Anion-induced adsorption of ferrocenated nanoparticles

Article abstract of DOI:10.1021/ja074161f

Au nanoparticles fully coated with ω-ferrocenyl hexanethiolate ligands, with average composition Au₂₂₅(ω-ferrocenyl hexanethiolate)₄₃, exhibit a unique combination of adsorption properties on Pt electrodes. The adsorbed layer is so robust that electrodes bearing submonolayer, monolayer, and multilayer quantities of these nanoparticles can be transferred to fresh electrolyte solutions and there exhibit stable ferrocene voltammetry over long periods of time. The kinetics of forming the robustly adsorbed layer are slow; monolayer and submonolayer deposition can be described by a rate law that is first order in nanoparticle concentration and in available electrode surface. The adsorption mechanism is proposed to involve entropically enhanced (multiple) ion-pair bridges between oxidized (ferrocenium) sites and certain specifically adsorbed electrolyte anions on the electrode. Adsorption is promoted by scanning to positive potentials (through the ferrocene wave) and by high concentrations of Bu ₄N⁺X^- electrolyte (X^- = ClO ₄^-, PF₆^-) in the CH ₂Cl₂ solvent; there is no adsorption if X^- = p-toluenesulfonate or if the electrode is coated with an alkanethiolate monolayer. The electrode double layer capacity is not appreciably diminished by the adsorbed ferrocenated nanoparticles, which are gradually desorbed by scanning to potentials more negative than the electrode's potential of zero charge. At very slow scan rates, voltammetric current peaks are symmetrical and nearly reversible, but exhibit E_fwhm considerably narrower (typically 35 mV) than ideally expected (90.6 mV, at 298 K) for a one-electron transfer or for reactions of multiple, independent redox centers with identical formal potentials. The peak narrowing is qualitatively explicable by a surface-activity effect invoking large, attractive lateral interactions between nanoparticles and, or alternatively, by a model in which ferrocene sites react serially at formal potentials that become successively altered as ion-pair bridges are formed. At faster scan rates, both ΔE_peak and E_fwhm increase in a manner consistent with a combination of uncompensated ohmic resistance of the electrolyte solution and of the adsorbed film, as distinct from behavior produced by slow electron transfer.

Full text of DOI:10.1021/ja074161f

Published on Web 01/17/2008
Anion-Induced Adsorption of Ferrocenated Nanoparticles
Rebecca L. Stiles,† Ramjee Balasubramanian,† Stephen W. Feldberg,‡ and
Royce W. Murray*,†
Kenan Laboratories of Chemistry, UniVersity of North Carolina, Chapel Hill, North Carolina
27599, and Chemistry Department, BrookhaVen National Laboratory, Upton, New York 11973
Received June 7, 2007; E-mail: rwm@email.unc.edu
Abstract: Au nanoparticles fully coated with ω-ferrocenyl hexanethiolate ligands, with average composition
Au225(ω-ferrocenyl hexanethiolate)43, exhibit a unique combination of adsorption properties on Pt electrodes.
The adsorbed layer is so robust that electrodes bearing submonolayer, monolayer, and multilayer quantities
of these nanoparticles can be transferred to fresh electrolyte solutions and there exhibit stable ferrocene
voltammetry over long periods of time. The kinetics of forming the robustly adsorbed layer are slow;
monolayer and submonolayer deposition can be described by a rate law that is first order in nanoparticle
concentration and in available electrode surface. The adsorption mechanism is proposed to involve
entropically enhanced (multiple) ion-pair bridges between oxidized (ferrocenium) sites and certain specifically
adsorbed electrolyte anions on the electrode. Adsorption is promoted by scanning to positive potentials
(through the ferrocene wave) and by high concentrations of Bu4N+X- electrolyte (X- ) ClO4-, PF6-) in the
CH2Cl2 solvent; there is no adsorption if X- ) p-toluenesulfonate or if the electrode is coated with an
alkanethiolate monolayer. The electrode double layer capacity is not appreciably diminished by the adsorbed
ferrocenated nanoparticles, which are gradually desorbed by scanning to potentials more negative than
the electrode’s potential of zero charge. At very slow scan rates, voltammetric current peaks are symmetrical
and nearly reversible, but exhibit Efwhm considerably narrower (typically 35 mV) than ideally expected (90.6
mV, at 298 K) for a one-electron transfer or for reactions of multiple, independent redox centers with identical
formal potentials. The peak narrowing is qualitatively explicable by a surface-activity effect invoking large,
attractive lateral interactions between nanoparticles and, or alternatively, by a model in which ferrocene
sites react serially at formal potentials that become successively altered as ion-pair bridges are formed. At
faster scan rates, both ∆Epeak and Efwhm increase in a manner consistent with a combination of
uncompensated ohmic resistance of the electrolyte solution and of the adsorbed film, as distinct from
behavior produced by slow electron transfer.
Introduction
electrode to a nanoparticle-free CH2Cl2 electrolyte solution, well-
defined stable voltammetry of the ferrocene ligands (ca. 43 per
Adsorption of ions and molecules from solutions onto
electrodes and electrified interfaces is a common chemical
phenomenon. During a study1 of the electrochemistry of Au
nanoparticles capped with monolayers of solely ferrocene
alkanethiolate ligands (monolayer protected clusters, MPCs),
we noticed an uncommon resultsa nearly irreversible adsorption
of a ca. monolayer of ω-ferrocenyl hexanethiolate-functionalized
MPCs from CH2Cl2 solutions onto Pt electrodes. Further study
of this adsorption has been carried out and the present report
will show that the adsorption of the MPC (Au225[(S(CH2)6Fc)43])
is a slow, cooperative process, requiring the presence of high
concentrations of particular electrolytes.
While studying the adsorption of the ferrocenated MPCs,
several unusual properties of the electroactive film were
discovered. First, the monolayer is robust, persisting after
rinsing, soaking, or sonication of the electrode in solvent (with
or without dissolved electrolyte). After transferring the filmed
nanoparticle) can be observed. The full-width-half-maximum
(Efwhm) of the voltammetric peak (for monolayer and submono-
layer coverage, ca. 35 mV) is considerably narrower than the
value expected (90.6 mV at 298 K) for an electroactive
monolayer. Analysis of anodic/cathodic peak potential separa-
tions (∆Epeak) reveals an uncompensated ohmic resistance of
the adsorbed MPC layer, while the electrode’s double layer
capacitance remains virtually unchanged. The adsorbed layer
is slowly removed by applying potentials estimated to be more
negative than the electrode’s potential of zero charge, and the
adsorption is nearly completely blocked by the prior formation
of a dodecanethiolate self-assembled monolayer. This paper will
analyze these collectively unusual effects.
The charge under the peaks of a cyclic voltammogram of an
electroactive film on an electrode, that is, a chemically modified
electrode,2-9 gives the total surface coverage, ΓT (mol/cm2).
(2) Chidsey, C. E. D.; Murray, R. W. Science 1986, 231, 25-31.
(3) Murray, R. W.; Ewing, A. G.; Durst, R. A. Anal. Chem. 1987, 59, A379.
(4) Wrighton, M. S. Science 1986, 231, 32-37.
† University of North Carolina.
‡ Brookhaven National Laboratory.
(1) Wolfe, R. L.; Balasubramanian, R.; Tracy, J. B.; Murray, R. W. Langmuir
2007, 23, 2247-2254.
(5) Faulkner, L. R. Chem. Eng. News 1984, 62, 28.
(6) Murray, R. W. Electroanal. Chem. 1984, 13, 191-368.
