Monolayer-Protected Clusters of Gold Nanoparticles - American

Jan 26, 2010 - †Department of Chemistry, University of Pretoria, Pretoria 0002, South ... Council for Scientific and Industrial Research (CSIR), Pre...
0 downloads 0 Views 3MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Monolayer-Protected Clusters of Gold Nanoparticles: Impacts of Stabilizing Ligands on the Heterogeneous Electron Transfer Dynamics and Voltammetric Detection Jeseelan Pillay,†,‡ Kenneth I. Ozoemena,*,†,§ Robert T. Tshikhudo,‡ and Richard M. Moutloali‡ †

Department of Chemistry, University of Pretoria, Pretoria 0002, South Africa, ‡Advanced Materials Division, Mintek, 200 Hans Strijdom Drive, Randburg 2125, South Africa, and §Energy and Processes Unit, Materials Science & Manufacturing, Council for Scientific and Industrial Research (CSIR), Pretoria 0001, South Africa Received November 25, 2009. Revised Manuscript Received January 4, 2010

Surface electrochemistry of novel monolayer-protected gold nanoparticles (MPCAuNPs) is described. Protecting ligands, (1-sulfanylundec-11-yl)tetraethylene glycol (PEG-OH) and (1-sulfanylundec-11-yl)poly(ethylene glycol)ic acid (PEG-COOH), of three different percent ratios (PEG-COOH:PEG-OH), 1:99 (MPCAuNP-COOH1%), 50:50 (MPCAuNP-COOH50%), and 99:1 (MPCAuNP-COOH99%), were studied. The electron transfer rate constants (ket/s-1) in organic medium decreased as the concentration of the surface-exposed -COOH group in the protecting monolayer ligand is increased: MPCAuNP-COOH1% (∼5 s-1) > MPCAuNP-COOH50% (∼4 s-1) > MPCAuNPCOOH99% (∼0.5 s-1). In aqueous medium, the trend is reversed. The surface pKa was estimated as ∼8.2 for the MPCAuNP-COOH1%, while both MPCAuNP-COOH50% and MPCAuNP-COOH99% showed two pKa values of about 5.0 and ∼8.0. These results have been interpreted in terms of the quasi-solidity and quasi-liquidity of the terminal -OH and -COOH head groups, respectively. MPCAuNP-COOH99% excellently suppressed the voltammetric response of the ascorbic acid but enhanced the electrocatalytic detection of epinephrine compared to the other MPCAuNPs studied. This study reveals important factors that should be considered when designing electrode devices that employ monolayer-protected gold nanoparticles and possibly for some other redox-active metal nanoparticles.

1. Introduction The science of gold nanoparticles (AuNPs) has emerged as an important broad new subdiscipline in the domain of colloids and surfaces.1 The AuNPs are easily modified with thioate ligands to provide superior stability and tune them for specific applications. Since the development of the Au55 cluster by Schmid et al.,2 these monolayer-protected clusters of gold nanoparticles (MPCAuNPs) have been a subject of intense fundamental investigation and practical applications in nanotechnology fields.1-4 Recent years have witnessed an influx of studies relating to the electron transport behavior of MPCAuNPs immobilized on gold electrode.5-9 Most MPCAuNPs are insoluble in water; thus, for electro-bioanalytical applications, it is essential to develop MPCAuNP systems that are stable, water-soluble, and capable of molecular recognition in aqueous media.10 This work, for the first time, reports the surface electrochemistry of stable, yet chemically versatile water-soluble novel MPCAuNPs self-assembled on gold electrode in terms of their (i) heterogeneous electron transfer (HET) dynamics in aqueous *Corresponding author: E-mail: [email protected]. (1) Sardar, R.; Funston, A. M.; Mulvaney, P.; Murray, R. W. Langmuir 2009, 25, 13840. (2) Schmid, G.; Pfeil, R.; Boese, R.; Bandermann, F.; Meyers, S.; Calis, G. H. M.; Vandervelden, W. A. Chem. Ber. 1981, 114, 3634. (3) Chen, S.; Murray, R. W.; Feldberg, S. W. J. Phys. Chem. B 1998, 102, 9898. (4) Pietron, J. J.; Murray, R. W. J. Phys. Chem. B 1999, 103, 4440. (5) Baum, T.; Bethell, D.; Brust, M.; Schiffrin, D. J. Langmuir 1999, 15, 866. (6) Chen, S. J. Phys. Chem. B 2000, 104, 663. (7) Bharathi, S.; Nogami, M.; Ikeda, S. Langmuir 2001, 17, 1. (8) Maye, M. M.; Luo, J.; Lin, Y.; Engelhard, M. H.; Hepel, M.; Zhang, C.-J. Langmuir 2003, 19, 125.  Heras, M. A.; Lopez-Palacios, J. Electrochem. Com(9) Ruiz, V.; Colina, A.; mun. 2006, 8, 863. (10) Tshikhudo, T. R.; Demuru, D.; Wang, Z.; Brust, M.; Secchi, A.; Arduini, A.; Pochini, A. Angew. Chem., Int. Ed. 2005, 44, 2.

