ANALYST. JANUARY 1993. VOL. 118 53 Voltammetric Determination of Gold Using a Carbon Paste Electrode Modified With Thiobenzanilide Xiaohua Cai College of Science and Technolog y, Hainan University, Haikou 570028, People's Republic of China Kurt Kalcher,* Christian Neuhold and Wolfgang Diewald lnstitut fur Analytische Chemie, Karl-Franzens Universitat, Universitatsplatz I , A-80 10 Graz, Austria Robert J. Magee La Tro be University, Bundoora, Melbourne, Victoria, Australia Gold(ii1) can be preconcentrated from acidic solution on a carbon paste electrode chemically modified with thiobenzanilide under open-circuit conditions. Using cathodic differential pulse voltammetry, 0.05-6 pg ml-1 of gold can be determined after preconcentration for 1 or 2 min. Except for mercury(ii) and platinum group metals, common ions have little effect on the determination of gold(iii).The preparation and regeneration of the modified electrode as well as the various methodical parameters for the preconcentration and measurement of gold(1li) were investigated. Keywords: Modified carbon paste electrode; thiobenzanilide; voltammetric determination of gold Since the first application of the direct mixing technique to the modification of a carbon paste electrode (CPE) by Ravichan- dran and Baldwin, many publications on chemically modified carbon paste electrodes (CMCPEs) have appeared.2 The advantage of this technique is its simplicity. The disadvantage, however, lies in the difficulty in obtaining homogeneous carbon pastes, when using solid modifiers as additives. For this reason, the 'solvent volatilization' technique coupled with ultrasonic vibration was developed.334 This special form of direct mixing is suitable only for modifiers which can be adsorbed on the carbon powder.This paper presents a new strategy for the modification of CPEs with thiobenzanilide (TBA, CBHS-CS-NH-ChHS) for the preconcentration and subsequent determination of gold. Determination of gold with CPEs can be traced back to the first application of such an electrode to conventional stripping analyskS More recently, determination of gold with CPEs containing various modifiers such as anion exchangers,Gg chelating resins,q Rhodamine BlO and dithizone" has been reported. Thiobenzanilide has a structure similar to that of dithizone. 12 Many of the methods employing CMCPEs for the determi- nation of gold have some disadvantages, mainly poor detec- tion limits of about 1 pg ml-l of gold(iii), or inhomogeneities of the electrode material due to particulate modifiers, or lack of chemical or electrochemical regenerability of the electrode surface.Thus, it seems necessary to design improved modified electrodes which overcome these difficulties. In this work, we have tried to overcome the problem ot inhomogeneity when mixing particdate modifiers directly with the carbon paste. A common procedure for producing a highly homogeneous electrode surface is to dissolve the additive in the pasting liquid of the electrode material. However, this method is only applicable if the modifier possesses a highly lipophilic character, which is not the case for TBA.Thus, it seemed promising to dissolve it in a more polar liquid, i.e., bromoform, which can actually be used as a pasting liquid. 13 Since carbon pastes containing only bromo- form as liquid paste component exhibit poor performance owing to adverse physico-chemical characteristics (low viscos- ity, volatility of CHBr3), a new approach to circumvent this problem by dissolving the modifiers in bromoform and mixing * To whom correspondence should be addressed. this solution with a common pasting liquid, i.e., paraffin oil, was tried. An attempt was also made to regenerate the electrode by a non-mechanical method, which is not possible for most of the CPEs modified with sulfur-containing organic com- pounds. 