9
1856
J. AM. CHEM. SOC. 2008, 130, 1856-1865
10.1021/ja074161f CCC: $40.75 © 2008 American Chemical Society

Adsorption of Ferrocenated Nanoparticles
A R T I C L E S
Many adsorbed species display near-ideal behavior, but reactions
with slow monolayer electron-transfer kinetics are also
known,10-12 as are unusual voltammetric waveshapes attributed
to nonideal surface activities and lateral interactions,13-15
environmental or dipolar heterogeneity of redox sites,16-18 ion-
pairing between redox species and electrolyte counterions,19,20
and interfacial potential distribution.21 In the particular case of
ferrocene, there are examples of adsorption from solutions that
are either concentrated22,23 or involve a poorly solvating medium
(aqueous),24 adsorption of ferrocenated dendrimers,25 precipita-
tion of poly(vinylferrocenium),15,26-29 electrodeposition of bi-
ferrocene derivative-attached gold nanoparticles,30-35 and bind-
ing through alkanethiolate16 or siloxane linkages.16,36 Among
all these reports, however, there is none of a slow, electrolyte-
induced, irreversible adsorption of monolayer quantities of a
ferrocenated species that yield voltammetric surface waves more
narrow than the ideal Efwhm ) 90.6 mV.
nanoparticles, the currents associated with the quantized charg-
ing are dwarfed by the much larger and more quantitatively
definable currents associated with the oxidation or reduction
of the multiple (ca. 43) ferrocene sites in the protecting
monolayer surrounding each Au225 core. We focus on the latter
currents in this study of ferrocenated MPC adsorption, which
has also lead us to a further study48 conducted in light of the
implied possibilities of pseudocapacitive charge (energy) storage.
Indeed, we find that analogous adsorption occurs on glassy and
mesoporous carbon electrode materials. Additionally, provided
the ion-induced adsorption mechanism can be extended to other
nanoparticles (redox coated or not), it may be useful for their
immobilization and formation of modified surfaces.
The ferrocenated nanoparticles referred to in the paper are
fully (solely) capped with ω-ferrocenyl hexanethiolate ligands
and are abbreviated Au225(SC6Fc)43. In one instance, these
ligands are diluted with hexanethiolate ligands.
The Au MPCs used in these experiments have an average
Au225(ω-ferrocenyl hexanethiolate)43 composition and are one
of a series1 of MPCs having average 75, 140, 225 and 314 Au
atom core sizes and ca. 37, 39, 43 and 48 ferrocenyl ligands,
respectively. Their non-ferrocenated analogues display size-
dependent voltammetry with quantized one-electron double layer
charging37-47 for the three larger cores and a molecule-like
energy gap43 for the Au75 case. However, for the ferrocenated
Experimental
Chemicals. Hexanethiol (HSC6, >99%), dodecanethiol (HSC12,
>99%), t-octylammonium bromide (Oct4NBr, >98%), sodium boro-
hydride (NaBH4, >98%), t-butylammonium perchlorate (Bu4NClO4,
>99%), t-butylammonium p-toluenesulfonate (Bu4NC7H7SO3, puress),
and t-butylammonium hexafluorophosphate (Bu4NPF6, puress) from
Aldrich and toluene (reagent grade), acetonitrile (Optima), methylene
chloride (HPLC grade), tetrahydrofuran (HPLC grade), and ethanol
(HPLC grade) from Fisher were used as received. HAuCl4‚xH2O (from
99.999% pure gold) was synthesized using a literature procedure49 and
stored in a freezer at -20 °C. Water was purified using a Barnstead
NANOpure system (18 MΩ).
Ferrocene hexanethiol (HSC6Fc) was synthesized by refluxing a
mixture of (1.11 g, 3.17 mmol) ω-bromohexane ferrocene (prepared
by a published method50) and thiourea (0.600 g, 7.88 mmol) in ethanol
(50 mL) overnight. The reaction mixture was neutralized with NaOH
(aq), refluxed for a further 3 h, and then acidified with HCl to pH ≈
2, diluted with water, and extracted with CH2Cl2, washing the organic
extract phase copiously with water. The material obtained after rotary
evaporation of the CH2Cl2 product solution was chromatographed on
(7) Murray, R. W. Acc. Chem. Res. 1980, 13, 135-141.
(8) Bard, A. J. J. Chem. Educ. 1983, 60, 302-304.
(9) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and
Applications, 2nd ed.; Wiley: New York, 2001.
(10) Li, J. H.; Schuler, K.; Creager, S. E. J. Electrochem. Soc. 2000, 147, 4584-
4588.
(11) Brevnov, D. A.; Finklea, H. O. J. Electrochem. Soc. 2000, 147, 3461-
3466.
(12) Creager, S. E.; Wooster, T. T. Anal. Chem. 1998, 70, 4257-4263.
(13) Matsuda, H.; Aoki, K.; Tokuda, K. J. Electroanal. Chem. 1987, 217, 15-
32.
(14) Brown, A. P.; Anson, F. C. Anal. Chem. 1977, 49, 1589-1595.
(15) (a) Peerce, P. J.; Bard, A. J. J. Electroanal. Chem. 1980, 114, 89-115. b)
Henning, T. P.; Bard, A. J. J. Electrochem. Soc. 1983, 130, 613-621. (c)
Henning, T. P.; White, H. S.; Bard, A. J. J. Am. Chem. Soc. 1982, 104,
5862-5868. (d) Henning, T. P.; White, H. S.; Bard, A. J. J. Am. Chem.
Soc. 1981, 103, 3938.
(16) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am.
Chem. Soc. 1990, 112, 4301-4306.
(17) Albery, W. J.; Boutelle, M. G.; Colby, P. J.; Hillman, A. R. J. Electroanal.
Chem. 1982, 133, 135-145.
(18) Gerischer, H.; Scherson, D. A. J. Electroanal. Chem. 1985, 188, 33-38.
(19) Rowe, G. K.; Creager, S. E. Langmuir 1991, 7, 2307-2312.
(20) Acevedo, D.; Abruna, H. D. J. Phys. Chem. 1991, 95, 9590-9594.
(21) Smith, C. P.; White, H. S. Anal. Chem. 1992, 64, 2398-2405.
(22) Rivera, I. M.; Cabrera, C. R. J. Electrochem. Soc. 1993, 140, L36-L38.
(23) Daschbach, J.; Blackwood, D.; Pons, J. W.; Pons, S. J. Electroanal. Chem.
1987, 237, 269-273.
(24) Bond, A. M.; McLennan, E. A.; Stojanovic, R. S.; Thomas, F. G. Anal.
Chem. 1987, 59, 2853-2860.
(25) Takada, K.; Diaz, D. J.; Abruna, H. D.; Cuadrado, I.; Casado, C.; Alonso,
B.; Moran, M.; Losada, J. J. Am. Chem. Soc. 1997, 119, 10763-10773.
(26) Flanagan, J. B.; Margel, S.; Bard, A. J.; Anson, F. C. J. Am. Chem. Soc.
1978, 100, 4248-4253.
1
silica gel with ethyl acetate/hexanes. H NMR (400 MHz, CD2Cl2) of
the thiol gave the appropriate NMR peaks: δ ) 4.0 (m, 9 H), 2.49 (q,
J ) 7.2 Hz, 2 H), 2.30 (t, J ) 7.6 Hz, 2 H), 1.56 (m, 2 H), 1.46 (m,
2 H), and 1.32 (m, 5 H) ppm with no dithiol peaks present and no
significant broadening, indicating that the majority of the ferrocenyl
groups were in the reduced state.
MPC Synthesis. Au225(SC6Fc)43 was synthesized as previously
described.1 Briefly, vigorous mixing of 3.19 g HAuCl4‚xH2O in 100
(39) Ingram, R. S.; Hostetler, M. J.; Murray, R. W.; Schaaff, T. G.; Khoury, J.
T.; Whetten, R. L.; Bigioni, T. P.; Guthrie, D. K.; First, P. N. J. Am. Chem.
Soc. 1997, 119, 9279-9280.
(40) Chen, S. W.; Murray, R. W.; Feldberg, S. W. J. Phys. Chem. B 1998, 102,
9898-9907.
(27) Peerce, P. J.; Bard, A. J. J. Electroanal. Chem. 1980, 112, 97-115.
(28) Peerce, P. J.; Bard, A. J. J. Electroanal. Chem. 1980, 108, 121-125.
(29) Merz, A.; Bard, A. J. J. Am. Chem. Soc. 1978, 100, 3222-3223.
(30) Yamada, M.; Nishihara, H. Chem. Phys. Chem. 2004, 5, 555-559.