Langmuir 2010, 26(11), 9061–9068

and nonaqueous media, (ii) surface ionization properties, and (iii) voltammetric recognition properties toward epinephrine and ascorbic acid. Here, new MPCAuNPs were prepared by varying the ratio of their two different stabilizing ligands: (1-sulfanylundec-11-yl)poly(ethylene glycol)ic acid (PEG-COOH) and (1-sulfanylundec-11-yl)tetraethylene glycol (PEG-OH). The extent to which these mixtures of protecting ligands influence the HET and ionization of the surface-exposed functional groups of the protecting ligands (surface pKa) is crucial for the potential applications of such platforms in many areas such as molecular electronics and chemical as well as biological sensing. The importance of catecholamine neurotransmitters (e.g., epinephrine) in the body has been well documented.11-13 However, the most common challenge in the electroanalytical detection of epinephrine in body fluids is the interference by ascorbic acid.

2. Experimental Section Material and Reagents. 2-(Dimethylamino)ethanethiol (DMAET) was obtained from Sigma-Aldrich. Citrate-stabilized gold nanoparticles (14 nm diameter) were prepared using the wellknown Turkevich-Frens procedure.14,15 The PEG-OH and PEG-COOH were purchased (Prochimia, Poland) or prepared using the established procedure.16 Three different percent ratios of PEG-COOH to PEG-OH (1:99, 50:50, and 99:1) were used to stabilize the gold nanoparticles. For example, the 50:50 (PEG-COOH/PEG-OH) was prepared as follows. The Ethanolic (11) Hernandez, P.; Sanchez, O.; Paton, F.; Hernandez, L. Talanta 1998, 46, 985. (12) Banks, W. A. Brain Res. 2001, 899, 209. (13) Schenk, J. O.; Milker, E.; Adam, R. N. J. Chem. Educ. 1983, 60, 311. (14) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discuss. Faraday Soc. 1951, 11, 55. (15) Frens, G. Nat. Phys. Sci. 1973, 241, 20. (16) Pale-Grosdemange, C.; Simon, E. S.; Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1991, 113, 12.

Published on Web 01/26/2010

DOI: 10.1021/la904463g

9061

Article

Pillay et al.

Scheme 1. Cartoon Illustrating the Functionalization of Gold Nanoparticles with Mixed Monolayers of PEG-COOH and PEG-OH Ligands

Scheme 2. Schematic of the Self-Assembly Process via Electrostatic Interaction between the Positively Charged DMAET Monolayer and the Negatively Charged Monolayer-Protected Clusters of Gold Nanoparticles

solutions of PEG-OH (1 mg, 0.5 mL) and PEG-COOH (1 mg, 0.5 mL) were mixed and added simultaneously under stirring into the citrate-stabilized gold nanoparticles (20 mL, 2 nM). The reaction mixture was stirred for 3 h and filtered using a 0.45 μm Millipore filter paper. The filtered particles were purified by repeated centrifugation and redispersion in distilled deionized water. This solution is abbreviated as MPCAuNPCOOH50%. The preparation procedure for the MPCAuNPs is shown as a cartoon in Scheme 1. The same procedure was used for the preparation of the other ratios; the ratio of 1:99 (PEG-COOH: PEG-OH) is abbreviated as MPCAuNP-COOH1%, while that of 99:1 (PEG-COOH:PEG-OH) is MPCAuNP-COOH99%. The final concentration of each solution mixture was 1.8 nM (20 mL), obtained by using a molar absorption coefficient of 4.2  108 M-1 cm-1 (at 526 nm) based on gold nanopartilces of 15 ( 1.2 nm diameter.17 High-purity pure water of resistivity 18.2 MΩ 3 cm was obtained from a Milli-Q Water System (Millipore Corp., Bedford, MA) and was used throughout for the preparation of solutions. Phosphate buffer solutions (PBS, pH 7.4) were prepared with appropriate amounts of K2HPO4 and KH2PO4. A 1 mM K4Fe(CN)6 and 1 mM K3Fe(CN)6 (1:1) mixture was prepared in 0.1 M KCl. The pKa experiment was performed with a PBS of potassium ferricyanide/ferrocyanide solution, pH adjusted appropriately with NaOH and HCl, in accordance with previous reports. All other reagents were of analytical grades and were used as received from the suppliers without further purification. Apparatus and Procedure. Electrochemical experiments were carried out using an Autolab potentiostat PGSTAT 302 (Eco Chemie, Utrecht, the Netherlands) driven by the GPES and FRA softwares version 4.9. Electrochemical impedance spectro(17) Demers, L. M.; Mirkin, C. A.; Mucic, R. C.; Reynolds, R. A.; Letsinger, R. L.; Elghanian, R.; Viswanadham, G. Anal. Chem. 2000, 72, 5535.

