dithiol self-assembled monolayers

Electrochimica Acta 56 (2011) 4361–4368

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Charge transfer and SEIRAS studies of 1,4-benzoquinone functionalized mixed monothiol/dithiol self-assembled monolayers Scott M. Rosendahl, Ian J. Burgess ∗ Department of Chemistry, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N 5C9

a r t i c l e

i n f o

Article history: Received 23 September 2010 Received in revised form 20 December 2010 Accepted 5 January 2011 Available online 31 January 2011 Keywords: Benzoquinone modified self-assembled monolayers Dithiol SAMs Place exchange Chronocoulometry Proton coupled electron transfer Surface enhanced infrared reflection absorption spectroscopy (SEIRAS) Standard heterogeneous rate constant

a b s t r a c t Although substantial information can be obtained from electrochemical measurements, much greater detail concerning molecular structure can be obtained by coupling such measurements with molecular spectroscopy. To this end, electrochemical and in situ surface enhanced infrared spectroscopy (SEIRAS) was performed to analyze 1,4-benzoquinone (BQ) terminated self-assembled monolayers. Monolayers were derived via the Michael addition of BQ to a pre-formed mixed monolayer composed of methyl and thiol terminated functionalities. This approach resulted in relatively robust but non-ideal redox-active monolayers. Spectroscopic and electrochemical measurements have allowed us to determine the pH dependence of both the apparent formal potential and the heterogeneous standard rate constant for proton coupled electron transfer (PCET) for this 2e− /2H+ redox system. While the former is in excellent agreement with predictions of step-wise PCET, the latter deviates from the expected kinetic response. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction Redox active self-assembled monolayers (SAMs) provide an excellent platform for the study of electrochemical charge transfer processes [1]. Thiols functionalized with redox centres can be synthesized a priori and subsequently self-assembled on electrode surfaces. Advantages of studying surface confined redox centres include the removal of mass transfer limitations and the ability to manipulate heterogeneous electron transfer rates through control of the spacer thickness separating the redox centre from the electrode surface. Modifying the length [1–6] and chemical nature of the spacer [7–12] has been found to be particularly useful for determining the mechanism of electronic coupling between substrate and redox centre as well as verifying Marcus theory predictions [3,4,13]. Benzoquinone (BQ)/hydroxybenzoquinone (HBQ) SAMs have been extensively studied [8–11,14–27] and are of particular fundamental interest owing to the importance of quinone moieties in biological processes [28]. Unlike simple electron transfer species, the reduction of BQ to HBQ in aqueous solutions involves the coupled transport of both two electrons and two protons (2e2H). Theo-

∗ Corresponding author. Tel.: +1 306 966 4722; fax: +1 306 966 4730. E-mail address: [email protected] (I.J. Burgess). 0013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2011.01.013

retical treatments of the observed apparent heterogeneous rate constants dependence on pH are based on either a step-wise mechanism (discrete proton and electron transfer steps) [29–34] or concerted mechanisms (simultaneous transfer of both forms of charge) [35–39]. Modifications of step-wise proton coupled electron transfer (SW-PCET) theory to describe surface confined processes have been provided in detail by Finklea [40]. Based on the previous work of others [41,42] we have recently outlined a procedure to form a nearly ideal electroactive BQ monolayer using the strategy of covalently attaching 1,4-benzoquinone (BQ) to a preformed amine terminated monolayer via Michael addition [43]. This system provides minimal heterogeneity and Nernstian responses at very slow voltammetric sweep rates and we have recently demonstrated that the resulting aminobenzoquinone monolayer exhibits two-electron, three-proton SW-PCET [44]. Although the aminobenzoquinone system is an excellent model system for electrochemical determination of formal potentials and charge transfer rate constants, we found that the aminobenzoquinone monolayer suffered from a gradual loss of electrochemical activity, presumably due to C–N bond hydrolysis. A more robust linkage is preferable for biosensing applications such as the confinement of quinone containing enzyme cofactors to electrode surfaces. In this manuscript we report upon our efforts to use nonanedithiol (NDT) diluted in a monolayer of octanethiol (OT) as

4362

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

the nucleophile in the surface tethering Michael addition. Highly facile and nearly quantitative yields have recently been reported for similar reactions involving freely dissolved thiols and benzoquinone in aqueous solution [45,46]. We are able to demonstrate that this reaction can be extended to surface reactions and provides a more stable redox active monolayer, albeit with an increase in the heterogeneity of the redox centre microenvironments. Furthermore, unlike the amino functionalized monolayers, the reaction with the thiol terminated SAM does not require elevated reaction temperatures. The greater robustness of this preparative method has allowed us to extend our kinetic and thermodynamic methodologies beyond voltammetry and chronocoulometry. We have also used surface enhanced infrared absorption spectroscopy (SEIRAS) to characterize our redox active monolayers. Ye et al. [26] and Port et al. [47] have previously reported FT-IRRAS (Fourier transform infrared reflection absorption spectroscopy) studies of BQ/HBQ monolayers. However, a major limitation of FT-IRRAS is the highly resistive, thin-cell configuration which prevents meaningful kinetic data from being obtained. The use of the internal reflection geometry in SEIRAS removes this limitation and has been successfully implemented by Osawa [48] and others [49–52] to study kinetic processes occurring on electrode surfaces. Herein we report upon the agreement between thermodynamic (apparent formal potentials) and kinetic (apparent heterogeneous rate constants) obtained from purely electrochemical (cyclic voltammetry, chronocoulometry) and spectroelectrochemical (SEIRAS) techniques.

