3786
Langmuir 2007, 23, 3786-3793
Covalent Grafting of Glassy Carbon Electrodes with Diaryliodonium Salts: New Aspects Karina Højrup Vase, Allan Hjarbæk Holm, Kion Norrman,† Steen Uttrup Pedersen,* and Kim Daasbjerg* Department of Chemistry, UniVersity of Aarhus, Langelandsgade 140, DK-8000 Aarhus, Denmark, and Danish Polymer Centre, Risø National Laboratory, DK-4000 Roskilde, Denmark ReceiVed October 5, 2006. In Final Form: December 6, 2006 The applicability and versatility of the recently communicated procedure for the grafting of conducting carbon substrates by diaryliodonium salts is expanded. We have found that several types of organic arylic layers can be formed on the carbon surface and that the chemical functionalities of the thus formed layers can be varied extensively over electron withdrawing (for example, -NO2) to electron donating (for example, -OMe) groups. A comparative study involving the grafting of aryldiazonium salts reveals that, despite the two approaches being similar, iodonium salts exhibit spontaneous grafting to a significantly lower extent. Nevertheless, the grafted layer becomes less accessible to proton transport as visualized from a greater reluctance toward the reduction of surface-confined nitro groups to amino groups in acidic medium. Employment of unsymmetrical iodonium salts opens up the interesting possibility of forming organic films consisting of a mixture of two different aryl groups. Alternatively, such composite layers may be prepared by selecting iodonium and diazonium salts with comparable reduction properties. Analysis of the surfaces is carried out by means of cyclic voltammetry, X-ray photoelectron spectroscopy, and ToF-SIMS (timeof-flight secondary-ion mass spectrometry). The ToF-SIMS analysis primarily serves to provide unambiguous evidence for the covalent attachment of the organic layers to the surface.
Introduction Derivatization of carbon surfaces is an important tool in the design of new functional materials. The most widely employed electrochemical procedures for derivatizing carbon surfaces consist of the electrochemical oxidation of amines1 and alcohols2 or the electrochemical reduction of aryldiazonium salts.3 In particular, the latter approach is straightforward, as the aryl radicals generated upon reduction of the aryldiazonium salts attack the surface and form covalent C-C or C-O bonds. This ensures a high stability of the modified electrode, and applications have already been reported in fields as diverse as material chemistry,1a electrocatalysis,4 combinatorial chemistry,5 and analytical sensors.6 The further development of convenient approaches to fabricate such covalently modified surfaces is therefore important. Recently, our research group discovered a new method for the electrochemically assisted grafting of carbon materials.7 The approach is based on the reduction of iodonium salts and allows * To whom correspondence should be addressed. E-mail: sup@ chem.au.dk (S.U.P.);
[email protected] (K.D.). † Risø National Laboratory. (1) (a) Barbier, B.; Pinson, J.; Desarmot, G.; Sanchez, M. J. Electrochem. Soc. 1990, 137, 1757. (b) Deinhammer, R. S.; Ho, M.; Anderegg, J. W.; Porter, M. D. Langmuir 1994, 10, 1306. (c) Adenier, A.; Chehimi, M. M.; Gallardo, I.; Pinson, J.; Vila`, N. Langmuir 2004, 20, 8243. (2) Maeda, H.; Yamauchi, Y.; Ohmori, H. Curr. Top. Anal. Chem. 2001, 2, 121. (3) (a) Delamar, M.; Hitmi, R.; Pinson, J.; Save´ant, J.-M. J. Am. Chem. Soc. 1992, 114, 5883. (b) Allongue, P.; Delamar, M.; Desbat, B.; Fagebaume, O.; Hitmi, R.; Pinson, J.; Save´ant, J.-M. J. Am. Chem. Soc. 1997, 119, 201. (c) Pinson, J.; Podvorica, F. Chem. Soc. ReV. 2005, 34, 429. (d) Liu, Y.-C.; McCreery, R. L. J. Am. Chem. Soc. 1995, 117, 11254. (4) (a) Bourdillon, C.; Delamar, M.; Demaille, C.; Hitmi, R.; Moiroux, J.; Pinson, J. J. Electroanal. Chem. 1992, 336, 113. (b) Yang, X.; Hall, S. B.; Tan, S. N. Electroanalysis 2003, 15, 885. (5) (a) Coulon, E.; Pinson, J.; Bourzat, J.-D.; Commerc¸ on, A.; Pulicani, J.-P. Langmuir 2001, 17, 7102. (b) Coulon, E.; Pinson, J.; Bourzat, J.-D.; Commerc¸ on, A.; Pulicani, J.-P. J. Org. Chem. 2002, 67, 8513. (6) (a) Downard, A. J.; Roddick, A. D.; Bond, A. M. Anal. Chim. Acta 1995, 317, 303. (b) Zen, J.-M.; Kumar, A. S.; Tsai, D.-M. Electroanalysis 2003, 15, 1073.