1 1 I * 4-16 Experimental Apparatus For the voltammetric measurements, a Princeton Applied Research (PAR) 264A polarograph was used in combination with a laboratory-constructed electrode assembly of Plexi- glas.I7 The cell consisted of a titration vessel of glass (Cat.No. 6.1415.220 from Metrohm) with a platinum wire as the counter electrode and a saturated calomel electrode (SCE, Ingold 30W-NS) as reference. The latter was in contact with the solution over a salt bridge with a Vycor frit, filled with 1 mol 1-1 KCl solution. A Teflon tube allowed purging of the solution with nitrogen, which was carried out for at least 5 min each time a fresh solution was put into the vessel. During the measurement nitrogen was passed over the solution. Voltammetric curves were registered on a two-channel recorder PAR Model R E 0089 and evaluated manually by the tangent-fit method.Cyclic voltammograms (CVs) were re- corded using an appropriate interface for A/D-conversion of the data in combination with a personal computer.*s The receptacle for preconcentrating the analyte onto the electrode was a SO ml glass beaker equipped with a poly- (tetrafluoroethylene) (PTFE)-coated stirring bar (length 30 mm, 7 mm c1.d.); the beaker was placed under a suitable holder for the electrode. Working Electrode The body of the working electrode was a Teflon rod ( I 1 mm 0.d.) with a hole (7 mm diameter) bored at one end for the carbon paste filling. Contact was made with a platinum wire through the centre of the rod. The modified paste was prepared by dissolving 0.40 g of TBA in 0.80 g of bromoform.The resulting solution was mixed with spectral carbon powder (2.0 g RWB, Ringsdorff- Werke) and paraffin oil (0.50 g, Uvasol, Merck). The carbon paste was packed into the hole of the electrode and smoothed off with a PTFE spatula.54 ANALYST, JANUARY 1993, VOL. 118 The plain carbon paste electrode was prepared in a similar way, without adding the modifier solution. Reagents De-ionized water was distilled twice in a quartz still and then purified by a cartridge de-ionization system (Nanopure, Barnstead). Hydrochloric acid was of Suprapur grade (Merck). Thiobenzanilide was of analytical-reagent grade (Eastman Kodak). A gold stock solution (1000 pg ml-1) was prepared in 0.01 mol 1- 1 hydrochloric acid from potassium tetrachloro- aurate(m) (pro analysi, Merck).Solutions of lower concen- trations were prepared by dilution of this stock solution just before use. Stock solutions of salts used for investigations on their interference had a concentration of 1 x 103 or 1 x 104 pg ml-' with respect to the ion. Procedure Activation of the electrode The freshly prepared electrode was exposed to a stirred analyte solution (0.01 mol 1-1 HCl) containing 10 pg ml-1 of gold(ii1) for 30 s. After rinsing with water for a short time, the electrode was dipped into dilute hydrochloric acid (7.5 x 10-3 moll-1) and scanned between 0.5 and -0.5 V versus SCE for five complete cycles at a scan rate of 50 mV s-l. This procedure was repeated. Then, the electrode was polarized for 5 min at -0.40 V. Preconcentration After activating the electrode, it was dipped into the stirred (300 rev min-1) analyte solution containing gold(rr1) for the required time.The electrode was removed, rinsed with water for a short time, placed into the voltammetric cell and connected to the polarograph. Voltammetry The supporting electrolyte for the voltammetric measure- ments was 7.5 x 10-3 mol 1-1 HCI. Quantitative determina- tions were performed in the differential pulse voltammetry (DPV) mode. The potential range was set from 0.5 to -0.4 V versus SCE in the cathodic direction. An equilibration period of 1.5 s, with the initial potential applied, was required in order to settle the solution. The pulse height was 50 mV and the scan rate 10 mV s-1 with an increment of 0.2 mV per data point. The current range was set according to the concentration of gold(iIi). Cyclic voltammograms were recorded with a scan rate of 50 mV s-1.The potential range was from 0.5 to -0.5 V. Other parameters were the same as for DPV. Regeneration After recording the voltammogram, the electrode was elec- trolysed at -0.38 V for 2 min. The renewed electrode usually did not show any peak within the potential range but if it did, the regeneration step was repeated. Results and Discussion Composition and Electrochemical Behaviour of the Modified Electrode In preliminary experiments, TBA was mixed directly with the conventional carbon