Article pubs.acs.org/ac
Distance-Dependent Electron Transfer at Passivated Electrodes Decorated by Gold Nanoparticles Abbas Barfidokht, Simone Ciampi, Erwann Luais, Nadim Darwish, and J. Justin Gooding* School of Chemistry, The University of New South Wales, Sydney NSW 2052, Australia S Supporting Information *
ABSTRACT: The phenomenon of nanoparticles attached to an electrode passivated by an organic layer allowing efficient electron transfer between redox species in solution and the underlying electrode to be restored has resulted in Chazalviel and Allongue proposing a theory [Chazalviel, J.-N.; Allongue, P. J. Am. Chem. Soc. 2011, 133, 762−764] to explain this phenomenon. The theory suggests that with electrode-organic layer-nanoparticle constructs, high exchange current densities, compared with when the nanoparticles are absent, results in the rate of electron transfer being independent of the thickness of the organic layer until a threshold thickness is exceeded. Thereafter, the thicker the organic layer, the slower the rate of electron transfer. Herein we provide the first experimental data to support this theory using a single experimental system that can show the transition from thickness independent electron transfer kinetics to distant dependent kinetics. This was achieved using ethylenediamine electrodeposited on a glassy carbon electrode. Different numbers of deposition cycles were applied in order to fabricate different thicknesses of the organic film. The deposited films showed progressively greater blocking abilities toward ruthenium hexamine, as a redox active probe in solution, as the films got thicker. Electron transfer kinetics of nanoparticle-decorated surfaces showed a change from thickness independent to thickness dependent as the organic layer exceeded an average thickness of 20 Å. Electrochemical impedance spectroscopy, cyclic voltammetry, scanning electron microscopy, ellipsometry, and atomic force microscopy were used to characterize the fabricated surfaces.
T
SAM (J0) assembly. This much higher current density, J1, is because of the high density of states on the other metal, to which the potential is not applied, in this case the nanoparticle. The consequence of this is, in the metal/organic layer case, the potential drops exponentially with distance across the insulating organic layer as in eq 1.17
he generation of effective analytical devices based on nanoparticles attached to electrodes using organic molecules has recently begun to attract attention.1 Such modified surfaces provide advantageous properties such as (a) low capacitances, as most of the electrode is passivated with a nonconducting organic layer, (b) a simple way of making nanoelectrode arrays with controllable active area,2 and (c) catalytic or electrocatalytic capabilities. These features result in electrode-organic layer-nanoparticle assemblies finding application in electrochemical sensing,3 electrocatalysis,4 optoelectronics,5 and photovoltaics,6,7 all of which arise from the catalytic properties of the nanoparticles employed. However, an understanding of electron transfer through such nanoparticle assemblies is required. The main point of interest is the distance dependence of electron transfer between the organic layer and nanomaterials. We8−10 and others11−16 have recently shown that decorating electrodes, passivated by an organic layer, with metal nanoparticles restores the electrochemistry. Moreover, the rate of electron transfer was surprisingly shown to be independent of the thickness of the organic film separating the nanoparticles from the underlying bulk electrode. These studies provoked Chazalviel and Allongue to very recently report a theoretical description of the mechanism of the nanoparticle-mediated electron transfer.17 The theory suggested effective nanoparticle-mediated electron transfer is expected, with distant independent rate of electron transfer as long as exchange current density across the SAM on the metal/ SAM/metal (J1) assembly is much larger than that on a metal/ © 2012 American Chemical Society
V=
⎛ ⎞ 2kBT J ⎟⎟ sinh−1⎜⎜ q ⎝ 2J0 exp( −βd) ⎠
(1)
where V is the overpotential, T is absolute temperature, kB the Boltzmann’s constant, q the elementary charge, J the current density, β the attenuation factor, and d the thickness of the organic layer. However, in the case of the metal/organic layer/ NP constructs, there are two contributions to the potential, the electrochemical interface and the potential drop across the insulating layer, eq 2:17 V=
⎤ ⎛ J ⎞ kBT ⎡⎢ J ⎥ 2 sinh−1⎜⎜ ⎟⎟ + q ⎢⎣ J1 exp( −βd) ⎥⎦ ⎝ 2J0 ⎠
(2)
Because of the large value of J1, the second term due to the organic layer remains negligible provided J1 exp(−βd) is greater than J and J0. If such a situation exists, then the result of metal/ Received: October 11, 2012 Accepted: December 10, 2012 Published: December 10, 2012 1073
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
Article
alkanethiols8 have not yet achieved a thickness in which there is a distance dependence of nanoparticle-mediated electron transfer. Therefore, a method of deposition that allows the constructs to go from the transfer unaffected into the transfer hindered regime is required to qualitatively and/or quantitatively test the theory. Since SAMs are not achievable for long alkyl chains upon 18 carbon units, because of difficulty in obtaining longer monolayer forming molecules, in this study electrochemical deposition of ethylenediamine was achieved to build organic layers on the electrode, as illustrated in the Scheme 1. Ethylenediamine is an ideal monomer for the fabrication of a suitable matrix for nanoparticle immobilization as it can be both deposited into an electrode surface18 and attach the AuNP via the free amine groups on the polymer backbone as well. With regards to electrochemical grafting of ethylenediamine, it has been shown that the electrochemical oxidation of ethylenediamine led to the passivation of various electrode materials.18,19 More importantly, the electrochemical quartz crystal microbalance technique provided compelling evidence for the oxidation of ethylenediamine and the formation of a multilayer deposit at the electrode surface.20 It is important to recognize; however, that unlike SAMs, polymer layers are much less welldefined with regards to uniformly thick organic layers. Therefore, what we seek to achieve in this experimental study is a qualitative agreement with the Chazalviel and Allongue model. The purpose of this paper is to verify the theoretical model by using the one type of nanoparticle-decorated assembly that is capable of exhibiting both thickness dependent and thickness independent electron transfer behavior. For this purpose, glassy carbon (GC) surfaces can be first modified with ethylenediamine by applying cyclic voltammetry (CV). To build different thicknesses of the organic film, the number of CV cycles was varied. Then the distal amine groups of the deposited ethylenediamine could be used to immobilize AuNPs. The electron transfer behavior of electrode/poly(ethylenediamine)/ nanoparticle construct was assessed using ruthenium(III) hexamine in aqueous solution as redox active species. Apart from using CV, the modified surfaces were characterized by electrochemical impedance spectroscopy, atomic force microscopy, ellipsometry, and scanning electron microscopy.
