Alicia
Gomis-Berenguer
a,
Maria
Gómez-Mingot
a,
Vicente
Montiel
a,
Antonio
Canals
b,
Thies
Thiemann
c,
Rashid O.
Kadara
d,
Craig E.
Banks
*d and
Jesús
Iniesta
*a
aPhysical Chemistry Department and Institute of Electrochemistry, University of Alicante, San Vicente del Raspeig, Alicante, 03690, Spain. E-mail: jesus.iniesta@ua.es; Fax: +34 965903537; Tel: +34 965903536
bAnalytical Chemistry, Nutrition and Bromatology Department, University of Alicante, San Vicente del Raspeig, 03690, Alicante, Spain
cDepartment of Chemistry, Faculty of Science, United Arab Emirates University, Al Ain, PO Box 17551, United Arab Emirates
dFaculty of Science and Engineering, School of Science and the Environment, Division of Chemistry and Environmental Science, Manchester Metropolitan University, Chester Street, Manchester M1 5GD, United Kingdom. E-mail: c.banks@mmu.ac.uk; Fax: +44 (0) 1612476831; Tel: +44 (0) 1612471196
First published on 21st June 2012
Screen printed graphite electrodes (SPGEs) have been characterized in the room temperature ionic liquid 1-hexyl-3-methylimidazolium hexafluorophosphate [C6MIM][PF6] by studying pertinent electrochemical parameters of the electroactive species ferrocene (Fc), 1,4-benzoquinone (BQ), 1,4-diphenyl-9,10-anthraquinone (AQ), tetracyclone (TC) and benzophenone-3 (BZ-3). Diffusion coefficients and kinetics calculations together with Digisim simulations for comparison were performed. Additionally, the reductive cyclovoltammograms of organic carbonyl containing BQ, AQ, TC and BZ-3 derivatives provided valuable information and comparison of the electrochemical reduction of the CO functional group in model molecules of different size, solubility and π-aryl aromaticity. Finally, potentialities for the electroanalytical measurement of BZ-3 at SPGEs via the room temperature ionic liquid [C6MIM][PF6] have been evaluated.
Carbonaceous materials are important in the electroanalytical and bioelectroanalytical fields for their acceptable conductivity, low cost and high versatility for electrochemical miniaturization. Recently, Pauliukaite et al.14 characterized carbon film electrodes in different RTILs and found applications for the electrochemical investigation of two ferrocene derivatives that are insoluble in water. Interestingly, RTILs media have been used by Maleki et al. which effectively act as binders for the fabrication of new carbon composite electrodes.15 Compton's group has performed an elegant comparison of electron transfer kinetics between different carbon materials such as basal and edge plane pyrolytic graphite and glassy carbon in RTILs and acetonitrile media using ferricinium ion/ferrocene (Fc+/Fc) as a redox couple.16
The effective performance of SPGEs has led to their widespread acceptance in environmental, biomedical and the major fields of analytical chemistry.17 Typically SPGEs are exclusively used in aqueous solutions and recent efforts have tried to make non-aqueous systems, thereby it is useful for SPGEs to operate in this type of media.18 Here, we aim to present the first investigation of SPGEs in an ionic liquid by investigating a range of probe molecules. Although we believe that this is the first such study of SPGEs in an ionic liquid, numerous other studies have successfully utilized ionic liquid layers on screen printed carbon electrodes which are then employed in aqueous systems. Hence, it is worth pointing out several works dedicated to the use of RTILs for the modification of the SPGE surface. In that respect, Ren et al.19 reported a chronocoulometric DNA sensor based on SPGEs doped with ionic liquids and polyaniline nanotubes. Further efforts have been focused on the determination of dopamine with the development of ionic liquid modified SPGEs.20,21 In the present work, electron transfer kinetic studies are performed using a wide variety of redox couples at SPGEs in RTIL 1-hexyl-3-methylimidazolium hexafluorophosphate [C6MIM][PF6]. This RTIL is one of the typical ionic liquids which is reported to have large viscosity, low miscibility with water and in which organic compounds are moderately soluble.22 The main aim of this work consists of making comparisons of two SPGEs, which have electrochemical behavior similar to that of edge plane or basal plane pyrolytic graphite where the electron kinetics in the former are significantly faster than that observed at the latter. Using these electrodes termed basal SPGE and edge SPGE, using [C6MIM][PF6] as the solvent, the graphite film electrochemical stability window and the voltammetric behavior of Fc, BQ, AQ, TC and BZ-3 were determined. We finally present the electrochemical reduction of BZ-3 in RTIL using cyclic voltammetry (CV) and differential pulse voltammetry (DPV) in order to gain insight into the electroanalytical determination of BZ-3.
