Formation of Multinuclear Metal− Terpyridyl Complexes Covalently

Nov 5, 2009 - Canon Inc., 3-30-2, Shimomaruko, Ohta-ku, Tokyo 146-8501, Japan. Langmuir , 2009, 25 (23), pp 13340–13343. DOI: 10.1021/la902055h...
0 downloads 0 Views 2MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Formation of Multinuclear Metal-Terpyridyl Complexes Covalently Bound to Carbon Substrates Wataru Kubo,* Masashi Nagao, Yoichi Otsuka, Tsutomu Homma, and Hirokatsu Miyata Canon Inc., 3-30-2, Shimomaruko, Ohta-ku, Tokyo 146-8501, Japan Received June 9, 2009. Revised Manuscript Received October 26, 2009 Multinuclear complexes consisting of metal ions and a bis(terpyridyl) ligand were covalently bound to carbon substrates. The bonding of the complexes is initiated by the bonding of phenylterpyridine (PT) on the substrates using its in-situ-generated diazonium derivative, followed by stepwise coordination of the metal ions and the ligand on it. The bonding of the PT and the formation of the multinuclear complexes were confirmed by XPS, AFM, and CV measurements. The heterogeneous rate constant (k) at the Co complex-substrate interface was evaluated by chronoamperometry (CA). The estimated high k=(2.9-3.6)  103 s-1) would be attributed to the C-C bond at the interface without interrupting the conjugation. These multinuclear complexes bound to the carbon substrates can facilitate electron transfer from redox species such as enzymes.

Introduction Carbon is commonly used as an electrode for bioelectrochemical devices because of its chemical stability, wide potential window, and small nonspecific adsorption of proteins. A glucose sensor using an enzyme electrode consisting of a redox enzyme, a redox polymer, and a carbon substrate is a representative of such applications.1 In the enzyme electrode, metal complexes, which are responsible for the electron transport, are bound to the side chains of the redox polymer. The total efficiency of the electron transport from the redox center of the enzyme to the substrate is limited by the frequency of the collisions of the complexes, which is induced by the segmental motion of the polymer.2 Therefore, the design of a new structure, which allows more efficient electron transport, can improve the total performance of the device. A multinuclear complex consisting of metal ions connected by conjugated ligands having two coordination sites on the opposite side of a ligand molecule has been investigated as a molecular wire because the complex forms a linear supramolecular structure and exhibits excellent electron transport.3 The use of the multinuclear complex bound to the substrate without breaking the conjugation is expected to improve the electron transport from the enzyme to the substrate. In a previous report, the electropolymerization of the diazonium salt of a metal complex was used to fix the metal complex to carbon substrates.4 However, the complex prepared, using this method, grows randomly. A fixing method, which can control the growth of the complex more finely, is needed because direction and distance play critical roles in facilitating electron transport from the enzyme to substrate.5 *Corresponding author. E-mail: [email protected]. (1) Heller, A. J. Phys. Chem. 1992, 96, 3579. (2) (a) Aoki, A.; Heller, A J. Phys. Chem. 1993, 97, 11014. (b) Heller, A. Phys. Chem. Chem. Phys. 2004, 6, 209. (3) (a) Maskus, L. M.; Abru~na, H. D. Langmuir 1996, 12, 4455. (b) Kanaizuka, K.; Murata, M.; Nishimori, Y.; Mori, I.; Nishio, K.; Masuda, H.; Nishihara, H. Chem. Lett. 2005, 34, 534. (c) Kosbar, L.; Srinivasan, C.; Afzali, A.; Graham, T.; Copel, M. Krusin-Elbaum, L. Langmuir 2006, 22, 7631. (d) Nishimori, Y.; Kanaizuka, K.; Murata, M.; Nishihara, H. Chem. Asian J. 2007, 2, 367. (4) Jousselme, B.; Bidan, G.; Billon, M.; Goyer, C.; Kervella, Y.; Guillerez, S.; Hamad, E. A.; Goze-Bac, C.; Mevellec, J.-Y.; Lefrant, S. J. Electroanal. Chem. 2008, 621, 277. (5) Willner, I.; Katz, E. Angew. Chem., Int. Ed. 2000, 39, 1180.