paste. The resulting electrode showed almost no ability to preconcentrate gold(ii1). The solvent volatilization technique was also used with alcohol or ether as a volatile solvent.It was shown that TBA does not mix well with carbon powder after solvent volatilization. The TBA is easily soluble in chloroform and bromoform, which have been used as pasting liquids.5 Bromoform was chosen because it has a lower volatility than chloroform. The bromoform containing dissolved TBA can be easily mixed, as a liquid modifier, with conventional carbon paste. A certain content of paraffin oil in the paste was found to be necessary, because it increases the viscosity of the carbon paste; thus the modified electrode can be used repeatedly for a longer time without any notable change in its electrochemical characteris- tics. The optimum mass ratio between bromoform and paraffin oil was found to be 1.6 : 1. The content of TBA in the carbon paste influences the height of the signal; an optimum concentration was found to be 10% m/m with respect to the analytical performance of the electrode.The CVs of the modified electrode in the absence of gold are shown in Fig. 1. A broad wave appears at about -0.6 V versus SCE when starting the scan at +1 V (curve A). This wave gradually decreases and finally disappears as the initial scan potential becomes less positive (curves B-D). It is known that TBA itself cannot be reduced directly in this potential range. But as the compound contains a C=S group in close vicinity to an NH group, it can be oxidized to its dimeric form via its tautomeric structure -C(SH)=N-. The resulting disul- fide can be easily reduced, as reported previously.19 Thus, the dependence of a voltammetric reduction on the initial potential clearly indicates that the wave must be assigned to the reduction of an oxidation product of TBA, which is produced at the more positive potentials.Therefore, in order to avoid oxidation of the modifier at the electrode surface, an initial scan potential less than +0.6 V should be used for practical applications of the modified electrode. Voltammetric Behaviour of Gold If a plain carbon paste clcctrode is used to preconcentrate gold under open-circuit conditions by interchanging media between accumulation and measurement, no signal response is obtained; this indicates that an unmodified carbon paste electrode does not adsorb gold(ii1). If gold(ii1) is present in the bulk solution of the measurement, the unmodified electrode gives responses due to electrochemical transformations of gold(ii1). Fig.2 shows the CVs. Reduction of gold(iri) to gold(o) occurs at 0.32 V versus SCE and re-oxidation occurs at 1.05 V. Evidently, for the reasons discussed above, only the current response of the reduction of gold(rr1) can be exploited analytically with modified electrodes. Fig. 3 shows the voltammetric behaviour of gold(iii) that had been accumulated externally (open circuit) onto the carbon paste electrode modified with TBA. During precon- centration, gold(iii) is accumulated by TBA onto the surface a E 2? 5 . c 0 1 0.5 0 -0.5 - 1 PotentialN versus reference Fig. 1 Cyclic voltammogram of a modified carbon paste electrode. Supporting electrolyte, HCI (0.0075 mol I-').Initial potentials: A , +1.0; B, +0.8; C, +0.6; and D, +0.5 V versus SCEANALYST, JANUARY 1993, VOL. 118 55 1.2 0.8 0.4 0 -0.4 PotentialN versus reference Fig. 2 Cyclic voltammogram of gold(1ii) on an unmodified carbon paste elcctrode. Supporting electrolyte, HCI (0.0075 mol 1- l ) . A, Blank; and B. 20 pg ml-' gold(iii) f 2 E 3 A I I I I 0.5 0.3 0.1 -0.1 -0.3 -0.5 PotentialN versus reference Fig. 3 Cyclic voltammogram of gold(m) accumulated with a modi- fied carbon paste electrode. Analyte solution: 0.01 moll-1 HC1 and 10 pg ml-1 gold(rr1); accumulation time, 1 min and supporting elec- trolyte, HCl (0.0075 mol 1-1). A, Blank; and B, 10 pg ml-1 gold(m) of the carbon paste electrode, forming a very stable complex Au3+(sol) + 3C6HS-C(SH)=N-C6H5(sur) -+ Au(C6HS-C(S-)=N-C6HS)3(sur) + 3H+(sol) (1) where (sol) refers to the analyte solution and (sur) to the electrode surface.The reactive species actually is the thiol- form of TBA. In