A SERS spectroelectrochemical investigation of the interaction of butylethoxycarbonylthiourea with copper surfaces

A SERS spectroelectrochemical investigation of the interaction of butylethoxycarbonylthiourea with copper surfaces

Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137 A SERS spectroelectrochemical investigation of the interaction of butylethoxyca...

184KB Sizes 0 Downloads 0 Views

Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

A SERS spectroelectrochemical investigation of the interaction of butylethoxycarbonylthiourea with copper surfaces G.A. Hope, R. Woods∗ , S.E. Boyd, K. Watling School of Science, Griffith University, Nathan, Queensland 4111, Australia Received 8 February 2003; accepted 21 October 2003

Abstract The interaction of the sulfide mineral flotation collector, n-butylethoxycarbonylthiourea (BECTU), with copper surfaces at pH 9.2 has been investigated by voltammetry and Raman scattering spectroscopy, NMR, Raman and FTIR spectroscopies have been applied to characterise BECTU and its copper(I) compound to provide a basis for interpreting the spectroelectrochemical data. In the copper compound, the organic molecule is bonded to metal atoms through both its sulfur atom and the nitrogen of the NHCO group of BECTU with the release of the proton. Voltammetry showed that the copper compound is formed above approximately −0.4 V and its presence on the surface inhibits oxide formation. Surface enhanced Raman scattering (SERS) spectra from copper electrodes in the presence of BECTU showed that the major surface species was the copper compound at potentials more than or equal to –0.4 V. The SERS spectrum at approximately −0.5 V was significantly different to that observed from for the bulk compound. This spectrum is interpreted in terms of charge transfer chemisorption of BECTU onto copper atoms in the metal surface. © 2003 Elsevier B.V. All rights reserved. Keywords: NMR spectroscopy; Raman spectroscopy; n-Butylethoxycarbonylthiourea

1. Introduction Butylethoxycarbonylthiourea (BECTU):

has recently been introduced by Cytec Industries, Inc. as a flotation collector (Aero 5500); it has proved particularly effective for the flotation separation of copper sulfide minerals. Fairthorne et al. [1] investigated the reaction between BECTU and aqueous copper(II) chloride and characterised the precipitated product using a range of spectroscopic techniques. They concluded that “the product has an extended structure with an average stoichiometry derived from a mixture of cuprous, CuBECTU, and cupric, Cu(BECTU)2 , species”. However, those authors interpreted their XPS data ∗

Corresponding author. Tel.: +61-738757550; fax: +61-738756572. E-mail address: [email protected] (R. Woods).

0927-7757/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2003.10.011

in terms of “cuprous being the bulk copper species”. Elemental analysis gave a stoichiometry of 1.5BECTU ligands per copper atom, but their NMR spectra indicate that a significant quantity of unreacted BECTU was present. Indeed Fairthorne et al. interpreted their FTIR spectra in terms of the precipitated product being a mixture of CuBECTU and unreacted BECTU and deposited it onto a copper oxide substrate to remove the BECTU. Thus, the elemental analysis could be explained by one-third of the solid compound being unreacted ligand. Fairthorne et al. [1] stated that, in the preparation of copper BECTU, Cu+ is obtained from copper(II) by “an electrochemical mechanism” and then complexes with the organic species, with the overall reaction being: Cu2+ + BECTU → CuBECTU + H+ + e−

(1)

where BECTU is BECTU with a proton removed from one of the NH groups. Those authors did not explain how the electrochemical mechanism operated, but reduction of copper ions but would have to result from concomitant oxidation of BECTU to its thiouram disulfide. There is no evidence, however, for such a compound in those authors’ spectroscopic analyses.