(31) Yamada, M.; Nishihara, H. C. R. Chim. 2003, 6, 919-934.
(32) Yamada, M.; Nishihara, H. Langmuir 2003, 19, 8050-8056.
(33) Yamada, M.; Tadera, T.; Kubo, K.; Nishihara, H. J. Phys. Chem. B 2003,
107, 3703-3711.
(41) Wolfe, R. L.; Murray, R. W. Anal. Chem. 2006, 78, 1167-1173.
(42) Kim, Y. G.; Garcia-Martinez, J. C.; Crooks, R. M. Langmuir 2005, 21,
5485-5491.
(43) Balasubramanian, R.; Guo, R.; Mills, A. J.; Murray, R. W. J. Am. Chem.
Soc. 2005, 127, 8126-8132.
(44) Chaki, N. K.; Kakade, B.; Sharma, J.; Mahima, S.; Vijayamohanan, K. P.;
Haram, S. K. J. Appl. Phys. 2004, 96, 5032-5036.
(45) Chaki, N. K.; Singh, P.; Dharmadhikari, C. V.; Vijayamohanan, K. P.
Langmuir 2004, 20, 10208-10217.
(34) Yamada, M.; Quiros, I.; Mizutani, J.; Kubo, K.; Nishihara, H. Phys. Chem.
Chem. Phys. 2001, 3, 3377-3381.
(46) Quinn, B. M.; Liljeroth, P.; Ruiz, V.; Laaksonen, T.; Kontturi, K. J. Am.
Chem. Soc. 2003, 125, 6644-6645.
(35) Men, Y.; Kubo, K.; Kurihara, M.; Nishihara, H. Phys. Chem. Chem. Phys.
2001, 3, 3427-3430.
(47) Yang, Y.; Pradhan, S.; Chen, S. J. Am. Chem. Soc. 2004, 126, 76-77.
(48) Stiles, R. L.; Balasubramanian, R.; Murray, R. W. Unpublished results,
UNC-CH, 2007.
(36) Lenhard, J. R.; Murray, R. W. J. Am. Chem. Soc. 1978, 100, 7870-7875.
(37) Chen, S. W.; 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.
(49) Brauer, G. Handbook of PreparatiVe Inorganic Chemistry; Academic
Press: New York, 1965.
(38) Hicks, J. F.; Templeton, A. C.; Chen, S. W.; Sheran, K. M.; Jasti, R.;
Murray, R. W.; Debord, J.; Schaaf, T. G.; Whetten, R. L. Anal. Chem.
1999, 71, 3703-3711.
(50) Yu, C. J.; Wang, H.; Wan, Y. J.; Yowanto, H.; Kim, J. C.; Donilon, L. H.;
Tao, C. L.; Strong, M.; Chong, Y. C. J. Org. Chem. 2001, 66, 2937-
2942.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 1857

A R T I C L E S
Stiles et al.
mL deionized water with 5.20 g Oct4NBr in 200 mL toluene, gave a
clear aqueous phase and an orange-brown toluene phase. Adding
ω-ferrocenyl hexanethiol to the isolated organic phase (2:1 ligand-to-
Au mole ratio) gave a colorless reaction mixture that was stirred for
20 min and then cooled to 0 °C. A 3.8 g portion of NaBH4 in 10 mL
water was added with very rapid stirring and reacted for 1 h at 0 °C,
after which the dark organic phase was collected and the solvent
removed on a rotary evaporator at room temperature. The black solid
suspension was stirred in 400 mL of acetonitrile for 6 h, and the solid
product collected and washed with acetonitrile on a fine glass frit.
Determination of the average Au core size and of the number of ligands
presented elsewhere1 was based on transmission electron microscopy,
UV-vis absorbance spectroscopy, cyclic and differential pulse volta-
mmetry, and constant potential coulometry.
1.0 M Bu4NClO4 or 1.0 M Bu4NPF6 electrolyte and one of eight
different concentrations of ferrocenated MPC ranging from 0.030 to
1.0 mM. The electrode potential was cyclically scanned between 0 and
0.8 V (vs Ag/AgCl) 25 times at 0.50 V/s (a total scanning time of 80
s); the electrode was then rinsed with MPC-free 1.0 M electrolyte and
placed in MPC-free 1.0 M electrolyte/CH2Cl2. Five preliminary CV
scans at 0.10 V/s were done to remove any loosely adsorbed MPCs.
Surface coverages were assessed from scans over the ferrocene wave
in the MPC-free electrolyte between 0 and 0.80 V versus Ag/AgCl at
six different scan rates (0.025-2.0 V/s), as described for the survey
experiments, again subtracting background currents. The measured
coverages were independent of potential scan rate. The Pt working
electrode was then repolished, rinsed, and exposed to the next
concentration of ferrocenated MPCs.
A mixed monolayer MPC with a lower ferrocene population (average
composition, by NMR, Au225(SC6)58(SC6Fc)17) was prepared by a
ligand place exchange between Au225(SC6)75 (prepared as previously
described41) and HSC6Fc at a 1:1 mole ratio of ω-ferrocenyl hexanethiol
to hexanethiolate ligand. A 10-mL tetrahydrofuran solution of 0.020 g
Au225(SC6)75 and 0.0089 g HSC6Fc was stirred for 1 h, the solvent
removed, and the mixed monolayer MPCs rinsed with acetonitrile to
remove excess thiols. The relative number of the two thiolate ligands
In a second adsorption-kinetics protocol, called the no-scan kinetics
protocol, a clean Pt disk electrode was exposed at open circuit to
unstirred 0.05 mM Au225(SC6Fc)43 in 1.0 M Bu4NClO4/CH2Cl2 for
chosen periods of time. (The typical open circuit electrode potential
was 0.2 to 0.3 V vs Ag/AgCl). The electrode was rinsed with and
immersed in MPC-free 1.0 M Bu4NClO4/CH2Cl2, and ΓMPC was
determined by scanning over the ferrocene wave at 0.025 V/s (after
preliminary scans as above). The electrode was then repolished and
returned to the MPC solution for another period of (no-scan) exposure.
In the third adsorption kinetics protocol, called the scanning kinetics
protocol, the potential of the Pt working electrode, in 0.05 mM Au-
(SC6Fc)43 in 1.0 M Bu4NClO4/CH2Cl2, was cyclically scanned at 0.50
V/s between 0 and 0.80 V versus Ag/AgCl for chosen periods of time.
Then the electrode was rinsed with and immersed in MPC-free 1.0 M
Bu4NClO4/CH2Cl2 solution, and the surface coverage was determined
as above. The electrode was then returned to the MPC solution for a
further period of potential scanning time. (In this last protocol, the
electrode was not polished in between exposures to the MPC solution,
so the adsorption coverages represent an accumulated quantity. Several
examples of doing the no-scan kinetics experiment in this way gave
results agreeing with those involving polishing the electrode in between
solution exposures.)
1
on the exchanged product was analyzed in CD2Cl2 solution with H
NMR on a Bruker 400 MHz Avance spectrometer with a 5 s relaxation
delay time.
Electrochemistry. Voltammetry of typically 0.1 mM MPC degassed
CH2Cl2 solutions containing varying concentrations of electrolyte was
done at 1.5 mm diameter Pt disk (working), Pt wire (counter), and
Ag/AgCl/3 M KCl (aq) electrodes using a model 100B Bioanalytical
Systems (BAS) electrochemical analyzer. The Pt working electrode was
polished (0.25 µm diamond, Buehler) and cleaned electrochemically
by potential cycling in 0.10 M H2SO4 solution. Its roughness factor
(2.9) was determined from the charge under the hydrogen desorption
peak in voltammetry of 0.10 M H2SO4, using the standard9 210 µC/
cm2.