organic layer/NP is an effective short circuit where the potential applied to the electrode is located on the NP. The result is a distant independent electron transfer process is expected until this condition fails, whereupon the electron transfer behavior will be distant dependent. Distant dependent behavior is expected with very small nanoparticles or thick organic layers (Figure 1). The transition from distant
Figure 1. Critical thickness of an organic film above which a layer coated with gold nanoparticles is expected to lead to a change in the voltammograms of a reversible redox system in solution. The black solid line shows the boundary between two regimes, and the equation for this line is eq 3, where λ is the reorganization energy, d the thickness of the organic layer, and a β value of 1 for saturated chains. The brown solid line shows a transfer hindered regime is expected when the thickness of organic film goes above 27.5 Å, if the NP diameter is 27 nm. Reprinted from ref 17. Copyright 2011 American Chemical Society.
independent to distant dependent electron transfer behavior in these systems (the solid black line in Figure 1) is determined by eq 3.17 J1 exp( −λ /4kBT ) exp(−βd) = J0
(3)
By estimating J1 being in the order of 10 −10 A/cm and J0 in the order of 1−10 mA/cm2, a nanoparticle-mediated electron transfer could be categorized into two different regimes (Figure 1) referred to as transfer unaffected and transfer hindered, respectively, and the boundary region (solid line in the Figure 1) between two regimes is defined for the case of J1 = J0. To date, all the reported results that explore the distance dependence of the charge transfer through nanoparticleassemblies have been conducted in the thickness range predicted by the theory to be in the transfer unaffected regime.8,14 Deviations from this behavior (to reach the transfer hindered regime) require thicker organic layers. In other words, the current methods of depositing layers such as layer-by-layer assembly of oppositely charged species14 and applying SAMs of 9
10
2
■
EXPERIMENTAL SECTION Chemicals, Solvents, and Materials. Acetonitrile from Sigma-Aldrich (HPLC grade) was distilled before use. Ethylenediamine (99% extra pure) was purchased from Acros Organics (Australia). Tetrabutylammoniumtetrafluoroborate (NBu4BF4) (dried by heating at 80 °C for 12 h prior to use), ruthenium(III) hexamine chloride (98% purity, Ru(NH3)63+),
Scheme 1. Schematic of the Fabrication of GC/Poly(ethylenediamine) and GC/Poly(ethylenediamine)/NP Surfacesa
a
AuNPs are not to scale. 1074
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
Article
water after polishing, followed by drying under a stream of nitrogen gas. AuNPs were synthesized according to the method of Frens.23 Briefly, to a 1 mM boiling solution of HAuCl4 in Milli-Q water, trisodium citrate was added while constant stirring was applied. A color change started to occur 10 s after the addition of citrate solution. The resulting aquasol had a concentration of ∼17 nM and a diameter of 27.5 ± 5 nm as determined by SEM (Figure S-1 in Supporting Information). Poly(ethylenediamine) films were grown on the GC working electrodes using electrochemical deposition by performing CV (the modified electrodes henceforth referred to as GC/ poly(ethylenediamine). The electrodeposition of poly(ethylenediamine) was conducted in a solution containing 7 mM ethylenediamine and 0.1 M NBu4BF4 in acetonitrile by sweeping the potential at 40 mV s−1 over the potential range of 0−1.5 V versus Fc/Fc+. In order to change the thickness of the deposited film, different numbers of scans (x = 1, 2, 3, 4, 5, 7, 10, and 15) were applied (where x is the number of CV scans). After surface modification, the electrodes were rinsed with copious amounts of acetonitrile and Milli-Q water and finally dried under a stream of nitrogen. The same procedure was performed on GC plates for surface characterization by ellipsometry. The GC/poly(ethylenediamine) surfaces used for AFM analysis were prepared using two cyclic voltammetry scans (x = 2). To achieve the AuNP modified surfaces (henceforth referred to as GC/poly(ethylenediamine)/NP), the GC/poly(ethylenediamine) surfaces were immersed in a 27 nm gold colloidal dispersion for 1 h. In some studies, in order to change the surface coverage of AuNPs, different deposition times including 2 min, 10 min, 30 min, 60 min, 3 h, and 6 h were applied for GC/poly(ethylenediamine) surfaces modified using 4, 5, and 7 scans (x = 4, 5, and 7). The same procedure was performed on modified GC plates to enable the surfaces to be imaged using SEM. Prior to any characterization technique, the modified electrodes were rinsed with Milli-Q water and then dried under nitrogen. Solutions were also purged with nitrogen gas prior to obtaining voltammograms and EIS spectra.