2. Experimental 2.1. Chemicals, reagents and gold substrate preparation Octanethiol (95%), 1,9-nonanedithiol (95%), NaClO4 ·H2 O (>99%), Na2 HPO4 (ACS Grade), NaH2 PO4 (ACS Grade) and HClO4 (Ultrahigh Purity Grade) were all purchased from Sigma–Aldrich and were used as received. Ethanol (95%) was purchased from Commercial Alcohols Inc. (Brampton ON, CA). 1,4-Benzoquinone (Alfa Aesar, <98%) was purified by sublimation at a slightly elevated temperature (37 ◦ C) before use. All aqueous solutions were formed from Milli-Q (>18.2 M cm) water. The procedure for fabricating the polycrystalline bead electrodes from gold wire (Alfa Aesar, 99.99%) has been described elsewhere [53]. Prior to incubation, the polycrystalline gold bead working electrodes were cleaned by immersion in Piranha solution (3:1 H2 SO4 :H2 O2 ) for a minimum of 30 min, followed by copious rinsing with Milli-Q water. The electrodes were then flame annealed and quenched with Milli-Q water. For deposition of gold onto the silicon hemispherical prism, the original procedure reported by Osawa et al. [54] was followed with the modification described by Delgado et al. [55]. Briefly, the reflecting plane of a 25 mm diameter, non-doped, silicon hemispherical prism (Harrick Scientific Products, Pleasantville NY, US) was successively polished with finer grade diamond suspensions (Leco Corporation, St. Joseph, MI, US) down to 0.5 ␮m. The prism was then degreased by sonication in ethanol and finally rinsed in Milli-Q water before deposition. To remove the oxide layer and to terminate the silicon surface with hydrogen, the reflecting surface was left in contact with 40% (w/w) solution of NH4 F for 4 min. The gold deposition was done at 55 ◦ C by dropping a solution containing 5 mg HAuCl4 , 0.3 M Na2 SO4 , 0.1 M Na2 S2 O3 , 0.1 M NH4 Cl and 2% HF (2:1 in volume) directly onto the polished face of the silicon hemispherical prism. After 80 s the gold deposition was quenched by rinsing the prism with copious amounts of Milli-Q water. The gold film was then electrochemically cycled into the gold oxida-

tion region using 0.5 M H2 SO4 until stable measurements were obtained. 2.2. Self-assembled monolayer preparation The cleaned gold substrate was incubated in a 3 mM octanethiol (OT) ethanolic solution for 5 h before being rinsed thoroughly with ethanol. Subsequently, the gold substrate was incubated in an ethanol solution containing 3 mM 1,9-nonanedithiol (NDT) to afford partial place-exchanged monolayers. After 15 min of placeexchange, the substrate was rinsed with ethanol, then rinsed with Milli-Q water before being exposed to an aqueous solution of ∼10 mM 1,4-benzoquinone for 18 h. The 1,4-benzoquinone solution was prepared from freshly sublimed 1,4-benzoquione at 37 ◦ C. Finally, the modified gold substrate was rinsed with Milli-Q water before performing a given experiment. 2.3. Electrochemical measurements A three-electrode arrangement was used to analyze the electrochemical features of the modified gold substrate SAMs. These measurements were performed in an all-glass sealed cell, which was connected to the external reference electrode (Ag/AgCl, saturated KCl) via a salt bridge. The counter electrode was a loop of gold wire which was flame annealed before every experiment. All the glassware used for an experiment was heated in a mixture of H2 SO4 and HNO3 (2:1 by volume) and then rinsed copiously and soaked overnight in Milli-Q water prior to every experiment. The various electrolytes used were de-oxygenated with argon before the introduction of the working electrode and a continual blanket of argon was maintained above the electrolyte during all electrochemical experiments. Cyclic voltammetric and chronocolubomic measurements were done using a computer controlled system, using in-house software written in the LabVIEW (National Instruments Corporation, Austin, TX, USA) environment, consisting of a HEKA Potentiostat PG590 (HEKA, Mahone Bay, NS, Canada) with data collected using a multifunction DAQ card (PCI 6251 M Series, National Instruments Corporation, Austin, TX, USA). 2.4. In situ spectroelectrochemical measurements All in situ measurements were performed in a Teflon spectroelectrochemical cell arranged using a Si hemisphere in the inverted Kretschmann attenuated total internal reflection (ATR) configuration. The cell was constructed in-house and was equipped with a reference electrode (Ag/AgCl, saturated KCl) connected via a glass salt bridge. Electrical contact was made to the working electrode by pressing a conductive spring against the Au-plated surface of the Si hemisphere outside the electrolyte solution. The counter electrode was a coil of gold wire, flame annealed before immersing into the working cell and electrolyte. Potential control was maintained using a HEKA PG590 potentiostat and custom software written in LabVIEW. The electrolyte was deaerated by purging with argon for 30 min and a continual blanket of argon was maintained over the electrolyte throughout the experiment. All ATR-FTIR spectra were measured using p-polarized incident radiation at 60◦ with respect to the normal of the ATR element. A Nicolet Nexus 870 Fourier Transform Infrared (FTIR) Spectrometer equipped with a mercury cadmium telluride (MCT) liquid nitrogen cooled detector was used to make the IR measurements at a resolution of 4 cm−1 . This spectrometer is capable of working in rapid scan modes with an approximate minimum time per interferogram of 77 ms at 4 cm−1 resolution. The sample chamber of the spectrometer was purged throughout the experiment using CO2 and H2 O free air supplied by a Parker Balston FT-IR Purge Gas Generator 75-62 (Parker Hannifin Corporation, Haverhill, MA, US).

