Multiple Electron Tunneling Paths across Self-Assembled Monolayers of Alkanethiols with Attached Ruthenium(II/III) Redox Centers

Article abstract of DOI:10.1021/jp962831q

Alkanethiol monolayers with pendant redox centers are deposited on gold electrodes by selfassembly.The monolayers are composed of both an electroactive thiol, HS(CH2)nC(O)NHCH2pyRu(NH3)5(2+/3+), with 10-15 methylene groups, and a diluent thiol, HS(CH2)mCOOH, also with 10-15 methylene groups.The monolayers are classified as "matched" (n = m), "exposed" ( n = 15, m = 10-14), and "buried" (n = 10, m = 11-15) according to the relative position of the redox center.Cyclic voltammograms in aqueous Na2SO4 indicate that the monolayers are close-packed with the redox centers residing in the aqueous phase in all but the most buried cases.Measurements of electron transfer kinetics by several methods (cyclic voltammetry, ac impedance spectroscopy, chronoamperometry) yield an internally consistent set of kinetic parameters, the standard rate constant k^o, and the reorganization energy λ of the redox centers.The reorganization energies are in good agreement with the theoretically predicted value of 1.0 eV for the pyRu(NH3)5 redox centers.Plots of ln(k^o) vs m are linear in all three cases.The slopes of the linear regression fit provide tunneling parameters (β, where k^o ca. e^-βm) of 0.97 +/- 0.03 (matched cases), 0,83 +/- 0.03 (exposed cases) and 0.16 +/- 0.02 (buried cases) per methylene.This pattern of β's is interpreted in terms of electronic coupling between the redox center and the electrode via both the redox thiol and the proximate diluent thiols, with the coupling via the diluent thiols dominating in the exposed cases.

Full text of DOI:10.1021/jp962831q

18852
J. Phys. Chem. 1996, 100, 18852-18858
Multiple Electron Tunneling Paths across Self-Assembled Monolayers of Alkanethiols with
Attached Ruthenium(II/III) Redox Centers
Harry O. Finklea,* Luna Liu, Melissa S. Ravenscroft, and Sesto Punturi
Department of Chemistry, West Virginia UniVersity, Morgantown, West Virginia 26506-6045
ReceiVed: September 13, 1996X
Alkanethiol monolayers with pendant redox centers are deposited on gold electrodes by self-assembly. The
monolayers are composed of both an electroactive thiol, HS(CH2)nC(O)NHCH2pyRu(NH3)52+/3+, with 10-
15 methylene groups, and a diluent thiol, HS(CH2)mCOOH, also with 10-15 methylene groups. The
monolayers are classified as “matched” (n ) m), “exposed” (n ) 15, m ) 10-14), and “buried” (n ) 10, m
) 11-15) according to the relative position of the redox center. Cyclic voltammograms in aqueous Na2SO4
indicate that the monolayers are close-packed with the redox centers residing in the aqueous phase in all but
the most buried cases. Measurements of electron transfer kinetics by several methods (cyclic voltammetry,
ac impedance spectroscopy, chronoamperometry) yield an internally consistent set of kinetic parameters, the
standard rate constant k°, and the reorganization energy λ of the redox centers. The reorganization energies
are in good agreement with the theoretically predicted value of 1.0 eV for the pyRu(NH3)5 redox centers.
Plots of ln(k°) vs m are linear in all three cases. The slopes of the linear regression fit provide tunneling
parameters (â, where k° ≈ e-âm) of 0.97 ( 0.03 (matched cases), 0.83 ( 0.03 (exposed cases) and 0.16 (
0.02 (buried cases) per methylene. This pattern of â’s is interpreted in terms of electronic coupling between
the redox center and the electrode via both the redox thiol and the proximate diluent thiols, with the coupling
via the diluent thiols dominating in the exposed cases.
Introduction
the potential dependence of the rate constant at constant
temperature and the temperature dependence of the standard
rate constant k° (defined at η ) 0 V).
Electroactive or blocking self-assembled monolayers (SAMs)
of alkanethiols on electrodes have proven to be potent tools for
studying long-range electron transfer between an electrode and
redox centers at a controlled distance from the electrode. In a
landmark paper, Chidsey1 demonstrated that a self-assembled
monolayer of alkanethiols with a low density of attached
ferrocenes could be used to probe electron transfer kinetics as
a function of both overpotential η (the potential relative to the
formal potential of the redox center) and temperature T. The
data fit a Marcussian model in which, for example, the cathodic
rate constant depends on the overlap of metal occupied states
and the redox center acceptor states:1
Since Chidsey’s report, a number of researchers have
examined the electron transfer kinetics of redox centers
bonded, electrostatically adsorbed, or freely diffusing to a tightly
packed alkanethiol monolayer (see Table 1 in the Supporting
Information).3-20 Rate constants extracted from a variety of
electrochemical methods (chronoamperometry (CA), cyclic
voltammetry (CV), ac impedance spectroscopy (ACIS)) are
converted to reorganization energies via Tafel plots, fits of peak
potentials and peak currents to theoretical working curves,
conversion of linear sweep voltammograms to k vs η plots, and
Arrhenius plots (ln(k°) vs 1/T). All of the data analyses are
based on versions of eq 1. For aqueous electrolytes, the
experimental λ’s are in good agreement with the λ’s predicted
by the dielectric continuum model:
kcat ) A n(E,T)D (E,T,η,λ) dE
(1)
∫
ox
n(E,T) is the Fermi function for occupied states in a continuum
of states in the metal and Dox is the density of states for the
oxidized form of the redox centers. The parameters E and λ
are the energy with respect to the Fermi energy of the metal
and the reorganization energy of the redox center. Dox is a
Gaussian function whose width and position with respect to the
formal potential are both a function of λ. This model is based
on several assumptions; the density of states in the metal is
independent of the energy, the tunneling probability function
is essentially independent of energy of electron transfer2 (and
hence the electrode potential), and double layer effects are
negligible. Equation 1 yields theoretical Tafel plots (ln(k) vs
η) which are symmetrical (equal slope magnitude at the same
positive and negative overpotential) and convergent at large
overpotentials (η > λ) to rates that are independent of both the
overpotential and the temperature. The experimental rate
constants for ferrocene oxidation and ferricenium reduction fit
the model.1 The fit yielded the reorganization energy from both
λ ≈ λos ) (e2NA/2a)(1/ꢀop - 1/ꢀs)
os is the outer-sphere component of the reorganization energy.