the cathodic scan gold(I1i) is reduced to gold(o) at -0.25 V versus. SCE, which cannot be re-oxidized to gold(iii) within the applied potential range. Preconcentra- tion with the modified electrode produces a high concentra- tion of gold(ii1) at the surface of the electrode. As a result of this type of accumulation, the reduction peak with the modified electrode is much more sensitive for gold than an unmodified one in direct voltammetry. A large shift of the peak potential in the cathodic direction, compared with an unmodified electrode, indicates that the reduction of gold(rr1) in its complex with TBA is harder to effect than in its free form.Differential-pulse voltammograms of gold(1ii) accumulated at the modified carbon paste electrode can be seen in Fig. 4. As can be expected, the signals are much broader than for mercury electrodes, but they are well shaped to be exploited for analytical determinations. Therefore, DPV was used throughout this work. [eqn * (1) 1 Activation and Renewal of the Modified Electrode If the freshly prepared, unactivated electrode is used to accumulate gold, the resulting current response is less sensitive to the concentration of gold, and the peak height increases when the procedure is repeated with the same k 0.5 0.2 -0.1 -0.4 PotentialN versus reference Fig. 4 Differential pulse voltammograms of gold(iri) with a modified carbon paste electrode.Supporting electrolyte, HCI (0.0075 moll-*); analytc solution, 0.01 mol 1-1 HCl; accumulation time, 5 min. A, Blank; B, 1 pg ml-I gold(ii1); and C, 2 pg ml-l gold(iii) electrode filling. Therefore, it may be concluded that the complexing groups at the surface of the electrode have not been fully activated. In order to avoid a pre-treatment dependent gold signal and to improve the analytical perfor- mance of the electrode, activation of the electrode is necessary. The optimum method involves exposure of the original electrode to a solution containing gold, because a small amount of gold(ii1) seems to be irreversibly adsorbed on the electrode material. The resulting electrode has good stability, sensitivity and reproducibility. Similar processes have been reported for carbon paste electrodes with other modifiers .2",21 For repetitive use, the regeneration of the reactive func- tional groups on the electrode surface is very important. Although this can be achieved simply by replacing the used carbon paste, it requires a highly reproducible treatment of the electrode, otherwise the physical and chemical properties will not remain constant during repetitive measurements. Therefore, chemical or electrochemical regeneration is prefer- able. Potassium cyanide and thiourea were used to regenerate the electrode surface. Although they are effective in removing gold from the electrode surface by forming complexes with gold(iii), the resulting electrode has poor reproducibility during the ensuing measurements.As can be seen from Fig. 3, the reduction peak for gold(i1i) gradually decreases and finally disappears during repetitive scans. This provides the possi- bility of regenerating the electrode electrochemically. Further experiments showed that after each electrochemical measure- ment, a potential more negative than the peak potential must be applied to the electrode for some time so that gold(ii1) can be completely reduced. As a result, the electrode does not display any DPV peak and can be used again to preconcen- trate gold(1Ii) without further activation. To characterize the reproducibility of the modified elec- trode by regenerating it electrochemically, repetitive precon- centration-measurement-regeneration cycles were carried out. The result of eight ensuing measurements showed a relative standard deviation of 4.9% for 1 pg ml-1 of gold(iii) with a preconcentration time of 1 min.Thus, electrode renewal gives a good reproducible surface. Usually, after eight measurements, the current response for gold begins to decrease gradually, indicating that the preconcentration ability of the electrode decreases. This phenomenon may be due to residues of elemental gold adsorbed at the electrode surface, which decrease the effective area of