9062 DOI: 10.1021/la904463g

scopy (EIS) measurements were performed between 10 mHz and 10 kHz using a 5 mV rms sinusoidal modulation. The FRA software allowed the fitting of the raw EIS data to equivalent circuit models using a complex nonlinear least-squares (CNLS) routine, with Kramers-Kronig rule check. The electrolytes used for the electrochemistry were (i) chloroform containing 0.1 M tetrabutylammonium perchlorate (TBAP), (ii) a 0.5 M H2SO4, and (iii) a 1 mM of K4Fe(CN)6 and 1 mM K3Fe(CN)6 (1:1) mixture containing 0.1 M KCl. Gold electrode (BAS, r = 0.8 mm) modified with the MPCAuNPs was used as the working electrode, with a roughness factor of 1.34. A Ag|AgCl wire and Ag|AgCl saturated KCl were used as pseudoreference and reference electrodes for the nonaqueous and aqueous electrolytes, respectively. Platinum rod was used as counter electrode. Prior to the formation of the positively charged DMAET SAM as before,18-20 the gold electrode was first cleaned by repetitive cyclic voltammetry.21-23 The Au-DMAET electrode was then dipped in a water solution of MPCAuNPs stabilized with the required mixture of PEG-COOH and PEG-OH ligands (2 mL) for 18 h in an airtight container to form the required MPCAuNP-COOH-based electrode. The physically adsorbed species were removed by rinsing in a copious amount of blank SAM solvent prior to electrochemical experiments. The self-assembly strategy is summarized in Scheme 2. All solutions were deaerated with pure nitrogen (Afrox) prior to each. All experiments were performed at 25 ( 1 °C. (18) Caruso, F.; Rodda, E.; Furlong, D. N.; Haring, V. Sens. Actuators, B 1997, 41, 189. (19) Pillay, J.; Agboola, B. O.; Ozoemena, K. I. Electrochem. Commun. 2009, 11, 1292. (20) Pillay, J.; Ozoemena, K. I. Electrochim. Acta 2009, 54, 5053. (21) Ozoemena, K. I.; Nkosi, D.; Pillay, J. Electrochim. Acta 2008, 53, 2844. (22) Nkosi, D.; Ozoemena, K. I. Electrochim. Acta 2008, 53, 2782. (23) Agboola, B. O.; Ozoemena, K. I. Phys. Chem. Chem. Phys. 2008, 10, 2399.

Langmuir 2010, 26(11), 9061–9068

Pillay et al.

Article

Figure 1. (a) Cyclic voltammograms of the electrodes obtained in CH2Cl2 containing 0.1 M TBAP at a scan rate of 25 mV s-1. Curves (b), (c), and (d) represent the amplified portion (0-1.2 V, vs AgAgCl) of the Au-DMAET-MPCAuNP-COOH1%, Au-DMAET-MPCAuNPCOOH50%, and Au-DMAET-MPCAuNP-COOH99%, respectively.

3. Results and Discussion Spectroscopic and Microscopic Characterization. The preparation of the MPCAuNPs (summarized in Scheme 1) was adopted from previous work.10 The PEG-stabilized MPCAuNPs are highly stable in water and can be centrifuged, dried, and resuspended in aqueous solution without any loss of materials. Unlike their citrate-stabilized counterparts that changed from red to colorless solution after ∼2 months storage, the PEG-stabilized MPCAuNPs have not shown any detectable change in their rubyred solutions even after 8 months. Also, unlike most other hydrosols, the PEG-stabilized MPCAuNPs do not show any detectable aggregation in 2 M NaCl solution. Their UV-vis spectra (not shown) were characterized by the plasmon absorption band at 526 nm, with size similar to their precursor citratestabilized gold nanoparticles (14.0 ( 1.2 nm) as confirmed by their TEM images (Figure ESI 1). AFM images (Figure ESI 2) show that MPCAuNPs assembled as bundles with needlelike protrusions, presumably due to the strong van der Waals attractive forces existing between carbon chains. Cyclic Voltammetric Evolution and Electron Transfer in Nonaqueous Solution. Figure 1a shows the CV profiles of the electrodes in CH2Cl2 containing 0.1 M TBAP. Some well-defined voltammetric peaks are observed in the potential range of -0.8 and þ1.1 V (vs Ag|AgCl). The activities in the 0.0 to -0.8 V region may be attributed to the presence of the DMAET molecule since the bare gold electrode is the only electrode that shows no peak at -0.45 V. For clarity, the voltammetric evolutions in the 0.0 to þ1.1 V region are shown for the activities of the Au-DMAET-MPCAuNP-COOH1% (Figure 1b), Au-DMAETLangmuir 2010, 26(11), 9061–9068