The initial experiments presented will be referred to as surVey
experiments, that probe the general features of the adsorption. Freshly
polished and electrochemically cleaned Pt electrodes were exposed to
unstirred 0.05 mM Au225(SC6Fc)43 MPC solutions in CH2Cl2 containing
given concentrations of Bu4N+X- electrolyte (X- ) ClO4-, PF6-, or
p-toluene sulfonate-), while cyclically scanning the electrode 25 times
(0.50 V/s, total scanning time 80 s.) between 0.2 and 1.0 V (vs Ag/
AgCl). The electrode was then rinsed and placed in an MPC-free
solution of the same electrolyte concentration. Electrodes were also
exposed to unstirred 0.05 mM MPC solutions in 1.0, 0.10, or 0.010 M
Bu4NClO4/CH2Cl2 for 3 min without potential scanning (no-scan
surVey); the electrode was then rinsed with and immersed in MPC-
free solutions having the same Bu4NClO4 concentration.
Self-Assembled Monolayer Formation. Self-assembled monolayers
(SAM) of dodecanethiolate were prepared by soaking a cleaned Pt
electrode for 20 h in 1 mM thiol in degassed, absolute ethanol.
Results and Discussion
Adsorption of Fully Ferrocenated Au225 MPCs: Survey
Experiments. Cyclic voltammetry (Figure 1A-C) of 0.1 mM
solutions of Au225(SC6Fc)43 MPCs in 1.0 M Bu4NPF6/CH2Cl2
shows obvious product adsorption effects; the reduction peak
currents are larger than the oxidation currents. The adsorption
proves to be very robust; after exposure to the MPC-containing
electrolyte solution, the electrode can be removed, rinsed with
MPC-free 1.0 M Bu4NPF6/CH2Cl2, and placed in MPC-free 1.0
M Bu4NPF6/CH2Cl2 to produce the very stable voltammetry
shown in Figure 1D-F. The currents in the MPC-free electrolyte
solution decay slightly upon initial potential scanning, but they
quickly stabilize (see Supporting Information Figure S-2). The
tenacity of the adsorption is illustrated by its nearly quantitative
persistence after rinsing, soaking, or sonication in electrolyte-
containing or electrolyte-free CH2Cl2; with minor variations,
the same stable voltammetry (Figure 1D-F) was seen thereafter
in MPC-free electrolyte solutions. The adsorbed MPCs are
gradually removed by prolonged potential scanning and, natu-
rally, by polishing the electrode. Prolonged potential scanning
results, the narrowed waveshape in Figure 1D, and implied
lateral attractive interactions are discussed later.
Surface coverages, ΓMPC (mol MPC/cm2), of adsorbed MPCs were
calculated from the charge, Q, under the voltammetric current peaks
(corrected for clean-electrode background currents) in MPC-free
electrolyte solution, using
Q ) nFAmΓMPC
(1)
where n is the number of ferrocenes per MPC (average 43, by previous
coulometry1) and Am is the Pt electrode area corrected for roughness.
The experimentally determined values of ΓMPC are compared to
estimates of a model MPC monolayer, based on an overall 2.4 nm
MPC radius (summing Au core radius and length of the monolayer
chain, including the ferrocene), that give 18.1 nm2/MPC, or ΓMONO,MPC
≈ 9 × 10-12 mol MPC/cm2.
Adsorption Kinetics. Three different protocols were employed to
measure the rates of forming robustly adsorbed MPC films. In one,
called the fixed-time kinetics experiment, a clean Pt electrode was
exposed to an unstirred CH2Cl2 electrolyte solution containing either
While the adsorption of the ferrocenated Au225 MPCs
produces robust monolayers under the above circumstances, the
9
1858 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Adsorption of Ferrocenated Nanoparticles
A R T I C L E S
Table 2. Effect in Survey Experiments of Potential Scanning
(between 0.2 and 1.0 V at 0.50 V/s), or Not Potential Scanning
(i.e., Open Circuit) on Adsorption of Ferrocenated MPCs from
Bu4NClO4/CH2Cl2 Solution
ΓMPC
ΓMPC
[Bu4NClO4]
M
with potential scanning
mol MPC/cm2
without potential scanning
mol MPC/cm2
1.0
0.10
0.010
2.9 × 10-12
1.4 × 10-12
no film formed
1.0 × 10-12
3.0 × 10-13
no film formed
a precipitated film; charges equivalent to many MPC monolayers
were passed during potential scanning in the MPC-containing
electrolyte solution (i.e., under the waves of voltammograms
like Figure 1B), yet less than one monolayer is robustly retained
on the electrode after rinsing and transfer to fresh electrolyte.
The importance of the chosen electrolyte anion is seen by
comparing results when the electrolyte is 1.0 M Bu4NPF6, Bu4-
NClO4, or Bu4NC7H7SO3. While more adsorption is seen after
transfer when using Bu4NPF6 than with Bu4NClO4, no adsorp-
tion is seen when the electrolyte is Bu4NC7H7SO3.
Table 1 also shows that the adsorption is promoted by larger
concentrations of electrolyte; increasing Bu4NClO4 concentration
from 0.1 to 1.0 M in the MPC-containing solution from which
adsorption is induced increases ΓMPC by ca. 2-fold (as observed
after transfer to MPC-free solution). Adsorption from MPC-
containing 0.010 M Bu4NClO4 was weak and nonpersistent; the
ferrocene currents dissipated quickly after a few potential scans
in MPC-free 0.010 M Bu4NClO4/CH2Cl2.
Figure 1. Cyclic voltammetry of 0.1 mM Au225(SC6Fc)43 in 1.0 M Bu4-
NPF6/CH2Cl2 solution at potential scan rates 0.025, 0.20, and 2.0 V/s in
curves A, B, and C, respectively, and of an adsorbed Au225(SC6Fc)43 film
on electrode (ΓMPC ) 7.1 × 10-12 mol/cm2) formed by potential scanning
in 0.1 mM MPC in 1.0 M Bu4NPF6/CH2Cl2 solution and transferred to an
MPC-free 1.0 M Bu4NPF6/CH2Cl2 solution, at potential scan rates of 0.025,
0.20, and 2.0 V/s in curves D, E, and F, respectively. Results in MPC-free
1.0 M Bu4NPF6/CH2Cl2 solution at additional potential scan rates are given
in Supporting Information, Figure S-1.
While adsorption does not require scanning of potential, its
amount is enhanced by scanning to positive potentials, through
the ferrocene wave. Using the surVey experiment procedure (see
Experimental), clean Pt electrodes were exposed, either at open
circuit (for 3 min, no-scan surVey) or with potential scanning
(see Experimental), to a 0.05 mM MPC solution of varying
concentrations of Bu4NClO4 in CH2Cl2. The electrode was
rinsed with and then immersed in the corresponding MPC-free
electrolyte solution to assess the extent of adsorption by cyclic
voltammetry. Table 2 shows that a smaller surface coverage is
observed in the no-scan experiment; the resulting voltammetry
is however, equally persistent in MPC-free electrolyte solution.
One may infer from the no-scan surVey result (Table 2) that
ferrocene sites need not be oxidized for robust, persistent
adsorption to occur. Ferrocenated dendrimers are also known
to adsorb without oxidation.25 Also, analysis of the (narrowed)
voltammetric waveshape (vide infra) is consistent with attractive
lateral interactions between both oxidized and reduced absorbing
nanoparticles, which would encourage adsorption. Further, a
small ambient fraction of the ∼43 ferrocene sites in the native
MPC solution is oxidized, and cationic, given that the 0.20-
0.30 V rest potentials of ferrocenated MPC solutions are not
far from the ferrocene formal potential (0.40 V). Those oxidized
ligand sites may aid adsorption by providing the bridging ion-
pair interactions described in the ion-induced adsorption model
discussed below.
Table 1. Effect of Type and Concentration of Supporting
Electrolyte on Coverage ΓMPC of Ferrocenated MPCs Adsorbed on
Pt Electrode from a 0.05 mM MPC Solution, during Potential
Scanning Survey Experiment (Scan between 0.2 and 1.0 V at 0.50
V/s)
supporting
electrolyte
[supporting electrolyte]
M
ΓMPC
mol MPC/cm2
Bu4NC7H7SO3
Bu4NPF6
Bu4NClO4
Bu4NClO4
Bu4NClO4
1.0
1.0
1.0
0.10
0.010
no film detecteda
6.3 × 10-12 a
2.9 × 10-12 a
1.4 × 10-12 a
no film detecteda
a Following rinsing and transfer to fresh electrolyte.
adsorption process is sensitive to a variety of factors, including
the choice and concentration of electrolyte anion, the state of
the Pt electrode surface, the population of ferrocene sites in the
MPC monolayer, and whether the electrode’s potential is
scanned or not during exposure to MPC solution. Also, there is
an uncompensated ohmic resistance associated with the adsorp-
tion that is too large to be explained by solution resistance alone.