13340 DOI: 10.1021/la902055h

We report the stepwise formation of multinuclear complexes, which consist of metal ions connected by conjugated ligands, bound to carbon substrates. The formation of the complexes is initiated by the bonding of phenylterpyridine (PT) to the substrates6 using its in-situ-generated diazonium derivative, followed by the coordination of a bis(terpyridyl) ligand via metal ions. This method binds, in a controlled way, a π-conjugated multinuclear metal complex to a carbon substrate and can facilitate electron transfer at the interface.

Experimental Section 0

4 -(4-Aminophenyl)-2,20 :60 ,200 -terpyridine (APT, Chart 1a) was synthesized on the basis of reported procedures.7 All other chemicals were used as purchased. Co(BF4)2 and Fe(BF4)2 were used as metal ion sources. Alumina-polished glassy carbon plates (GC, BAS Inc.) for X-ray photoelectron spectroscopy (XPS) and cyclic voltammogram (CV) experiments or freshly cleaved highly oriented pyrolytic graphite plates (HOPG, SPI) for atomic force microscopy (AFM) measurements were used as carbon substrates. Phenylterpyridine (PT)-bound carbon substrates were prepared under an N2 atmosphere as follows. A 0.1 M HCl solution (0.96 mL) containing 1.3 mg of APT was mixed with 44 μL of a 0.1 M NaNO2 aqueous solution and kept in an ice bath for 30 min. The resulting mixture was added to 29 mL of sodium phosphate buffer (20 mM, pH 7.0) containing 1.2 mg of sodium dodecyl sulfate (SDS) and kept in an ice bath for 1 h. The carbon substrates were dipped into the prepared solution in an ice bath for 3 h, followed by washing in water. To form metal complexes, the PT-bound substrates were dipped into a 0.1 M metal salt ethanol solution and a 0.5 mM 40 ,40000 -(1,4-phenylene)bis(2,20 :60 ,200 -terpyridine) (TPT, Chart 1b) CHCl3 solution for 3 min (Co) or 3 h (Fe) each and washed in ethanol or CHCl3 between dipping steps. An ultrasonic cleaner (100 W) was used in all washing processes of the GC substrate. Multinuclear complexes were prepared by repeating the ion and the ligand coordinating processes. The complex-forming processes were carried out at 25 °C in air. (6) Haight, R.; Sekaric, L.; Afzali, A.; Newns, D. Nano Lett. 2009, 9, 3165. (7) (a) Storrir, G. D.; Colbran, S. B.; Craig, D. C. J. Chem. Soc., Dalton Trans. 1997, 3011. (b) Yutaka, T.; Kurihara, M.; Nishihara, H. Mol. Cryst. Liq. Cryst. 2000, 343, 193.

Published on Web 11/05/2009

Langmuir 2009, 25(23), 13340–13343

Kubo et al.

Letter

Chart 1. Chemical Structures of (a) APT and (b) TPT

XPS was carried out using an ESCALAB 200i-XL system (VG) equipped with a monochromatic Al KR X-ray source. CV and CA measurements were carried out using a three-electrode cell configuration, which consists of the complex-bound GC as a working electrode, a Pt wire as a counter electrode, and an Ag/Agþ electrode for acetonitrile electrolyte solution or an Ag/AgCl (NaCl) electrode for aqueous electrolyte solution as a reference under an N2 atmosphere. The electrode area of the GC estimated from CV using K4[Fe(CN)6] was 0.48 cm2. CVs were measured using a PS-08 multipotentiostat system (Toho Technical Research). CA was performed using a PS-2000 potentiostat system (Toho Technical Research) equipped with an HP54540A oscilloscope (Hewlett-Packard) on the basis of reported procedures.3d,8 AFM images of HOPG substrates were obtained using a NanoNavi scanning probe microscope (SII) equipped with an easy PLL system (Nanosurf) and a SI-DF-20 cantilever (SII) operated in frequency modulation mode. The thickness of PT and the complex on the HOPG was estimated using a scratching technique described in a previous report.9