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

3. Results and discussion 3.1. Mixed monolayers To prepare a thiol terminated SAM on gold we looked at a variety of different techniques reported in the literature including passive incubations of neat 1,9-nonanedithiol (NDT) solutions in various solvents [56–58], potential assisted deposition [59–62], and passive place exchange incubations resulting in mixed monolayer systems [63]. Dithiol monolayers formed from passive incubation and subsequently exposed to BQ were found to exhibit poor electrochemical responses. Although voltametric signals were weakly discernable, peak separations were very large and increased with successive potential cycles. Furthermore, the intensity of the peak currents decayed rapidly with potential scanning. This is interpreted as evidence of the physisorption of BQ in the aliphatic hydrocarbon matrix of the SAMs rather than the desired chemical attachment of BQ to the ␻-functionalized monolayer. The absence of chemical linkage to the BQ most likely results from the propensity of both thiol endgroups of NDT to attach to the gold substrate as has been reported elsewhere [64]. Only minimal improvement was observed by using deoxygenated organic solvents [58] for the passive incubation of NDT monolayers. Although the electrochemical approach reported by Rifai et al. [59,60] was successful, it was quite cumbersome to implement and the best results for quinone attachment were found for dithiol monolayers formed by a passive place-exchange method. This involved the initial incubation of the gold substrate in an alkane thiol (octanethiol, OT) for 5 h, followed by a short incubation in 1,9-nonanedithiol (NDT). OT was chosen on the basis of the relative lengths of OT and NDT. Assuming a fully extended hydrocarbon region, NDT is longer than OT by the length of one methylene unit and one thiol unit which ensures that the resulting mixed SAM has an exposed thiol group readily accessible for the addition of the quinone. As this method provided monolayers with the best reactivity toward BQ, it was used exclusively in the electrochemical and IR studies reported below. To analyze the SAM during the incubation process, and to confirm that the place-exchange reaction occurred, IR spectra were measured before and after the incubation of the dithiol. Fig. 1 is the resulting plot of the normalized difference spectrum, S = (SVAR − SREF )/SREF , with SREF and SVAR representing the single beam signals for the monolayer before and after the place exchange reaction. In this representation, an increase in the intensity of an IR absorption between the sample and the reference will result in a downward going feature. An upward going feature indicates a loss of absorption signal intensity. Inspection of Fig. 1 indicates that there are four major IR bands present, two upward going bands (2874 cm−1 and 2972 cm−1 ) and two downward going bands

-1

0.0020 0.0015

2972cm CH3 (asym)

-1

2874cm CH3 (sym)

0.0010

ΔS/S

To perform kinetic SEIRAS experiments, it was necessary to trigger the start of the infrared spectra measurements simultaneously with the change in the working electrode potential upon a potential  step to the formal potential, E0 . This was accomplished using the Nicolet Start-Accessory (to trigger the FTIR spectrometer) and custom written LabVIEW software to synchronize the potential step and commencement of IR data collection. In our kinetic experiments, eight interferograms were measured at 4 cm−1 resolution and binned into 1.3 s intervals with total data acquisition times lasting up to 120 s. In order to build the signal-to-noise ratio of the IR measurements, this process was repeated 128 times for a total of 1024 co-added spectra for each 1.3 s time interval. It is of importance to note that as the mirror drive of an FT instrument is increased the signal-to-noise is decreased. This effectively requires a higher number of scans to achieve suitable signal-to-noise from the collected data.

4363

0.0005 0.0000 -1

-0.0005 -0.0010 2700

-1

2853cm CH2 (sym)

2922cm CH2 ((asym) y )

2800

2900

3000

3100

Wavenumber (cm-1) Fig. 1. IR difference spectrum of the CH stretching region indicating the vibrational changes as a result of a passive place exchange of an OT-SAM with 1,9-NDT (reference spectrum is the OT-SAM before place-exchange).

(2853 cm−1 and 2922 cm−1 ). The upward going bands are associated with the symmetric and asymmetric C–H stretching modes in CH3 and the downward bands for similar modes for CH2 [65–67]. Fig. 1 clearly demonstrates that during the place exchange process there is a loss of signal from CH3 groups and a gain in signal for the CH2 groups. This qualitatively confirms the success of the placeexchange reaction as dithiol molecules displacing OT will decrease the number density of methyl groups at the surface. Conversely, the place exchange process leads to a gain in the number of CH2 groups present at the surface. Quinone functionalization of the free thiol endgroup in the mixed NDT:OT monolayers was achieved through Michael addition of 1,4-benzoquinone. The product of this reaction should be the reduced (i.e. dihydroxybenzoquinone) form of the redox centre. However, Budavari et al. have shown that the negative shift in the reduction potential of the tethered species results in its oxidation by excess free benzoquinone in the incubating solution [41,42]. Although we did not establish the open-circuit form of our modified monolayer, cyclic voltammetry clearly demonstrated that our system could be readily toggled between its two redox states and that it was amenable for various electrochemical and infrared experiments. In a series of independent experiments, the ratio of the charge associated with the reductive desorption of the mixed thiol monolayer (in the absence of quinone) and the charge passed in the quinone redox was determined to be (0.95 ± 0.05):1. Assuming one electron transfer for thiol reduction and two electron transfer for the quinone electrochemistry, we estimate the initial loading of benzoquinone on our monolayers to be approximately 50%. 3.2. Electrochemical characterization After devising a method to reproducibly build BQ modified SAM on gold substrates, electrochemical techniques were used to characterize these systems. To start, cyclic voltammetry (CV) of these systems in 5 mM phosphate buffers of various pH were measured. These buffers were prepared using an excess of NaClO4 (100 mM) to ensure common ionic strengths with each of the electrolyte solutions. Fig. 2(a) depicts the CVs of three separately prepared electrodes in the different pH electrolytes. The CVs in all cases were background corrected by removing capacitive currents measured due to the charging of the double layer. The CVs in Fig. 2(a) reveal a well-known shift in peak position as a function of pH. The kinetics of this BQ-SAM system are extremely slow, making it very difficult to ensure equilibrium is reached on the time scale of the potential sweep in a CV measurement. As an alternative to