Benzophenone derivatives, particularly BZ-3 are utilized in sunscreen agents. The concentration of BZ-3 (and other relevant additives) needs to be monitored to ensure they meet current legislation since there are undesirable dermatological side effects associated with the use of sunscreens. UV filters based on benzophenone derivatives can penetrate the human body through the skin, being excreted at trace levels in urine. Thus, the development of sensitive and selective analytical methods plays an important role.23
There are several works about the determination of benzophenone sunscreen using high-performance liquid chromatography (HPLC) 23–25 or gas chromatography (GC) coupled with ultraviolet spectroscopy (UV)26 and mass spectrometry,27 in cosmetics, there are limited reports using electrochemistry. To this end, mercury film electrodes,28 hanging drop mercury electrodes29 as well as epoxy-carbon composite electrodes30 were used for the electrochemical determination of BZ-3 in cosmetics. Recently Laranjeira et al.31 studied the electroanalytical determination of BZ-3 in sunscreen preparations using boron-doped diamond electrodes. Likewise Cardoso et al.29 studied the determination of three simultaneous sunscreens –included BZ-3- using differential pulse polarography. Therefore, it is clear that the electroanalytical determination of BZ-3 using SPGEs has not been explored.
Scanning electron microscopic (SEM) measurements of both basal and edge plane SPGE surfaces were carried out on a scanning electron microscope (JEOL JSM-6700F) at 15 kV. Raman spectroscopy of SPGE surfaces was carried out using a Fourier Transform Raman spectrometer Bruker RFS/100, with an Ar-Ne ion laser to induce the Raman spectrum. The laser beam was focused on the surface of the samples with the aid of a microscope, where three consecutive spots separated 100 μm each were illuminated for 60 s. The Raman spectrum was obtained from the average of three spectra at a basal and edge SPGE surface, respectively. XPS chemical analyses were performed with a VG-Microtech Multilab electron spectrometer, by using the Mg K−1//_/(1253.6 eV) radiation of twin anode in constant analyser energy mode with a pass energy of 50 eV.
CV, chronoamperometry (CA) and DPV measurements were carried out using a μ-Autolab III (Eco Chemie, The Netherlands) potentiostat/galvanostat and controlled by Autolab GPES software version 4.9 for Windows XP. All measurements were conducted using a three electrode configuration where the working and the counter electrode consist of graphite ink with a pseudo silver/silver chloride reference electrode and were fabricated in-house.33 The two types of fabricated SPGEs have a 3.1 mm diameter with a carbon counter and a silver/silver chloride pseudo-reference electrode on a flexible plastic base.16,34 The type of carbon ink used in the SPGE preparation was previously described elsewhere.24,26 For this study two types of SPGEs—named edge and basal—were used. All electrochemical experiments were carried out under a high purity argon flow (from Air-liquide) in order to keep an inert atmosphere and discard the electrochemical reduction of oxygen.
Connectors for the electrochemical connection of the SPGEs were purchased from Kanichi Research Services Ltd. (UK).35 SPGEs were used as received without further conditioning of the working electrode surface. CV experiments were carried out at 293 ± 2 K. CV simulations of the ferrocene/ferrocinium couple together with theoretical calculations of diffusion coefficients (D) and the apparent rate constants (k0) were obtained by Digisim (Digisim Professional version 3.0b) using a charge-transfer coefficient α of 0.5. All potentials were referred to the pseudo reference Ag/AgCl.
Glassy carbon (GC, Goodfellow, UK) and boron doped diamond (BDD, Windsor Diamonds, UK) electrodes with 3.0 mm diameter, were purchased for comparison of the electrochemical behavior in the RTIL for background response and the electrochemical reduction of BZ-3. Prior to use, GC and BDD electrodes were polished with an alumina slurry of successively smaller particles (1.00, 0.30 and 0.05 μm diameter), rinsed with water, sonicated in ultra pure water, then rinsed with acetonitrile and finally, dried under argon atmosphere. For both CV and DPV electrochemical experiments 1 mL of RTIL solution was thoroughly purged with high-purity argon to remove the residual oxygen. A gold wire was used as the counter electrode and a silver wire as the pseudo-reference electrode. At the end of the experiments, Fc was added to serve as an internal probe (0.53 V vs. Ag).