130

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

BECTU was considered by Fairthorne et al. [1] to bond to copper through both the sulfur and the carbonyl oxygen atoms. This conclusion was based on the observation of a similar difference between the position of the FTIR band due to the stretching vibration of the C=O group for BECTU and Cu+ to that reported by Basilio [2] and by Mielczarski and Yoon [3,4] for the copper/isobutylethoxycarbonylthionocarbamate system, and the fact that these authors had interpreted their FTIR data in terms of bonding of both S and O in the organic group to a copper atom. Bonding of thionocarbamates through sulfur and oxygen was also reported by Leppinen et al. [5]. Single crystal X-ray diffraction investigation of a CuBECTU crystal has shown [6], however, that copper atoms are bonded to S and to the N of a NHCO group. The nitrogen of one CuBECTU group is bonded to a copper atom which is also bonded through sulfur to two other CuBECTU groups. The CuBECTU structure contains Cu6 (BECTU )6 units in a three-dimensional arrangement analogous to the hexameric “paddle-wheel” structure observed previously in a number of copper complexes containing N,S-donor ligands such as thioamidates, dithiocarbamates, 2-pyridinethiones and isothiocyanates [7–10]. Fairthorne et al. [11,12] investigated the interaction of BECTU with chalcopyrite and found [11] flotation recovery to be in good agreement with the amount of BECTU adsorbed on the mineral surface. In that study, adsorption was determined by measuring the abstraction of the collector from solution into which the mineral was added. In the other study by those authors [12], adsorption was determined using XPS and SIMS. SIMS analysis provided strong evidence for the presence of a 1:1 CuBECTU species on the mineral surface. Later SIMS investigations [13] showed that negligible BECTU adsorbed on pyrite and this explained the excellent flotation selectivity for chalcopyrite against pyrite exhibited by BECTU. The SIMS and XPS data [12] did not distinguish the initial BECTU monolayer on chalcopyrite from a multilayer. The S 2p spectrum was composed of three peaks, but these were assigned to unreacted chalcopyrite, and oxidation products of the mineral, viz., a metal-deficient sulfide and polysulfide. A SERS spectroelectrochemical investigations of the interaction of flotation collectors with copper, silver and gold surfaces have been carried out in the School of Science at Griffith University. The objective of these studies has been to characterise the surface species formed and, in particular, to distinguish the initial monolayer from multilayer species. Collectors investigated have been ethyl xanthate [14,15], isopropyl, isobutyl and isoamyl xanthates [16], O-isopropyl-N-ethylthionocarbamate [17], 2-mercaptobenzothiazole [18], and diisobutyldithiophosphinate [19]. In the present paper the SERS spectroelectrochemical studies have been extended to the interaction of BECTU with copper. As in previous studies, SERS spectra are compared with Raman spectra from bulk species that have been characterised by NMR spectroscopy.

2. Experimental methods 2.1. Materials A sample of BECTU was obtained from Cytec Industries, Inc., and was used without further purification. The copper(I) compound, CuBECTU , was prepared by mixing BECTU and CuSO4 , both dissolved in 70:30 aqueous ethanol, and adding ascorbic acid dissolved in the same medium. The role of the ascorbic acid is to reduce copper(II) to copper(I). A similar approach was used to prepare the copper(I) compound of O-isopropyl-N-ethylthionocarbamate (IPETC) [4,17]. The precipitated CuBECTU was filtered, washed with a copious quantity of water to remove any unreacted BECTU, and recrystallized from chloroform–ethanol mixtures. Unsuccessful attempts were made to make CuBECTU using a variation of the method reported by Fairthorne et al. [1]. Those authors reacted aqueous solutions of BECTU and copper(II) chloride. The present authors replaced copper(II) chloride with copper(II) sulfate to avoid the possibility of forming a BECTU copper(I) chloride complex rather than the copper(I) compound. Such complexes were found to be formed with IPETC when reaction was carried out in chloride media [4,17]. 2.2. Voltammetry Voltammograms were recorded for a copper electrode in a 0.05 mol dm−3 sodium tetraborate solution of pH 9.2 using an ADInstruments MacLab/4e interfaced with a PC. A potential program was applied that consisted of switching the potential from the open circuit value to the lower potential limit, holding at this value for 2 s, and then applying a triangular potential scan. Voltammetry was carried out with the electrode in the buffer solution alone and in one containing 10−4 mol dm−3 BECTU. Potentials were referred to a Ag–AgCl–KCl (saturated) reference electrode and converted to the standard hydrogen electrode (SHE) scale taking the potential of the reference to be 0.20 V on this scale. All potentials in this paper are presented on the SHE scale. 2.3. NMR spectroscopy 1H