Results and Discussion The N 1s core-level photoelectron spectra of four GC substrates are shown in Figure 1. The GC treated with APT solution with NaNO2 (Figure 1a) shows a main peak at 400 eV and a shoulder at 401 eV; however, only a trace signal at 400 eV was observed for the sample treated with APT solution without NaNO2 (Figure 1b). The N 1s signal of 2,20 :60 ,200 -terpyridine was confirmed at 400 eV using a GC substrate on which the terpyridine ethanol solution was cast and dried (Figure 1c). The bare GC substrate (Figure 1d) shows no significant signal in the range of the bonding energy, 392-410 eV. The distinct N 1s peak observed for the GC treated with PT solution with NaNO2 indicates that PT is covalently bound to the GC substrate via a corresponding diazonium derivative formed by the reaction of APT with NaNO2. It is likely that PT is bound to the carbon surface with a C-C covalent bond, which is formed through the elimination of nitrogen by nucleophilic attack of the carbon on the diazonium cation of PT.10 The possibility of simple physical adsorption is excluded by the fact that the GC treated with APT solution without NaNO2 provides only a trace signal at 400 eV, as shown in Figure 1b. The N 1s peak recorded for the terpyridinecast GC and the PT-bound GC can be deconvoluted into two components, centered at 400 and 401 eV (Supporting Information Figure S1). The ratios of the two peaks are 1:0.15 for the terpyridine-cast GC and 1:0.55 for the PT-bound GC. The components can be assigned as the N atom of a neutral and protonated pyridine in a terpyridyl group, respectively. It is reported that the N 1s signal of protonated pyridine appears in the 1.2-2.2 eV higher bonding-energy region than that of nonprotonated pyridine.11 The relatively large signal corresponding to the protonated pyridine observed for PT-bound GC can be explained by the HCl solution used in the preparation process. (8) Chidsey, C. E. D. Science 1991, 251, 919. (9) Anariba, F.; DuVall, S. H.; McCreery, R. L. Anal. Chem. 2003, 75, 3837. (10) Toupin, M.; Belanger, D. Langmuir 2008, 24, 1910. (11) Lindberg, B. J.; Hedman J. Chem. Scr. 1975, 7, 155.

Langmuir 2009, 25(23), 13340–13343

Figure 1. XPS N 1s core-level spectra of GC substrates. GC dipped in APT solution (a) with NaNO2 and (b) without NaNO2, (c) GC on which 2,20 :60 ,200 -terpyridine ethanol solution was cast and dried, and (d) bare GC. The vertical axis represents counts normalized using the integral of the C 1s signal attributed to the GC substrate.

The AFM image of the PT-bound HOPG substrate (Figure 2a) demonstrates the formation of a layer consisting of many particles of less than 10 nm diameter. The thickness of this layer is estimated to be 1.7 nm, which is 1.4 times lager than the calculated thickness of a PT monolayer, estimated using computational chemistry software (MOE, CCG). This suggests that surfacebound PT exists not only in a monomeric structure but also in a dimeric and/or a trimeric structure that is formed through the intermolecular reaction of the diazonium cation of PT (Scheme 1). There are some reports on the formation of the multimeric structure of aryl compounds on carbon substrates via their diazonium ions.9,12 Because we used a moderate reaction condition, such as low APT concentration and a nonelectrochemical deposition process, the formation of the multimeric PT should be considerably suppressed in this study. (A thickness of no more than 1.3 times that calculated for the monolayer of PT supports this explanation.) However, the formation of the multimeric structure would be inevitable because the C-C bond is formed via the reaction of radical species having low site specificity. It is evident from Figure 2b that the observed particles of PT-HOPG become large after sequential treatment with Fe2þ and TPT. The thickness of the layer also increases to 2.4 nm. This growth of particles can be attributed to the dendritic complexation of TPT to the PT on the substrate via an Fe ion. This explanation is supported by the further growth of particles with the deposition cycle (Supporting Information Figure S3). The possibility that the particles shown in Figure 2b are the precipitation of iron hydroxide is excluded by the fact that the PT-bound HOPG substrate treated with Fe2þ and TPT under an N2 atmosphere shows the same image as Figure 2b. Figure 3a shows the CVs of the PT-bound GC substrate after repeated sequential complexation with Fe ions and TPT. The half-wave potential of the observed CV curves is close to the potential reported for a bis(terpyridyl)-iron complex.13 The peak current increases with the number of deposition cycles, which proves the increased number of layers of the complex fixed on the firmly bound PT on the carbon substrate. A small increase in the peak separation with the number of the deposition cycles indicates that electron transfer from the complex to the substrate is hardly retarded with the increase in thickness. Replacing Fe ions with Co ions allows similar electrochemical behavior, such as an increase (12) Kariuki, J. K.; McDermott, M. T. Langmuir 2001, 17, 5947. (13) Prasad, R.; Scaife, D. B. J. Electroanal. Chem. 1977, 84, 373.