XPS spectroscopy reveals some differences in terms of functionalization of the SPGE surfaces. XPS of the basal SPGE surface, shown in the electronic supporting information, shows the presence of 87.5% atomic carbon in the basal SPGE surface (Fig. ESI3†), whereas the edge SPGE surface (Fig. ESI4†) presents a similar composition, 85.9%. Only graphitic and C–O groups at 284.2 (80.5%) and 285.6 (7.0%) respectively, are found at basal SPGEs (Fig. ESI3†). However, the de-convoluted XPS spectrum of the edge SPGE surface (Fig. ESI4†) revealed the presence of graphitic, C–O and carbonyl functional groups at 284.5 (65.3%), 285.7 (10.5%) and 286.6 (10.1%) eV. It is worth noting the remarkable increase in the oxygen functionalization at edge SPGEs which provide a more hydrophilic surface characteristic.
Fig. 1 Cyclic voltammograms of [C6MIM][PF6] at (A) basal SPGEs; scanning to negative potentials (solid trace) and positive potentials (dashed trace), (B) edge SPGEs; scanning to negative potentials (solid trace) and positive potentials (dashed trace). Scan rate 50 mV s−1. First scan recorded. |
However, the nature of the carbonaceous electrodes such as glassy carbon, carbon films or carbon nanotubes can determine the window limits11,12,41 when using for example [C4MIM][PF6] and [C4MIM][BF4]. Unlike the basal SPGE, the edge surface presented clear evidence of double layer capacitance of about ~35 μF. This difference between the basal and edge SPGE surfaces is likely due to the micro scale surface roughness (Fig. ESI1A and ESI1B†) and the higher incorporation of site like-basal planes on the edge SPGE surface rather than on the basal surface, providing a higher surface functionality on the edge SPGE surfaces,42 as demonstrated in this work by the XPS results. Fig. 1A and 1B display several anodic and cathodic waves, but it is not clear whether they stem from the carbon electrode or the ionic liquid. Therefore, we performed CVs of more conventional electrodes such as GC and BDD to compare with SPGEs. Fig. ESI5A† shows the first scan of the CV trace of [C6MIM][PF6] at GC in the absence of oxygen and scanned first to negative potentials, it shows a similar pattern as Fig. 1A, where the stability window is close to 5 V. Anodic peaks obtained at GC at −0.89 and +0.75 V vs. Fc+/Fc are similar to those obtained at basal SPGEs, −0.95 and +0.82 V vs. Fc+/Fc. Moreover, GC electrode did not reveal a cathodic current associated to the oxygen reduction. We also investigated the electrochemical behavior at BDD in [C6MIM][PF6] under de-aerated conditions, as shown in Fig. ESI5B.† In this case, no dioxygen reduction wave is observed but an anodic pre-wave is observed at +1.61 V vs. the pseudo-Ag (+1.08 V vs. Fc+/Fc); we presume that dioxygen reduction at basal and edge SPGEs could be related to either physisorbed oxygen on the carbon or an oxygen-containing functionality on the carbon.