and 13 C NMR spectra were recorded on a Varian INOVA400 spectrometer at 298 K from BECTU and CuBECTU dissolved in CDCl3 . The spectra were referenced to solvent residuals (1 H: CHCCl3 , 7.23 ppm; 13 C: CDCl3 , 77 ppm). gCOSY, gHSQC and gHMBC spectra employed the standard sequences provided in the VNMR6.1C software package. gCOSY (1 H/1 H) spectra were acquired with 400 increments of 2048 data points. Sinebell weighting functions were applied in both dimensions prior to Fourier transformation. gHSQC (1 H/13 C) and gHMBC (1 H/13 C) spectra were acquired with 200 and 400 increments of 2048 data points, respectively. Gaussian weighting functions were applied in both dimensions prior to Fourier transformation.

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

2.4. Raman spectroscopy Raman spectra from BECTU, CuBECTU and electrode surfaces in solutions containing BECTU, were obtained using a System 100 Renishaw Raman Spectrograph (Multi Channel Compact Raman Analyser) that has a rotary encoded grating stage, and an internal two stage Peltier cooled (−70 ◦ C) CCD detector. The spectral resolution was 5 cm−1 and the wavenumber resolution was 1 cm−1 . The incident radiation was conveyed from a Renishaw HeNe laser of 633 nm excitation, through a fibre optic Raman probe. Raman spectra from BECTU, CuBECTU and material extracted from electrodes with chloroform, were also obtained with a Renishaw RM2000 Raman spectrometer equipped with a computer controlled stage and a Leica metallurgical microscope incorporating a range of objectives. The FWHM of silicon calibration band at 520 cm−1 was 5 cm−1 and the wavenumber resolution was 1 cm−1 . The He–Cd laser 442 nm excitation line provided the incident radiation. CuBECTU was found to strongly fluoresce red under the blue laser, but an acceptable background was obtained when small, isolated particles were examined under the microscope. SERS spectra were obtained from copper electrodes under potential control. The electrochemical cell was as described previously [20]. It was constructed of borosilicate glass with a flat window at one end. Experiments were carried out with copper electrodes mounted on an assembly constructed from PTFE and positioned close to the window. This design follows that devised by Fleischmann et al. [21]. Copper electrodes of 6 mm diameter were prepared from metal of 99.99% purity and the surface was electrochemically roughened prior to obtaining SERS spectra by oxidationreduction cycling in 2 mol dm−3 H2 SO4 [22]. This procedure involved the application of 4–5 cycles between −0.3 and 1.2 V with a polarisation period of approximately 30 s before reversing the polarity. Experiments were also carried out in a cell of similar design, but the working electrode was a copper flag attached to a copper wire. Spectra were recorded for a copper electrode in 10−4 or 10−5 mol dm−3 BECTU in a 0.05 mol dm−3 sodium tetraborate solution of pH 9.2 that had been deoxygenated with ‘high purity’ nitrogen. For the lower BECTU con-

131

centration, the working electrode was held at a potential just above the hydrogen evolution potential in the buffer solution alone to remove any surface oxides and then a 10−4 mol dm−3 solution of BECTU was added to make the solution 10−5 mol dm−3 in BECTU. In the other experiments, a freshly activated copper flag electrode was introduced into the cell containing 10−4 mol dm−3 BECTU and a potential of −0.5 V applied simultaneously with the electrode entering the solution. SERS spectra were recorded in situ under potential control. 2.5. FTIR spectroscopy FTIR spectra from BECTU and CuBECTU were obtained using a Thermo Nicolet NEXUS spectrometer running OMNIC 6A. Chloroform solutions of the compounds were evaporated onto KBr discs and spectra accumulated for 50 scans between 4000 and 400 cm−1 at a 2 cm−1 resolution.