4364

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

a

1.0

pH 9.0

pH 5.0

pH 7.0 p

iP/iP0

0.9

Amino-BQ Thio-BQ 0.8

0.7

pH 5.0 pH 7.0 pH 9.0

0.6 0 0.0

0.2

0.4

E / V vs Ag/AgCl 1.0

qF (Normalize ed)

0.8

0.6

Evar,3 Evar,2 Evar,1

0.2

Eref

300s

0.0 -0.4

300s

300s

time

-0.2

0.0

0.2

10

15

20

Scan Number Fig. 3. Relative stability of electrochemical peaks measured in CV experiments at 10 mV/s for amino-quinone and thio-quinone SAM systems. The electrolyte used in these measurements was 5 mM NaPBS (pH 5), 100 mM NaClO4 . In both cases the potential was scanned ±500 mV from the formal potential.

b

0.4

5

0.6

0.4

0.6

E / V vs Ag/AgCl Fig. 2. (a) CVs of BQ-SAM system in 5 mM NaPBS, 100 mM NaClO4 buffer solutions run at 10 mV/s. CVs have been normalized by peak area. (b) Normalized Faradaic charges of BQ SAM system in 5 mM NaPBS, 100 mM NaClO4 buffer solutions. Inset in (b) shows the potential step sequence with the downward arrows indicating where the current transients were measured.

ultraslow potential scanning, the following potential step sequence was employed. The working electrode was initially biased at a rest  potential, EREF , ∼200 mV negative of the formal potential, E0 , for 5 s to ensure all the BQ moieties were fully reduced. The potential was then stepped to a variable potential, EVAR , and held at this value for 300 s, well beyond the time required for the system to reach equilibrium. After this time, the potential was immediately stepped back to the EREF and the current transient was measured and numerically integrated to provide the difference in the total charge, Q, between EREF and EVAR . The procedure was repeated using increasingly positive values of EVAR . From the measurements of Q, the background charge associated with any capacitive processes was subtracted leaving only the Faradaic charge information for the BQ moiety. A detailed description of the background correction procedure is provided elsewhere [43]. The results are depicted in Fig. 2(b). Similar to the CV results shown in Fig. 2(a), there is a shift in the formal potential to more positive values with decreasing pH. The experimental data obtained from the chronocoulometry measurements were fit to a sigmoidal function and the resulting curve was numerically differentiated to provide the apparent formal potential. As mentioned previously, the stability of the amino-quinone SAM was a concern for longer duration IR experiments but could

be greatly alleviated by using thio-quinone linkages. To compare the stability of the two systems, a series of CVs were measured for both amino-quinone and thio-quinone SAM modified electrodes. The procedure for the preparation of the amino-quinone modified electrode has been described elsewhere [43]. From the collected CVs for each electrode, a ratio of the peak height for a given scan to the first scan is plotted in Fig. 3. It can be seen in both cases that there is some quinone lost during each CV scan cycle; however, this loss is more substantial for the amino-quinone system. After nearly 1 h of continuous potential scanning, the loss of signal from the thio-quinone was ∼5% compared to the nearly one-third attenuation observed in the amino-quinone system. We speculate that the loss of electrochemical signal is due to the hydrolysis of the C-N bond removing the quinone moiety from the SAM.  Fig. 4 plots the shift in formal potential, E0 , as a function of electrolyte pH for BQ modified SAMs using various phosphate buffered electrolytes and different electrochemical and spectroelectrochemical techniques. From this plot the equation of the line fitted through the data points has a slope of −63 mV/pH in the range of pHs 1–10. This result is comparable (within 5%) between the three different techniques used to determine the formal potential. For completeness we note the presence of an inflection in the data

500

CV CC IR

400

0'

-0.2

E / V vs Ag/AgCl

-0.4

300 200 100 0 -100 -200 0

2

4

6

8

10

12

pH Fig. 4. Formal potential of BQ-SAM system measured in 5 mM NaPBS (variable pH), 100 mM NaClO4 solutions using different techniques. CV values were determined from measurements made with a sweep rate of 10 mV/s. Chronocoulometry values were determined from differentiation of the baseline corrected charge data. IR values were obtained from the integrated peak area of the carbonyl band (∼1660 cm−1 ).