MPCAuNP-COOH50% (Figure 1c), and Au-DMAET-MPCAuNP-COOH99% (Figure 1d). All the three PEG-stabilized MPCAuNPs exhibit three well-defined redox processes at equilibrium potential (E1/2 ≈ 0.53, 0.78, and 0.96 V vs Ag|AgCl wire), attributed to the discrete charging of the adsorbed particle double layers.24,25 The observation of the discretized double-layer charging is dependent on the nanoparticle potential change (ΔV) incurred upon a single electron transfer to/from the working electrode or any other electron donor/acceptor. The charging peak around 0.5 V (I) (Figure 1) appeared broad for all the MPCAuNP-based electrodes, suggesting the occurrence of a twoelectron transfer process arising from two close ΔVs—more pronounced at higher scan rate (Figure ESI 3). Thus, the two close peaks are assigned to MPC1þ/0 and MPC2þ/1þ, while processes II and III may be ascribed to the MPC3þ/2þ and MPC4þ/3þ, respectively. Impedance spectroscopy represents an important technique for interrogating the HET kinetics at gold electrodes modified with self-assembled molecular layers of several species,21-23,19,20,26-29 including the MPCAuNPs.6,24,25 Figure 2 presents typical comparative Nyquist plots obtained for the three modified (24) Chen, S.; Murray, R. W. J. Phys. Chem. B 1999, 103, 9996. (25) Chen, S.; Murray, R. W. Langmuir 1999, 15, 682. (26) Creager, S. E.; Wooster, T. T. Anal. Chem. 1998, 70, 4257. (27) Mathebula, N. S.; Pillay, J.; Toschi, G.; Verschoor, J. A.; Ozoemena, K. I. Chem. Commun. 2009, 3345. (28) Ozoemena, K. I.; Mathebula, N. S.; Pillay, J.; Toschi, G.; Verschoor, J. A. Phys. Chem. Chem. Phys. 2010, 12, 345. (29) Nkosi, D.; Pillay, J.; Ozoemena, K. I.; Nouneh, K.; Oyama, M. Phys. Chem. Chem. Phys. 2010, 12, 604.

DOI: 10.1021/la904463g

9063

Article

Pillay et al.

Figure 2. Comparative Nyquist plots of the (a) Au-DMAET-MPCAuNP-COOH1%, (b) Au-DMAET-MPCAuNP-COOH50%, and (c) AuDMAET-MPCAuNP-COOH99% in CH2Cl2 containing 0.1 M TBAP at a scan rate of 25 mV s-1. (d) is the equivalent circuit used in fitting the impedance spectra. The symbols in (a)-(c) represent the experimental data, while solid lines are the fitted curves using the equivalent circuit. Table 1. Comparative EIS Data Obtained for the Electrodes in CH2Cl2 Containing 0.1 M TBAP electrochemical impedance spectroscopy dataa electrode modifier

E (V)

Rs/kΩ cm2

Cdl/μF cm-2

Cads/μF cm-2

Rct/kΩ cm2

ket/s-1

MPCAuNP-COOH1%

0.53 3.06 (0.87); 9.50 (4.16); 8.56 (8.96); 12.43 (7.84); 0.78 3.01 (0.83); 9.93 (4.27); 7.38 (8.93); 13.05 (7.02); 5.19 ( 0.47 0.96 2.96 (0.76); 11.07 (3.64); 9.57 (9.37); 11.79 (8.01); 4.43 ( 0.41 0.53 3.34 (0.64); 10.01 (3.11); 11.56 (9.43); 8.67 (6.76); MPCAuNP-COOH50% 0.78 3.25 (0.65); 10.98 (3.29); 9.49 (8.16); 12.05 (6.92); 4.37 ( 0.42 0.96 3.25 (0.58); 12.83 (3.02); 11.13 (9.32); 10.98 (7.75); 4.09 ( 0.38 0.53 9.67 (1.11); 10.50 (3.68); 56.65 (8.61); 14.54 (7.67); MPCAuNP-COOH99% 0.78 9.61 (1.07); 11.02 (3.48); 63.08 (8.84); 13.94 (8.04); 0.57 ( 0.06 0.96 9.56 (1.08); 12.83 (3.68); 60.31 (9.77); 15.09 (8.78); 0.55 ( 0.08 a The values in brackets are the estimated error percentages obtained from the fitting using the electrical equivalent circuit shown in Figure 2d.

electrodes, based at different potentials (∼0.53, 0.78, and 0.96 V vs Ag|AgCl). The experimental data were satisfactorily fitted with the modified Randles electrical equivalent circuit (Figure 2d). In the model, the Rs is the solution or electrolyte resistance, Rct represents the electron-transfer resistance, CAu is double-layer capacitance of the gold electrode, and Cads is the capacitance of the adsorbed MPCAuNP species (values summarized in Table 1). The Bode plots (exemplified in Figure ESI 4) indicate that these MPCAuNP-modified gold electrodes exhibit pseudocapacitive behavior; the slopes of the log |Z| vs log f plots were similar (ca. -0.6) at the midfrequency region, while the phase angles were in the 49°-60° range. Since these values are less than the -1.0 or 90° expected of an ideal capacitive behavior, the CAu and Cads used in the fitting suggest constant phase elements (CPE), not true double-layer capacitances. 9064 DOI: 10.1021/la904463g