hydrogen tetrachloroaurate(III) (HAuCl4), and trisodium citrate (99% purity) were all from Sigma-Aldrich (Sydney, Australia). Phosphate buffer solutions (pH 7.0) used in this work contained 0.1 M KCl and 0.05 M K2HPO4/KH2PO4. All reagents were used as received, and aqueous solutions were prepared with purified water (18 MΩ cm, Millipore, Sydney, Australia). Electrochemical Measurements. The electrochemical characterization techniques of CV and electrochemical impedance spectroscopy (EIS) were performed in a conventional three-electrode system, comprising a GC working electrode, a platinum foil as the auxiliary electrode and Ag| AgCl|3 M KCl as the reference electrode in aqueous solution and ferrocene/ferricenium couple (Fc/Fc+) with a platinum wire as pseudoreference electrodes in nonaqueous media. Therefore, all potentials in aqueous and nonaqueous solutions are reported vs Ag|AgCl and Fc/Fc+, respectively. CV was conducted using an Autolab (PGSTAT12 Instrument, Sydney, Australia) potentiostat. The EIS measurements were performed using a Solartron SI 1287 electrochemical interface coupled with an SI 1260 frequency response analyzer (Solartron Analytical, Hampshire, England). The 1 mM Ru(NH3)63+ redox species in phosphate buffer solution (pH 7.0) was used as the electrolyte solution. EIS measurements were recorded at room temperature within the frequency range of 10−1−105 Hz superimposed on a dc potential of −0.202 V, with ac of 10 mV peak to peak amplitude, and 10 points per decade of frequencies. The Z-view software was used for electrical equivalent circuits modeling. Surface Characterization Methods. Scanning electron microscopy (SEM) was conducted using a Hitachi S-900 SEM (Berkshire, England) with a 4 kV field emission source. A J. A. Woollam Co. Inc. spectroscopic ellipsometer was used to measure the thickness of the ethylenediamine film. The wavelength was varied between 250 and 1000 nm in steps of 1.57 nm, and three different incidence angles of 65°, 70°, and 75° were applied. WVASE 32 software was used to model and fit the acquired data. First, the average optical constants of the GC substrate were determined with four pieces of unmodified substrates. The surface thickness was modeled as a single absorbing layer atop a 2.0 mm thick substrate. The polymer was modeled using the Cauchy approximation.21 Refractive indexes used in fitting are 1.45 for organic layer,22 1 for the ambient, air, and the refractive index of the GC substrate was determined experimentally. At least three ellipsometric measurements were made per sample, and more than three samples of each type were prepared. The observed uncertainty in repeated measurements of the same spot was typically less than 5 × 10−3 nm. The surface morphology was determined with an atomic force microscope (AFM) NanoScope (Veeco Instruments Inc., Australia) using normal silicon nitride tips (OTESPA 42 N/m) in Tapping Mode. In the present study, the Rz roughness (10point roughness height) is defined as Rz = 1/10 (Σ |Rpi| + |Rvi|, where Rpi is maximum profile peak height and Rvi is maximum profile valley depth. The Rz roughness was evaluated using the NanoScope Software 6.13. Electrode Preparation and Modification. GC electrodes were purchased as 3-mm-diameter disks from Bioanalytical Systems Inc. The electrodes were polished successively with 1.0, 0.3, and 0.05 μm alumina slurries made from dry Buehler alumina and Milli-Q water on microcloth pads (Buehler, Lake Bluff, IL). The electrodes were thoroughly rinsed with Milli-Q
■
RESULTS Polymer Film Deposition and Characterization. In order to form nanoparticle-assemblies, the GC/poly(ethylenediamine) modified surfaces were first prepared by electrochemical deposition of ethylenediamine. Figure 2 shows
Figure 2. Cyclic voltammograms of a GC electrode in 7 mM ethylenediamine/0.1 M NBu4BF4 in acetonitrile at a scan rate of 40 mV s−1. The numbers in the curve indicate the order of the applied CV cycles. 1075
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
Article
Table 1. Electrochemical Impedance Spectroscopy, Cyclic Voltammetry, and Ellipsometry Data for GC-Modified Surfaces with only Ethylenediamine (GC/Poly(ethylenediamine)) and Ethylenediamine and Gold Nanoparticles (GC/ Poly(ethylenediamine)/NP) CV cycles 1 2 3 4 5 7 10 15
thickness/Å 6.6 12.6 16.8 20.2 24.1 27.7 32.4 37.7
± ± ± ± ± ± ± ±
1.3 2.1 0.9 1.3 2.3 1.0 1.3 1.2
Ip/μA GC/poly(ethylenediamine)/NP 10.4 10.1 9.9 9.9 8.5 6.6 3.6 1.1
± ± ± ± ± ± ± ±
0.5 0.5 1.0 1.0 1.5 0.8 0.3 0.1
ΔEp/mV GC/poly(ethylenediamine)/NP 70 70 70 70 90 110 121 130
± ± ± ± ± ± ± ±
RCT/Ω GC/poly(ethylenediamine)
2 2 2 2 5 9 13 10
5.4 1.0 2.2 2.7 4.1 8.4 2.2 4.9
(±1.0) (±0.2) (±0.2) (±0.5) (±1.1) (±2.4) (±0.4) (±2.1)
× × × × × × × ×
104 105 105 105 105 105 106 106
RCT/Ω GC/poly(ethylenediamine)/NP 723 739 788 832 3 399 17 398 57 558 99 010
± ± ± ± ± ± ± ±
85 72 96 67 130 4 356 8 031 15 000
ket/cm s−1 GC/poly(ethylenediamine)/NP 3.9 3.8 3.6 3.4 8.4 1.6 4.9 2.9
(±0.3) (±0.4) (±0.2) (±0.3) (±1.2) (±0.7) (±0.3) (±0.4)
× × × × × × × ×
10−1 10−1 10−1 10−1 10−2 10−2 10−3 10−3