3. Results and discussion 3.1. Characterisation of BECTU and CuBECTU 3.1.1. NMR spectroscopy NMR spectroscopy confirmed the purity of the sample of BECTU and the prepared CuBECTU , and provided information characterising these compounds. The 1 H and 13 C NMR spectra are given in Table 1 together with assignments for each of the C and H atoms in the molecules. Our assignments of the N–H protons differ from those made by Fairthorne et al. [1]. Furthermore, we are able to assign the chemical shifts corresponding to the protons on each of the carbon atoms in the two alkyl chains of BECTU. When BECTU bonds to copper, it is the proton on the NHCO entity that is released to solution as a hydrogen ion. There is a significant change in the chemical shifts for both the carbon atom of the NHCS and that in the NHCO group. This is consistent with the crystal structure of CuBECTU [6] that shows copper bonds to both the S and N. 1 H NMR spectra obtained for BECTU and the CuBECTU are presented in Fig. 1. It can be seen that the spectrum from the copper compound (Fig. 1(b)) consists of resonances that

Table 1 1 H and 13 C chemical shift data for BECTU and CuBECTU Assignment

BECTU δ(1 H) (ppm)

CuBECTU δ(1 H) (ppm)

Assignment

BECTU δ(13 C) ppm

CuBECTU δ(13 C) (ppm)

NHCS NHCO H1 H1 H2 H3 H2 H4

9.63 8.04 4.18 3.61 1.61 1.37 1.27 0.92

10.38 – 4.04 3.62, 3.56 1.62–1.56 1.35 1.22 0.89

CS CO C1 C1 C2 C3 C2 C4

178.93 152.78 62.56 45.34 30.24 20.02 14.10 13.62

180.98 160.73 61.00 46.21 31.21 20.06 14.52 13.69

132

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

Fig. 1. 1 H NMR spectra (400 MHz, CDCl3 , 298 K) of (a) BECTU and (b) CuBECTU . Each resonance is identified with the protons associated with the carbon and nitrogen atoms marked in the insert. The resonance at 7.23 ppm arises from the solvent.

are sharp and well resolved. This supports the copper being in the diamagnetic univalent oxidation state and the absence of paramagnetic copper(II) impurities. The gCOSY spectrum obtained for BECTU is presented in Fig. 2. Crosspeaks are revealed correlating the doublet of triplet resonance attributable to the methylene protons adjacent to the N atom attached to the butyl substituent (H1) at 3.61 ppm, and the NH proton on the NHCS entity resonating at 9.63 ppm. These correlations firmly establish the relative proximity of the nuclei with respect to the bonding framework. Consequently, the change in chemical shift of the NHCS proton upon copper compound formation is somewhat less dramatic than previously reported [1], δ being 0.75 ppm rather than 2.2 ppm [1]. Nevertheless, the further deshielding of the NHCS proton in the copper compound is still significant, as it is consistent with a structure in which the NHCS hydrogen is internally H-bonded and the NHCS nitrogen retains significant sp2 character. Fig. 3 presents relevant expansions of the room temperature 1 H NMR spectrum obtained for BECTU and the series of 1 H NMR spectra obtained for CuBECTU over an extended range of temperatures. Only a single species is apparent in the spectra of the copper compound over the temperature range examined.1 Examination of the resonances 1 Chivers et al. [7] and others have noted that they have observed equilibrating mixtures of oligomeric species in analogous N,S-donor ligand

