be distinguishable from adjacent stable core dimensions (28,
55 atoms). It is possible that the difference in protecting ligands
1
, Figure 1). The appearance of the second distinct electronic
is a factor, although further investigation has not to date
supported this thesis. Adsorbed bromide ligands on the level
level below the HOMO signals that the latter is indeed a
condensed and discrete electronic level and not merely the edge
-
of one Br per MPC core or less cannot be cleanly ruled out,
and this possible impurity in some of the samples we prepare
(
2
19) (a) The UV-vis absorption spectrum of Au38(PhC S)24 is fitted to a sum
is another concern. The energy gap observed here matches that3b
of five Gaussians in Figure S-1 (Supporting Information). The fifth Gaussian
at the highest energy (dashed line) is to take the UV rise in the absorption
feature into account. (b) Shim, M.; Wang, C.; Guyot-Sionnest, P. J. Phys.
Chem. B 2001, 105, 2369.
(21) (a) Shim, M.; Guyot-Sionnest, P. Nature 2000, 407, 981. (b) Wang, C.;
Shim, M.; Guyot-Sionnest, P. Science 2001, 291, 2390.
(
20) Banin, U.; Cao, U.; Katz, D.; Millo, O. Nature, 1999, 400, 542.
6196 J. AM. CHEM. SOC.
9
VOL. 126, NO. 19, 2004

Molecule-like Au38 Nanoparticles
A R T I C L E S
a
Table 2. Formal Potential Spacings (V) for Ligand Exchanged
Au38 MPCs in CH2Cl2 Containing 0.1 M TBAP
MPC
T (°C)
ox3−ox2
ox2−ox1
ox1−re1
Au38(PEG135S)13(PhC2S)11
Au38(PEG135S)13(PhC2S)11
Au38(C6S)22(PhC2S)2
Au38(C10S)19(PhC2S)5
Au38(C10S)19(PhC2S)5
25
-70
25
25
-70
0.68
0.68
0.80
0.80
0.81
0.26
0.19
0.34
0.34
0.26
1.65
1.69
1.68
a
Formal potentials are averages of reduction and oxidation peak
potentials in the DPV potential scans.
The Au38(C6S)22(PhC2S)2 solution exhibited a rather positive
electrochemical rest potential, indicating that the nanoparticle
was in an oxidized state, which could account for the bleaching
around 700 nm. Generally, Au38 cores with alkanethiolate
ligands seem less stable than those with PhC2S ligands.
Solvent Effects on Au38 MPC Properties. When a large
molecule (or a molecule-like nanoparticle) is electrochemically
oxidized or reduced, the observed electrochemical potentials are
affected by the relative degrees of solvent dipole and electrolyte
ion pairing stabilization of the electron donor and acceptor states.
This effect is collectively called the “charging energy” and has
been examined by many workers, including experiments on C60
Figure 4. Differential pulse voltammograms (DPVs) at 25 °C and 20 mV/s
of (a) Au38(PEG135S)13(PhC2S)11, (b) Au38(C6S)22(PhC2S)2, and (c) Au38-
(C10S)19(PhC2S)5 in 0.1 M Bu4NPF6 in CH2Cl2 at 0.4 mm diameter Pt
22
in different solvents and electrolytes by Kadish et al. and
working, AgQRE, and Pt wire counter electrode.
1
0
calculations on 3.5 nm semiconductor nanoparticles bearing
coatings of different dielectric constant but otherwise in a
of a glutathione-stabilized Au28 MPC, but we have no basis to
question the analytical evidence supporting that formulation.
We have no clear explanation for the difference between the
presently and previously reported gap energies at this time.
Ligand-Exchanged Au38 MPCs. The Au38 core readily
undergoes thiolate ligand place-exchange reactions, as shown
previously in a study18 in which a pegylated nanoparticle melt,
Au38(PEG365S)19(PhC2S)5 (PEG365 ) a MW 365 thiolated poly-
2
3
24
vacuum. Evans et al. and Weaver et al. have discussed, in
different ways, the substantial effects of electrolyte/solvent
media on charging energies attendant to reductions of organic
molecules and C60, respectively. Charging energy normally
enhances the energy difference between electrochemical oxida-
tion and reduction potentials, relative to the optical band gap
energy.
(
ethylene glycol)), was prepared from the Au38(PhC2S)24 MPC.