Cads is defined as eq 1:26 Cads ¼

F 2 AΓ 4RT

ð1Þ

where F is the Faraday constant, A is the area of the electrode, Γ is the surface coverage, R is the gas constant, and T is the Kelvin temperature. The values of surface coverage (determined using the one-electron processes II and III) are Au-DMAETMPCAuNP-COOH1% (ca. 9.2  10-12 mol cm-2 or 5.54  1012 molecules cm-2), Au-DMAET-MPCAuNP-COOH50% (ca. 8.5  10-12 mol cm-2 or 5.12  1012 molecules cm-2), and AuDMAET-MPCAuNP-COOH99% (ca. 10.7  10-12 mol cm-2 or 6.44  1012 molecules cm-2). Assuming a closed-packed structure of the surface MPCAuNP assembly, this corresponds to a Langmuir 2010, 26(11), 9061–9068

Pillay et al.

Article

Figure 3. Cyclic voltammetric profiles of the bare and MPCAuNP-modified electrodes in 0.5 M H2SO4 at a scan rate of 25 mV s-1.

center-to-center internanoparticle distance of ∼4 nm. This value is smaller than the physical diameter of 14 nm (core þ mixed ligand monolayers) and represents a surface coverage of about a magnitude higher than the expected monolayer for 14 nm (i.e., 8.47  10-12 mol cm-2 or 5.10  1011 molecules cm-2). The reason for this higher coverage is possibly due to relatively higher assembling time used in this work. However, attempts at using short assembling time did not produce a noticeable voltammeric response. The electron transfer rate constant (ket/s-1) of each of the electrodes was obtained from eq 2:26 ket ¼

1 2Rct Cads

ð2Þ

From Table 1, it is seen that the average ket value decreases as the concentration of the surface-exposed -COOH group in the protecting monolayer ligand increases: Au-DMAETMPCAuNP-COOH1% (∼5 s-1) > Au-DMAET-MPCAuNPCOOH50% (∼4 s-1) . Au-DMAET-MPCAuNP-COOH99% (∼0.5 s-1). Also, given that the ionization constant (pKa) values of alkanols are inherently higher than their corresponding alkanoic acids.30 Thus, this trend may be interpreted in terms of the hydrophobicity or affinity of the terminal functional groups (-COOH and -OH) with organic solvent such as CH2Cl2. Considering that the pKa of -OH-based monolayer ligands will be higher that their -COOH counterparts, the extent to which these MPCAuNPs will associate with the organic solvents will decrease as MPCAuNP-COOH1% > MPCAuNP-COOH50% > MPCAuNP-COOH99%. Next, we carried out this same cyclic voltammetry experiment in 0.5 M H2SO4 for the three MPCAuNPs (Figure 3). As expected, the quantized charging processes of the MPCAuNPs seen in the organic solution were not seen or clearly defined in aqueous solution. This observation agrees with the literature6,24 and is interpreted using the proposed equivalent circuit (Figure 2d). In organic solutions, the overall electrode doublelayer capacitance is governed by the adsorbed MPCAuNPs (i.e., Cads > CAu, Table 1),6,24 while the measured current response is the collective quantized charging of individual surface-confined (30) Bruice, P. Y. Organic Chemistry, 5th ed.; Prentice Hall: Englewood Cliffs, NJ, 2007.

Langmuir 2010, 26(11), 9061–9068

Figure 4. Cyclic voltammetric profiles of the bare and MPCAuNP-modified electrodes in 0.1 M KCl containing equimolar (1 mM each) mixture of K4Fe(CN)6 and K3Fe(CN)6 at a scan rate of 25 mV s-1.