attributable to the methylene protons adjacent to the nitrogen on the butyl substituent in the protonated free ligand, reveals the expected doublet of triplet multiplicity. At room temperature, however, there is an apparent decrease in symmetry of the organic backbone of the ligand when it is bonded to copper. The first-order triplet of doublets observed for BECTU becomes a second order AB system in the CuBECTU compound. Similarly, in d8 -toluene solution, a second order AB system is observed for the methylene protons of the ethyl substituent, δ ∼ 4 ppm, instead of the expected first order quartet. On warming, however, the resonances of both sets of methylene protons are seen to broaden and eventually coalesce. At higher temperatures, exchange between the chemically inequivalent sites become fast on the 400 MHz NMR timescale and the expected first order doublet of triplet and quartet resonances are observed. The observed dynamic behaviour is consistent with a structure in which the organic backbone of the ligand is confined sterically and flips between allowed conformations. At room temperature the rate of conversion between the conformations is slow on the 400 MHz NMR timescale and thus discrete environments are observed for the backbone methylene protons. As the temperature of the system increases, the rate of exchange increases and thus an averaged, symmetric environment is observed, as expected. A full analysis of the dynamic behaviour observed in the solution NMR spectra of the CuBECTU complex will be presented in a subsequent manuscript [6]. 3.1.2. Raman spectroscopy Raman spectra for BECTU and CuBECTU , recorded using 633 nm radiation, are presented in Fig. 4 and the characteristic bands listed in Table 2. The assignment of Raman bands was made on the basis of previous reports of vibrational spectra of isobutyl compounds and of thiourea compounds. Many of these are combinations bands, but only the primary vibration is presented here to assist the reader. The Raman spectrum from BECTU shows two bands near 3200 cm−1 arising from vibrations of the N–H stretch vibrations of the two N–H groups in the molecule. The band at the higher wavenumber must arise from the proton on the NHCO group since it is absent from the spectrum from CuBECTU and NMR spectroscopy showed that it is this proton that is ionised when CuBECTU is formed (reaction (1)). The bands corresponding to the stretching vibrations of the carbonyl group in BECTU is blue-shifted by 59 cm−1 when the copper compound is formed. Such a shift in the FTIR band arising from the carbonyl stretching vibration has been interpreted [1] as indicating that Cu is bonded to the carbonyl O as well as the S in BECTU. As pointed out above, the crystal structure [6] shows, however, this conclusion to complexes in solution. In the case of the CuBECTU complex at hand, however, we observe no evidence of the appearance of other species either on heating of the sample or on standing for prolonged periods in CDCl3 or d8 -toluene solution.

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

133

Fig. 2. 400 MHz 1 H gCOSY spectrum of BECTU (CDCl3 ; 298 K); the crosspeaks correlating the NHCS proton with the protons of the adjacent methylene group are highlighted.

be invalid. It can be seen from Table 2 that there is also a significant red shift in the C–N stretching vibration bands (13–44 cm−1 ) on going from BECTU and CuBECTU and this is consistent with copper bonding to N as shown by the crystal structure. The shift in the carbonyl stretching band can be explained by the formation of bonding conjugation within the BECTU group on copper compound formation resulting from the loss of the proton from the NHCO group and the co-ordination of N and S to copper(I) species. Carbonyl bands are known to be sensitive to changes in electron density in the molecule. This structural change is supported by the gCOSY and H, H1 temperature dependence NMR results. 3.1.3. FTIR spectroscopy The N–H bands are weak in Raman, but strong in infrared spectra. To corroborate the interpretation of the Raman spectra from BECTU and CuBECTU , FTIR spectra were also recorded for these two compounds and these are presented

in Fig. 5. The spectra substantiate the conclusions reached from the NMR and Raman spectroscopic investigations. It can be seen from Fig. 5 that the N–H stretch region above 3000 cm−1 has a much higher intensity than in the corresponding Raman spectra shown in Fig. 4. BECTU gives rise to a broad band with peaks at 3296 and 3181 cm−1 , and these can be assigned to the two N–H groups in the molecule. The spectrum from CuBECTU shows only a peak at 3176 cm−1 ; the peak at the higher wavenumber observed with BECTU is absent. This correlates with the Raman spectra in which two N–H stretch bands were observed for BECTU and only the one at the lower frequency apparent with CuBECTU . Fairthorne et al. [1] also observed a narrower peak in the FTIR spectrum of CuBECTU compared with that for BECTU, but their spectra show the intensity at the lower wavenumbers, rather than the higher, being absent. The peaks at 3100 and 2800 cm−1 in Fig. 5 arise from C–H vibrations and their positions are similar to those observed with Raman spectroscopy (Fig. 4 and Table 2). The

134

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

Fig. 3. Expansions of the H1 and H1 region of the 1 H NMR spectra of (a) CuBECTU (400 MHz, d8 -toluene, temperature as indicated) and (b) BECTU (400 MHz, CDCl3 , 298 K). Splitting diagrams are provided as qualitative illustrations only.