The effect of solvent polarity on the HOMO-LUMO gap
and charging energies was examined by using 2:1 to 1:2 (v:v)
mixtures of toluene and acetonitrile solvents (ꢀS ) 2.4 and 37,
respectively). Voltammetry results at 25° and -40 °C for Au38-
(PhC2S)24 MPCs (Figure 5, Table 1) show that the differences
between formal potentials of the first oxidation and reduction
steps, at both temperatures, decrease modestly (0.05 to 0.08 V)
as the solvent mix becomes more polar. Table S-1 shows that
We have gone on to prepare other Au38 nanoparticles in a similar
manner, as detailed elsewhere,9 and describe here the elec-
trochemical and spectral properties of the exchange products
Au38(PEG135S)13(PhC2S)11, Au38(C6S)22(PhC2S)2, and Au38-
,18
(C10S)19(PhC2S)5.
DPV voltammetry of these mixed monolayer Au38 MPCs in
CH2Cl2 at 25 °C, shown in Figure 4, is qualitatively similar to
that of Au38(PhC2S)24 in Figure 1. Au38(PEG135S)13(PhC2S)11
and Au38(C6S)22(PhC2S)2 show a doublet of oxidations (Figure
1
+/0
0/1-
while both Au38
and Au38
formal potentials shift posi-
tively in a more polar solvent environment, the latter shifts more;
i.e., the reduced form of Au38(PhC2S)24 is slightly stabilized
relative to its oxidized form. The larger sensitivity of the Au380
reduction potential to the change in solvent may not of course
be simply due to solvent polarity but may include other effects,
4
a,b), a more positive oxidation step, and, like Figure 1, a first
/1-
reduction widely separated from the first oxidation. These
features appear, but less distinctly, in the DPV of Au38(C10S)19-
(PhC2S)5 (Figure 4c), especially the reduction step which is only
-
a weak shoulder at ca. -1.7 V. (The broad reoxidation peak at
ca. -1.2 V is possibly due to degraded reduced product of this
nanoparticle.) Data for the differences between the formal
potentials of these reactions are given in Table 2 and for the
actual potentials in Table S-2.
for example, a stabilizing solvation of Au38 by the hydrophobic
electrolyte cation.
The difference between the Au38+ (ox1) and Au382+/+ (ox2)
formal potentials was used above to estimate the charging energy
contribution to the electrochemical energy gap. Table 1 shows
that in 2:1 and 1:2 toluene/CH3CN mixtures, this energy varies
from 0.44 to 0.25 V and from 0.38 to 0.22 V at 25 and -40
/0
The electronic spectrum of the exchanged Au38(PEG135S)13-
(PhC2S)11 MPC (Figure S-3) is, like the previously reported
18
0/1-
result for Au38(PEG365S)19(PhC2S)5, essentially identical to that
of the parent Au38(PhC2S)24 nanoparticle. This suggests that the
Au38 core is substantially preserved during the PEG-thiolate
exchange reactions. The spectral definition of the Au38(C6S)22-
°C, respectively. The analogous spacing between the Au38
1
-/2-
(re1) and Au38
(re2) formal potentials varies less, but in
(
22) Dubois, D.; Moninot, G.; Kutner, W.; Jones, M. T.; Kadish, K. M. J. Phys.
Chem. 1992, 96, 7137-7145
(PhC2S)2 sample was on the other hand degraded, showing a
(23) Evans, D. H.; Hu, K. J. Chem. Soc., Faraday Trans. 1996, 92, 3983-
3990.
less detailed steplike structure (as if it were less monodisperse)
and a greatly diminished absorption intensity around 700 nm.
(24) Roth, J. D.; Lewis, G. J.; Safford, L. K.; Jiang, X.; Dahl, L. F.; Weaver,
M. J. J. Am. Chem. Soc. 1992, 114, 6159.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 19, 2004 6197

A R T I C L E S
Lee et al.