The adsorption does not appreciably change the Pt electrode
double layer capacitance at potentials on the wings of the
ferrocene wave. These features, the evidence for which follows,
collectively point to an electrolyte anion-induced, entropically
enhanced adsorption mechanism.
The extent of ferrocenated MPC adsorption, ΓMPC, is strongly
dependent on the particular electrolyte anion used, as shown
by the surface coverage data in Table 1. The ΓMPC results in
Table 1, taken under surVey experiment conditions (see Ex-
perimental) with potential scanning, are all less than the model
estimate for a full MPC monolayer, ca. 9 × 10-12 mol/cm2.
This result shows that the adsorption phenomenon is not trivially
The 3 min exposure in the no-scan surVey experiments
reported in Table 2 is more than sufficient (assuming D ≈ 3 ×
10-6 cm2/s)51 to produce a complete MPC monolayer, were
adsorption to occur at a diffusion-controlled rate. (No attempt
(51) Green, S. J.; Pietron, J. J.; Stokes, J. J.; Hostetler, M. J.; Vu, H.; Wuelfing,
W. P.; Murray, R. W. Langmuir 1998, 14, 5612-5619.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 1859

A R T I C L E S
Stiles et al.
Figure 2. Oxidation peak (solid blue line) from cyclic voltammetry (2.0
V/s) of partly ferrocenated, adsorbed Au225(SC6)58(SC6Fc)17 in 1.0 M Bu4-
NClO4, after transfer to MPC-free electrolyte. The dashed pink line is a
simulated waveshape for a reversible one-electron transfer with no lateral
interactions and with Eo′ the same for all ferrocenes (eq 3). ΓMPC ) 2 ×
10-13 mol/cm2.
Figure 3. Cyclic voltammetry (curve A, 0.20 V/s) of an adsorbed Au225
-
(SC6Fc)43 film formed on a clean Pt electrode from a 0.1 mM MPC, 1.0 M
Bu4NPF6/CH2Cl2 solution and transferred to an MPC-free 1.0 M Bu4NPF6/
CH2Cl2 solution. Curve B is the cyclic voltammetry (0.20 V/s) of an
identically treated Pt electrode, except that the Pt electrode had been first
coated with a dodecanethiolate self-assembled monolayer (SAM).
was made to control inadvertent convective stirring.) This low
coverage and the apparent irreversibility of the adsorption
suggest slow kinetics of forming the robustly adsorbed film, as
reported in a later section.
The ferrocenated MPC adsorption is weaker when there are
fewer ferrocene sites in the MPC monolayer. Evidence for
adsorption has been noticed previously52,53 in voltammetry (in
0.10 M Bu4NPF6/CH2Cl2) of hexanethiolate-coated MPCs
(average 4.8 nm core diameter) where a small fraction (10-
15%) of the original thiolate ligands had been replaced (by
ligand exchange) by ferrocenated ligands. Those adsorbed MPCs
could be removed from the Pt electrodes simply by rinsing with
fresh CH2Cl2 solvent. This observation was probed further here,
using a higher electrolyte concentration (1.0 M Bu4NClO4/CH2-
Cl2) and Au225 MPCs bearing ca. 17 ferrocene alkanethiolate
ligands (see Experimental). Figure 2 shows the oxidation peak
of the cyclic voltammogram of this adsorbed MPC in MPC-
free 1.0 M Bu4NClO4/CH2Cl2 solution. The surface coverage
is smallsabout 10% of that seen in Figure 1 and less persistent
than that of fully ferrocenated MPCs; the adsorption redox peaks
gradually decreased during voltammetry in MPC-free electrolyte
solution. Also, the Figure 2 peaks, unlike the Figure 1D-F
voltammetry of adsorbed fully ferrocenated MPCs, are not
narrowed and in fact show nearly ideal Efwhm. These results show
that the robust adsorption of Figure 1D-F is facilitated by a
dense coverage of ferrocene sites on the Au MPC.
The Table 1 ΓMPC results suggest that the adsorption of fully
ferrocenated Au MPCs depends on specific adsorption of the
electrolyte anion on the Pt surface in the CH2Cl2 medium.
Indeed, coating the Pt electrode surface with a self-assembled
monolayer (SAM) of dodecanethiolate ligands, almost com-
pletely eliminated the MPC adsorption in 1.0 M Bu4NPF6/CH2-
Cl2 (Figure 3).
In the course of evaluating values of ΓMPC, cyclic voltam-
metric background currents were recorded on clean Pt electrodes
in Bu4NClO4/CH2Cl2 and Bu4NPF6/CH2Cl2 solutions. Remark-
ably, the double-layer charging currents at potentials more
positive and negative of the ferrocene wave are substantially
unaffected by the ferrocenated MPC adsorption (see Figure S-3
in Supporting Information). Non-ferrocenated alkanethiolate-
coated MPCs in CH2Cl2 solutions readily physisorb on elec-
trodes and substantially suppress (by a factor of 2 or 3) double
layer charging currents, presumably by forming a hydrocarbon-
like film.52 The adsorbed, fully ferrocenated Au225 MPCs are
unique in our experience with MPC voltammetry in that they
hardly disturb the electrode double layer capacitance. An
additional implication is that small electrolyte ions can access
the electrode surface through the interstices in the layer of (avg
diameter near 5 nm) adsorbed MPCs.
The longer term persistence of Au225(SC6Fc)43 MPC adsorp-
tion in MPC-free 1.0 M Bu4NClO4/CH2Cl2 was evaluated by
cyclically scanning the potential of an adsorbate-coated electrode
between 0.20 and 0.55 V at 0.20 V/s. Over 1840 scans of the
ferrocene redox states provoked only a 16% loss in MPC
coverage. On the other hand, scanning to a more negative
potential limit (-0.80 V) caused a rapid decrease in ΓMPC
(Figure S-4); after only 280 scans ΓMPC decreased ca. 90% of
its original value. The -0.8 V limit of the voltammetric scan is
more negative than the ca. -0.1 V potential of zero charge
(EPZC) for Pt in an organic medium.54 The significance of this
statement is that anions should be driven off the Pt electrode at
such potentials, an observation in harmony with the results in
Figure S-4. At the same time, we note that the MPCs are not
immediately desorbed at negative potentials, which is consistent
with the slow adsorption kinetics discussed later and the thus
inferred lateral organization of the adsorbed film.
∆Epeak values and peak-shapes in voltammograms like Figure
1D-F reveal that the adsorbed MPC layer offers a resistance
to current flow. Figure 4A shows that ∆Epeak (at constant scan
rate) increases gradually at sub- and monolayer MPC coverages.
(Varied coverage is obtained in scanning kinetics protocol
experiments, described later.) This is a defining characteristic
of an uncompensated ohmic resistance effect; ∆Epeak should not
vary with coverage (at a given scan rate) if the adsorbed layer’s
resistance involved slow heterogeneous electron transfers. This
assignment of the adsorbed layer’s resistance is supported by
comparisons of experimental and simulated waveshapes (see
Supporting Information Figure S-5).
Figure 4B provides a further analysis of the film’s resistance,
showing ∆Epeak versus current results at varied scan rate and
constant (submonolayer) coverage (open circles). The ∼2000
(52) Miles, D. T.; Murray, R. W. Anal. Chem. 2003, 75, 1251-1257.
(53) Green, S. J.; Stokes, J. J.; Hostetler, M. J.; Pietron, J.; Murray, R. W. J.
Phys. Chem. B 1997, 101, 2663-2668.
(54) (a) Marinkovic, N. S.; Hecht, M.; Loring, J. S.; Fawcett, W. R. Electrochim.