Fig. 4. Raman spectra of BECTU and CuBECTU (442 nm laser excitation).

Fig. 5. FTIR spectra of BECTU and CuBECTU .

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

135

Table 2 Characteristic bands in Raman spectra from BECTU and CuBECTU and in SERS spectrum of initial layer BECTU

CuBECTU

3237 3169 2973 2933 2906 2877 1712 1567 1447 1329 1305 1277 1248 1228 1203 1156 1119 1089 1006 942 912 895 812 737 420 314

3174 2979 2935 2907 2875 1653 1559 1445 1373 1347 1297 1264 1248 1264 1154 1114 1079 1043 992 908 864 814 774 403 330

vw vw vs vs vs vs s m m

m m m m vw m w m w m w w m m m

vw vs vs vs vs vs m m

s vs s m w w w m m w w w w w m

SERS

Assignment

2965 s 2932 s 2903 s 2873 s 1620 s 1589 m 1441 m 1368 1348 1292 s 12648 m m 1173 1114 1067 1044 990 914 878 808 726 421 330

m w w w m m w m w w sh

NH stretch NH stretch CH3 antisymmetric stretch CH2 antisymmetric stretch CH3 symmetric stretch CH2 symmetric stretch C=O stretch CNH:CNC stretch CH3 antisymmetric deformation C–N stretch C–N stretch C–N stretch C–N stretch C–N stretch COC stretch CCC skeletal stretch CH3 rock; CH2 rock CCC antisymmetric stretch C–N stretch, C–C stretch

CNC stretch CH3 rock; CH2 rock C–S stretch NCS deformation; CS stretch

very strong, sharp peak at 1718 cm−1 due to the C=O stretch vibration in the spectrum from BECTU in Fig. 5 is shifted to 1648 cm−1 in that from CuBECTU . This is in agreement with the FTIR spectra reported by Fairthorne et al. [1] and with the Raman data reported above, where a corresponding shift from 1712 to 1653 cm−1 was observed for the C=O stretch band. As pointed out above, this shift can be explained by the backbone conjugation which occurs in the BECTU group on copper compound formation. The peaks in the region between 1150 and 1400 cm−1 can be assigned to C–N vibrations and, like the Raman spectra, the peaks from CuBECTU appear at significantly greater wavenumbers than those from BECTU. Again, this is consistent with copper bonded to both N and S. 3.2. Interaction of BECTU with copper electrodes 3.2.1. Voltammetry Linear potential sweep voltammograms for a copper electrode in 0.05 mol dm−3 sodium tetraborate solution (pH 9.2) containing 0 and 10−4 mol dm−3 BECTU are presented in Fig. 6. The lower potential limit of the scan corresponds to the onset of hydrogen evolution. In the absence of the collector, the anodic current above −0.1 V is due to oxidation of the metal to Cu2 O [23] and the cathodic current peak to reduction of the oxide back to the metal. In the presence of BECTU, it can be seen that oxidation of the metal is inhibited and an anodic wave observed com-

Fig. 6. Linear potential sweep voltammograms at 5 mV s−1 for a copper electrode in 0.05 mol dm−3 sodium tetraborate solution (pH 9.2) containing 0 (dashed line) and 10−4 mol dm−3 BECTU (solid line): scans commenced at lower potential limit.

mencing at approximately −0.4 V due to an oxidation reaction involving the collector. The anodic current continues at the beginning of the following negative-going scan and there is no evidence for the reverse process occurring within the potential limits used. This voltammetric behaviour is consistent with the irreversible formation of CuBECTU : Cu + BECTU → CuBECTU + H+ + e−

(2)