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

3.3. In situ spectroelectrochemical results Two different in situ spectroelectrochemical experiments were used to examine the thermodynamic and kinetic aspects of the prepared BQ modified SAMs. To determine the formal potential as a function of pH, a set of normalized difference spectra were calculated for the BQ modified SAM by using a constant reference potential and varying the stepped potential. The value of the reference potential was chosen to be several hundred millivolts negative of the formal potential to ensure only the reduced form of BQ existed on the electrode surface. After the application of the potential step, a delay of appropriate duration (usually 5 min) was implemented before collecting the IR spectra to ensure the interface was at equilibrium. Representative data are shown in Fig. 5(a) for experiments performed at pH 3. When BQ undergoes a PCET reaction, its molecular structure changes in accordance with Scheme 1. Upon oxidation of 1,4-dihydroxybenzoquinone to 1,4benzoquinone there is a loss of two hydroxyl (OH) groups and a gain of two carbonyls (C O). The carbonyl IR stretch is a very strong absorber in the mid-IR, ∼1660 cm−1 , and provides an intense signal in the monolayer system. Fig. 5(a) shows the carbonyl stretch progressively increasing in intensity (i.e. becoming a more pronounced negative-going peak) with increasing potential. It is important to point out that this region of the mid-IR spectrum is usually complicated by the presence of a strong water absorption band due to an HOH deformation stretch, ∼1640 cm−1 . A convenient method to eliminate the possibility of spectral overlap in this region arising from the solvent is to substitute water for deuterated water (D2 O). An experiment of this nature was performed (results not shown) and indicated that the IR stretch in this region is the carbonyl stretch and any IR signal associated with H2 O is removed by calculation of a difference spectrum. Concomitant with the changes in the carbonyl signal, a decrease in the benzene ring stretches and single bond C–O stretches are expected upon oxidation of the redox centre. These spectral features should appear as positive-going bands at approximately 1450 cm−1 and 1200 cm−1 , respectively [26]. While the latter is below the frequency cut-off for our Si substrate, inspection of Fig. 5(a) shows that only at the most positive potentials do very weak upward signals at ∼1460 cm−1 appear. It is possible that the lack of pronounced signals attributable to HBQ arises from changes in the molecular orientation upon oxidation of the redox centre. If the reduced HBQ is aligned such that its principal C2 molecular axis is roughly parallel to the reflecting plane, then surface selection

a

-150 -50 150 250

ΔS/S

350 450 550 650

E / mV vs Ag/AgCl

0

750

x5

0.005 1400

1600

1800

2800

2900

3000

3100

-1

Wavenumber (cm )

pH 11.0 1.0

pH 1.0

b

-1

Normalized Peak Area (1660 cm )

at pH ≥ ∼10 which is consistent with a change in the number of protons transferred as predicted by Laviron for quinone systems [34]. As a point of comparison, the amino-quinone system was completely unstable at pH > 9. This finding further suggests that the C–S bond in the thio-quinone SAM has a greater degree of stability to hydrolysis compared to that of the C–N bond in the amino-quinone SAM. There are no data points for the chronocoulometry results below pH 5. This is because as the pH is decreased, the formal potential shifts to more positive potentials. Eventually, the potential is shifted to the point where there are significant interferences due to background currents which can not be accurately removed. This was also true, to a lesser extent, for the CVs measured at these pHs and resulted in slightly less reliable results. It should be noted that the IR measurements are not hindered by these limitations and provide an alternate approach to calculate the formal potential in this system, the details of which are described in-depth in the next section. The shift in formal potential can be explained using Laviron’s theory of coupled proton electron-transfer [33], later modified by Finklea for monolayer systems [40]. The slope calculated from the points in Fig. 4 is consistent with the expected shift in the formal potential (−60 mV/pH).

4365

0.8

0.6

pH 1.0 pH 3.0 pH 5.0 pH 7.0 pH 9.0 pH 11.0

0.4

0.2

0.0 -400

-200

0

200

400

600

800

E / V vs Ag/AgCl Fig. 5. (a) IR difference spectra data for the BQ-SAM system in 5 mM NaPBS (pH 3), 100 mM NaClO4 . Co-addition of 1024 scans where used to generate each spectrum. Reference potential was −250 mV vs. Ag/AgCl. (b) Normalized peak areas for the carbonyl group (C O) associated with the quinone moiety at different pH (5 mM NaPBS, 100 mM NaClO4 electrolytes) and potentials.

rules render both the benzene ring stretch and the C–O stretching modes IR inactive. Correspondingly, there will be no apparent loss in these signals upon oxidation. The strong signals arising from the carbonyl stretches indicate that the BQ must be oriented perpendicular to the gold surface. The weak changes observed in Fig. 5(a) for the CH stretching region provide ancillary evidence of potential dependent distortion of the hydrocarbons which would be consistent with our inference of molecular rearrangement. The intensity of the carbonyl band increases until reaching a plateau at 550 mV. To quantify this result, integrated IR bands associated with the C O stretch are plotted as a function of potential

O

OH

-2H+, -2e+2H+, +2eRS

RS

O

OH Scheme 1.

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

(Fig. 5(b)). From this figure, the data points for a given pH fit to  a sigmoidal function and the formal potential, E0 , can be determined and plotted as a function of pH (Fig. 4). As the potential is stepped through the formal potential there is a rapid change in the C O stretching region similar in shape to the calculated charge passed described in the previous section. Between 1 < pH ≤ 11 there is some variation in the width of the sigmoidal curves which is due to different extents of heterogeneity from monolayer to monolayer. However, the data for pH 1 has an appreciably much larger width indicating that our wait conditions were insufficiently long enough to ensure that the system was measured at equilibrium conditions. As will be shown below, this is consistent with the decrease in the standard rate constant with decreasing pH.

pH 5.0 (IR) pH 9.0 (IR) pH 5 5.0 0 (CC) pH 9.0 (CC)