(2)
λ
Since the inner-sphere component is on the order of 0.1 eV for
most redox centers examined, the total λ is approximately equal
to λos. NA is Avogadro’s constant, a is the radius of the redox
center, and ꢀop and ꢀs are the optical and static dielectric
constants of the electrolyte. Equation 2 assumes that the image
charge effects are negligible and that the redox center is
completely solvated by the electrolyte. All of the monolayers
thus far have been designed to fix the locus of the redox center
at the external surface of the monolayer where, apparently, they
are fully solvated by the contacting electrolyte. The precision
of the experimental λ’s is as yet insufficient to test more
sophisticated theoretical predictions of λ based on an inhomo-
geneous environment.21 In nonaqueous electrolytes, the ex-
perimental λ’s agree less well with the predictions of eq 2,
presumably because of the increased disorder in the monolayer
in those electrolytes.11,13,14,17
X Abstract published in AdVance ACS Abstracts, November 1, 1996.
S0022-3654(96)02831-6 CCC: $12.00 © 1996 American Chemical Society

Alkanethiols with Attached Ruthenium(II/III) Redox Centers
J. Phys. Chem., Vol. 100, No. 48, 1996 18853
orientations. In the matched case, the redox center projects
slightly into the electrolyte because of the extra NHCH2 linkage.
Even in the most buried case examined, part of the redox center
is exposed to the electrolyte. In all cases, hydrogen-bonding
interactions are possible between the coordinated amines on the
redox couple and the terminal carboxyls on the diluent chains.
Three electrochemical methods are used to measure the standard
rate constant, while λ is obtained from both Tafel and Arrhenius
plots. The importance of proper iR drop correction in chrono-
amperometry experiments is emphasized. The apparent â’s for
the three cases demonstrate the presence of multiple electronic
coupling paths for electron tunneling. In particular, the coupling
paths via the diluent thiol chains can be dominant even with
the presence of a noncovalent step in the path.
Experimental Section
Syntheses. Chain lengths of 10, 11, and 15 methylenes for
the HS-Cm-COOH and the corresponding Ru derivatives were
prepared from commercially available starting materials.15,28
A
general scheme for synthesizing the intervening chain lengths
is shown below:29
Br-Cn-Br + 2CH2dCHCH2MgCl f
CH2dCHsCn+2sCHdCH2 + 2MgClBr (i)
Figure 1. Relative chain lengths and orientations for HS-C15-Ru (top),
HS-C10-Ru (bottom), and HS-Cm-COOH (m ) 10-15) in a well-
oriented monolayer.
-
B2H6 H2O2/OH
CH2dCHsCn+2sCHdCH2
8
8 HO-Cn+6-OH
(ii)
By systematically varying the number of methylene spacers
n in the alkanethiol (both the electroactive thiol and the diluent
thiol), the rate of electron transfer can be examined as a function
of distance between the redox center and the electrode. Since
all tunneling models for these systems predict an exponential
decay of the rate constant with distance, a plot of ln(k°) vs n is
fit to a straight line whose slope yields â, the tunneling
parameter. For alkanethiol spacers, â has been found to be 1.1
( 0.1 per CH2 independent of the redox center (which includes
ferrocene, Ru(NH3)5py, cytochrome c, and freely diffusing redox
couples).3,4,6,9,14,15,19 Assuming that the alkanethiol SAM adopts
an all-trans conformation of the hydrocarbon chain with an
average tilt of 30° from the surface normal,22 then â can be
HO-Cn+6-OH + HBr f Br-Cn+6-OH
Br-Cn+6-OH + CrO3 f Br-Cn+5-COOH
(iii)
(iv)
-
CH3COS
OH
Br-Cn+5-COOH
8
-8 HS-Cn+5-COOH (v)
HS-Cn+5-COOH + NH2CH2pyRu(NH3)5 DCC8
2+
HS-Cn+5-Ru (vi)
Commercially available alkane dibromides with 7-9 methylenes
were reacted with 2 equiv of allyl magnesium chloride to
produce the diene. The diene was converted to the R,ω-diol
via hydroboration followed by oxidation. A two-phase (octane-
water) reaction with HBr yielded the monobromo derivative.
Jones oxidation with chromate afforded the ω-bromocarboxylic
acid. The bromide was converted to the thiol by reaction with
thioacetate followed by base hydrolysis. Finally, the desired
Ru compound was synthesized by dicyclohexylcarbodiimide-
catalyzed coupling of the ω-thiol carboxylic acid with pen-
taammine N-(4-(aminomethyl)pyridine)Ru(II) (prepared by re-
ducing pentaamminechloroRu(III) in the presence of excess
4-(aminomethyl)pyridine). Confirmation of the desired structure
for all but the last step was provided by NMR spectra, and
occasionally GC-MS, of the product. Details of the syntheses
and characterizations are provided in the Supporting Information.