MPCAuNP. In the aqueous solution, however, the inverse becomes the case (i.e., Cads , CAu).6,24 The electron transfer constants, also estimated from eq 2 of the EIS data (Figure ESI 5 and Table ESI 1) and fitted with circuit Figure 2d, decreased as follows: Au-DMAET-MPCAuNP-COOH99% (23.1 s-1) > AuDMAET-MPCAuNP-COOH50% (21.8 s-1) > Au-DMAETMPCAuNP-COOH1% (17.9 s-1). Note that the capacitance of any thiol-SAM is dependent on its terminal functional group and increases as -COOH > -OH > -CH3.31 In addition, the hydrophilic terminal groups are by nature quasi-liquids, while the hydrophobic groups are quasi-solids,31-33 meaning that the SAMs of the -COOH terminal groups (in our case, the MPCAuNP-COOH99%) should be more permeable to solution ions than those of the -OH-terminated MPCAuNPs (notably the Au-DMAET-MPCAuNP-COOH1%). Thus, the higher ket value of the MPCAuNP-COOH99% may also be explained by the quasiliquidity of this materials that allow the penetration of the solution species. Such penetration may be enhanced by (i) the electrostatic interactions between the negatively charged carboxylic headgroup of the MPCAuNP-COOH99% and the H3Oþ of the electrolyte solution and/or (ii) the repulsive interactions between the neighboring ionized -COOH head groups that could create some interparticle voids or pinholes that permit the penetration of the solution ions.34 As a contrast, the relatively lower ket of the more hydrophobic MPCAuNP-COOH1% is due to the unfavored interaction of the solution ions with the unionized, quasi-solid terminal -OH groups. Electron Transfer Kinetics in an Aqueous Solution of a Redox Probe. Electron transport properties of the electrodes were studied in 0.1 M KCl containing equimolar (1 mM) mixture of K4Fe(CN)6 and K3Fe(CN)6. Typical comparative CVs are shown in Figure 4. The modified electrodes exhibited stable electrochemistry as the voltammograms recorded did not change after several repetitive cyclings. From the CV, the current responses of the modified electrode are essentially the same as that of the bare gold electrode. This type of behavior has been elegantly described by the theoretical framework of Davies and (31) Finklea, H. O. In Bard, A. J., Rubinstein, I., Eds.; Electroanalytical Chemistry; Marcel Dekker: New York, 1996; Vol. 19, pp 109-335. (32) Chidsey, C. E. D.; Loiacono, D. N. Langmuir 1990, 6, 682. (33) Ozoemena, K. I.; Nyokong, T.; Westbroek, P. Electroanalysis 2003, 15, 1762. (34) Jun, Y. Y.; Beng, K. S. Electrochem. Commun. 2004, 6, 87.

DOI: 10.1021/la904463g

9065

Article

Pillay et al.

Table 2. Comparative EIS Data Obtained for the Electrodes in 0.1 M KCl Containing Equimolar Mixture (1 mM Each) of K4Fe(CN)6 and K3Fe(CN)6 electrochemical impedance spectroscopy dataa electrode modifier

Rs (Ω cm2)

CPE (μF cm-2)

n

Rct (Ω cm2)

103Zw

103kapp (cm s-1)

bare Au 4.96 (0.85); 47.26 (3.31); 0.86 (0.47); 46.50 (0.70); 0.12 (0.55); 5.71 ( 0.04 6.17 (1.37); 84.89 (4.99); 0.83 (0.79); 96.67 (1.58); 0.10 (1.56); 2.70 ( 0.04 MPCAuNP-COOH1% 4.76 (1.99); 18.91 (6.10); 0.89 (0.78); 50.0 (1.02); 0.11 (0.96); 5.31 ( 0.05 MPCAuNP-COOH50% 4.44 (1.74); 20.24 (6.91); 0.89 (0.88); 23.11 (1.03); 0.12 (0.54); 12.20 ( 0.12 MPCAuNP-COOH99% a The values in brackets are the estimated error percentages obtained from the fitting using the electrical equivalent circuit shown in Figure 5c.

Compton35,36 as the type 4 behavior (i.e., planar/linear diffusion, wherein the diffusion layer thickness, δ, is much larger than the insulating layer leading to a complete or heavily overlapping of the adjacent diffusion layers and a linear concentration profile). The peak-to-peak separation potential (ΔEp) approximately follows as MPCAuNP-COOH1% (0.081 V)>MPCAuNPCOOH50% (0.079 V) > MPCAuNP-COOH99% (0.070 V), suggesting that the electron transport at the MPCAuNP-COOH99% is fastest compared to the other electrodes. Figure 5a shows typical Nyquist plots and the corresponding Bode-phase angle plot (Figure 5b) obtained for the electrodes in 0.1 M KCl containing the [Fe(CN)6]4-/[Fe(CN)6]3- solution at the formal potential of the electrodes (0.22 V vs Ag|AgCl saturated KCl). The EIS data were satisfactorily fitted with the modified Randles equivalent circuit model (Figure 5c), wherein the true capacitance is replaced by the CPE. In this model the Zw is the Warburg impedance, while other parameters retain their usual meaning. The apparent electron transfer rate constant (kapp/ cm s-1) of each of the electrodes was obtained from eq 3:37 kapp ¼