0.5

Fraction of oxidized product

4366

0.4

0.3

Eo/+0.2V

0.2 Eo/

0.1 Eref

0.0

time

3.4. Proton coupled electron transfer kinetics Proton-coupled electron transfer mechanisms have been debated for the last few decades and include step-wise [29–35,40] and concerted mechanisms [35–38]. For a two electron Nernstian system, the full-width half maximum (FWHM) of a redox peak measured in a CV should be 45 mV. In the case of our prepared BQ-SAMs, a FWHM of ∼100–150 mV was typically measured. This indicates that there are a multitude of local environments surrounding each quinone moiety having various Frumkin interactions. We suspect that during the place exchange process dithiol molecules tend to form island domains in the alkane SAM [63]. These dithiol domains increase the interactions between neighbouring redox centres and even allow for possibile double-tethering of the quinone centres to the monolayer [41,42]. This in turn provides a variety of local microenvironments for the redox moieties, resulting in heterogeneous kinetic behaviour as demonstrated by electrochemical measurements. As this electrochemical system does not behave in a Nernstian manner, obtaining details of the PCET mechanism is complicated. However, the pH dependence of the apparent rate constant, ks,app , for this reaction can be determined through the use of a variety of electrochemical and spectroelectrochemical techniques. Laviron showed that the apparent rate constant can be determined by studying the redox peak separation at varying scan rates [68]. This technique is very useful and has been used extensively in the literature as a means to measure ks,app . Laviron’s method necessarily assumes that the transfer coefficient for the process is potential independent which Finklea has argued is not the case for PCET occurring in monolayer systems [40]. To prevent errors in determining rate constants via extrapolation of data measured at significant overpotentials, we employed three techniques based on (i) chronocoulometry, (ii) cyclic voltammetry and (iii) in situ spectroelectrochemistry. All three methods measure the apparent standard rate constant of our BQ-SAM system based on measure ments at, or very near E0 . To study the kinetics, the double-step chronocoulometric method described above was modified as follows. As before, an initial potential, EREF , ∼200 mV more negative than the formal potential was applied to the working electrode to fully reduce the BQ molecules. The potential was then stepped to the formal poten tial, E0 , and held for a variable amount of time, tVAR . The potential is then stepped back to EREF and the current transient measured. The potential is then stepped to a potential ∼200 mV more positive than the formal potential to fully oxidize all the BQ molecules. Finally, the potential is stepped back to EREF and the transient measured once again. This process is repeated numerous times changing tVAR to longer times up to 300 s. The purpose of measuring the transient  from E0 + 200 mV to EREF is to account for the slow loss of quinone during the overall measurement. The collected transient data has the background charge subtracted and the ratio of charge measured at the formal potential to the total charge is plotted (Fig. 6) as a func-

0

20

40

60

80

100

120

time / s Fig. 6. Comparison of double-step chronocoulometry and rapid-scan in situ spectroelectrochemical time dependent measurements used in the determination of the heterogeneous rate constant, ks,app . Electrolytes were prepared using 5 mM NaPBS and 100 mM NaClO4 at the indicated pHs. Inset shows the modified double potential step sequence used for the chronocoulometric measurements with the downward arrows indicating where the current transients were measured.

tion of time. From this data, an exponential fit can be used to extract the average apparent standard rate constant of the redox centres. In order to obtain the rate kinetics from the CV measurements, we used a technique described by Finklea and co-workers [69,70]. Briefly, the Faradaic current, iF , is isolated from the double-layer charging currents and the charge obtained by integrating the corrected CVs. At the apparent formal potential, the apparent standard rate constant can be calculated using ks,app =

iF qF,tot (1 − 2=0 )

(1)

where  = 0 is the fraction of the BQ moieties reduced or oxidized in the cathodic or anodic sweep to the formal potential. For each CV, ks,app can be calculated for both the forward and reverse potential scans and the average value reported. Finally, as an alternative to using purely electrochemical techniques to measure the heterogeneous rate constant, a fast-scan IR spectroscopic method was used. Exponential fitting of the normalized integrated intensity of the 1660 cm−1 IR band plotted as a function of time was used to measure the rate of formation of BQ. A comparison of kinetic data obtained with the chronocoulometric and the IR measurements is shown in Fig. 6. There is some discrepancy in the data from the two techniques which probably arises from a slow loss in the quinone signal in the IR experiments. Whereas, the modified double-step chronocoulometric step sequence accounts for the slow loss of quinone from the monolayer, no such correction was applied for the IR kinetic measurements. In independent experiments we have measured CVs before and after subjecting the electrode to the potential step sequence used in the IR measurements. We determined that roughly 15–20% of the quinone is lost over this time period which will perturb the signal transients such as those shown in Fig. 6. However, in general, the evolution of the spectroscopic signal followed that of the charge-based signal fairly closely and representative curves are plotted in Fig. 6. The heterogeneous rate constants obtained for the three methods as a function of pH have been superimposed in Fig. 7. Above pH > 5 the agreement between the three techniques is good but at low pH there is more scatter in the data. It becomes difficult to accurately subtract the background signal for both the CV and chronocoulometry method at pH < 5 as the redox potentials of the quinone centre begin to overlap with non-analytical signals most likely arising from the onset of the oxidation of the Au-thiol bond.

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

4367

proton coupled electron transfer. More subtle changes in the rate constant dependence are most likely obscured by the kinetic heterogeneity of the system.

-1.25

-1.50

Acknowledgements

kapp) log(k

-1.75

This work has been funded by NSERC (Canada). S.M.R. would like to thank NSERC for a Post Graduate Scholarship (PGS-D).

-2.00

References

CV CC IR

-2.25

-2.50

-2.75 0

2

4

6

8

10

12

pH Fig. 7. Heterogeneous rate constant, ks,app , as determined by the different electrochemical and in situ spectroelectrochemical techniques at different pHs (5 mM NaPBS, 100 mM NaClO4 electrolytes) for the BQ-SAM system.