Presumably, cathodic reduction of the products of oxidation requires the potential to be in the hydrogen evolution region and hence does not occur within the potential range explored in Fig. 6. The formal potential of reaction (2) could not be obtained from rest potential measurements, even when a high impedance voltmeter was employed, since the potential drifted with time. This behaviour is similar to that observed with MBT [18] and DIBDTPI [19]. Another problem with BECTU is its low solubility in aqueous media and this means that determining a Nernst dependency is not possible. No prewave is apparent in Fig. 6; this voltammetric behaviour contrasts with that of xanthate which gives rise to a chemisorption prewave, but is similar to that found for copper in the presence of MBT [18] or DIBDTPI [19]. The absence of a discernible prewave cannot be taken as proof that charge transfer chemisorption does not occur at underpotentials to the formation of the CuBECTU . As pointed

136

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

Fig. 7. SERS spectra from a copper electrode in 10−4 mol dm−3 BECTU solution of pH 9.2 held at different potentials recorded in situ; a Raman spectrum of the residue extracted from the electrode after polarisation for 21 h at −0.098 V; and a Raman spectrum from CuBECTU ; 633 nm radiation.

out previously [24], chemisorption may occur at potentials below that of the negative limit of the scan and hence be present before the potential sweep is applied. Indeed, Fig. 6 shows that hydrogen evolution is inhibited in the presence of BECTU even before the initial positive-going scan was initiated and this indicates that a BECTU species was present at the interface. Spectroelectrochemical investigations are required to determine if the BECTU is chemisorbed or specifically adsorbed in this potential region. 3.2.2. SERS spectra SERS spectra were observed for a copper electrode in 10−4 or 10−5 mol dm−3 BECTU at pH 9.2 at all potentials investigated. It was found that the intensity of the SERS bands were relatively weak, and surface enhancement needed to be maximised in order to properly identify SERS bands. The intensity was greatest in the middle wavenumber range (1000–2000 cm−1 ) and all experiments showed the characteristic bands in this region. However, a spectrum of sufficient intensity to identity all bands throughout the extended range of 200 and 4000 cm−1 was only recorded in one experiment. The characteristic bands observed in this spectrum are listed in Table 2 where they are compared to the Raman spectra from BECTU and CuBECTU . Even for this spectrum, no bands could not be discerned in the NH stretch region, no doubt due to these bands being very weak in Raman spectra. SERS spectra recorded after holding a copper electrode in a 10−4 mol dm−3 BECTU solution of pH 9.2 for different times at various potentials are shown in Fig. 7. At −0.5 V, the SERS spectrum is very different from that from BECTU. It is similar to the Raman spectrum from CuBECTU , but displays a number of significant differences. As the potential was increased, the SERS spectrum became identical to that from CuBECTU . Confirmation that the overall process at

−0.094 V is the formation of CuBECTU was obtained by extracting into chloroform the product formed after 21 h at this potential, evaporating the solution on a glass slide, and recording a Raman spectrum. It can be seen from Fig. 7 that the spectrum is the same as that from CuBECTU . The SERS data showed that the formation of bulk CuBECTU on a copper surface is quite facile and occurs at potentials more than or equal to −0.4 V. This conclusion is in agreement with that derived from interpretation of the voltammogram (Fig. 6). As pointed out above, the spectrum in Fig. 7 recorded after 5 min at −0.495 V differs significantly from that from CuBECTU . The positions of the bands for this spectrum in Fig. 7 are the same as those for the SERS spectrum in Table 2. It can be seen that the CO stretch vibration appears at ∼92 cm−1 less than that from BECTU. Indeed, this band is shifted 33 cm−1 from its position from CuBECTU . The CN stretch vibrations occur at similar wavenumbers at those from CuBECTU , whereas the CS stretch lies between the values for BECTU and CuBECTU . The surface species is clearly not a specifically adsorbed, unaltered BECTU molecule. The SERS characteristics are consistent with a charge-transfer chemisorbed BECTU in which both the sulfur and nitrogen atoms of the organic compound are bonded to copper atoms in the surface. The difference between the spectrum of this species and that from CuBECTU can be explained by a difference in orientation between a chemisorbed species bonded to a copper atom remaining in its position in the surface, and a compound in which copper atoms can move into sites of minimum energy.