the covalent immobilization not only of aryl groups, such as phenyl or nitrophenyl, but also of alkynyl groups. In particular, the immobilization of alkynyl groups may turn out to be important, as such graftings cannot be accomplished using any other known reductive procedure. Related approaches using diaryliodonium salts have recently been employed in the modification of platinum surfaces.8 In this paper, the applicability and versatility of the iodonium salt approach in the aryl derivatization of carbon surfaces is addressed. As chemical systems, we selected aryl-containing substrates having the following substituents: nitro, carboxylic acid, bromine, chlorine, hydrogen, methyl, and methoxy. The profound feature of essentially all of these substituents is that they will be detectable by either cyclic voltammetry, X-ray photoelectron spectroscopy (XPS), or ToF-SIMS (time-of-flight secondary-ion mass spectrometry) analysis. In particular, we present comparisons between selected diaryliodonium salts and aryldiazonium salts to elucidate the similarities as well as the differences of the two approaches. The possibility of forming organic films consisting of a mixture of two different aryl groups will also be addressed. Finally, unambiguous evidence of the covalent attachment of the aryl moieties is presented. Experimental Section Chemicals. Iodonium salts were prepared and purified according to published procedures,9 dried under vacuum, and then stored cold (-16 °C). Acetonitrile (MeCN, anhydrous, 99.9%) was purchased from Lab-Scan and used as received. Tetrabutylammonium tetrafluoroborate (Bu4NBF4) was prepared using standard procedures. Water was triple distilled. All other compounds were commercial and used in the highest grade available. A detailed description of (7) Vase, K. H.; Holm, A. H.; Pedersen, S. U.; Daasbjerg, K. Langmuir 2005, 21, 8085. (8) Dirk, S. M.; Pylypenko, S.; Howell, S. W.; Fulghum, J. E.; Wheeler, D. R. Langmuir 2005, 21, 10899. (9) Beringer, F. M.; Falk, R. A.; Karniol, M.; Lillien, I.; Masullo, G.; Mausner, M.; Sommer, E. J. Am. Chem. Soc. 1959, 81, 342.
10.1021/la0629227 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/20/2007
Grafting of Electrodes with Diaryliodonium Salts
Langmuir, Vol. 23, No. 7, 2007 3787
Scheme 1. Investigated Diaryliodonium Salts
our electrochemical instrumentation and XPS equipment was provided previously.7 ToF-SIMS. Mass spectrometric analysis was performed using a ToF-SIMS IV instrument (Ion-Tof GmbH, Mu¨nster, Germany) operated at a pressure of 10-8 mbar (with sample). Mass spectra were obtained employing 25 ns pulses of 25 keV Bi+ ions (primary ions), which were bunched to form ion packets with a nominal temporal extent of 5000. Procedure. Glassy carbon electrodes were carefully polished before any modification by successive treatments with diamond suspensions (Struers, grain size: 9, 3, 1, and 0.25 µm) followed by a thorough rinse with water and ethanol and finally 10 min of ultrasonic cleansing in ethanol. The electrochemical modification was performed via potentiostatic electrolysis, without any compensation of the solution resistance, of a 2 mM solution of the iodonium salt in 0.1 M Bu4NBF4/MeCN for 300 s at a potential 200 mV negative to the voltammetric peak (determined by an initial cyclic voltammetric sweep) unless otherwise specified in the text. After electrolysis the electrodes were thoroughly rinsed and ultrasonicated for 10 min in MeCN. Direct electrochemical analysis