Monolayer Deposition. All monolayers were deposited on
a polycrystalline gold ball electrode (geometric area 0.13 cm2)
attached to a fine gold wire. Prior to each monolayer deposition,
the gold electrode was cleaned by heating to incandescence in
converted to 1.1 Å-1 23 We15 have argued that this value of â
.
is more consistent with through-bond rather than through-space
coupling of electron tunneling. For example, Liang and Newton
have calculated â’s between 0.7 and 1.2 per CH2 for an all-
trans alkane spacer.25 It is interesting to note that â’s more
indicative of through-space tunneling have been observed for
redox centers linked to the electrode via short-chain (1-5
carbons) sulfide derivatives26 (1.45 Å-1) and bis(4-pyridyl)
derivatives27 (0-3 methylenes) (1.6 Å-1). It is unclear whether
there is a change in tunneling mechanism for the shorter
distances, since â remains close to 1.0 Å-1 for alkanethiol SAMs
with pendant ferrocenes with as few as 5 methylenes in the
chain.6
The focus of this work is the evaluation of thermodynamic
and kinetic behavior of mixed monolayers containing
HS(CH2)nC(O)NHCH2pyRu(NH3)52+/3+ (abbreviated HS-Cn-Ru,
where n ) 10-15) and HS(CH2)mCOOH (abbreviated HS-Cm-
COOH, where m ) 10-15) in aqueous electrolyte. The mixed
monolayers permit the formation of a densely packed SAM
while maintaining a low density of the ruthenium redox centers.
The monolayers are categorized according to the location of
the redox center: matched (n ) m), exposed (n ) 15, m )
10-14) and buried (n ) 10, m ) 11-15). Figure 1 illustrates
the relative dimensions and location of the redox center for the
three cases using normal assumptions about bond angles and
a gas-air flame. Deposition solutions were typically 10-4
M
total thiol in acetonitrile, with the molar ratio of HS-Cn-Ru to
HS-Cm-COOH being 1:2 or 1:3. The initial deposition lasted
15-20 min, whereupon the electrode was rinsed with acetoni-
trile and water and then examined by cyclic voltammetry. The
gold electrode was subsequently immersed for 1 min in a 10-4

18854 J. Phys. Chem., Vol. 100, No. 48, 1996
Finklea et al.
M acetonitrile solution of the corresponding diluent thiol at 45-
50 °C. The last step was repeated if a CV revealed nonideal
behavior of the monolayer (see below).
Electrochemistry. All electrochemical experiments were
performed in a two-compartment cell at room temperature (ca.
22 °C). The working gold electrode was centered with respect
to a cylindrical coil gold counter electrode. A wide bore (2
mm) Luggin capillary with an unconnected gold wire30 provided
an electrolyte path from the SCE reference electrode to a point
ca. 1 mm below the gold ball. All experiments were performed
in 1 M Na2SO4 adjusted to pH 4 with H2SO4. The uncompen-
sated resistance (measured by ACIS) was typically 10-20 Ω.
Cyclic voltammograms were acquired between +0.5 and
-0.5 V vs SCE at scan rates from 0.1 to 2 V/s using a
Bioanalytical Systems CV-27 potentiostat. Faster scan rates
were avoided due to distortions introduced into the CV by the
relatively slow time constant of the current follower circuit in
this potentiostat.
Chronoamperometry experiments were performed using a
Princeton Applied Research Model 273 potentiostat, a function
generator, and a Tektronix Model 2232 100 MHz digital storage
oscilloscope. The gold electrode was subjected to a square wave
of amplitude 2η centered on the formal potential of the Ru redox
center. The frequency of the square wave was adjusted to allow
the current to decay to background levels before the next
potential step. To lower the noise level, 32 current transients
were co-added for each potential step.
Figure 2. Cyclic voltammograms taken during the deposition of a HS-
C14-Ru/HS-C14-COOH SAM. All scans are at 0.1 V/s. The CVs are
offset for clarity. (A) After initial deposition in an acetonitrile solution
containing HS-C14-Ru and HS-C14-COOH in a 1:2 mole ratio. (B) After
immersion of the electrode in a warm acetonitrile solution of HS-C14-
COOH for 1 min. (C) After a second immersion in the warm
acetonitrile solution of HS-C14-COOH for 1 min.
TABLE 1: Mean Reversible and Quasi-Reversible CV
Potential Parameters for HS-Cn-Ru/HS-Cm-COOH SAMsa
n
m
E°′ (V vs SSCE)
∆Ep (mV)
∆Efwhm (mV)
Matched
Ac impedance experiments also used the Model 273 poten-
tiostat, a PAR Model 5210 lock-in amplifier, and PAR M388
software. The gold electrode was subjected to a dc potential
equal to the formal potential of the Ru redox center and an ac
potential of 3 mV amplitude and frequencies between 0.1 and
105 Hz. Below 5 Hz, the M388 software generates a multifre-
quency potential waveform and uses the Fourier transform to
extract impedance data from the current signal.
10
11
12
13
14
15
10
11
12
13
14
15
-0.023
-0.024
-0.024
-0.024
-0.020
-0.020
3
6
4
14
31
76
98
96
98
101
102
126
Exposed
Buried
15
15
15
15
15
10
11
12
13
14
-0.022
-0.021
-0.020
-0.017
-0.017
5
3
10
22
52
96
95
94
102
122
Results and Discussion
Monolayer Formation. The method for depositing the
mixed SAM’s is designed to maximize order in the monolayer.