Fig. 2 Cyclic voltammograms for the electrochemical behavior of 5.8 mM Fc in [C6MIM][PF6] obtained at basal SPGEs (A), and edge SPGEs (B) using 35 μL of sample solution. Scan rate 50 mV s−1. First scan (solid line) and third scan (dashed line). Inset figures, calibration plots of Fc (anodic peak current). |
Moreover, chronoamperometry experiments showed that the slope of plots of log10 of current vs. log10 of time were about 0.5 denoting that the results fitted the Cottrell equation for all tested Fc concentrations (experiments not shown). Accordingly, inset figures in 2A and 2B depict the linear plots of Ipvs. Fc concentration using the Randles–Sevcik equation44 providing the following fitting equations Ip (μA) = (0.77 μA m3 mol−1)·[Fc] − 0.20 μA and Ip (μA) = (1.31 μA m3 mol−1)·[Fc] + 0.23 μA for basal and edge SPGEs, respectively. Our results confirm that D values of Fc are independent of Fc concentration in the range tested values between 1.79 and 11.75 mM. Furthermore, single step CA experiments also showed no variation of the DFc with concentration, as previously stated by Rogers et al.45
Fig. 3A shows a comparison between the experimental and Digisim simulated CVs at a scan rate of 50 mV s−1. DFc in [C6MIM][PF6] at basal SPGEs was calculated by CV and single step CA experiments, whereas DFc+ was obtained from Digisim. DFc was 4.5 × 10−12 m2 s−1 using the Cottrell equation44 and DFc+ was 4.0 × 10−12 m2 s−1 obtained from Digisim. The theoretical DFc value using Randles–Sevcik equation was 2.0 × 10−12 m2 s−1, which is lower than the calculated (5.9 × 10−12 m2 s−1) using a shorter alkyl chain such as [C4MIM][PF6].16,45 This difference in the diffusion coefficient could be explained by the lower viscosity of [C4MIM][PF6].22
Fig. 3 (A) Experimental Fc 5.8 mM CVs in [C6MIM][PF6] at basal SPGEs (dashed trace) compared to simulation with Digisim model (solid trace). First scan recorded, scan rate 50 mV s−1. (B) Anodic and cathodic peak potentials plots vs. log10 scan rate (V s−1), 5.8 mM Fc in [C6MIM][PF6] at basal SPGEs compared to Digisim simulation. |
Fig. 3B shows the experimental Ep values for the anodic and cathodic peaks compared to Digisim values. ΔEp separation and the kinetic parameters are in a good agreement to those obtained by Digisim. The heterogeneous rate constant (k0) from Digisim was 2.8 × 10−6 cm s−1 which is similar to the experimental one, 2.6 × 10−6 cm s−1, indicating good fitting is obtained between Digisim simulation and experimental data. Moreover, the experimental k0 value was in agreement with that presented by Compton's group16 where they studied the electron transfer kinetics of Fc in 1-butyl-3-methylimidazolium hexafluorophosphate [C4MIM][PF6], using different carbonaceous materials such as basal plane pyrolytic graphite or GC.
Fig. 4 CVs of 5.2 mM BQ in [C6MIM][PF6] at basal SPGEs (A) and edge SPGEs (B). Solid line (first scan), dashed line (third scan). Scan rate 50 mV s−1. |
Next, we further investigated the electrochemical behaviour of other carbonyl containing compounds such as AQ. Fig. 5 depicts the voltammetric response of 5.8 mM AQ at basal and edge SPGEs in [C6MIM][PF6]. During the first scan in Fig. 5A, a reduction pre-wave appeared at a potential of −0.72 V which could correspond either to the electrochemical reduction of residual dioxygen dissolved in the RTIL as shown in the blank (Fig. 1A) or a pre-adsorption process of AQ. After the second scan, the reductive pre-wave disappeared and reproducible CVs were obtained. Two well-defined, chemically reversible redox couples were observed. This corresponds to the reduction of neutral AQ to the radical anion, AQ·− (first step), and the reduction of the anion radical to AQ2−. This is consistent with the literature on aprotic solvents48,49 or more recently 9,10-anthraquinone in the RTIL, 1-butyl-3-methylimidazolium hexafluorophosphate [C4MIM][PF6].50 The electron transfer process was diffusion-controlled (log10Ipcat/μA = 0.40 log10(v/V s−1) + 0.79) as shown in Fig. ESI7,† and the estimated DAQ for the neutral species was 2.68 × 10−13 m2 s−1 from the Randles–Sevcik equation. Furthermore, we also investigated the electrochemical behavior of AQ in non de-aerated solutions and the onset potential of oxygen reduction under our experimental conditions. Fig. ESI8† depicts the voltammetric response of 5.5 mM AQ at basal SPGEs in [C6MIM][PF6] under air-saturated conditions. It appears that the electrochemistry of AQ (Fig. 5, ESI8†) may be significantly influenced by oxygen reduction and therefore faradaic currents of oxygen reduction-oxidation are comparable with those of Fig. 1A. We have also found that under air-saturated conditions the first reduction potential wave involved an enhancement of about 1 μA compared to argon-saturated conditions and a similar increase in current was found for the second reduction wave. It is remarkable that oxygen reduction starts at similar potentials as those in Fig. 1A and this potential refers to pre-wave potential shown for the electro-reduction of tetracyclone (vide infra). Comparison of the electrochemical behavior of AQ was made to a more conventional electrode such as the BDD electrode (results not shown) in an acetonitrile/tetrabutylammonium hexafluorophosphate system to determine if the reductive pre-wave occurring at about −0.70 V vs. standard reference electrode Ag/AgCl/(3 M KCl) was stemming from the SPGE or the RTIL itself. Results revealed an identical CV pattern with a reduction pre-wave before the appearance of the first reduction peak (results not shown) suggesting it is not presumably a contribution from the RTIL.