Figure 3. Cyclic voltammograms recorded in phosphate buffer solution containing 1 mM Ru(NH3)63+, presenting blocking behavior of (a) GC/ poly(ethylenediamine) surfaces (compare to the bare GC, dotted line) and (b) restoring of the electrochemistry (by adding AuNPs) at GC/ poly(ethylenediamine)/NP surfaces at a scan rate of 100 mV s−1. The numbers in the curve indicate the number of applied CV cycles in order to deposit ethylenediamine.
Figure 4. Nyquist plots recorded in phosphate buffer solution containing 1 mM Ru(NH3)63+, for (a) GC/poly(ethylenediamine) and (b) GC/ poly(ethylenediamine)/NP surfaces. The numbers in the curves indicate the number of applied CV cycles in order to deposit ethylenediamine. Insert in part b shows a Randles equivalent circuit used to fit the data for both GC/poly(ethylenediamine) and GC/poly(ethylenediamine)/NP surfaces. Dots show the real data, and lines show the fitted data. As a comparison, the Rct value for the bare GC electrode is 4.5 Ω.
a typical cyclic voltammogram for the electropolymerization of ethylenediamine onto GC electrodes, which is similar to that reported in the literature.19 The single broad irreversible peak was located at about +1.1 V vs Fc/Fc+. The oxidation mechanism of forming poly(ethylenediamine) has been reported previously.20 As the number of oxidation cycles was increased, the low conductivity of the deposited films resulted in passivation of the electrode and hence a substantial decrease in electrodeposition current. After several cycles, the electrode was sufficiently blocked such that only very small oxidation currents were observed. With the aim of changing the thickness of the deposited film, different CV cycles were applied to the electrodes, including x = 1, 2, 3, 4, 5, 7, 10, and 15 cycles. As it
can be seen in Table 1, ellipsometric experiments show that the film thickness is increasing gradually from 6.6 Å to 37.7 Å as more scans are applied to the electrode. The passivation properties of the modified electrodes by the polymer film were also investigated. Figure 3a shows the cyclic voltammograms of the modified GC electrodes compared to the bare electrode. All cyclic voltammograms were conducted in 1 mM ruthenium(III) hexamine in phosphate buffer, in a potential range of −450 to +200 mV, at a scan rate of 100 mV s−1 and were referenced relative to Ag|AgCl. Clearly, the modified GC electrodes exhibit a significant suppression of the Faradaic electrochemistry of the Ru(NH3)63+. Furthermore, the deposited films exhibited progressively greater blocking abilities 1076
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
Article
charge transfer process. These results are consistent with the CVs, where corresponding surfaces have identical voltammograms (Figure 3b). However, when the organic layer exceeded a thickness of 20 Å (x = 5, 7, 10, and 15), the RCT values increase with further growth of the film. This indicates a thickness of 20 Å as a threshold value above which the system steps into the distance dependent charge transfer process. According to Fermin and co-workers,25 the apparent rate constant of electron transfer, kapp et at a single gold nanoparticle can be determined using the equation:
toward ruthenium(III) hexamine as the number of cycles increased, which was an indication of an increase in thickness of the organic film. The voltammograms in Figure 3a suggest that the deposited layers of poly(ethylenediamine) are highly compact, as, for example, a 6 Å thick film (Table 1) significantly suppresses the Faradaic signal of the redox reaction, Ru(NH3)63+. Results consistent with cyclic voltammetry were also obtained using EIS. The impedance spectra were run over a frequency range of 0.01−100000 Hz and an applied dc potential of −0.204 V (vs Ag|AgCl) in 1 mM Ru(NH3)63+ solution (Figure 4a), and the circuit used to fit the data can be found in the insert of Figure 4b (This circuit was used to fit the data for both GC/poly(ethylenediamine) and GC/poly(ethylenediamine)/NP surfaces. Also, in order to account for all possible resistances for NP-modified surfaces, a different equivalent circuit consisting of other resistances (including Rnano that accounts for transfer from the redox species to the particles and Rfilm, which is related to the transfer from the particles to the electrode, similar to that of ref 14, was applied and similar values for resistance were obtained). According to the conventional Randles circuit,24 the high frequency loop is attributed to the charge transfer resistance, RCT, for the carbon electrode only modified with the poly(ethylenediamine) film. Increase in the RCT values means the passivation of the electrodes is more efficient as the organic layer gets thicker. The fitted charge transfer resistance of modified electrodes, GC/poly(ethylenediamine), varied from 104 to 106 Ω (Table 1). Electrochemistry of GC/Poly(ethylenediamine)/NP Surfaces. The presence of the free amine groups on the polymer backbone can be utilized for the attachment of AuNPs (Scheme 1). For this purpose, AuNPs