In a reversible CV,31 the diagnostics for a highly ordered SAM
are a minimal and flat charging current baseline, and zero peak
splitting and 90 mV peak half-width for the faradaic current
peaks of a 1-electron redox center. Figure 2 shows CV’s
obtained at several stages during monolayer deposition. The
10
10
10
10
10
11
12
13
14
15
-0.024
-0.023
-0.012
-0.003
-0.007
2
3
3
3
2
98
98
102
102
109
a The scan rate is 0.1 V/s in all CV’s. The reference electrode is a
sodium-saturated calomel electrode. Each value is the average of at
least two independently prepared SAMs. The uncertainty of the
potential parameters in a single measurement is (4 mV; the standard
deviation of repeat measurements is no more than 8 mV and is typically
less than 5 mV.
first CV (after the initial deposition) contains two waves, the
2+ 3+
expected wave near 0 V vs SCE due to the pyRu(NH3)5
/
redox center, and a second wave near +0.2 V vs SCE. The
latter wave is believed to be caused by either a dipyridyl Ru
complex or a Ru dimer generated during the synthesis of the
NH2CH2pyRu(NH3)5 complex.32 Effective removal of the
impurity wave is achieved by soaking the electrode in warm
acetonitrile solution of the diluent thiol. The rate of loss of the
impurity wave greatly exceeds the rate of loss of the desired
pyRu(NH3)5 wave; also, charging current is reduced to a
minimum by the annealing step (Figure 2). Coverages are
typically adjusted to 3-6 µC/cm2 (15-30% of maximum
possible coverage) to minimize iR drop and double layer effects
while maintaining sufficient faradaic current for data analysis.
Reversible Voltammetry. Aqueous sodium sulfate adjusted
to pH 4 provides an electrochemical window extending 0.5 V
in both the positive and negative directions relative to the formal
potential of the Ru redox centers; hence, reorganization energies
can be evaluated for both the reduced and the oxidized form of
the redox centers with equal accuracy. At this pH, the prevailing
evidence in the literature indicates that the terminal carboxylic
acids are largely un-ionized.33-35 The high ionic strength is
intended to minimize double layer effects, i.e., variations in the
electrostatic potential at the plane of the redox centers as a
function of the electrode potential and the oxidation state of
the redox centers. Theoretical treatments36,37 indicate that, for
cationic redox centers, double layer effects lead to a broadening
of the reversible CV and a positive shift of the peak potentials.
The peak broadening is not evident in the HS-Cn-Ru/HS-Cm-
COOH SAMs (Figure 2 and Table 1); matched or exposed
SAMS for which the standard rate constant is greater than ca.
10 s-1 exhibit peak half-widths less than 100 mV. It is possible
that stronger ion pairing of the Ru(III) centers could lead to a
smaller change in the surface charge density during the redox
process and hence an absence of peak broadening. In any event,
the nearly ideal peak splitting and peak half-widths demonstrate
that the redox centers are thermodynamically homogeneous; the

Alkanethiols with Attached Ruthenium(II/III) Redox Centers
J. Phys. Chem., Vol. 100, No. 48, 1996 18855
entire population of redox centers exhibits a single formal
potential. As noted previously,28 the formal potential is within
60 mV of the formal potential (+0.04 V vs SCE) of a solution
analog at a bare gold electrode. The redox centers are believed
to be fully solvated by the aqueous electrolyte.
For the buried cases, significant positive shifts in the formal
potential appear when the diluent chain is more than 2
methylenes longer than the redox chain (Table 1); also, the peak
half-width broadens noticeably. The first effect can be rational-
ized by two mechanisms: either the effective dielectric constant
around the redox center is decreasing as the redox center moves
toward the hydrocarbon domain (which decreases the stability
of the Ru(III) state relative to the Ru(II) state),38,39 or double
layer effects are becoming more conspicuous.36 The latter
mechanism also accounts for the peak broadening.
Kinetic Measurements with iR Drop Correction. The
standard rate constant is obtained from all three kinetic
measurement methods. Peak splittings from fast scan CV’s are
converted to k° values using the working curve given by
Laviron40 for a transfer coefficient of 0.5. For peak splittings
less than 200 mV, the working curve is relatively insensitive to
the potential dependence of the transfer coefficient predicted
by Marcus theory.7,10 ACIS data are corrected for the uncom-
pensated resistance and interfacial capacitance. At zero over-
potential, a plot of the cotangent of the phase angle of the
faradaic current vs the radial frequency (cot(γ) vs ω) is linear
with a slope of 1/(2k°).41
Figure 3. Tafel plots of three matched SAMs. Open squares are rate
constants for 50% conversion and solid triangles are rate constants for
90% conversion to the final oxidation state. Solid and dashed lines
are theoretical fits to the data for 50 and 90% conversion, respectively.
The relevant parameters are (top to bottom): HS-C12-Ru/HS-C12-
COOH, anodic branch (λ(50%) ) 1.1 eV, k° ) 26, λ(90%) ) 0.8 eV,
k° ) 26), cathodic branch (λ(50%) ) 1.3 eV, k° ) 29, λ(90%) ) 0.8
eV, k° ) 29); HS-C13-Ru/HS-C13-COOH, anodic branch (λ(50%) )
1.2 eV, k° ) 13, λ(90%) ) 1.1 eV, k° ) 10), cathodic branch (λ(50%)
) 1.1 eV, k° ) 9.0, λ(90%) ) 1.1 eV, k° ) 8.6); HS-C14-Ru/HS-C14-
COOH), anodic branch (λ(50%) ) 1.0 eV, k° ) 3.9, λ(90%) ) 1.0
eV, k° ) 3.2), cathodic branch (λ(50%) ) 1.0 eV, k° ) 4.5, λ(90%) )
1.0 eV, k° ) 4.1).