potentials of the Au38+ (ox1) and Au382+/+ (ox2) couples of
the Au38(PhC2S)24 nanoparticles. In the case of alkanethiolate-
protected Au140 MPCs, ∆V was roughly independent of the
/0
25
solvent medium. For the Au38 MPCs, however, Table 1 shows
that the solvent appears to make a difference. If the solvent
dependency is expressed in terms of effectiVe values of ꢀ of the
Au38(PhC2S)24 MPC’s monolayer, the Table 1 ox2-ox1 results
would be satisfied by eq 1 using r ) 0.55 and d ) 0.682 by
taking ꢀ ) 3.3, 4.5, and 5.8 in the 2:1, 1:1, and 1:2 mixed
toluene/acetonitrile solvents, respectively. When it is considered
that toluene is probably strongly favored in the Au38(PhC2S)24
solvation shell, these results are quite reasonable and suggest
that the effective capacitance of the dielectric surrounding the
Au38 core must be a combination of that of the ligand and of
solvent that to some degree penetrates the ligand shell. The
important conclusion is that the phenylethylthiolate monolayer
ligand shell of the Au38 nanoparticle is somewhat open to
intrusion by other species, which is geometrically unsurprising
considering the sharp curvature of the Au38 surface.
5,26
The UV-vis spectra of Au38(PhC2S)24 solutions on the other
hand reveal no obvious solvent effects (Figure S-4). The spectra
are essentially identical in CH2Cl2, toluene, and toluene/
acetonitrile mixtures. Either the above-mentioned solvent intru-
sion is insufficient to provoke a measurable effect or the excited
state nanoparticle is relatively nonpolar with a consequent weak
dependence on the dielectric environment.
Figure 5. Differential pulse voltammograms (DPVs) at -40 °C and 20
mV/s of Au38(PhC2S)24 in degassed 0.1 M Bu4NPF6 in (a) 2:1, (b) 1:1, and
(c) 1:2 toluene/acetonitrile at 0.4 mm diam Pt working, AgQRE, and Pt
wire counter electrode. The currents at ∼ -0.8 V are due to residual oxygen
that was not completely removed by the degassing.
the opposite direction. (Table S-1 shows the actual potentials
and that the different formal potentials have decidedly different
solvent sensitivities.) If one estimates charging energy from the
difference between the Au38+/0 (ox1) and Au382+/+ (ox2) formal
potentials, over the range of solvents the electrochemically
determined HOMO-LUMO gap energy varies from 1.24 to 1.35
eV; use of the analogous reductions to estimate charging energy
gives gap energies of 1.44 to 1.30 eV. The message here is that
our estimate of charging energy is quite rough but that most
data indicate it as about 0.3 eV, and the corresponding
electrochemical gap energy is about 1.3 eV as shown in Figure
Photoluminescence of Au38 MPCs. The luminescence
spectra of Au38 MPCs raise significant new issues in under-
standing these molecule-like nanoparticles. Au38(PhC2S)24 nano-
particles luminescence only weakly, but that of ligand-
exchanged versions is more intense and is illustrated by the
Figure 6a emission spectra for Au38(PEG135S)13(PhC2S)11.
Spectra excited at 400 and 685 nm both show broad emission
peaks around 1000 nm and can be fitted by two Gaussian-shaped
emissions centered at 902 nm (1.38 eV) and 1025 nm (1.2 eV),
shown as dashed lines. The excitation spectrum (inset, Figure
6a) shows that the emission intensity roughly follows the profile
of absorbance (Figure 2), with the emission peak at 685 nm
(1.8 eV) lying near the absorbance maximum at 670 nm. Figure
6b compares emission spectra taken at several different low-
energy wavelength excitations: 600, 650, 680, and 750 nm. The
emission differs substantially, being very weak when excited
at 650 and 750 nm, being comparatively strong in both the 1.38
and 1.2 eV bands when excited at 680 nm, and favoring the
higher energy component at 1.38 eV when excited at 600 nm.
The sharply selective excitation behavior shows that emission
from the higher energy channel (1.38 eV) is favored by higher
energy excitation. It is also another indication of strongly
condensed electronic energy levels.
3
. The second message is that, as exemplified by studies by
2
2
Kadish et al., multiple factors come into play in stabilization
of differently charged molecule-like species in organic solvents,
and an extremely detailed study is required to unravel the main
ones.
We have previously25 employed a concentric sphere capacitor
model to estimate potential differences for quantized double
layer charging of nanoparticles such as Au140 alkanethiolates.