were immobilized by immersing GC/poly(ethylenediamine) surfaces in AuNP solution for 1 h. SEM images (Figure S-2 in the Supporting Information) reveal a uniform distribution of AuNPs on the surfaces. Cyclic voltammograms of such modified electrodes in 1 mM Ru(NH3)63+ solution are shown in Figure 3b. Compared to the voltammograms obtained with GC/poly(ethylenediamine) surfaces, the voltammograms (Figure 3b) are dramatically affected by the presence of nanoparticles and more importantly they behave differently as the organic film gets thicker. For the first four modified surfaces (x = 1, 2, 3, and 4), the redox peaks of Ru(NH3)63+ are identical to those observed at a bare GC electrode, see Figure 3b, (showing a peak-to-peak separation, ΔEp, of 70 ± 2 mV compared with ΔEp of 69 mV and 68 mV of bare GC and bare Au, respectively). From the fifth cycles onward, the voltammograms have a gradual decrease in peak current and increase in ΔEp as the number of cycles for electropolymerization of poly(ethylenediamine) increases (Figure 3b and Table 1). That is, the GC/poly(ethylenediamine)/NP surfaces behave like a bare GC when the thickness of the polymer is below 20 Å, whereupon the kinetic of electron transfer became sensitive to the thickness of the polymer. EIS measurements were performed to further characterize the AuNP-modified surfaces. Figure 4b shows that the resistance to charge transfer, RCT, was reduced by orders of magnitude for the NP-modified electrodes. The variations in charge transfer resistance with thickness of the polymer before and after AuNPs immobilization are shown in the Table 1. The RCT values are all similar for the organic films thinner than 20 Å (x = 1, 2, 3, and 4), which highlights the distance independent
ketapp =
RT n F Γπ (rNP)2 R CT,NPAC 2 2
(4)
where Γ is the surface coverage of particles, A is the geometrical area of the electrode, rNP is the particle radius, and RCT,NP is the charge transfer resistance via the nanoparticle mediated pathway. RCT,NP contains elements associated with the kinetics of electron exchange between the particles and the redox couple as well as between the particle and the electrode substrate. The other terms have their usual meaning. Hence using eq 4, and the values of Γ of 2.2 (±0.3) × 109 particles per cm2 and r = 13.6 (±0.2) × 10−7 cm as determined from the SEM images (Figures S-1 and S-2 in the Supporting Information), the rate constants for different film thicknesses were calculated as listed in Table 1. Furthermore, the change in the rate of electron transfer for GC/poly(ethylenediamine)/NP surfaces is clearly shown in Figure 5.
Figure 5. Variation in the rate of electron transfer to ruthenium(III) hexamine with the thickness of the organic film for GC/poly(ethylenediamine)/NP surfaces.
These rate constants for electrodes modified by AuNPs (x = 1, 2, 3, and 4) are all similar to the rate constant electron transfer for ruthenium(III) hexamine reported previously.26 Thus, it seems the rate constant of electron transfer at a single AuNP is similar to that observed at bare GC electrode despite the presence of an intervening organic film between the bulk electrode and the nanoparticle. However, the thickness dependence of electron transfer is emphasized again by a significant decrease in the rate of electron transfer from 3.9 × 10−1 cm s−1 to 2.9 × 10−3 cm s−1, as the polymer gets thicker (Table 1). The value of the attenuation factor (β) that was obtained from the slope of the line in Figure S-4 in the Supporting Information and was 0.16 (±0.02) Å−1. This value is much lower than the normally observed β value of 0.8 Å−1 for aliphatic monolayer systems. The β value however depends on the molecular structure of the spacer. Perhaps the electropolymerization of ethylenediamine is normally an aliphatic oxidative process but not as structurally well-defined system as the SAMs. First, there is some variability in the chemical 1077
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
Article
Table 2. Electrochemical Impedance Spectroscopy, Cyclic Voltammetry, and SEM Data for (GC/Poly(ethylenediamine)/NP) Surfaces as a Function of the AuNPs Deposition Time AuNP deposition time/min AuNP surface coverage/cm−2 ket/cm s−1 Ip/μA ΔEp/mV
2 1.1 (±0.2) × 108 0.34 (±0.01) 1.1 ± 0.2 173 ± 6
10 3.8 (±0.4) × 108 0.34 (±0.02) 4.2 ± 0.5 162 ± 8
30 6.7 (±0.3) × 108 0.34 (±0.02) 7.5 ± 0.4 113 ± 2
60 2.2 (±0.3) × 109 0.34 (±0.01) 12 ± 1 70 ± 5
180 2.4 (±0.4) × 109 0.34 (±0.01) 13 ± 0.9 70 ± 3
360 2.5 (±0.3) × 109 0.34 (±0.02) 12 ± 0.7 70 ± 2
Figure 6. (a) Cyclic voltammograms and (b) Nyquist plot recorded in phosphate buffer solution containing 1 mM Ru(NH3)63+ for GC/ poly(ethylenediamine)/NP surfaces as a function of the AuNPs deposition time including 2, 10, 30, 60, 180, and 360 min at a scan rate of 100 mV s−1. The response from bare GC is also displayed for comparison (dotted line in part a). In part b, the points are the experimental data and lines are the fitted data. The Randles equivalent circuit used to fit the data is shown in the inset of Figure 4b.