of the surface-modified electrodes was performed in either 0.1 M Bu4NBF4/MeCN or 0.1 M H2SO4. Alternatively, the blocking properties of the modified surfaces were investigated against two fundamentally different solution probes, that is, K3Fe(CN)6 in either 0.1 M H2PO4-/HPO42- (pH 7) or 0.1 M KCl/0.1 M HCl (pH 1) or ferrocene in 0.1 M Bu4NBF4/MeCN. All solutions were carefully deaerated with argon before use. The total charge, Q, used for the reduction of surface-grafted electroactive groups was obtained by coulombmetric integration of the background-subtracted electrochemical response recorded in cyclic voltammetry at a sweep rate of 0.1-2 V s-1. Two types of backgrounds were employed: a simple version in which a linear background was assumed and manually adjusted under the faradaic peak or an actual voltammetric signal recorded in the absence of a faradaic signal. The surface coverage, Γ, was then calculated using Faraday’s law, that is, Γ ) Q/nFA, where n is the number of electrons transferred and A represents the geometric area of the electrode. A more advanced fitting procedure involving the use of Gaussian or Lorentzian shapes for reversible faradaic signals or the extreme function for irreversible signals was also carried out in a few cases.10 The differences compared to the simple integration method using a linear background were, however, found to be minimal (∼5%), and all results reported herein were accordingly obtained with this (10) Origin 6.1, Originlab Corporation.
technique (in these calculations, we do not account for the average surface roughness of the electrodes). The systematic error is estimated to be of the order of 5-10%. Glassy carbon plates for XPS and ToF-SIMS analysis were rinsed ultrasonically for 10 min in hexane before use. The grafting was accomplished by using the plates as working electrodes in a conventional three-electrode setup in a similar manner as described previously.7 No supporting electrolyte was added to the iodonium salt solution. To ensure that both sides had been grafted to the same extent, the plates were turned over halfway during a 2 × 10 min grafting period. The glassy carbon plates were ultrasonicated in MeCN and hexane for 10 min before analysis.
Results and Discussion Grafting and Mechanistic Aspects. In our previous communication, a new procedure for functionalizing carbon surfaces with aryl groups using diphenyliodonium and 3,3′-dinitrodiphenyliodonium as the specific grafting agents was described.7 In the present work, the versatility of this method will be investigated through the study of an extended series of diaryliodonium salts, as depicted in Scheme 1. These particular iodonium salts were prepared to cover substituents going over electron withdrawing (for example, -NO2) to electron donating (for example, -OMe) groups. A few unsymmetrical salts (2, 5, and 10) were included to reveal the prospects of using such compounds in the formation of mixed layers. Figure 1 offers a representative example of the assembling of a film onto glassy carbon electrodes through the electrochemical reduction of 4 in 0.1 M Bu4NBF4/MeCN. The first cyclic
Figure 1. Cyclic voltammograms of 2 mM 4 in 0.1 M Bu4NBF4/ MeCN recorded at a freshly polished glassy carbon electrode at a sweep rate of 0.2 V s-1; sweep no. 1 (a), 2 (b), 3 (c), 4 (d), 5 (e), and after potentiostatic electrolysis at Ep,c - 0.2 V for 300 s (f). The solution was stirred between cycles.
3788 Langmuir, Vol. 23, No. 7, 2007
Vase et al.