This is particularly problematic in the CV method where the slow electron transfer kinetics mandate that relatively large overpotentials are required to observe the quinone electrochemistry even at the slowest scan rates employed (1 mV/s). Accounting for the background is further compromised by the asymmetry of the apparent transfer coefficient which leads to a significantly broader and less sharp anodic peak compared to its cathodic analogue and thus the error in the rate constants is quite large for this method at low pH. The IR technique is not affected by this limitation and we were able to measure the apparent rate constant for the PCET process with equal degree of accuracy over the entire pH range. In general, the heterogeneous rate constants measured using the electrochemical techniques closely follow those measured spectroscopically. This reveals that the in situ molecular spectroscopic technique is as sensitive as the electrochemical measurements in determining ks,app . Fig. 7 also demonstrates that the rate constant appears to have an overall linear dependence on pH. This strongly deviates from the expected theory for two-electron, two-proton step-wise PCET which predicts that Fig. 7 should provide a “W shaped” plot. Our previous studies of analogous aminobenzoquinone monolayers showed comparably slow rate constants that exhibited distinctly more complex pH dependence [43]. The apparent linear trend reported herein may result from the larger heterogeneity of our thiobenzoquinone monolayers which provides greater distributions (larger error bars) in the measured values reported in Fig. 7. 4. Conclusions We have outlined a place-exchange method to afford a mixed monolayer which includes dithiol molecules covalently attached to gold surfaces through only one thiol end group. It has been shown that these monolayers can be chemically modified in aqueous solution at room temperature by the addition of 1,4benzoquinone. With this chemically modified electrode substrate, we have been able to study the thermodynamics and kinetics of the quinone REDOX moiety using electrochemical and in situ spectroelectrochemical techniques. Both the formal potential and the heterogeneous rate constant have been measured using these techniques over a wide range of pHs and are in good agreement. The formal potential is found to shift to more positive potentials with increasing electrolyte acidity at a rate of ∼60 mV pH−1 as is expected for a two-electron, two-proton system. The logarithm of the apparent standard rate constant decreases linearly with pH, which is not predicted by models of either step-wise or concerted

[1] H.O. Finklea, Electroanalytical Chemistry, Plenum Press, New York, 1996. [2] C.E.D. Chidsey, C.R. Bertozzi, T.M. Putvinski, A.M. Mujsce, J. Am. Chem. Soc. 112 (1990) 4301. [3] H.O. Finklea, D.D. Hanshew, J. Am. Chem. Soc. 114 (1992) 3173. [4] R.J. Forster, L.R. Faulkner, J. Am. Chem. Soc. 116 (1994) 5444. [5] E. Katz, N. Itzhak, I. Willner, Langmuir 9 (1993) 1392. [6] Y.T. Long, E. Abu-Irhayem, H.B. Kraatz, Chem. – A Eur. J. 11 (2005) 5186. [7] J.F. Smalley, S.B. Sachs, C.E.D. Chidsey, S.P. Dudek, H.D. Sikes, S.E. Creager, C.J. Yu, S.W. Feldberg, M.D. Newton, J. Am. Chem. Soc. 126 (2004) 14620. [8] S.A. Trammell, D.S. Seferos, M. Moore, D.A. Lowy, G.C. Bazan, J.G. Kushmerick, N. Lebedev, Langmuir 23 (2006) 942. [9] S.A. Trammell, D.A. Lowy, D.S. Seferos, M. Moore, G.C. Bazan, N. Lebedev, J. Electroanal. Chem. 606 (2007) 33. [10] S.A. Trammell, M. Moore, D. Lowy, N. Lebedev, J. Am. Chem. Soc. 130 (2008) 5579. [11] S.A. Trammell, M. Moore, T.L. Schull, N. Lebedev, J. Electroanal. Chem. 628 (2009) 125. [12] H.D. Sikes, J.F. Smalley, S.P. Dudek, A.R. Cook, M.D. Newton, C.E.D. Chidsey, S.W. Feldberg, Science 291 (2001) 1519. [13] C.E.D. Chidsey, Science 251 (1991) 919. [14] A.D. Abhayawardhana, T.C. Sutherland, J. Phys. Chem. C 113 (2009) 4915. [15] A. Bulovas, N. Dirvianskyte, Z. Talaikyte, G. Niaura, S. Valentukonyte, E. Butkus, V. Razumas, J. Electroanal. Chem. 591 (2006) 175. [16] H.-G. Hong, W. Park, Langmuir 17 (2001) 2485. [17] H.G. Hong, W. Park, E. Yu, J. Electroanal. Chem. 476 (1999) 177. [18] A.G. Larsen, K.V. Gothelf, Langmuir 21 (2004) 1015. [19] J. Lukkari, K. Kleemola, M. Meretoja, J. Kankare, Chem. Commun. (1997) 1099. [20] G. March, S. Reisberg, B. Piro, M.C. Pham, M. Delamar, V. Noel, K. Odenthal, D.B. Hibbert, J.J. Gooding, J. Electroanal. Chem. 622 (2008) 37. [21] M. Nagata, M. Kondo, Y. Suemori, T. Ochiai, T. Dewa, T. Ohtsuka, M. Nango, Colloids Surf. B 64 (2008) 16. [22] K. Nakano, K. Ohkubo, H. Taira, M. Takagi, N. Soh, T. Imato, J. Electroanal. Chem. 623 (2008) 49. [23] W. Park, J. Ahmed, S. Kim, Colloids Surf. B 68 (2009) 120. [24] S. Sarkar, S. Sampath, Langmuir 22 (2006) 3388. [25] Y. Sato, M. Fujita, F. Mizutani, K. Uosaki, J. Electroanal. Chem. 409 (1996) 145. [26] S. Ye, A. Yashiro, Y. Sato, K. Uosaki, J. Chem. Soc. Faraday Trans. 92 (1996) 3813. [27] L. Zhang, T. Lu, G.W. Gokel, A.E. Kaifer, Langmuir 9 (1993) 786. [28] M. Gunner, J. Madeo, Z. Zhu, J. Bioenerg. Biomembr. 40 (2008) 509. [29] E. Laviron, J. Electroanal. Chem. 109 (1980) 57. [30] E. Laviron, J. Electroanal. Chem. 124 (1981) 1. [31] E. Laviron, J. Electroanal. Chem. 124 (1981) 9. [32] E. Laviron, J. Electroanal. Chem. 146 (1983) 1. [33] E. Laviron, J. Electroanal. Chem. 146 (1983) 15. [34] E. Laviron, J. Electroanal. Chem. 164 (1984) 213. [35] C. Costentin, Chem. Rev. 108 (2008) 2145. [36] C. Costentin, M. Robert, J.M. Savéant, J. Am. Chem. Soc. 128 (2006) 8726. [37] C. Costentin, M. Robert, J.M. Savéant, J. Am. Chem. Soc. 129 (2007) 5870. [38] C. Costentin, M. Robert, J.M. Savéant, J. Electroanal. Chem. 588 (2006) 197. [39] S. Hammes-Schiffer, A.V. Soudackov, J. Phys. Chem. B 112 (2008) 14108. [40] H.O. Finklea, J. Phys. Chem. B 105 (2001) 8685. [41] V. Budavári, Á. Szucs, C. Somlai, M. Novák, Electrochim. Acta 47 (2002) 4351. [42] V. Budavári, Á. Szucs, A. Oszkó, M. Novák, Electrochim. Acta 48 (2003) 3499. [43] W. Zhang, S.M. Rosendahl, I.J. Burgess, J. Phys. Chem. C 114 (2010) 2738. [44] W. Zhang, I.J. Burgess, Phys. Chem. Chem. Phys. (2010). doi:10. 1039/C0CP01251C. [45] J.S. Yadav, T. Swamy, B.V.S. Reddy, D.K. Rao, J. Mol. Catal. A: Chem. 274 (2007) 116. [46] V.K. Tandon, H.K. Maurya, Tetrahedron Lett. 50 (2009) 5896. [47] S.N. Port, D.J. Schiffrin, T. Solomon, Langmuir 11 (1995) 4577. [48] M. Osawa, in: R.C. Alkire, D. Kolb, P.N. Ross, J. Lipkowski (Eds.), Diffraction and Spectroscopic Methods in Electrochemistry, Wiley-VCH, Weinheim, 2006 (Chapter 9). [49] B. Han, Z. Li, T. Wandlowski, A. Baszczyk, M. Mayor, J. Phys. Chem. C 111 (2007) 13855. [50] S. Pronkin, T. Wandlowski, J. Electroanal. Chem. 550–551 (2003) 131. [51] Z.Y. Zhou, S.C. Lin, S.P. Chen, S.G. Sun, Electrochem. Commun. 7 (2005) 490. [52] Z.Y. Zhou, S.G. Sun, Electrochim. Acta 50 (2005) 5163. [53] I. Burgess, B. Seivewright, R.B. Lennox, Langmuir 22 (2006) 4420. [54] H. Miyake, S. Ye, M. Osawa, Electrochem. Commun. 4 (2002) 973. [55] J.M. Delgado, J.M. Orts, J.M. Pérez, A. Rodes, J. Electroanal. Chem. 617 (2008) 130.