the electrode and, thus, may block binding sites for tetrachloroaurate. Optimum Conditions for Analysis Acetate buffer, phosphate buffer and hydrochloric acid were investigated as media for the analyte solution. Dilute HCI was56 ANALYST, JANUAKY 1993, VOL. 118 1.80 I 0.90 ' I I I I I 0.005 0.010 0.015 0 020 0.025 Concentration of HCl/mol 1-1 Fig.5 Dependence of the peak current on the acidity of the analyte solution. Supporting electrolyte, HCI (0.0075 rnol 1-1); analyte solution, I pg ml-1 gold(iu); and accumulation time, I min 2.1 a g z 3 1.5 1.2 0 0.004 0.008 0.012 0.016 Concentration of HCVmol I 1 Fig. 6 Dependence of the peak current on the concentration of the supporting electrolyte (HCl). Analyte solution, 0.01 rnol I-' HCI, 1 pg ml -1 gold(iii); and accumulation time, 1 min found to be preferable, because it precipitates, and thus separates, any silver interferent present in the analyte solution. The depcndcnce of the peak height on the acidity of the analyte solution is displayed in Fig. 5 ; a steady signal can be obtained if the concentration of HCl in the analyte solution lies between 5 X 10-3 and 1 X 10-2 rnol I-'.For our investigations, a concentration of 1 X 10-2 mol 1 - 1 HCI in the analyte solution was used. Dilute HCI is also best suited as supporting electrolyte. As can be seen from Fig. ti, the current response in DPV is maximum with a concentration of HCI of 7.5 x 10-3 rnol 1-1. This concentration was used for the subsequent analyses. The dependence of the peak current on the preconcentra- tion time is displayed in Fig. 7. An exponential increase of the peak with increasing preconcentration time is observed for the modified electrode, resulting in a constant value for longer accumulation periods. The overall curve shape of the current- preconcentration time diagram is very typical for this type of modified electrode.1 1 The manifestation of a limiting value for the current at longer periods of time is due to reaching equilibrium conditions for the reaction between complexing reagent groups at the surface and analyte ions in the solution. With suitable preconcentration times, a linear ratio between peak height and concentration of gold(ii1) exists for 0.05-6 pg ml- I as shown in Fig. 8 when a proper accumulation time is chosen, i.e., 2 min for concentrations below 1 pg ml-1 of gold(Ir1) and 1 rnin above this concentration. The detection limit is 0.02 pg ml-1 when the preconcentration time is 5 min. For the analysis of 1 pg ml-1 of gold, the relative standard deviation is 1.9% for five determinations when using an internal standard additions methods.3.5 ' Accumulation time/min Fig. 7 Effect of the accumulation time on the pcak currcnt. Analyte solution, 0.01 mol I - ' HCI, 1 pg m1-I gold(ii1); and supporting electrolyte, HCI (0.075 rnol 1 1) 10 8 5. g 3 =; c 6 4 2 0 2 4 6 8 10 Concentration range Fig. 8 Dependence of the signal current on the concentration of gold(ii1). Supporting electrolyte, HCI (0.075 rnol 1 I ) ; analytc solution, 0.01 tnol I- 1 HCI; and concentration rangcs (0 and 10 of x-axis correspond to minimum and maximum of given range). Accumulation times: A, 0-1.0 pg ml-1 gold(iii), 2 min; and B, 0-10 pg ml-1 gold(iri), 1 min Table 1 Interferences with the determination of gold; analyte solution: 1 pg ml-1 gold, 0.01 mol 1-1 HCl; supporting electrolyte 0.0075 rnol 1-1 HC1; accumulation time: 1 rnin Peak change (%) Concentration of interferent Interferent HglI CU" Ag' ZnlI Cd" In111 Pb" A P As" Fe'II CO" NilI Pd" IF1 IP' Pt" BiIII Added as Hg C12-HC1 CuC12-H35 AgN03-H20 ZnC12-HC1 BiCl,-HCI In(N03)3-HCl AszO3-NaOH FeCI3-HCl NiC12-HCI Na2PdC14-HC1 IrC13-HC1 H2IrCl6-HC1 CdCI2-HC1 Pb(N03)2-H20 As~OS-H~O CoCIz-HCI (NH4)2PtC14 20 pg ml-1 -55 -4.9 -13.3 +11.1 +35.6 -4.6 -11.1 -5.4 -13.2 -11.7 -44.5 -5.0 - 10.0 -6.7 - 13.8 -5.0 -15.2 -9.4 -22.3 -7.5 -11.0 -6.3 -12.0 -7.6 - 17.4 120 pg ml-1 - 90 -26.6 - 100 -33.3 - 100 +4.6 +14.5 Interferences Various common ions were examined with respect to their interference in the determination of gold (Table 1).Most of the species analysed have only a slight effect on the determina- tion of gold(r1i) even up to a 120-fold excess with respect