however, a doubling in ΔEp (from 63 to 125 mV in the simulation, 70 to 130 mV experimentally) is observed. We suggest the experimentally observed decrease in the peak current (Figure 3b) as a function of the thickness of the organic film is indicative of a decrease in electrochemically active surface area. However, as the number of NPs of each surface is determined by SEM to be similar across the different thickness monolayers, the decrease in the electrochemical surface suggests that the number of nanoparticles on the surface that allow efficient electron transfer to the underlying electrode changes. A consequence of fewer electrochemically active NPs would be a greater distance between them, on average, such that the diffusion layers to each particle no longer overlap and the CVs change from peak shaped to sigmoidal (Figure 3b). The AFM images in Figure S-3 in the Supporting Information shows the topography of the electrodeposited polymer is not uniform, with peaks and some troughs. Thus NPs that are located in the troughs will be closer to the underlying electrode surface and may be dominating the electrochemistry, compared with NPs on the peaks. Electron Transfer as a Function of the NP Surface Coverage. In order to confirm that changes in the shape and magnitude of the voltammograms in the Figure 3b is due to different numbers of NPs that are communicating with the underlying electrode, the deposition time of NPs onto a polymer film of the thickness 20.2 Å was varied. SEM images showed an increase in the number of NPs as the deposition time increased from 2 to 360 min (Figure S-6 in the Supporting Information and Table 2). The corresponding cyclic voltammograms and impedance spectra of such GC/poly(ethylenediamine)/NP surfaces in 1 mM Ru(NH3)63+ solution after various times of NP deposition are presented in Figure 6. Each measurement was performed on a freshly modified GC electrode. Figure 6 suggests that AuNPs density at the surface of the electrode is affected by the deposition time (as evident by an increase in the peak current and a decrease in the peak-to-peak
composition of the resulting compound such as formation of aziridinium molecules or even formation of some CN and CO bonds (the presence of oxygen has been shown to have a significant effect on conductance of SAMs of alkanethiols27) as reported in ref 20. Second and more significantly, the electrodeposited layers are not completely uniform in thickness, as shown using AFM analysis of the GC/poly(ethylenediamine) surface (Figure S-3 in the Supporting Information). The AFM images confirmed that a nonuniform organic film is produced with the Rz value of 6.0 ± 0.6 nm, compared to that of bare GC (Figure S-3 in the Supporting Information) with Rz = 3.1 ± 0.2 nm. In an organic layer of nonuniform thickness, the electrochemistry will be dominated by the depression rather than the average thickness, as determined by ellipsometry. Hence the β value of 0.16 Å−1 estimated from Figure S-4 in the Supporting Information is not a true reflection of the attenuation factor. This is important for attempts to correlate the experimental data with the theoretical model of Chazalviel and Allongue as it indicates any correlation will only be qualitative and not quantitative. Hence Figure 5 shows the transition from transfer unaffected to transfer hindered occurs once a threshold polymer thickness is achieved in qualitative agreement with the model. An important observation from the voltammograms in Figure 3b is the decrease in peak current as the thickness of the organic film increases. Further, the shape of the CVs changes from the classical peak shape of linear diffusion to a bulk electrode to the sigmoidal shape indicative of convergent diffusion to microelectrodes. Using DigiElch 7, digital simulation of the cyclic voltammetric behavior of GC/ poly(ethylenediamine)/NP surfaces for all thicknesses were performed by applying all similar experimental conditions and using the electron transfer rates presented in the Table 1 (see Figure S-5 in the Supporting Information). The simulated CVs show that as the rate of the electron transfer decreases, only a very small decline in the peak currents occurred and no change in shape to sigmoidal. Similar to the experimental data, 1078
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
Article
separation) with an increase in the number of NPs on the surface (Table 2). However, similar electron transfer rates per particle (ket) for different NP densities is observed. This is in contrast to the effect of polymer thicknesses (Table 1) which changes the electron transfer rate (ket) by 2 orders of magnitude. Further, the CV shape changes from sigmoidal to peak shaped as the number of particles increases, which is consistent with the supposition above that as the polymer film gets thicker, less NPs are communicating with the underlying electrode. Consistent results for the different polymer thicknesses (24.1 Å and 27.7 Å) were also obtained (Figure S-7 and Table S-1 in the Supporting Information).