Scheme 2. Electrochemical Cleavage Reactions of Diaryliodonium Salts (Y ) X or H)
voltammogram exhibits a broad and irreversible wave with a peak potential, Ep,c, of -0.86 V versus the saturated calomel electrode (SCE). On subsequent cycles, the currents are lower and the peak is shifted in the negative direction. Eventually, the voltammograms become featureless in the investigated region. Such a behavior is consistent with progressive electrode passivation as a consequence of electrode derivatization.3b In general, the same behavior is seen for all diaryliodonium salts studied, although the grafting efficiency is somewhat lower for 9.11 This is expressed by the facts that the peak current, ip,c, on the first sweep in this case is almost twice as large as that observed for the other salts (6 µA versus 3 µA) and that twice as many cycles are required to obtain an effective passivation of the surface.12 A similar observation, but to an even larger extent, was noted by Be´langer et al. in their study of a series of substituted aryldiazonium salts on gold and carbon surfaces.13 To have a better mechanistic understanding of the grafting processes, we carried out more fundamental studies of the electrochemical reduction process of the iodonium salts. Scheme 2 (Y ) X or H) outlines the expectancy that electrochemical reduction of iodonium salts produces aryl radicals which then may react with the electrode surface. For the unsymmetrical iodonium salts 2, 5, and 10 (Y ) H), two mechanistic pathways (a) and (b) may be predicted. While pathway (a) leads to the formation of a substituted aryl radical, pathway (b) produces the corresponding phenyl radical. We will demonstrate that whereas the first route dominates in the case of electron withdrawing groups, a mechanistic crossover takes place when going to electron donating groups. Since the Ep,c’s of the iodonium compounds are significantly lower by several hundred millivolts compared to those of the related diazonium salts,3,7 the aryl radicals might be further reduced to the corresponding anions. This situation may be investigated through a determination of the electron stoichiometry for the reduction process. It is important to ensure that such measurements are carried out under conditions characterized by ip,c ∝ ν1/2C*, where ν is the sweep rate and C* is the substrate concentration.14 Figure 2 illustrates the correlation between ip,c and C* for 1 as well as the 3-nitrophenyldiazonium salt. In the case of 1, we observe an approximate linear relation for 0.2-1 mM while ip,c levels off to become only marginally (11) An appropriate definition of the grafting efficiency should include the ratio of the surface coverage, Γ, and the consumed faradaic charge, Q. Because a precise experimental determination of Γ and Q can be quite tedious, it is, in most cases, convenient to obtain a rough estimate of the relative grafting efficiency simply from a comparison of the peak currents obtained on the first sweep in cyclic voltammetry or from the number of cycles required to completely passivate the electrode surface. (12) The grafting of all iodonium salts but 9 is characterized by relatively low and comparable peak currents on the initial sweep and by the same number of sweeps required to effectively passivate the surface. In general, ip,c can vary up to 1 µA (2 mM concentration; sweep rate ) 0.2 V s-1), depending on the salt and electrode, but there appears to be no simple correlation between ip,c and the substituent parameter, σ. (13) (a) Laforgue, A.; Addou, T.; Be´langer, D. Langmuir 2005, 21, 6855. (b) Baranton, S.; Be´langer, D. J. Phys. Chem. B 2005, 109, 24401. (14) Andrieux, C. P.; Pinson, J. J. Am. Chem. Soc. 2003, 125, 14801.
Figure 2. Plots of cathodic peak currents, ip,c, for 1 (b) and the 3-nitrophenyldiazonium salt (O) as a function of substrate concentration, C*. Data are extracted from cyclic voltammograms recorded at a sweep rate of 0.2 V s-1 in 0.1 M Bu4NBF4/MeCN.