to gold.The influence of weakly interfering components can easily be eliminated by applying the standard additions method for the evaluation of the concentration of gold(ii1). Mercury(i1) and platinum group elements interfere signifi- cantly, as they can also form complexes with TBA. Other thiophilic elements such as silver, bismuth and arsenic(v) interfere severely at higher concentrations.ANALYST, JANUARY 1993, VOL. 118 57 Table 2 Recovery of gold in an artificial mining waste; standard deviations for five determinations Concentrationhg ml- 1 Foundhg ml- 1 50 49.7 k 1.0 100 99.1 * 1.9 500 496 * 7 Mining waste water An artificial sample with representative concentrations of constituents22: 20 mg 1 - 1 manganese(lI), 100 mg 1-1 iron(IIi), 100 pg 1 - 1 chromium(rii), 2.5 mg 1 - 1 cobalt(ii), 5 mg 1-1 nickel(ir), 1.5 mg 1-1 copper(Ii), 5 mg 1-1 zinc(Ii), 10 pg 1-1 cadmium(Ii), 100 pg I--’ lead(ii) was spiked with different amounts of gold.The pH was adjusted to 2.1 with HCI, and the recovery rate of gold was determined by the method presented above (Table 2). Pharmaceutical Gold was analysed in tablets containing Auranofin (Ridaura from Smith Kline Dauelsberg, Gottingen, Germany). One tablet was dissolved in a mixture of 3 ml of concentrated HN03 and 0.2 ml of concentrated HC104, and the liquid was evaporated to dryness. The residue was dissolved in 250 pl of HCI (10 moll-1) and made up to 20 ml. An aliquot of 2 ml was diluted to 20 ml and used for the preconcentration of gold.The gold content was found to be 0.86 k 0.01 mg of gold per tablet (six determinations; reference value 0.87 mg). The method presented here has a much lower detection limit than similar methods with modified carbon paste electrodes. Whereas phosphor-organic compounds or ion exchangers can be used as modifiers for CPEs to determine gold in concentrations higher than 1 pg ml-1 of gold (= 5 x 10-6 mol I-’), TBA is suitable for concentrations down to 50 ng ml-1 of gold (= 2.5 x 10-7 moll-’). Thus, the detection limit is higher than for atomic absorption spectroscopy (AAS) when using acetylene-dinitrogen oxide ( 5 X 10-8 mol 1 - I ) , but comparable when using acetylene-air to avoid some matrix effects. Therefore, the method presented here may be a reasonable alternative to AAS.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 The authors wish to acknowledge financial support of this work by the Austrian “Fonds zur Fiirderung der Wissenschaft- lichen Forschung” (Project No. P8539-CHE). X.C. acknow- ledges a scholarship from the North-South Dialogue Program of the Austrian government. References Ravichandran, K., and Baldwin, R. P., J . Electroanal. Chem., 1981, 126, 293. Kalcher, K., Electroanalysis, 1990, 2, 419. Ravichandran, K., and Baldwin, R. P., Anal. Chem., 1983,55, 1586. Prabhu, S. V., Baldwin, R. P., and Kryger, L., Electrounalysis, 1989, 1. 13. Jacobs, E. S . , Anal. Chem., 1963,35, 2L13. Kalcher. K., Anal. Chim. Acta, 1985, 177, 175. Kalcher. K., Greschonig. H., and Pietsch, R., Fresenius Z. Anal. Chem., 1987, 327, 513. Peng, T., Shi, Q., and Lu, R . , J. Appf. Sci., 1990, 8, 361. Gao, Z.. Li, P., Dong, S . , and Zhao, Z., Anal. Chim. Acta, 1990, 232, 367. Kolbl, G., Kalcher, K., and Voulgaropoulos, A., Fresenius 2. Anal. Chem., 1992, 342, 83. Kalcher, K., FreseniuJ Z. Anal. Chem.. 1986, 325, 181. Beilstein Handbuch der Organischen Chemie, Frankfurt, 2nd Supplement, 1950, vol. XII, p. 154. Adams, R. N., Anal. Chem., 1958, 30, 1576. Kalcher, K., Fresenius 2. Anal. Chem., 1986, 325, 186. Tanaka, S . , and Yoshida, H., Talanta, 1989, 36, 1044. Sagawara, K.. Tanaka, S . , and Taga, M. ~ J. Electroanal. Chem., 1991, 304. 249. Kalcher, K., Fresenius Z. Anal. Chem.. 1986, 323, 238. Kalcher, K., and Jorde, C., Comput. Chem., 1986, 10, 201. Cai, X., Pei, J., Zhou, X., and Zhao, Z., Anal. Sci., 1991, 7, 109. Amine. A., Kauffmann, J . M., and Patriarche, G . J., Talanta, 1991, 38, 107. Wang, J . , Taha, Z . , and Nasser, N., Talanta, 1991, 308, 81. Fiirstner, U. and Wittmann, G . T. W., Metal Pollution in the Aquatic Environment, Springer, Berlin, 2nd edn., 1981. Paper 2103.5986 Received July 8, 1992 Accepted October 12, 1992