RT n2 F 2 ARct C

ð3Þ

where n is the number of electron transferred (1), Rct value is obtained from the fitted Nyquist plots, and C is the concentration of the [Fe(CN)6]3- (in mol cm-3, the concentration of [Fe(CN)6]3- and [Fe(CN)6]4- are equal), while other terms retain their usual meaning. From Table 2, the kapp value decreases as the concentration of the surface-exposed -COOH group in the protecting monolayer ligand decreases: Au-DMAET-MPCAuNPCOOH99% (12.2  10-3 cms-1) > Au-DMAET-MPCAuNPCOOH50% (5.3  10-3 cms-1) > Au-DMAET-MPCAuNPCOOH1% (2.7  10-3 cms-1). Also from the Bode plot, high frequency and low phase angle (which correspond to faster rate of reaction) decreased as follows: Au-DMAET-MPCAuNPCOOH99% (2.512 kHz at 41.2°) > Au-DMAET-MPCAuNPCOOH50% (1.514 kHz at 50.8°) > Au-DMAET-MPCAuNPCOOH1% (0.316 kHz at 51°), perfectly corroborating the kapp trend. The kapp for this outer-sphere redox probe in aqueous solution follows the same trend as in the 0.5 M H2SO4, but the reverse of the results obtained in the nonaqueous electrolyte already discussed. The high kapp value for the MPCAuNPCOOH99% is interpreted as for the experiment in the 0.5 M H2SO4, that is, in terms of its quasi-liquid nature as opposed to the quasi-solid nature of the MPCAuNP-COOH1%. The values of the Zw and Rs are essentially the same for the electrodes, indicating that ohmic resistance of the solutions was not significantly affected by any modification of the electrode surface. (35) Davies, T. J.; Compton, R. G. J. Electroanal. Chem. 2005, 585, 63. (36) Compton, R. G.; Banks, C. E. Understanding Voltammetry; World Scientific Publishing Co.: London, 2007. (37) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; John Wiley and Sons: New York, 2001.

9066 DOI: 10.1021/la904463g

Figure 5. (a) Comparative Nyquist plots resulting from the bare and MPCAuNP-modified electrodes obtained in 0.1 M KCl containing equimolar (1 mM each) mixture of K4Fe(CN)6 and K3Fe(CN)6. (b) Bode plots showing the (a) phase angle vs log frequency. (c) Modified Randles circuit used to fit the EIS data.

Surface pKa of the MPCAuNPs. The pKa of a surfaceimmobilized species is the value of the pH in contact with monolayer when half of the functional groups have been ionized. Surface pKa is easily determined with EIS strategy using solutions of [Fe(CN)6]4-/[Fe(CN)6]3- of different pH values.19-23 Figure 6 represents plots of the Rct vs pH (pH 1.91-10.0 range) for the three MPCAuNPs obtained in the [Fe(CN)6]4-/[Fe(CN)6]3solution as obtained from the impedance spectral profiles Langmuir 2010, 26(11), 9061–9068

Pillay et al.

Article

Figure 6. Comparative plots of the Rct vs pH for the three MPCAuNP-modified electrodes obtained in PBS solutions of [Fe(CN)6]4-/[Fe(CN)6]3- (pH 1.91-10.0 range).

(Figure ESI 6). There are four main findings in this experiment. First, at the pH < 8.0, the resistance to electron transport (Rct) follows as MPCAuNP-COOH1% > MPCAuNP-COOH50% > MPCAuNP-COOH99%, which means that at low pH the redox species experiences more difficulty in penetrating the MPCAuNP monolayer as the concentration of the -COOH group decreases. Second, the electron transport of the MPCAuNPs are much higher at pH > 8.0 than at pH < 8.0, indicating that the surface groups are more deprotonated at pH > 8.0, resulting in electrostatic repulsion between these negatively charged head groups and the negatively charged [Fe(CN)6]4-/[Fe(CN)6]3-. Note that at the pH > 8.0 the resistance to electron transport is more difficult at the MPCAuNP-COOH1% compared to the MPCAuNP-COOH50% and MPCAuNP-COOH99%, which means that the penetration of the redox probe at the terminal -OH group is more difficult compared to the -COOH groups. The efficient electron transport observed for the MPCAuNPCOOH50% and MPCAuNP-COOH99% may be related to the enhanced repulsive interactions among the neighboring deprotonated groups that create wider spaces (pinholes) for the penetration of the redox probe into the films. Third, unlike the MPCAuNP-COOH1%, both MPCAuNP-COOH50% and MPCAuNPCOOH99% exhibit a sigmoidal shape with a midpoint at ∼pH 5, signifying an initial pKa of ∼5. Fourth, at pH > 8, the three MPCAuNPs gave well-defined sigmoidal curves; pKa of ∼8.2 for the MPCAuNP-COOH1%, while both MPCAuNP-COOH50% and MPCAuNP-COOH99% showed two pKa values of ∼5.0 and ∼8.0. These two pKa’s may be related to two possible locations of the -COOH groups: presumably the well surface-exposed -COOH groups that easily access the electrolyte and the slightly “buried” -COOH groups that are somewhat less easily accessible to the electrolyte solution. In general, the surface pKa study confirms that there is higher hydrophilicity of the MPCAuNPCOOH99% and MPCAuNP-COOH50% compared to their MPCAuNP-COOH1% counterpart and hence better electron transport properties in the aqueous and nonaqueous electrolytes. Unfortunately, there is no accessible literature for surface pKa of MPCAuNP with which to compare our present data. Voltammetric Detection of Ascorbic Acid and Epinephrine. Figure 7a shows comparative cyclic voltammetric evolutions at the MPCAuNP electrodes in a PBS solution (pH 7.4) containLangmuir 2010, 26(11), 9061–9068