DOI: 10.1021/la902055h

13341

Letter

Kubo et al.

Figure 2. Noncontact AFM images of (a) PT-bound HOPG and (b) PT-bound HOPG after Fe2þ and TPT complexation. (Height scale: (a) 1.6 nm and (b) 2.4 nm.) Scheme 1. Schematic Description of the Formation of a Multinuclear Complex on a Carbon Substrate

in the current with the number of deposition cycles (Supporting Information Figure S4). Figure 3b shows the charge density (or density of the attached complex), estimated from the CV curves for the Fe and Co complexes, plotted against the number of deposition cycles. The values of the charge density are the average of six samples for every point. The charge density estimated for the first deposition cycle of Fe and Co complexes gives a density of the attached complex of (1.0-1.1)  10-10 mol cm-2. This corresponds to about 70% of the value reported for a similar complex that is close packed on a gold substrate.14 Therefore, the observed increase in charge density in Figure 3b with the deposition cycle evidently proves the multilayered growth of the complexes, which is consistent with the AFM result. For the Fe complex, the charge density increases linearly with the number of deposition cycles up to four and then saturates around the seventh cycle. In contrast, the Co complex does not show a linear increase in charge density, and the saturation level is lower than that of the Fe complex. Note that the dipping times are optimized and the extension of the times did not increase the redox currents. The observed saturation of the charge density of the Fe complex would be ascribed to the suppression of growth of the Fe complex caused by the steric hindrance of the radially grown complexes formed from the multimeric PT bound to a substrate. This explanation is supported by the narrower space between the particles shown in the AFM image of the substrate after six dipping cycles than that after four cycles (Supporting Information Figure S3). Unlike the (14) Arana, C. R.; Abru~na, H. D. Inorg. Chem. 1993, 32, 194.

13342 DOI: 10.1021/la902055h

Figure 3. (a) CV curves of the Fe complex-GC recorded in 0.1 M Bu4NBF4 acetonitrile solution under an N2 atmosphere at a scan rate of 0.2 V s-1. The number shows the deposition cycles. (b) Charge density (left axis) or density of complex attached on the GC (calculated using charge density and electrode area: right axis) of the Fe (2) and Co (b) complexes plotted against the deposition cycles. (One cycle: one sequential dipping of a GC substrate in the metal ion and TPT solutions.)

result of this study, it is reported that an Fe-TPT multinuclear complex grew linearly from a monomeric terpyridine-thiol derivative ligand on an Au substrate.3b,d This difference supports the fact that the suppression of the growth of the Fe complex is caused by the multimeric PT bound to a substrate. The suppression of the growth of Co can be explained by the low compatibility of Co2þ ions with multinuclear octahedral coordination. Kosbar et al. reported that a similar Co multinuclear complex on an Au substrate did not form more than five layers and that the ionic radius of Co2þ does not match the range for consistent growth.3c This effect of ionic radius would be reproduced in this study. Langmuir 2009, 25(23), 13340–13343

Kubo et al.