Upon normalization to the electrode surface area, the exchange current densities were estimated from eq 5 for GC/ poly(ethylenediamine) and GC/poly(ethylenediamine)/NP for x = 4, 5, 6, where the polymer films were 20 Å, 24 Å, and 27 Å thick. The theoretical model predicts that the J1/J0 value changes as the system steps into different electron transfer regimes. The high value of J1/J0 in the transfer unaffected regime decreases to 1 (J1/J0 = 1) in the boundary region (solid line in Figure 1). Thereafter, a value of less than 1 is expected for the transfer hindered regime. From the experimental data, the transition from the transfer unaffected regime to the transfer hindered regime is reflected in a change in the value of J1/J0 from 323 ± 58 (20 Å) to 0.56 ± 0.13 (24 Å), as would be predicted by the Chazalviel and Allongue theory. The main finding in this paper is that the simple theory presented by Chazalviel and Allongue qualitatively describes the charge transfer behavior of electrode-organic layer-nanoparticle constructs. Such knowledge is important with regards to developing devices that exploit this type of system. This type of construct is being explored for a range of applications including photovoltaic devices,6,7 electrocatalysis,4 and most importantly from our perspective electroanalytical chemistry. The analytical attractiveness of the construct is that (1) a low background capacitance modified electrode can be constructed where the organic layer can also be used to resist nonspecific adsorption of species to the electrode surface,28−30 and (2) nanoelectrode arrays can easily be fabricated with a high degree of control over the size, spacing, and crystal face of the nanoelectrodes (the latter provided by the incredible control we now have over the synthesis of nanoparticles10). In order to fully understand and evaluate various aspects of the theoretical model, still more experimental variables such as applying different organic films, redox probes with different charges, different NP sizes and surface coverage, and diverse electrode materials need to be studied.
■
DISCUSSION This study shows that nanoparticle-mediated electron transfer at passivated surfaces changes from thickness independent to thickness dependent as the thickness of the polymer layer exceeds 20 Å. For the sake of consistency with the theory, the terms “transfer unaffected regime” and “transfer hindered regime” will be used for differentiating between x ≤ 4 (thickness ≤20 Å) and x > 4 (thickness ≥24 Å), respectively. The range of thicknesses of the organic layer that encompasses the transfer unaffected regime data presented here are completely consistent with the previous results in our group8 and with the observation of Fermin and co-workers.14 We previously observed no chain-length dependence from alkanethiols of different lengths, from 2 Å to 11 Å (C = 2, 6, 8, 11), assuming −CH2 units of 1 Å. The group of Fermin also reported an unaffected electron transfer when the thickness was less than ∼12 Å, assuming the values of 2 Å for alkanethiol with two methylene units and about 10 Å for poly-L-lysine (PLL) films. However, Fermin and co-workers 14 reported a suppression of Faradaic current for organic layer thicker than 15 Å, assuming 5 Å for five methylene units and 10 Å for PLL film. According to the Chazalviel and Allongue theory,17 by considering a nanoparticle of 27 nm in diameter, if the thickness of the organic film is greater than 27.5 Å, nanoparticle-mediated electron transfer is expected to be affected (Figure 1). Experimentally we see a transition from transfer unaffected to transfer hindered with an average thickness of 20.2 ± 2 Å. However, as discussed above, the polymer film is not completely uniform in thickness and the kinetic analysis is expected to be dominated by NPs located in thinner regions of the polymer film. Hence direct quantitative comparisons based on the thickness of the organic layers are not valid, and only the correlation in the general trend in electrode-organic layer-NP assemblies going from a transfer unaffected regime to transfer hindered as the organic layer gets thicker is important. As mentioned in the introduction, high exchange current densities across the organic film in nanoparticle-assemblies, relative to when nanoparticles are not present, plays a crucial rule in electron transfer. Therefore, effective nanoparticlemediated electron transfer is expected as long as J1 is bigger than J0. Then, to compare the experimental results to the theoretical model, first exchange current densities for GC/ poly(ethylenedimine) and GC/poly(ethylenedimine)/NP surfaces were experimentally determined by EIS from RCT, and the exchange current densities could be calculated using the following equation,24 R CT = RT /nFi0
■
CONCLUSIONS In conclusion, this study provides experimental agreement with the current theoretical study of charge transfer in nanoparticleassemblies and shows that electron transfer changes from thickness independent to thickness dependent as the average thickness of the organic layer exceeds 20 Å. Moreover, it demonstrates that an efficient electron transfer at electrode/ organic layer/nanoparticle constructs proceeds as long as the exchange current density between the underlying substrate and nanoparticle is larger than that of between the substrate and the passivating film. These results provide important insight into designing analytical devices based on nanoparticle-assemblies.31,4
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
(5) 1079
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080
Analytical Chemistry
■
Article
(31) Jensen, P. S.; Chi, Q.; Grumsen, F. B.; Abad, J. M.; Horsewell, A.; Schiffrin, D. J.; Ulstrup, J. J. Phys. Chem. C 2007, 111, 6124−6132.