In CA (and, to a lesser extent, CV) kinetic measurements, iR
drop causes the true overpotential to deviate from the applied
overpotential:
TABLE 2: Average Kinetic Parameters for HSCnRu/
HSCmCOOH SAMs in Aqueous Electrolytesa
k° (s-1
)
η(t) ) ηappl-i(t)Ru
(3)
CAb
λ (eV)
50%
90%
CAb
where Ru is the uncompensated resistance. iR drop distorts the
rate constant data and adds an extra charging current component
not readily extrapolated from the pure charging current domains.
CA transients are corrected for iR drop using the following
equation:17,42
n
m
an cat an cat ACISb CVb 50% 90% Arrheniusc
Matched Cases
10 10 220 200 150 130
160 130 1.1 0.8
58 52 1.1 0.9
32 34 1.2 0.9
12 10 1.2 1.1
0.7
0.7
11 11 67
12 12 26
13 13 12
74
30
10
6
41
26
10
3
54
26
10
4
k(t) ) (1 - (i(t + ∆t)/i(t))/(1 + R(η(t + ∆t) -
η(t))/0.0257))/∆t (4)
14 14
4
6
1
4
1
1.0 1.0
1.0 0.9
15 15 1.5 1.0 1.0 0.7
0.7
Exposed Cases
The current transient and applied potential transient are both
digitized and smoothed; smoothing is necessary since eq 4 in
effect contains time derivatives of the overpotential and current.
An iR drop-corrected overpotential transient is obtained using
eq 3. The transfer coefficient R is dependent on η, but digital
simulations show that a constant value of 0.4 provides a
reasonable correction for iR drop in the accessible range of
overpotentials.
We arbitrarily record η(t) and k(t) at times corresponding to
50 and 90% conversion of the redox centers to the final
oxidation state (as determined by integration of the current
transient). Figure 3 exhibits Tafel plots for three matched SAMs
with 12, 13, and 14 methylenes. It is clear that the data at 50
and 90% conversion represent populations of redox centers with
faster and slower rate constants, respectively, a much-discussed
phenomenon known as kinetic heterogeneity.10,12,17 Other
manifestations of kinetic heterogeneity include unusually broad
CV peaks under fast scan conditions and curvature in the ACIS
plots of cot(γ) vs ω.16 Since the redox centers appear to be
thermodynamically homogeneous, the kinetic heterogeneity is
attributed to variations in the electronic coupling and reorga-
nization energies within the population of redox centers due to
diverse chain conformations and loosely packed domains; other
possible causes include the presence of lattice steps and different
15 10 81
15 11 22
15 12 12
74
20
14
6
48
18
8
3
1.5
54
17
10
4
62 50 1.1 0.8
24 25 0.9 0.8
14 16 0.9 0.9
6
1
0.5
0.5
15 13
15 14
4
3
6
0.9 0.9
3
1
1.5 0.9 0.9
Buried Cases
10 11 130 120 100 100
105 82 1.1 0.7
93 88 0.8 0.6
90 99 1.1 0.7
57 67 0.8 0.5
74 70 1.2 0.8
0.6
0.5
10 12 140 110 110
90
67
52
67
10 13 130
10 14 89
10 15 81
80
70
90
99
60
68
a Each datum is the average of at least two independently prepared
SAMs. b Electrolyte 1 M Na2SO4. c From Arrhenius plots based on k°
values from CV’s in 2 M NaCl, 0-40 °C.
crystallographic faces on the electrode surface.42 Data points
for both branches and both conversion %’s are fitted separately
with Marcus theory lines from eq 1 to yield k° (by extrapolation)
and λ. The uncertainties in the resulting λ’s are estimated to
be (0.1 eV.43
Because of kinetic heterogeneity, standard rate constants from
ACIS plots are obtained from the limiting slope at the lowest
frequencies.
Electron Transfer Kinetics. Table 2 summarizes the
average k° and λ data for all cases and all electrochemical

18856 J. Phys. Chem., Vol. 100, No. 48, 1996
Finklea et al.
methods. The data are averages of at least two independent
experiments for each combination of chain lengths. The
standard rate constants are reproducible within a factor of 2
and the λ’s within 0.1 eV for repeat experiments with a given
combination of chain lengths.
We consider first the apparent reorganization energies. For
a given % conversion, the same λ fits both the anodic and the
cathodic branches of the Tafel plot within experimental uncer-
tainty, so only one value is shown. As noted in the Introduction,
this observation implies that the tunneling parameter is inde-
pendent of the electron tunneling energy. In the matched and
exposed cases, the λ values generally fall in the 0.9-1.1 eV
range. Values of λ are higher than previously reported15,16
because iR drop systematically distorts the Tafel plots toward
greater curvature. For the pyRu(NH3)5 redox center in water,
the predicted value of λ based on eq 2 is 1.0 eV (0.92 eV for
λos), in good agreement with the experimental measurements.
The experimental uncertainty masks any change less than ca.
0.2 eV of λ with respect to chain length, so the predictions of
Liu and Newton21 cannot be tested.
Figure 4. Tunneling parameter plots for the matched (circles), exposed
(squares), and buried (triangle) cases. The symbols are the average
values for measurements on several SAMs, and the horizontal lines
mark the high and low values. A linear regression line for the average
values is shown for each case.
the formal potential, the reorganization energy, and the tunneling
distance on the apparent kinetic parameters obtained by CV and
by Tafel plots. Dispersion in the formal potential or the
reorganization energy can result in anomalously high k° values
and low λ values, while dispersion in the tunneling distance
only affects the apparent k° values. Since there is no dispersion
of the formal potential in our system, we conclude that the
anomalies in λ values imply a real dispersion in λ’s for the redox
centers. However, there is still no satisfying explanation for
the discrepancy between the Tafel and Arrhenius λ values.