concentration dependent at higher concentrations. The fact that the plot almost reaches a plateau reflects the formation of an organic layer on the surface and the progressive blocking of the electrode. At the same time, Ep,c is shifted negatively from -0.39 V versus SCE at 0.2 mM to -0.51 V versus SCE at 4 mM (recorded at a sweep rate of 0.2 V s-1). We may note that the same behavior is observed with the 3-nitrophenyldiazonium salt, although the transition to the plateaulike behavior now occurs much more abruptly and with substantially lower maximum attainable currents. We attribute these differences to possible preliminary adsorption and, in particular, spontaneous reaction of the diazonium salt occurring even prior to the potential sweep as well as a higher grafting efficiency of the diazonium salts during the sweep.15 Since the ip,c values of diaryliodonium salts at concentrations of 0.2 mM are essentially the same as those for the verified single electron reduction of aryldiazonium salts, we may state that the electron stoichiometry also is, by and large, one throughout the series of diaryliodonium salts.16 For instance, ip,c does not differ significantly for the easily reduced 1 with the nitro groups (Ep,c ∼ -0.39 V versus SCE at 0.2 mM; ν ) 0.2 V s-1) and 8 with the electron donating methyl groups (Ep,c ∼ -0.60 V versus SCE at 0.2 mM; ν ) 0.2 V s-1); that is, the substituent exerts no detectable influence on the stoichiometry of the mechanism. Accordingly, we suggest that Scheme 2 indeed provides an adequate picture of the grafting mechanism; that is, a one-electron reduction leads to the formation of an aryl radical, which may then attack the surface. A further reduction to the corresponding anion would give rise to a two-electron stoichiometry, and this aryl anion would not be able to graft to the surface. The possibility that the reduction of the iodonium salt first leads to the generation of an iodanyl radical prior to the expulsion of aryl iodide cannot (15) Although employment of low concentrations seems to prevent the grafting process, this is not necessarily true. In fact, a partial grafting is not expected to affect the electrochemical response, as long as the diffusion layer thickness is larger than the distance between the underivatized regions of the surface. Under such conditions, the grafted electrode will still behave as a normal disk electrode characterized by its original electrode area and with a square root dependency of the peak current on the sweep rate. On the other hand, a more extensive electrode derivatization will make the surface resemble an array of ultramicroelectrodes, now giving rise to sigmoidal-shaped voltammograms under steady-state conditions. Overall, the dependency of ip,c on ν can therefore turn out to be a rather complex function, depending on the progress of the electrode derivatization. (16) (a) The comparison is based on the assumption that both the transfer and diffusion coefficients of the various iodonium and diazonium salts are the same. However, this may turn out to be not completely true,16b considering that the smaller size of the diazonium salts would be expected to result in a relatively larger diffusion coefficient. (b) Bard, A. J.; Faulkner, L. R. Electrochemical Methods Fundamentals and Applications; Wiley: New York, 2001.
Grafting of Electrodes with Diaryliodonium Salts
be excluded. Actually, the very same discussion is valid in the case of the diazonium salt with the possible formation of a diazenyl radical as the intermediate prior to the expulsion of dinitrogen.14,17 On the basis of these measurements, we may deduce that the reduction potential of aryl radicals has to be more negative than -0.60 V versus SCE, noting that since aryl radicals are of the sigma-type, the substituent effect would be expected to be relatively small. Our estimated boundary value does not conflict with a previous determination of the Ep,c of the phenyl radical to a value of -0.64 V versus SCE in MeCN.14 However, in a forthcoming publication, we will present new results showing that the reduction potential of the phenyl radical is somewhat more negative than the published value.18 Analysis of the Grafted Electrodes. The blocking ability of the grafted electrodes toward the redox probes K3Fe(CN)6 and ferrocene was tested in aqueous solution and acetonitrile by cyclic voltammetry. Several examples of the resulting voltammograms are provided in the Supporting Information. These verify the formation of thick and insulating films on the electrodes. We have previously investigated to what