ACKNOWLEDGMENTS The authors thank the Australia Research Council (Grant LP100200593) and the Australian Government for funding. Thanks is also extended to Professor Brynn Hibbert for discussions and Dr. Leigh Aldous and Md. Mokarrom Hossain for help with the digital simulation analysis.
■
REFERENCES
(1) Katz, E.; Willner, I.; Wang, J. Electroanalysis 2004, 16, 19−44. (2) Cheng, W. L.; Dong, S. J.; Wang, E. K. Anal. Chem. 2002, 74, 3599−3604. (3) Raj, C. R.; Okajima, T.; Ohsaka, T. J. Electroanal. Chem. 2003, 543, 127−133. (4) Hernandez-Santos, D.; Gonzalez-Garcia, M. B.; Garcia, A. C. Electroanalysis 2002, 14, 1225−1235. (5) Willner, I.; Willner, B. Pure Appl. Chem. 2002, 74, 1773−1783. (6) Kamat, P. V. J. Phys. Chem. C 2008, 112, 18737−18753. (7) Su, B.; Fermin, D. J.; Abid, J. P.; Eugster, N.; Girault, H. H. J. Electroanal. Chem. 2005, 583, 241−247. (8) Shein, J. B.; Lai, L. M. H.; Eggers, P. K.; Paddon-Row, M. N.; Gooding, J. J. Langmuir 2009, 25, 11121−11128. (9) Le Saux, G.; Ciampi, S.; Gaus, K.; Gooding, J. J. ACS Appl. Mater. Interfaces 2009, 1, 2477−2483. (10) Dyne, J.; Lin, Y. S.; Lai, L. M. H.; Ginges, J. Z.; Luais, E.; Peterson, J. R.; Goon, I. Y.; Amal, R.; Gooding, J. J. ChemPhysChem 2010, 11, 2807−2813. (11) Grabar, K. C.; Allison, K. J.; Baker, B. E.; Bright, R. M.; Brown, K. R.; Freeman, R. G.; Fox, A. P.; Keating, C. D.; Musick, M. D.; Natan, M. J. Langmuir 1996, 12, 2353−2361. (12) Bethell, D.; Brust, M.; Schiffrin, D. J.; Kiely, C. J. Electroanal. Chem. 1996, 409, 137−143. (13) Horswell, S. L.; O’Neil, I. A.; Schiffrin, D. J. J. Phys. Chem. B 2003, 107, 4844−4854. (14) Bradbury, C. R.; Zhao, J. J.; Fermin, D. J. J. Phys. Chem. C 2008, 112, 10153−10160. (15) Zhao, J. J.; Bradbury, C. R.; Fermin, D. J. J. Phys. Chem. C 2008, 112, 6832−6841. (16) Kissling, G. P.; Miles, D. O.; Fermin, D. J. Phys. Chem. Chem. Phys. 2011, 13, 21175−21185. (17) Chazalviel, J. N.; Allongue, P. J. Am. Chem. Soc. 2011, 133, 762− 764. (18) Herlem, G.; Goux, C.; Fahys, B.; Dominati, F.; Goncalves, A. M.; Mathieu, C.; Sutter, E.; Trokourey, A.; Penneau, J. F. J. Electroanal. Chem. 1997, 435, 259−265. (19) Barbier, B.; Pinson, J.; Desarmot, G.; Sanchez, M. J. Electrochem. Soc. 1990, 137, 1757−1764. (20) Herlem, G.; Reybier, K.; Trokourey, A.; Fahys, B. J. Electrochem. Soc. 2000, 147, 597−601. (21) Guide to Using WVASE32; J. A. Woollam Co. Inc.: Lincoln, NE, 1994. (22) CRC. Handbook of Chemistry and Physics, 61st ed.; CRC Press, Inc.: Boca Raton, FL, 1980. (23) Frens, G. Nature 1973, 241, 20−22. (24) Bard, A. J.; Faulkner, L. R., Electrochemical Methods: Fundamentals and Applications, 2nd ed.; Wiley: New York, 2001. (25) Zhao, J. J.; Bradbury, C. R.; Huclova, S.; Potapova, I.; Carrara, M.; Fermin, D. J. J. Phys. Chem. B 2005, 109, 22985−22994. (26) Elliott, J. M.; Cabuche, L. M.; Bartlett, P. N. Anal. Chem. 2001, 73, 2855−2861. (27) Scullion, L. E.; Leary, E.; Higgins, S. J.; Nichols, R. J. J. Phys.: Condens. Matter 2012, 24. (28) Liu, G. Z.; Iyengar, S. G.; Gooding, J. J. Electroanalysis 2012, DOI: 10.1002/elan.201200333. (29) Liu, G. Z.; Paddon-Row, M. N.; Gooding, J. J. Chem. Commun. 2008, 3870−3872. (30) Liu, G. Z.; Khor, S. M.; Iyengar, S. G.; Gooding, J. J. Analyst 2012, 137, 829−832. 1080
dx.doi.org/10.1021/ac3029486 | Anal. Chem. 2013, 85, 1073−1080