In general, standard rate constants from CV and ACIS
measurements are in reasonable agreement with the extrapolated
k° values from the Tafel plots (Table 2). Even with the spread
of k° values caused by both kinetic heterogeneity and variations
between SAMS of identical composition, it is clear that the main
factors that determine k° are the chain lengths of the redox and
diluent thiols. Consequently, all of the standard rate constants
for a given composition are combined into a single ln(k°) vs m
plot (Figure 4); average, high, and low values are shown. A
linear regression line based on the average k°’s is included for
each case. A linear dependence of ln(k°) with m is expected
for the matched case; the other lines are included primarily for
comparison.
There is greater disparity between the λ’s at 50 and 90%
conversion for the buried cases. Also, the λ’s at 50% conversion
for the C12 and C14 diluent chain lengths are significantly smaller
than the values for the surrounding chain lengths. We interpret
these results as evidence of a greater heterogeneity of environ-
ments around the redox centers, with a possible correlation of
that heterogeneity with diluent chain length, as the redox center
becomes embedded in the outer portions of the monolayer. The
cause of the smaller λ’s for the even diluent chain lengths is
not clear.45 There is no systematic decrease in λ with increasing
diluent chain length, as would be expected if the redox center
was moving into a less polar environment. Because of the
hydrophilicity of both the redox center and the terminal groups
on the diluent chains, it is conceivable that local pockets of
high polarity due to retained water and ions are created around
the redox centers as long as the redox center is not buried too
deeply.
In most cases, the population of redox centers with faster
rate constants (50% conversion) appear to have the larger λ’s
than the slower population (90% conversion). This trend is in
contradiction to expectations based on theory. A larger rate
constant implies greater electronic coupling, which in turn
suggests that the redox centers are closer to the electrode surface
or in better contact with the SAM. A closer redox center should
have a smaller reorganization energies because of image charge
effects and the less polar monolayer environment. Yet the data
indicates that the faster population of redox centers is located
in a more polar environment, i.e., out in the electrolyte.
Resolution of this contradiction will require a better understand-
ing of the factors controlling the standard rate constant as well
as the reorganization energies.
The slope of the line for the matched cases yields a â of
0.97 ( 0.03 per CH2, a value in excellent agreement with the
earlier reported slope15 and in good agreement with other
reported values for â using alkane chain spacers in SAMs.3,4,6,9,14,19
Extrapolation to m ) 0 provides a limiting k° of 3((1) × 106
s-1. With the same assumptions as Smalley et al.6 (m ) 0
corresponds to the plane of closest approach, â ≈ 1 Å-1, λ
invariant with distance), this number can be converted into an
estimate of the heterogeneous rate constant at the bare electrode
(k°het ≈ k°/â ) 3((1) × 10-2 cm/s). While a double-layer-
2+/3+
corrected value of the heterogeneous k° for pyRu(NH3)5
is not available, several reports indicate that k° for the smaller
is about 1 cm/s.48-50 Considering that the
2+/3+
Another anomaly is apparent in the values of λ determined
from Arrhenius plots. While the data set is not complete, and
the measurements are in a different electrolyte (which has no
effect on the kinetic parameters16), the available Arrhenius λ’s
are significantly smaller than the λ values from the Tafel plots.
λ’s for ferrocene redox centers do not show a discrepancy
between Tafel and Arrhenius measurements.1,5 We do not have
an explanation for this anomaly. It may be due to a shortcoming
in the theory;46 for example, Beratan et al.47 posit a temperature
dependence of the coupling factor due to molecular motion in
the bridging groups.
Ru(NH3)6
extrapolation corresponds to a redox center separated from the
electrode by a thiol, an amide, and a methylene, the limiting k°
is consistent with expectations.
The respective â’s for the exposed and buried cases are 0.83
( 0.03 per CH2 and 0.16 ( 0.02 per CH2. The large â for the
exposed cases is surprising if the dominant coupling path for
electron transfer is anticipated to be along the chain connecting
the redox center to the electrode. The increased motional
freedom of the component of the chain extending beyond the
close-packed domain defined by the diluent thiol would result
in higher degree of gauche defects. Coupling through gauche
conformations should decrease the rate constant;51-53 hence k°
should decrease as the diluent thiol chain shortens. Instead,
It is instructive to consider the possibility that kinetic
heterogeneity may be the root of the preceding anomalies.
Murray and co-workers12 simulated the effects of dispersion in

Alkanethiols with Attached Ruthenium(II/III) Redox Centers
J. Phys. Chem., Vol. 100, No. 48, 1996 18857
the pattern of â’s can be interpreted in terms of tunneling
coupling by multiple paths, i.e., via both the redox thiol chain
and the diluent thiol chains. In the exposed cases, chain motion
allows the redox centers to contact the terminus of nearby diluent
thiols long enough for electron transfer to occur. The coupling
via the diluent thiol must dominate in the exposed case in order
to account for the sensitivity of the rate constants to the diluent
chain length. The â value is diminished from the matched case
because the statistical fraction of redox centers actually in
contact with the adjacent diluent thiol chains at any time
decreases in proportion to the degree of mismatch between the
redox and diluent chains. In the buried cases, the degree of
contact between the redox center and the adjacent diluent thiols
is relatively insensitive to the diluent chain length. There is
little variation of coupling along either the redox chain or the
adjacent diluent chains, and hence â is small. It is also possible
that the apparent rate constant is partially dependent on the
mobility of the counterions in the most buried cases. In none
of the cases does the data scatter permit the observation of any
“odd-even” effect on â in either the redox chain length or the
diluent chain length.
A large â for the exposed cases and a small â for the buried
cases is interpreted in terms of multiple paths of electronic
coupling between the redox center and the electrode. Coupling
via the diluent thiol, with one through-space step, is believed
to be the dominant path in the exposed case.
Acknowledgment. This work was partially supported by a
grant from the National Science Foundation (CHE-9114626).