Figure 7. (a) Comparative cyclic voltammetric evolutions obtained for the three MPCAuNP-modified electrodes in 10 μM ascorbic acid (PBS, pH 7.4) at a scan rate of 25 mV s-1. (b) Comparative cyclic voltammetric evolutions of the Au-DMAETMPCAuNP-COOH1% and Au-DMAET-MPCAuNP-COOH99% in 10 μM epinephrine pH 7.4 and 9.68, respectively.

ing 10 μM ascorbic acid. Clearly, Au-DMAET-MPCAuNPCOOH1% shows an enhanced peak current response than the other MPCAuNPs with higher carboxyl contents. While the terminal -COOH group of the PEG ligands and ascorbic acid are expected to be fully deprotonated at this pH 7.4 (pKa of ascorbic acid = 4.17),38 the deprotonation of the terminal -OH group is highly unlikely considering the inherent high pKa values of alkanolic compounds. Thus, the excellent suppression of the voltammetric response of the ascorbic acid by the Au-DMAETMPCAuNP-COOH99% is ascribed to the repulsive interaction between the negatively charged ascorbic acid and the -COOH groups of the PEG ligands. We then investigated the response of epinephrine at the MPCAuNP-COOH1% and MPCAuNPCOOH99% based electrodes. Considering that epinephrine has different pKa’s (8.7, 9.9, and 12.0),39,40 we conducted identical experiments in pH 7.4 and pH 9.68 (Figure 7b) solutions containing 10 μM epinephrine. At both pH’s, the MPCAuNP-COOH99% gave better electrocatalytic response (low peak potential and high peak current) compared to the MPCAuNP-COOH1%. Also, (38) Pournaghi-Azar, M. H.; Ojani, R. Talanta 1995, 42, 1839. (39) Ciolkowski, E. L.; Maness, K. M.; Cahlil, P. S.; Wightman, R. M. Anal. Chem. 1994, 66, 3611. (40) Jewett, S. L.; Eggling, S.; Geller, L. J. Inorg. Biochem. 1997, 66, 165.

DOI: 10.1021/la904463g

9067

Article

Pillay et al.

unlike the MPCAuNP-COOH1%, the MPCAuNP-COOH99% showed two oxidation peaks for epinephrine in pH 9.68 conditions, possibly arising from the oxidation of the different forms of the epinephrine in this pH conditions. From the pH 7.4 results, it seems that for a simultaneous detection of the ascorbic acid and epinephrine in physiological pH medium, the Au-DMAETMPCAuNP-COOH1% may prove to be the best electrode. Further experiments are required to prove this concept and will be the subject of future engagement. However, the key finding in this experiment is that one can possibly tune the ability of the gold nanoparticles to electrochemically recognize ascorbic acid or epinephrine in aqueous medium by smart manipulation of the ratios of their monolayer-protecting PEG ligands.

4. Conclusions The key findings in this work should be emphasized. First, in both aqueous and nonaqueous solutions, there is electronic communication between the immobilized MPCAuNPs and the gold electrode, possibly from electron tunneling between these protecting ligands and the gold electrode. Second, the electronic communication is strongly influenced by the hydrophilicity of the head groups (-OH and -COOH); in aqueous solution the electron transport of the -COOH-based ligand is favored, while in the nonaqueous medium the electron transport of the -OH-

9068 DOI: 10.1021/la904463g

based ligands is favored. Third, the MPCAuNP with extreme -COOH content strongly suppressed the voltammetry of ascorbic acid but enhanced the detection of epinephrine. Simply put, this study has provided some useful physical insights into the impact of different ratios of the protecting -OH- and -COOHbased monolayer ligands of redox-active gold nanoparticles on the dynamics of electron transport between solution species, in organic and aqueous media, and the electrode surface. We envisage that these insights should provide some thoughts on the factors that must be considered when designing molecularscale electronics or electrocatalytic sensing devices that employ monolayer protected gold nanoparticles and possibly for some other redox-active metal nanoparticles. Acknowledgment. This work was supported by the National Research Foundation (NRF, South Africa) and MinTEK. J. P. thanks MinTEK and NRF for PhD bursaries. We thank N. Sosibo (MinTEK) for technical help and B. O. Agboola (CSIR) for valuable scientific discussions. Supporting Information Available: Additional figures (Figures ESI 1-6) showing typical TEM and AFM images, cyclic voltammograms at different scan rates, and impedance spectral profiles and data (Table ESI 1). This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(11), 9061–9068