Acta 1996, 41, 641-651. (b) Experimental conditions were 100 mM Et4-
NClO4/acetonitrile on Pt(100) and Pt(111) single crystals.
9
1860 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Adsorption of Ferrocenated Nanoparticles
A R T I C L E S
Figure 4. From cyclic voltammetry of adsorbed Au225(SC6Fc)43 films transferred to MPC-free CH2Cl2 electrolyte solution. Panel A: Values of peak
potential splitting ∆Epeak and formal potential as a function of accumulated amount of robust adsorption, in 1.0 M Bu4NClO4 electrolyte; adsorbed films
were prepared by scanning kinetic protocol. Panel B: Values of peak potential splitting ∆Epeak and peak current (summed oxidation and reduction) as a
function of potential scan rate, from 0.025 to 2.0 V/s for films adsorbed during survey experiments like that in Figure 1D-F, in 1.0 M Bu4NPF6/CH2Cl2
solution.
and in NMR57 and QCM22 experiments.The promotion of
adsorption by scanning to positive potentials is consistent with
increasing counterion adsorption as well as cationizing the
nanoparticles. The present results show the importance of (a)
the electrolyte anions and their concentrations, (b) their largely
unchanged retention in the electrode double layer following
MPC adsorption, (c) the population of ferrocene sites on the
MPC, and (d) the nullifying of anion specific adsorption on the
Pt surface with a hydrophobic SAM.
A likely essential aspect of Figure 5 is the multiplicity of the
ion-pair bridges made possible by the concentrated ferrocenium
population on the MPC. We propose that the strong persistence
Figure 5. Cartoon of ion-induced adsorption, where ferrocenium cations
of the adsorption is at least partially due to entropic stabilization,
on the MPC form ion-pair bridges with electrolyte anions specifically
akin to the well-known chelate effect.58 Formation of multiple
adsorbed to the Pt electrode. It can be imagined that the ligand shell may
lateral ferrocenium/X-/ferrocenium ion associations between
become deformed to form ion-pair bridges of similar dimensions. Formation
of ion-pair bridges shown stabilizes successive ion-pair bridges and can
neighbor MPC adsorbates, and of analogous ferrocene/ferrocene
cause a successive shift in Fc/Fc+ formal potentials (Ejmax scheme, see later).
interactions, would further contribute to entropic stabilization
There may also be imagined lateral ion-pair bridges between adsorbates.
and is consistent with the positive lateral interactions between
neighboring adsorbed nanoparticles revealed by analysis of
Ω total uncompensated resistance indicated by these data is
voltammetric waveshape (vide infra). The implied large col-
lective ion-pair formation constants would be sensitive to the
anion structure, the population of anions in the double layer,
and the population of ferroceniums on the MPCs; effects all
seen in the above experimental behavior. The differing ΓMPC
coverages in Bu4NPF6 and in Bu4NClO4 electrolytes (Table 1)
may reflect differences in such details. The slow robust
adsorption as well as the slow desorption are arguably indicative
of kinetic slowness of adsorption and may signal some lateral
ordering or organization of the MPC monolayer. Figure 5 is a
nanomaterials reincarnation of the “anion-induced adsorption”
model proposed many years ago by Anson, et al.59 to account
for the (much weaker) adsorption of small metal complexes on
Hg electrodes.
consistent with the Figure S-5 simulations and is much larger
than the ∼600 Ω RUNC expected at a naked 1.5 mm diameter
electrode. (This RUNC was estimated from resistivities 188 and
156 Ω cm of 1 M Bu4NPF6 and Bu4NClO4 solutions in CH2-
Cl2 as measured by ourselves and others, respectively.)55 It thus
appears that a monolayer of robustly adsorbed MPCs offers an
intrinsic ohmic resistance of the order of 1500-2000 Ω, which
perhaps reflects restrained ion movement through the adsorbed
layer.
While further aspects of ferrocenated MPC adsorption and
its voltammetry are evaluated below, we pause here to outline
how the above results suggest a mechanism for the ferrocenated
MPC adsorption. We propose that it involves electrolyte anions
in the electrochemical double layer forming bridging ion-pairs
between the electrode and ferrocenium sites on the MPC, as
cartooned in Figure 5. The persistent adsorption does not occur
when the electrolyte ion (e.g., p-toluenesulfonate) is less
structurally suited to serving as an ion bridge. Ion-pairing has
The above phenomena are not confined to Pt surfaces and
CH2Cl2 solvent. The persistent adsorption from CH2Cl2 also
occurs on glassy carbon and on Au electrodes. The ferrocenated
MPCs are insoluble in most electrochemically useful solvents
but (much less comprehensive) tests in tetrahydrofuran showed
-
been detected previously between ferrocenium and ClO4 and
-
PF6 in aqueous voltammetry of ferrocenated SAMs,19 in
(56) Hupp, J. T. Inorg. Chem. 1990, 29, 5010-5012.
(57) Yang, E. S.; Chan, M. S.; Wahl, A. C. J. Phys. Chem. 1980, 84, 3094-
3099.
(58) Schwarzenbach, G. HelV. Chim. Acta 1952, 35, 2344-2359.
(59) Anson, F. C.; Barclay, D. J. Anal. Chem. 1968, 40, 1791.
intervalent charge transfer of acetylene bridged biferrocene,56
(55) (a) Kadish, K. M.; Ding, J. Q.; Malinski, T. Anal. Chem. 1984, 56, 1741-
1744. (b) Newman, J, J. Electrochem Soc. 1966, 113, 501-502.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 1861

A R T I C L E S
Stiles et al.
adsorption effects similar to those in CH2Cl2. MPCs adsorbed
from 1.0 M Bu4NPF6/CH2Cl2 solutions could be rinsed with
MPC-free 1.0 M Bu4NPF6/CH2Cl2 and placed in 1.0 M Bu4-
NPF6/CH3CN where stable voltammetry similar to that discussed
above could be observed (results not shown).
We comment that Figure 5 is based on the implicit assumption
that when ΓMPC coverages do not exceed the model ΓMPC
monolayer coverage (see Experimental), multilayers and that
dendrites do not form. Reproducible observations of monolayer
quantities of nanoparticle adsorption on mechanically polished
surfaces over a range of experimental conditions make the latter
events implausible.
Adsorption Kinetics. The slow formation of films of robustly
adsorbed, fully ferrocenated Au MPCs prompted a study of
their adsorption kinetics. Three different protocols (see Experi-
mental section) were used in assessing adsorption kinetics.
The fixed-time kinetics protocol involved exposure, while
scanning the electrode potential for a fixed time, of a clean
Pt electrode to either 1.0 M Bu4NClO4 or 1.0 M Bu4NPF6/
CH2Cl2 solutions containing varied concentrations of ferro-
cenated MPCs. In the no-scan kinetics protocol, the Pt
electrode was exposed at open circuit to a 0.05 mM MPC
solution in 1.0 M Bu4NClO4/CH2Cl2 for varying periods of time.
A third, scanning kinetics protocol, involved cyclically scanning
the electrode potential in a 0.05 MPC/1.0 M Bu4NClO4/CH2-
Cl2 solution through the ferrocene wave for varying periods of
time.
The results of the above experiments are displayed in Figure
6. Panel A displays the data of the fixed-time kinetics protocol
(fixed 80 s exposure time) for the two electrolyte systems as
plots of fractional coverage, θ ) ΓMPC/ΓSAT, where ΓSAT is the
apparent saturation (plateau) monolayer coverage, versus MPC
concentration, c (lower x-axis), and versus time multiplied by
concentration, tc (upper x-axis). Panel B displays MPC cover-
ages, ΓMPC, for all protocols, the fixed-time data of panel A,
and no-scan, and scanning protocols at constant, 0.05 mM MPC
concentrations, plotted versus log [tc]. Approximate adsorption
rate constants can be derived from these experimental data by
fitting to the first-order rate law
Figure 6. Panel A: Fractional coverages (θ ) ΓMPC/ΓSAT) of MPCs
measured in fixed-time kinetics experiments at different MPC concentrations
(lower horizontal axis) in 1.0 M Bu4NClO4 (open black circles, ΓSAT ) 0.7
× 10-11 mol/cm2) and 1.0 M Bu4NPF6 (filled black circles, ΓSAT ) 1.3 ×
10-11 mol/cm2) electrolytes. The upper horizontal axis is normalized MPC
concentration multiplied by time (tc), where t ) 80 s for all data points.