Dr. Daniel Lo¨wy’s assistance in the preparation of the manu-
script is gratefully acknowledged.
Supporting Information Available: Description of the
syntheses of the ω-thiol-alkanecarboxylic acids and the corre-
sponding pyRu(NH3)5 derivatives, and a table summarizing
previous kinetic parameters obtained for self-assembled mono-
layers with either attached or freely diffusing redox couples (6
pages). This material is contained in many libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, can be ordered from the ACS, and can
be downloaded from the Internet; see any current masthead page
for ordering information and Internet access instructions.
Another factor in the overall rate is the degree of coupling
for the noncovalent interaction of the pyRu(NH3)5 redox centers
and the terminal COOH groups of the diluent thiols. Prelimi-
nary estimates suggest that â (in units of Å-1) for through-
space coupling is a factor of 3 larger than for through-bond
coupling;47,54 however, a more recent calculation by Liang and
Newton25 indicates that van der Waals contacts can lead to â’s
of the same magnitude as the through-bond couplings. We
surmise that hydrogen-bonding interactions between the diluent
thiol COOH groups and the coordinated amine groups may
contribute to the coupling.54,55 The magnitude of the nonco-
valent coupling cannot be extracted from this data set because
of the uncertain effects of chain mobility on the overall rate.
These results indicate that care will be needed when designing
SAMs to test the effects of molecular structure on electron
transfer rates. In particular, having a redox center tethered to
the electrode by one type of chain (e.g., an extended π system,
or a rigid rod structure) and using linear alkanes as a packing
component is a strategy to be avoided if unambiguous inter-
pretation of tunneling parameters is intended. On the other hand,
if, as the data suggest, the monolayer can be constructed so
that a single noncovalent step forms part of the electronic
coupling path between the redox center and the electrode, then
it should be possible to explore the effects of the essential
noncovalent interactions (van der Waals, dipole, hydrogen
bonds) on the electronic coupling through judicious choice of
the redox center and the terminal group on the diluent thiol.
References and Notes
(1) Chidsey, C. E. D. Science 1991, 251, 919-22.
(2) Evidence supporting this assumption has been reported.3,4
(3) Becka, A. M.; Miller, C. J. J. Phys. Chem. 1992, 96, 2657-68.
(4) Xu, J.; Li, H.-L.; Zhang, Y. J. Phys. Chem. 1993, 97, 11497-500.
(5) Dubois, L. H.; Nuzzo, R. G. Annu. ReV. Phys. Chem. 1992, 43,
437-63.
(6) Smalley, J. F.; Feldberg, S. W.; Chidsey, C. E. D.; Linford, M. R.;
Newton, M. R.; Liu, Y.-P. J. Phys. Chem. 1995, 99, 13141-9.
(7) Weber, K.; Creager, S. E. Anal. Chem. 1994, 66, 3164-72.
(8) Hockett, L. A.; Creager, S. E. Langmuir 1995, 11, 2318-21.
(9) Weber, K.; Hockett, L. A.; Steiger, S.; Creager, S. E. Presented at
the 188th Electrochemical Society Meeting, Chicago, IL, Oct. 8-13, 1995,
Abstract No. 1024.
(10) Tender, L.; Carter, M. T.; Murray, R. W. Anal. Chem. 1994, 66,
3173-81.
(11) Richardson, J. N.; Peck, S. R.; Curtin, L. S.; Tender, L. M.; Terrill,
R. H.; Carter, M. T.; Murray, R. W.; Rowe, G. K.; Creager, S. E. J. Phys.
Chem. 1995, 99, 766-72.
(12) Rowe, G. K.; Carter, M. T.; Richardson, J. N.; Murray, R. W.
Langmuir 1995, 11, 1797-806.
(13) Richardson, J. N.; Rowe, G. K.; Carter, M. T.; Tender, L. M.;
Curtin, L. S.; Peck, S. R.; Murray, R. W. Electrochim. Acta 1995, 40, 1331-
8.
(14) Carter, M. T.; Rowe, G. K.; Richardson, J. N.; Tender, L. M.;
Terrill, R. H.; Murray, R. W. J. Am. Chem. Soc. 1995, 117, 2896-9.
(15) Finklea, H. O.; Hanshew, D. D. J. Am. Chem. Soc. 1992, 114,
3173-81.
(16) Finklea, H. O.; Ravenscroft, M. S.; Snider, D. A. Langmuir 1993,
9, 223-7.
(17) Ravenscroft, M. S.; Finklea, H. O. J. Phys. Chem. 1994, 98, 3843-
50.
(18) Song, S.; Clark, R. A.; Bowden, E. F.; Tarlov, M. J. J. Phys. Chem.
1993, 97, 6564-72.
(19) Bowden, E. F. Presented at the 188th Electrochemical Society
Meeting, Chicago, IL, Oct. 8-13, 1995; Abstract No. 857.
(20) Terrettaz, S.; Becka, A. M.; Traub, M. J.; Fettinger, J. C.; Miller,
C. J. J. Phys. Chem. 1995, 99, 11216-24.
(21) Liu, Y.-P.; Newton, M. D. J. Phys. Chem. 1994, 98, 7162-9.
(22) Finklea, H. O. Electroanal. Chem. 1996, 19, 109-335.
(23) A similar value for â is found for electron hopping between gold
clusters coated with alkanethiol SAMs.24
(24) Terrill, R. H.; Postlethwaite, T. A.; Chen, C.; Poon, C.-D.; Terzis,
A.; Chen, A.; Hutchison, J. E.; Clark, M. R.; Wignall, G.; Londono, J. D.;
Superfine, R.; Falvo, M.; Johnson, C. S.; Samulski, E. T.; Murray, R. W.