Fig. 5 CVs of 5.5 mM AQ in [C6MIM][PF6] at basal SPGEs (A) and edge SPGEs (B). Solid line (first scan), dashed line (third scan). Scan rate 50 mV s−1. |
Fig. 6 depicts the electrochemical behavior of TC in [C6MIM][PF6] at basal SPGEs under de-oxygenation. Similarly to previous reports,51–53 we found two reduction waves which could correspond to two successive one-electron processes. For the first step we surmise that the formation of the relatively stable radical anion is taking place, which is resonance-stabilized, not only by the phenyl rings conjugated into the cyclic dienone, but also by the aromaticity of the cyclopentadienyl anion (potential of −0.77 V with a ΔEp separation of 85 mV). The second quasi-reversible electron transfer leads to a TC dianion described elsewhere54 with a potential of −1.14 V and a ΔEp separation of 110 mV. Separation between the redox potentials between the radical anion and the dianion was 347 mV which is similar to those presented for the observed when applied to the electrochemical reduction of BQ and AQ, see Fig. 4 and 5. As shown in Fig. 6B inset, a reduction pre-wave was observed at about −0.65 V previous to the first cathodic peak. After the fifth scan, stable CVs were obtained with a slight current reduction of the first cathodic peak. Interestingly, a successive scan up to −1.50 V revealed again the formation of that pre-wave at −0.65 V. The electron transfer process among the scan rates tested (20–400 mV s−1) was diffusion controlled (log10Ipcat/μA = 0.42 log10 (v/V s−1) + 0.21) as shown in Fig. ESI9,† and a DTC for the neutral species was estimated to be 9.87 × 10−14 m2 s−1 using Randles–Sevcik equation. When the potential was scanned from 0.00 V to−1.50 V and then scanned back (solid line, Fig. 6), a current decrease on the re-oxidation wave at a potential of −0.70 V, together with the appearance of a new oxidation peak at −0.31 V which is related to the second reduction peak, was observed. The nature of that species will be addressed in future work.
Fig. 6 CVs of 2.7 mM TC in [C6MIM][PF6] at basal SPGEs (A) and edge SPGEs (B). Solid line (first scan), dashed line (third scan). Inset figure in Fig. 6B shows five consecutive cycles between 0 and −0.9 V. Scan rate 50 mV s−1. |
Fig. 7 CVs of 6.2 mM BZ-3 in [C6MIM][PF6] at basal SPGEs (first scan). Scan rate 20 mV s−1. |
DPV was chosen as a methodological technique for the electroanalytical determination of BZ-3. Fig. 8 depicts the DPV measurements of a standard 6.2 mM solution at basal SPGE and shows two reductive peaks at −1.40 V vs. Ag/AgCl (−1.85 V vs. Fc+/Fc) and −1.83 V vs. Ag/AgCl (−2.28 V vs. Fc+/Fc) under de-aerated conditions. Furthermore, similar experiments in aprotic solvents such as acetonitrile were also made showing a reductive peak at −2.20 V vs. Fc+/Fc. The BZ-3 concentration dependence plot using the first reductive peak yielded a straight line (r2 = 0.991) with the following equation I (μA) = −0.88 [BZ-3] (mM) − 3.88 μA within a concentration range between 1 and 20 mM BZ-3. Taking into account these results, the electroanalytical determination of BZ-3 in sunscreen is being studied.
Fig. 8 DPV of 6.2 mM BZ-3 in [C6MIM][PF6] at basal SPGEs. Modulation amplitude 100 mV and step potential 50 mV. Modulation time 0.05 s. Interval time 0.5 s. |
The electrochemical reduction of different carbonyl containing organic compounds such BQ, AQ and TC was applied in RTIL at SPGEs. The carbonyl derivatives presented a two successive one-electron transfer with a similar behaviour as in common organic solvents. Electrochemistry of BZ-3 at SPGE in RTIL shed light on the applicability of screen printed platforms for the electroanalytical measurements in non-volatile ionic liquids.
Footnote |
† Electronic Supplementary Information (ESI) available. See DOI: 10.1039/c2ra20927f/ |
This journal is © The Royal Society of Chemistry 2012 |