Panel B: Coverage of MPCs (ΓMPC) measured as a function of log(tc) in
no-scan (open red circles), scanning (filled red circles), and fixed-time
(ClO4- filled black circles, PF6- circles) kinetics protocol experiments. No-
scan and scanning kinetics protocols used a fixed (0.05 mM) MPC
concentration in 1.0 M Bu4NClO4/CH2Cl2. The solid lines in both panels
are fits to the first-order rate expression, eq 2, for the rate constants displayed
in Table 3. No fit was performed on the scanning protocol as at long
scanning times it led to multilayers and no obvious saturation coverage.
θ ) 1 - exp(-ktc)
(2)
where k is the adsorption rate constant, t is the time the electrode
spent in the MPC solution, and c is the MPC concentration (mol/
cm3). The solid curves shown in Figure 6A,B are such fits. The
corresponding rate constants for the fixed-time kinetics protocol
are listed in Table 3. This analysis assumes that there is a maxi-
mum monolayer surface coverage, that adsorption is effectively
irreversible, and that there is no concentration polarization of
the MPC near the electrode. It is important to recognize that
since the adsorption coverages are not true equilibrium values
(any potential dependency of adsorption amount and rate is
obscured by the potential scanning during adsorption), the Table
3 rate constants represent an average over the scanning potential
range.
Figure 6A shows that in the fixed-time kinetics experiment,
while ΓSAT is somewhat larger in 1.0 M Bu4NPF6/CH2Cl2 (ΓSAT
) 1.3 × 10-11 mol/cm2) than in 1.0 M Bu4NClO4/CH2Cl2 (ΓSAT
) 0.7 × 10-11 mol/cm2), the overall rates of MPC adsorption
are identical in the two electrolyte media. A single value of k
(Table 3) thus serves to fit the results in both media.
Table 3. Rate Constants for MPC Adsorption Derived from
Fixed-Time and No-Scan Experimental Protocolsa
experimental protocol
k, cm3/mol
‚s
no-scan (1.0 M Bu4NClO4/CH2Cl2)
fixed-time (1.0 M Bu4NClO4/CH2Cl2)
fixed-time (1.0 M Bu4NPF6/CH2Cl2)
9 × 103
9 × 104
9 × 104
a ΓMPC was measured as a function of MPC concentration in fixed-time
experiments and as a function of time in no-scan protocol. (All scans were
taken between 0 and 0.8 V/s.)
Figure 6B plots the deposition data for all protocols as a
function of tc. The open and closed black circles are the fixed-
time kinetics data of Figure 6A. The no-scan kinetics results
are also well-fit (open red circles, solid line) by eq 2 and produce
an adsorption rate constant (Table 3) that is ca. 10-fold slower
than seen in the fixed-time experiment (Figure 6A). This rate
difference is consistent with the results of Table 2 and the model
of Figure 5 and reflects the influence of scanning to positive
potentials in the latter (fixed-time) experiment.
9
1862 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Adsorption of Ferrocenated Nanoparticles
A R T I C L E S
The results for the scanning kinetics protocol in Figure 6B
(closed red circles) reveal an additional aspect of the ferro-
cenated MPC adsorption, namely that robust multi-monolayer
deposition can occur given sufficiently long times (more than
3 h) of exposure to the MPC solution. The largest ΓMPC attained
was 2.9 × 10-11 mol/cm2, which is equivalent to about three
monolayers, and the adsorption does not appear to have reached
a saturation point. From the plot, it is evident that as long as
ΓMPC < ΓSAT, the rate of accumulation of ΓMPC is roughly similar
to that seen in the other experimental protocols, and presumably
occurs with a similar rate constant. However, when ΓMPC
>
ΓMONO, the adsorption rate profile changes. These effects, not
further analyzed here, clearly signal an increased complexity
of film properties upon multilayer formation.
The voltammetry of the adsorbed ferrocenated MPCs through-
out the preceding experiments continued to exhibit quite narrow
peaks (like those of Figure 1D) at all monolayer or submono-
layer coverages. However, when multilayers were formed in
the scanning kinetics protocol (Figure 6B), there is (Figure 4)
a small shift in formal potential and sharply increased ∆Epeak
values (i.e., in the film’s uncompensated resistance). The latter
effect enhances the Efwhm of the wave by ca. 2-fold, but it
remains more narrow than the ideal 90.6 mV. Additionally, the
behavior illustrated in Figure S-2, where a small loss of MPC
coverage occurred upon preliminary potential scanning after
transfer to MPC-free solution, was not seen when MPC
multilayers formed.
Figure 7. Ferrocene oxidation peak in experimental voltammetry (green
solid line, 0.10 V/s) of ferrocenated MPCs adsorbed from a 0.1 mM MPC
solution in 1.0 M Bu4NClO4/CH2Cl2 compared to waveshapes (pink dashed
lines) predicted from eqs 3 (curve A) and 4 (curve B for rox ) rred ) 2 ×
1011 mol-1cm2). Experimental data collected from 1.0 M Bu4PF6/CH2Cl2
(not shown) gave a similarly good fit to eq 5, using rox ) rred ) 6 × 1011
mol-1cm2. Predicted curves are normalized to the experimental Eo′ and peak
current.
There have been several theoretical analyses9,14,15,63-68 of how
activity effects change surface waveshapes using a quantity-
dependent activity, a parametrized interaction between ox and
red sites, or other theoretical models.62 A one-electron reaction,
representing interaction parameters between ox and red as rox
and rred yields the equation14
We will return to a discussion of the possible origin of the
slowness of the adsorption process later.
Voltammetric Waveshape Analysis. We turn last to the
narrowed voltammetric peaks noted above for adsorbed ferro-
cenated MPCs (Figure 1D, Efwhm ) 35 mV). A one-electron
reaction of an ideal adsorbed layer has, in contrast, a voltam-
metric Efwhm ) 90.6 mV and a waveshape following the
equation14,60-62
bF2ΓTνf(1 - f)
i ) -
(4)
RT[1 - fΓT(rox + rred)(1 - f)]
where f is the fraction of the molecules on the surface in their
oxidized form. For rox ) rred ) 0 (i.e., no interactions), eq 4
reduces to eq 3. The interaction parameter is signed positive
and negative for attractive and repulsive interactions, respec-
tively; attractive interactions in eq 4 lead to voltammetric peak
narrowing.
Comparisons of experimental data to this model support the
idea that the adsorbed layers of ferrocenated MPCs are stabilized
by lateral interactions that probably involve counterion bridges
that serve to screen ferrocenium charges from one another.
Figure 7 compares eqs 3 and 4 (dashed pink lines, curves A
and B, respectively) to MPC adsorption data (green solid lines)
taken in a surVey experiment in 1.0 M Bu4NClO4/CH2Cl2
solution. Curve A shows that the experimental oxidation peak
is substantially more narrow than predicted by eq 3, while curve
B shows an excellent fit for eq 4 using (quite large) attractive
interaction parameters rox ) rred ) 2 × 1011 mol-1 cm2. Fits
for eqs 3 and 4 to the experimental oxidation peak of
ferrocenated MPCs in 1.0 M Bu4NPF6/CH2Cl2 (not shown) gave
F
RT
exp
(E - Eo′)
bF2ΓTV
RT
[
]
i ) -
(3)
2
F
1 + exp
(E - Eo′)
(
[
]
)
RT
where b is the number of ferrocenes per MPC, ΓT is adsorbed
MPC coverage (ΓMPC,red + ΓMPC,ox) and ν is potential scan rate.