J. Am. Chem. Soc. 1995, 117, 12537-48.
Summary
A general method is given for the synthesis of arbitrary chain
lengths for molecules of the form HS-Cn-COOH; three previ-
ously unavailable chain lengths are synthesized to fill in the
gap in previously studied monolayer systems. Mixed mono-
layers with pendant Ru redox centers and diluent thiols are
prepared and their electron transfer kinetics measured by three
electrochemical methods. The monolayers appear to be well
ordered, with close-packed alkane chains and the redox centers
residing in or near the aqueous phase. Reorganization energies
from Tafel plots show no systematic trend with chain length
and are reasonably close to values predicted by the dielectric
continuum model for the pyRu(NH3)5 redox center. For
monolayer spacers with 10-15 methylenes, standard rate
constants decrease exponentially with chain length at a rate
consistent with through-bond coupling of the electron tunneling.
(25) Liang, C.; Newton, M. D. J. Phys. Chem. 1993, 97, 3199-211.
(26) (a) Li, T. T.-T.; Liu, H. Y.; Weaver, M. J. Am. Chem. Soc. 1984,
106, 1233-9. (b) Li, T. T.-T.; Weaver, M. J. J. Am. Chem. Soc. 1984,
106, 6107-8.
(27) Forster, R. J.; Faulkner, L. R. J. Am. Chem. Soc. 1994, 116, 5444-
52.
(28) Finklea, H. O.; Hanshew, D. D. J. Electroanal. Chem. 1993, 347,
327-340.

18858 J. Phys. Chem., Vol. 100, No. 48, 1996
Finklea et al.
(29) L. Liu, M. S. thesis, West Virginia University, 1994.
(30) Fletcher, S.; Horne, M. D. J. Electroanal. Chem. 1991, 279, 297-
9.
(31) CV’s for surface-bound redox centers are reversible at room
temperature if the standard rate constant in s-1 is more than 100 times larger
than the scan rate in V/s.
(32) Early efforts (crystallization, column chromatography) to remove
this impurity from the desired complex were not successful. More recently,
we have succeeded in separating the impurity and product using column
chromatography on reverse-phase C18 silica with 0.2% NH4PF6, adjusted
to pH 2.5 with CF3COOH.
(44) For a single Tafel plot, visual fits of the theoretical Tafel lines to
the data points have an uncertainty of (0.1 eV. Because of scatter in the
data points, the same uncertainty is obtained when a “least-squares”
minimization algorithm is applied to achieve an optimum fit between data
and theory.
(45) A parity rule has been invoked to account for large changes in
electronic coupling with the insertion of one bond in a multistrand link
between an electron donor and acceptor; see: Shephard, M. J.; Paddon-
Row, M. N. J. Phys. Chem. 1995, 99, 17497-500. Whether the parity
rule can be applied to a system consisting of many unlinked single strands
remains to resolved.
(33) Bain, C. D.; Whitesides, G. M. Langmuir 1989, 5, 1370-1378.
(34) Lee, T. R.; Carey, R. I.; Biebuyck, H. A.; Whitesides, G. M.
Langmuir 1994, 10, 741-9.
(46) Note that temperature is an explicit variable in just the Fermi
distribution and the density of redox donor or acceptor states; see eq 1. In
particular, the tunneling parameter â is assumed to be independent of
temperature.
(47) Beratan, D. N.; Onuchic, J. N.; Hopfield, J. J. J. Chem. Phys. 1987,
86, 4488-98.
(48) Gennet, T.; Weaver, M. J. Anal. Chem. 1984, 56, 1444-8.
(49) Iwasita, T.; Schmickler, W.; Schultze, J. W. Ber. Bunsenges. Phys.
Chem. 1985, 89, 138-142.
(50) Wipf, D. O.; Kristensen, E. W.; Deakin, M. R.; Wightman, R. M.
Anal. Chem. 1988, 60, 306-10.
(51) Hoffmann, R. Acc. Chem. Res. 1971, 4, 1-9.
(52) Jordan, K. D.; Paddon-Row, M. N. Chem. ReV. 1992, 92, 395-
410.
(35) Creager, S. E.; Clarke, J. Langmuir 1994, 10, 3675-83.
(36) Smith, C. P.; White, H. S. Anal. Chem. 1992, 64, 2398-405.
(37) Fawcett, W. R. J. Electroanal. Chem. 1994, 378, 117-24.
(38) Creager, S. E.; Rowe, G. K. Anal. Chim. Acta 1991, 246, 233-9.
(39) Rowe, G. K.; Creager, S. E. J. Phys. Chem. 1994, 98, 5500-7.
(40) Laviron, E. J. Electroanal. Chem. 1979, 101, 19-28.
(41) Laviron, E. J. Electroanal. Chem. 1979, 97, 135-149.
(42) Bowden has described a superior method for extracting rate
constants from current transients which does not introduce as much noise
and does not require an assumption about the transfer coefficient; see: Nahir,
T. M.; Clark, R. A.; Bowden, E. F. Anal. Chem. 1994, 66, 2595-8.
(43) A reviewer pointed out that solvent relaxation modes could be
slower when the redox center is partially buried in the monolayer and
suggested that this could contribute to the kinetic heterogeneity. However,
solvent relaxation times are anticipated to be unimportant in rate expressions
for strongly nonadiabatic electron transfer, such as is the case in this system;
see ref 17.
(53) Newton, M. D. Chem. ReV. 1991, 91, 767-92.
(54) Onuchic, J. N.; Beratan, D. N. J. Chem. Phys. 1990, 92, 722-33.
(55) de Rege, P. J. F.; Williams, S. A.; Therien, M. J. Science 1995,
269, 1409-13.
JP962831Q