Single-electron changes in the capacitor charge occur at formal
potentials spaced (∆V) by
∆
V ) e/CCLU ) ed/4πꢀꢀ r(r + d)
(1)
0
where e is electron charge, ꢀ0, the permittivity of free space, ꢀ,
the static dielectric constant of the monolayer medium around
the metal core, r, the radius of the gold core, and d, the thickness
of the monolayer medium. Smaller peak spacings are expected
when the dielectric constant of the medium surrounding the
metal core is higher. The model ignores the diffuse double layer.
Equation 1 is, in effect, a simple calculation of charging
energy for nanoparticles, so it is worth comparing it to the Au38
data in Table 1, in particular to the differences in formal
The 1.38 eV emission from Au38(PEG135S)13(PhC2S)11 nano-
particles matches very well the HOMO-LUMO gap energy
determined above and can be assigned as relaxed luminescence
across the HOMO-LUMO gap. On the other hand, the 1.2 eV
band, which extends to energies as low as 0.89 eV (Figure 3),
is obviously a sub-bandgap energy luminescence. The sub-gap
3a
emission energy is reminiscent of observations of emission
energies (1.55 and 1.15 eV) from a glutathione-protected Au28
MPC, where the latter is lower than the optical absorbance edge
energy (1.3 eV). The sub-gap emission was, in a solid-state-
(
25) Hicks, J. F.; Templeton, A. C.; Chen, S.; Sheran, K. M.; Jasti, R.; Murray,
R. W.; Debord, J.; Schaaff, T. G.; Whetten, R. L. Anal. Chem. 1999, 71,
3
703.
(26) Wuelfing, W. P.; Murray, R. W. J. Phys. Chem. B 2002, 106, 3139.
6198 J. AM. CHEM. SOC.
9
VOL. 126, NO. 19, 2004

Molecule-like Au38 Nanoparticles
A R T I C L E S
Figure 6. (a) Emission and excitation (inset) spectra of dilute Au38(PEG135S)13(PhC2S)11 at 25 °C in CH2Cl2 (UV absorbance is 0.34 at 300 nm). Excitation
and emission spectra were collected using λEM ) 1050 nm and λEX ) 400 and 680 nm, respectively. Dashed lines are Gaussian curvefits. Asterisks indicate
artifacts from second-order excitation peak (800 nm) and solvent absorption (1165 nm). (b) Emission at indicated excitation wavelengths, λEX.
like analysis, ascribed to an intra-band transition and, in a
molecular view, to emission from a possible triplet state. While
that analysis is appealing, we have further observed,2 and will
report elsewhere that the near-IR emissions of Figure 6 can also
be seen for a variety of Au MPCs of differing core dimensions
and that they are very dependent on the ligands present. The
overall picture is a complex one and probably must include
dipole excitations that are associated with defect Au-ligand sites
on the nanoparticle core surface.
Acknowledgment. This research was supported in part by
the STC Program of the National Science Foundation under
Agreement No. CHE-9876674 and in part by grants from the
U.S. Department of Energy, Division of Basic Sciences and the
National Science Foundation. The authors gratefully acknowl-
edge helpful discussions of this work with Dr. Stephen W.
Feldberg, and R.L.D. thanks the Natural Sciences and Engineer-
ing Research Council of Canada for a postdoctoral fellowship.
7
Finally, we note the report by Yam et al.3a,28 on the
luminescence of Au(I) sulfido complexes. Recognizing that Au-
Supporting Information Available: Synthesis of thiolated
poly(ethylene glycol) (PEG135SH) and DPV formal potentials
for Au38(PhC2S)24 in various solvents (Table S-1). This material
is available free of charge via the Internet at http://pubs.acs.org.
1
3c,29
(I) thiolates are present in MPC solutions exposed to air,
we examined solutions of Au(I) ligated to thiolates used here
but observed no emission in the visible or near-infrared.
JA049605B
(
27) Wang, G.; Murray, R. W. Unpublished results, University of North
Carolina-Chapel Hill, 2003.
(28) (a) Yam, V. W.-W.; Cheng, E. C.-C.; Zhou, Z.-Y. Angew. Chem., Int. Ed.
2
000, 39, 1683. (b) Yam, V. W.-W.; Cheng, E. C.-C.; Cheung, K.-K.
(29) Song, Y.; Huang, T.; Murray, R. W. J. Am. Chem. Soc. 2003, 125, 11694-
Angew. Chem., Int. Ed. 1999, 38, 197.
11701.
J. AM. CHEM. SOC.
9
VOL. 126, NO. 19, 2004 6199