Equation 3 assumes that all redox sites react independently, have
the same formal potential, Eo′, and have adsorbate activities
proportional to Γred and Γox. Equation 3 accurately represents
the waveshape of dilute ferrocene monolayers as shown by the
comparison in Figure 2 (pink dashed line), but for concentrated
populations of ferrocene/ferrocenium sites in the MPC ligand
shell, Figure 1D shows that the assumptions of eq 3 fail and
narrowed voltammograms result.
We offer two alternative interpretations of the narrow
voltammetric waveshape based on failures of the above as-
sumptions. The first is a surface activity model for narrowed
MPC voltammetric waveshapes that is based on attractive lateral
interactions between adsorbed nanoparticles that cause activities
of the adsorbates to scale nonlinearly with surface coverage.
(63) Chidsey, C. E. D.; Murray, R. W. J. Phys. Chem. 1986, 90, 1479-1484.
(64) Coleman, S. T.; McKinnon, W. R.; Dahn, J. R. Phys. ReV. B 1984, 29,
4147-4149.
(65) Ikeda, T.; Leidner, C. R.; Murray, R. W. J. Electroanal. Chem. 1982, 138,
343-365.
(60) Laitinen, H. A.; Vincent, C. A.; Bednarsk, T. J. Electrochem. Soc. 1968,
115, 1024-1028.
(61) Laviron, E. Bull. Soc. Chim. Fr. 1967, 3717-3723.
(62) Murray, R. W. Molecular Design of Electrode Surfaces; Wiley: New York,
1992.
(66) Laviron, E. J. Electroanal. Chem. 1981, 122, 37-44.
(67) Ellis, D.; Eckhoff, M.; Neff, V. D. J. Phys. Chem. 1981, 85, 1225-1231.
(68) Angersteinkozlowska, H.; Klinger, J.; Conway, B. E. J. Electroanal. Chem.
1977, 75, 45-60.
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 1863

A R T I C L E S
Stiles et al.
Figure 8. Simulations of the oxidative peak of a one electron transfer of an ideal adsorbed ferrocene monolayer (blue dashed line) and Ejmax scheme with
j ) 10 and varied ∆Eo′ (pink dashed line), compared to experimental voltammetry (green solid line, 0.025 V/s) of ferrocenated MPCs adsorbed from a 0.1
mM MPC solution in 1.0 M Bu4NPF6/CH2Cl2 (curves A-C) and 1.0 M Bu4NClO4/CH2Cl2 (curve D). Varied ∆Eo′ fits (blue dashed line) for curve A has
∆Eo′ ) -1 mV (fwhm ) 23 mV), for curves B and D ∆Eo′ ) -4 mV (fwhm ) 35 mV), and for curve C ∆Eo′ ) -6 mV (fwhm ) 53 mV); the best fit
for both electrolytes is at ∆Eo′ ) -4 mV. To compare the curves, all are normalized to the experimental Eo′ and peak current.
identical results but required slightly larger interaction param-
eters (rox ) rred ) 6 × 1011 mol-1cm2). These comparisons
show that an activity-based model successfully represents the
voltammetric peak narrowing.
number of steps that involve ∆Eo′ changes (j) is undoubtedly
only a fraction of the total ca. 43 ferrocenes and may involve
only those closest to the electrode (Figure 5), and perhaps next
to neighbor MPCs. The reverse sequence of events occurs during
reduction. Statistical aspects of many-electron-transfer reactions
of redox polymers9,26 have been presented, but do not predict
peak-narrowing, and the statistical shifts of E0/ predicted are
opposite to those proposed here.
Here it seems appropriate to summarize aspects of the
preceding results and analyses for robustly adsorbed ferrocenated
MPCs. The kinetic slowness of forming the MPC monolayers
is striking and is consistent with a gradual organization of MPCs
within two-dimensional films. At sub-overall-monolayer cover-
ages, these may be small, ordered rafts. Ideas explaining slow
adsorption have been discussed,69 that include formation of
ordered surface structures. Conversion of MPCs to their ferro-
cenium forms and lowered solubility in CH2Cl2, entropically
driven formation of multiple ion-pair bridges to the electrode
and neighbor MPCs, and the consequent strong interactions that
demand organization to more stable nanoparticle arrangements
must play important roles in the slow surface reaction(s). These
slowly generated ordered domains on the Pt electrode may be
the ones that survive solvent rinsing and transfer to MPC-free
electrolyte.
The voltammetric peak-narrowing of adsorbed MPCs in a
Ejmax scheme is illustrated in the Supporting Information in
Figures S-6 and S-7, respectively, by simulations for ∆Eo′
)
-4 mV and jmax ) 1 to 10 (curve 10 is thus an overall 10
electron wave) and by simulations that show the extent of peak
narrowing for jmax ) 10 to be larger for smaller ∆Eo′. While
the choices of jmax ) 10 and ∆Eo′ ) -4 mV to illustrate the
peak narrowing effect are nominally arbitrary, Figure 8 shows
that these parameters provide a good fit to the observed
narrowed adsorbed MPC voltammetry in both 1.0 M Bu4NPF6/
CH2Cl2 and 1.0 M Bu4NClO4/CH2Cl2 solutions (curves B, D).
We do not suggest that the Figure 8 comparisons are unique,
as other combinations of j and ∆Eo′ may fit the data equally
well. The main point is that a model based on progressive
formation of ion-pair bridges, causing formal potential changes,
can represent the peak narrowing experimental results. The Ejmax
scheme is physically appealing because it can be related to the
chemical binding interactions proposed in the model for ion-
induced adsorption (Figure 5). Its obvious weakness is the
presumption of surface activity-concentration proportionality.
It is possible that the observed peak narrowing may reflect a
convolution of both of the above explanations.
It is worthwhile in closing to contrast the fully ferrocenated
MPC adsorption described here to other examples of adsorption
of multiferrocenated entities. A close example is that of Abrun˜a,
et al.25 who observed significant adsorption from 0.10 M Bu4-
NClO4/CH2Cl2, of diaminobutane-based poly(propylene imine)
dendrimers bearing 8, 32, and 64 peripheral ferrocenyl moieties.
The adsorption followed a Langmuir isotherm and was strong,
being observable from sub-micromolar concentrations. The
voltammetric peak shapes were normal, however. A more distant
comparison can be made with adsorption of poly(vinylferrocene)
A second interpretation of the narrowed voltammetric wave-
shapes (Figure 1D) considers that the ferrocene formal potentials
are not all equal, but instead their Eo′ values gradually become
less positive (by increments ∆Eo′) as successive ferrocene sites
are oxidized. The potential shift is promoted by the successive
stabilization of the adsorbed ferrocenated MPC through ion-
pair bridging between ferrocenium sites, electrolyte anions, the
electrode, and neighbor MPCs. The cartoon in Figure 5 shows,
for example, five ion-pair bridges. As each ferrocene is oxidized,
consequent stabilization of the adsorbed MPC by formation of
an ion-pair bridge lowers the potential for oxidation of sub-
sequent ferrocene sites to slightly less positive Eo′. The scheme
comprises jmax sequential reversible electron transfers (Ejmax
scheme), in which Ejo′ - Eo′ is assumed negative and
j-1
constant.
In the Ejmax scheme, each sequential electrochemical reaction
is accompanied by formation of ion-pair bridges between the
electrode, counteranion, and individual ferroceniums. The
(69) Zhang, J. S.; Stanforth, R. Langmuir 2005, 21, 2895-2901.
9
1864 J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008

Adsorption of Ferrocenated Nanoparticles
A R T I C L E S
(PVF).15,26-29 These multilayer films showed interesting com-
plexities, including narrowed peak shapes. They have been
inferred to be caused by different environments of oxidized (and
reduced) sites within the film and conversions between them.
We thank Dr. Amala Dass for measuring the resistivity of 1 M
Bu4NPF6 in CH2Cl2.
Supporting Information Available: Cyclic voltammetry of
adsorbed MPCs and simulated cyclic voltammetry. This material
is available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. This research was supported by grants
from the National Science Foundation and Office of Naval
Research. Rebecca L. Stiles is the former Rebecca L. Wolfe.
JA074161F
9
J. AM. CHEM. SOC. VOL. 130, NO. 6, 2008 1865