4. Conclusions Differences between the NMR and Raman spectra of BECTU and CuBECTU are in accordance with the crystal structure of the copper compound that shows copper to be bonded to both sulfur and nitrogen atoms in the organic group. Voltammetric investigations of BECTU at copper electrodes show that CuBECTU is formed by a charge transfer process. SERS and normal Raman spectroscopy shows that oxidative charge transfer adsorption of BECTU occurs on copper surfaces with a multilayer of CuBECTU developing at potentials more than or equal to −0.4 V. At lower potentials, SERS spectra indicate the formation of a chemisorbed layer.

Acknowledgements This project was supported by the Australian Research Grants Scheme. The BECTU was generously provided by Cytec Industries Inc.

G.A. Hope et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 232 (2004) 129–137

References [1] G. Fairthorne, D. Fornasiero, J. Ralston, Anal. Chim. Acta 346 (1997) 237–248. [2] C.I. Basilio, Ph.D. Thesis, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA, 1989. [3] J.A. Mielczarski, R.-H. Yoon, J. Colloid Interface Sci. 131 (1989) 423. [4] J.A. Mielczarski, R.-H. Yoon, Langmuir 7 (1991) 101–108. [5] J.O. Leppinen, C.I. Basilio, R.H. Yoon, Colloids Surf. 32 (1988) 113–125. [6] R. Woods, G.A. Hope, S.E. Boyd, P.C. Healy, in press. [7] T. Chivers, A. Downard, M. Parvez, G. Schatte, Organometallics 20 (2001) 727–733. [8] K. Schulbert, R. Mattes, Z. Anorg. Allg. Chem. 621 (1995) 72–76. [9] R. Castro, M.L. Duran, J.A. Gargia-Vazquez, J. Romero, A. Sousa, E.E. Castellano, J. Zukerman-Schpector, J. Chem. Soc., Dalton Trans. (1992) 2559–2664. [10] N. Narasimhamurthy, A.G. Samuelson, H. Manohar, J. Chem. Soc., Chem. Commun. (1989) 1803–1804. [11] G. Fairthorne, D. Fornasiero, J. Ralston, Int. J. Miner. Process. 50 (1997) 227–242. [12] G. Fairthorne, J.S. Brinen, D. Fornasiero, D.R. Nagaraj, J. Ralston, Int. J. Miner. Process. 50 (1998) 147–163.

137

[13] D.R. Nagaraj, Int. J. Miner. Process. 50 (2001) 45–57. [14] R. Woods, G.A. Hope, G.M. Brown, Colloids Surf. A 137 (1998) 329–337. [15] R. Woods, G.A. Hope, G.M. Brown, Colloids Surf. A 137 (1998) 339–344. [16] G.A. Hope, K. Watling, R. Woods, Colloids Surf. A 178 (2001) 157–166. [17] R. Woods, G.A. Hope, Colloids Surf. A 146 (1999) 63–74. [18] R. Woods, G.A. Hope, K. Watling, J. Appl. Electrochem. 30 (2000) 1209–1222. [19] G.A. Hope, R. Woods, K. Watling, Colloids Surf. A 214 (2003) 87–97. [20] G.M. Brown, G.A. Hope, D.P. Schweinsberg, P.M. Fredericks, J. Electroanal. Chem. 380 (1995) 161–166. [21] M. Fleischmann, D. Sockalingum, M.M. Musiani, Spectrochim. Acta A46 (1990) 285–294. [22] G.A. Hope, D.P. Schweinsberg, P.M. Fredericks, Spectrochim. Acta A50 (1994) 2019–2026. [23] M. Pourbaix, Atlas D’Equilibres Electrochimiques, GauthierÄVillars, Paris, France, 1963, 644 pp. [24] R. Woods, in: J.O’M. Bockris, B.E. Conway, R.E. White (Eds.), Modern Aspects of Electrochemistry, No. 29, Plenum Press, New York, pp. 401–453.