The presence of surfactant and APT concentration are critical in forming a PT-metal complex on a carbon substrate. The redox current of the complex fixed on the substrate is difficult to measure when the substrate is treated with APT without using SDS. High APT concentration also causes the redox current to deteriorate. It is likely that both the lack of surfactant and the high PT concentration form covalently bound PT molecules on the substrate in the densely packed form through the hydrophobic or π-π interactions between planar PT molecules, such as thiopyridine molecules adsorbed on a gold substrate,15 and then the PT molecules sterically hinder the coordination of more metal ions. In addition to this, in the case of high APT concentration, a thick PT layer is formed on the substrate and would work as an insulator, which hinders electron transfer from the coordinated metal ions located far from the substrate. We choose nonelectrochemical deposition16 instead of electrochemical deposition,17 which is widely used to form a C-C bond between aryl derivatives and carbon substrates because the redox current observed using the substrate prepared by the nonelectrochemical deposition was larger than that produced by electrochemical deposition. The difference in the observed current can be explained by the concentration of the radical existing on the electrode surface. In the case of electrochemical deposition, the aryl radical is formed by the electroreduction of the diazonium ion addition to the radical existing in the case of the nonelectrochemical method. As a result, the same phenomenon is observed as in the case of the nonelectrochemical method using a high PT concentration. The complex bound to the carbon substrate without breaking the conjugation is expected to enhance the electron transfer at the interface. The CA measurement was carried out to evaluate k. Observed transient currents decayed single exponentially with time except for the initial time region affected by the capacitance current (Supporting Information Figure S5). Assuming the presence of nonmonomeric forms of PT on the substrate, this suggests that the timescale of electron transfer between a cobalt ion coordinated to a PT bound to GC via another PT molecule and GC is indistinguishable from that coordinated to a PT directly bound to GC. The k estimated from the transient current (15) Sawaguchi, T.; Mizutani, F.; Taniguchi, I. Langmuir 1998, 14, 3565. (16) Toupin, M.; Belanger, D. J. Phys. Chem. C 2007, 111, 5394. (17) (a) Delamar, M.; Hitmi, R.; Pinson, J.; Saveant, J. M. J. Am. Chem. Soc. 1992, 114, 5883. (b) Allongue, P.; Delamar, M.; Desbat, B.; Fagebaume, O.; Hitmi, R.; Pinson, J.; Saveant, J. M. J. Am. Chem. Soc. 1997, 119, 201. (c) Belmont, J. A. U.S. Patent 5,672,198, 1997.

Langmuir 2009, 25(23), 13340–13343

Letter

changes with the applied potential and takes values of (2.9-3.6)  103 s-1 in the observed potential region (Supporting Information Figure S6). Because the value of k is affected by many parameters, such as the dielectric constant of the solvent and the size of the redox species, the simple comparison of k to a reported value measured under a different condition is not a good approach. However, the k of this system shows an order-of-magnitude higher value than that for the pyrroloquinolinequinonecystamine-gold system (44.5 s-1),18 which is known as an efficient surface-confined mediator and is used in many enzyme electrode systems, an osmium complex covalently attached to a GC with conjugation breaking (90 s-1),19 and a π-conjugated iron-terpyridyl complex adsorbed on a gold substrate through a thiol group (>110 s-1).3d The fast electron transfer observed in this study is probably attributed to the bonding of the π-conjugated metal complex to the GC substrate, which also has a π-conjugated structure, without interrupting the conjugation. Because the value of k observed in this study is higher than the electron-transfer turnover rate for an excellent enzyme electrode system using a redox mediator (700 s-1)20 and is close to the highest turnover rate as far as we know (5  103 s-1),21 we believe that this system can be used as an alternate redox mediator for electrochemical devices such as enzyme electrodes. We report a method to bond multinuclear complexes covalently to carbon substrates through the nonelectrochemical deposition of PT followed by sequential complexation with a metal ion and TPT. The fast electron-transfer property of these substrates suggests the usefulness of these composite electrodes for bioelectrochemical applications. Acknowledgment. We are grateful to Prof. Hiroshi Nishihara and Mr. Yoshihiko Nishimori (The University of Tokyo) for their valuable suggestions. We thank Dr. Tomona Yutaka (Canon Inc.) for teaching us the synthesis of APT and Dr. Otto Albrecht for correcting the manuscript. Supporting Information Available: Deconvoluted XPS spectra, AFM image of bare HOPG, and CV curves of the Co-complex-bound GC. This material is available free of charge via the Internet at http://pubs.acs.org. (18) Willner, I.; Riklin, A. Anal. Chem. 1994, 66, 1535. (19) Boland, S.; Barriere, F.; Leech, D. Langmuir 2008, 24, 6351. (20) Zayats, M.; Katz, E.; Willner, I. J. Am. Chem. Soc. 2002, 124, 2120. (21) Xiao, Y.; Patolsky, F.; Katz, E.; Hainfeld, J. F.; Willner, I. Science 2003, 299, 1877.

DOI: 10.1021/la902055h

13343