extent the selected electrolysis potential affects the blocking behavior in a more systematic and detailed manner.7 Figure 3 illustrates the electrochemical signals of electrodes modified by 3-nitrophenyl, 4-bromophenyl, 4-chlorophenyl, and 4-methoxyphenyl groups. These signals not only confirm the presence of the organic modifiers on the electrode surface but also enable us to estimate the quantity of electroactive groups. In Table 1, the electrochemical reduction peak potentials, Ep,c, of the various iodonium salts are presented along with the electrochemically estimated surface coverages, Γ. We also analyzed the constitution of the modified surfaces using XPS. These results are presented in Table 2. 3-Nitrophenyl Films. Figure 3A illustrates the cyclic voltammogram of a 3-nitrophenyl-grafted glassy carbon electrode formed by potentiostatic electrolysis of 1. The voltammogram shows the chemically reversible one-electron reduction of the nitrophenyl group at ∼ -1.2 V versus SCE with ∼10 mV peak separation.3a,3b,19 A similar voltammogram could be observed if 2 was used as the precursor. The coulombmetric integration gave surface coverages of (5 ( 1) and (4 ( 1) × 10-10 mol cm-2, respectively. In entries 2 and 3 of Table 2, data from the XPS survey spectra are provided. The N/C ratios may be used to calculate approximate Γ values of 6 × 10-10 mol cm-2, if the simplest case of the carbon surface density being equal to that of HOPG is assumed.3d,20 Although such calculations are very dependent on the underlying carbon material, the result appears to be well in line with the electrochemical estimate of Γ. Table 3 outlines pertinent high-intensity ions for the negative ToF-SIMS analysis of the 3-nitrophenyl-grafted glassy carbon plates. The mass spectrometric analysis was first conducted on a clean plate to establish a background reference spectrum of the carbon material. A range of peaks originating from the underlying carbon framework could be identified, corresponding to a variety of hydrogenated and oxygenated fragments. (17) (a) Daasbjerg, K.; Sehested, K. J. Phys. Chem. A 2002, 106, 11098. (b) Daasbjerg, K.; Sehested, K. J. Phys. Chem. A 2003, 107, 4462. (18) Vase, K. H.; Rasmussen, S. W.; Dittmer, J.; Lund, H.; Pedersen, S. U.; Daasbjerg, K. In preparation. (19) Saby, C.; Ortiz, B.; Champagne, G. Y.; Be´langer, D. Langmuir 1997, 13, 6805. (20) As pointed out by the reviewers, the N/C ratio is a function of the XPS take-off angle and is effected by the underlying carbon framework. However, even if the actual N/C ratio should be significantly higher, the results can still be used for relative comparisons.
Langmuir, Vol. 23, No. 7, 2007 3789
Figure 3. Cyclic voltammograms of grafted films of (A) 3-nitrophenyl, (B) 4-bromophenyl, (C) 4-chlorophenyl, and (D) 4-methoxyphenyl on glassy carbon electrodes in 0.1 M Bu4NBF4/MeCN. Sweep rates were 1, 2, 2, and 1 V s-1, respectively. Electrodes were grafted by means of potentiostatic electrolysis of 2 mM solutions of the precursors 1, 4, 6, and 9, respectively, for 300 s at Ep,c - 0.2 V in 0.1 M Bu4NBF4/MeCN. When suitable, the plots include sweep no. 1 (a) and 2 (b).
It is important to emphasize that the mass resolution of the instrument is sufficiently good that the indicated ions and fragments can be assigned with a minimal risk of misinterpretation due to possible overlapping of peaks. Several nitro-containing
3790 Langmuir, Vol. 23, No. 7, 2007
Vase et al.
Table 1. Reduction Peak Potential (Ep,c) of 2 mM Solutions of Diaryliodonium Salts (First Sweep), Average Surface Coverages (Γ), and Peak Potentials (Ep) of Surface-Confined Electroactive Molecules and Freely Diffusing Analogues Measured by Cyclic Voltammetrya YC6H4-I+-C6H4X |-C6H4X |-C6H4X C6H5X Ep,c b Γ (×1010)c Epb Epb
salt 1 (Y ) X ) NO2) 2 (Y ) H, X ) NO2) 3 (Y ) X ) COOH) 4 (Y ) X ) Br) 5 (Y ) H, X ) Br) 6 (Y ) X ) Cl) 7 (Y ) X ) H) 8 (Y ) X ) Me) 9 (Y ) X ) MeO) 10 (Y ) H, X ) MeO)
-0.54 -0.67 -0.98 -0.86 -0.83 -0.88 -0.96 -0.99 -1.21 -0.93
5(1 4(1
-1.2d -1.2d
-1.2d -1.2d
5(1 3(1
-2.5e -2.5e -2.7e
-2.8e -2.8e -3.0f
1.7e
1.7e 1.7e
a
In each case, data are given as an average of the results obtained for at least five electrodes. b V vs SCE, in 0.1 M Bu4NBF4/MeCN; sweep rate ) 0.2 V s-1. c In mol cm-2. d Sweep rate ) 1 V s-1. e Sweep rate ) 2 V s-1. f Approximate value; the reduction occurs close to the solvent discharge. Table 2. XPS Surface Elemental Compositions of Modified Glassy Carbon Platesa,b entry 1 2 3 4 5 6 7
electrode
C
N
|-C6H4NO2d |-C6H4NO2e |-C6H4Brf |-C6H5 + Br- g |-C6H4Clh |-C6H5 + Cl- i
95 72 83 80 93 73 94