4368

S.M. Rosendahl, I.J. Burgess / Electrochimica Acta 56 (2011) 4361–4368

[56] P. Kohli, K.K. Taylor, J.J. Harris, G.J. Blanchard, J. Am. Chem. Soc. 120 (1998) 11962. [57] M.L. Carot, M.J. Esplandiu, F.P. Cometto, E.M. Patrito, V.A. Macagno, J. Electroanal. Chem. 579 (2005) 13. [58] H. Hamoudi, Z. Guo, M. Prato, C. Dablemont, W.Q. Zheng, B. Bourguignon, M. Canepa, V.A. Esaulov, Phys. Chem. Chem. Phys. 10 (2008) 6836. [59] S. Rifai, M. Laferrière, D. Qu, D.D.M. Wayner, C.P. Wilde, M. Morin, J. Electroanal. Chem. 531 (2002) 111. [60] S. Rifai, G.P. Lopinski, T. Ward, D.D.M. Wayner, M. Morin, Langmuir 19 (2003) 8916. [61] R. Meunier-Prest, G. Legay, S. Raveau, N. Chiffot, E. Finot, Electrochim. Acta 55 (2010) 2712.

[62] D. Garcia-Raya, R. Madueno, M. Blazquez, T. Pineda, J. Phys. Chem. C 114 (2010) 3568. [63] D.J. Fuchs, P.S. Weiss, Nanotechnology 18 (2007) 044021. [64] J.J. Calvente, G. Lopez-Perez, J.M. Jurado, R. Andreu, M. Molero, E. Roldan, Langmuir 26 (2009) 2914. [65] F. Bensebaa, R. Voicu, L. Huron, T.H. Ellis, E. Kruus, Langmuir 13 (1997) 5335. [66] R.G. Snyder, H.L. Strauss, C.A. Elliger, J. Phys. Chem. 86 (1982) 5145. [67] R.A. MacPhail, H.L. Strauss, R.G. Snyder, C.A. Elliger, J. Phys. Chem. 88 (1984) 334. [68] E. Laviron, J. Electroanal. Chem. 101 (1979) 19. [69] R.M. Haddox, H.O. Finklea, J. Electroanal. Chem. 550–551 (2003) 351. [70] H.O. Finklea, R.M. Haddox, Phys. Chem. Chem. Phys. 3 (2001) 3431.