J. Chil. Chem. Soc., 48, N 2 (2003)

PHOTOELECTROCHEMICAL CHARACTERIZATION OF NATURAL
COPPER SULPHIDE ELECTRODES IN ALKALINE MEDIA:
EFFECT OF XANTHATE ADSORPTION.

H. Gómez, A. Cortés, R. Henríquez, G. Riveros, R. Córdova and R. Schrebler.

Instituto de Química, Facultad de Ciencias Básicas y Matemáticas, Universidad Católica de Valparaíso,
Casilla 4059, Valparaíso, Chile.
( Received : December 12, 2002 ­ Accepted : Marzo 4, 2003 )

ABSTRACT

A photoelectrochemical approach is utilized for studying the processes that take place on copper sulfide mineral surfaces in alkaline solutions, either in absence or presence of potassium ethyl - xanthate. It is found that photoresponses are closely related to the mineral stoichiometric composition which also determines its semiconducting or metallic character. It appears that in the potential interval where xanthate anions can be adsorbed, surface composition and reactivity are closely linked.

Key Words: copper sulfides; non stoichiometry; photoelectrochemistry; semiconducting properties; flotation.

1. INTRODUCTION

It is well established that flotation of sulfide minerals -a procedure allowing concentration of minerals- is electrochemical in nature. The processes leading to a reagent adsorption at the mineral surface rendering it hydrophobic are very complex and depend on the nature of the sulphide. In spite the important number of contributions in this field, they remain not yet well understood. Xanthate (ROCS₂^-) homologues are the most commonly compounds used in flotation of copper sulfides and many studies have been carried out regarding the mechanism of adsorption on the mineral surface [1-24].

Voltammetric experiments performed with the cuprous sulphide - xanthate solution system in alkaline media show a multiplicity of oxidation peaks during positive scan, generally attributed to underpotential deposition of xanthate (chemisorption), followed of bulk copper xanthate formation and other cuprous sulfide oxidation products, respectively. However, important differences are found when the potential peak values reported for these processes are compared. In most of these studies, composition of the working electrode was specified as corresponding to chalcocite, Cu₂S, but natural specimens usually contain the non stoichiometric phase djurleite, Cu_1.95S. Accordingly, if composition is not well characterized, the peak assignment may remain ambiguous.

In a previous work [25], we studied the influence of surface composition on the voltammetric response of natural copper sulfide electrode whose stoichiometry was driven by a simple procedure describes therein. Differences, both in anodic as well as cathodic current peak contributions, either in presence or absence of xanthate anions were evidenced for the chalcocite and djurleite enriched phases. Furthermore, voltammograms recorded under chopped white light illumination showed photoresponses also dependent on surface stoichiometry. These results stimulates us to get more detailed information about the influence of light on the electrochemical behavior of copper sulfides. Photoelectrochemical studies of phenomena excited by illumination that are relevant to adsorptions of collectors reagents onto mineral surfaces have been mainly confined to galena [26,27], although there is a considerable literature related to the photoelectrochemical properties of metal sulfides and other chalcogenides in relation to solar energy conversion.

2. EXPERIMENTAL

The specimen utilized as working electrode were natural samples from El Teniente (Rancagua, Chile) mine. The XRD analysis reveals mainly the presence of djurleite enriched phase and no iron or others element peaks were detected. The samples were cut and then inserted in a cavity close to one end of a rectangular acrylic holder. The ohmic contact was made at the back-side of the holder attaching a copper wire with In-Ga eutectic and embedding both sides with epoxy resin. The exposed surface area was ca. 0.1 cm² and at least 1.0 mm layer was removed by grinding with SiC paper. Before each test run, the electrode was polished with 0.3 and 0.05 mm alumina powder, then rinsed and sonicated in deionized water. When was necessary, mineral electrode surface composition was modified towards chalcocite following the procedure describes in reference [25]. The electrolytic solution consists in sodium tetraborate of pH 9.2 . Stock solutions of sodium ethyl xanthate were prepared daily to ensure minimum decomposition. The reference electrode was a saturated calomel electrode and a large area platinum foil was used as counterelectrode.

Instrumentation for photoelectrochemical experiments consists of: a 100 W tungsten-halogen lamp, power supply, Standford Research chopper and chopper control unit, Jovin-Ivon f/3.4 monochromator and stepping motor control; a PAR 263A potentiostat and a PAR5210 phase sensitive lock-in amplifier, both from EG&G, all controlled via an IEE bus and interfaces by a Quick Basic program run on a PC microcomputer. The lock-in amplifier was adjusted so that in-phase currents corresponded to a phase angle of -90. Spectral response data were normalized respect to a UDT Sensors Inc., calibrated silicon detector.

3. RESULTS AND DISCUSSION

Behavior in sodium tetraborate pH 9 2 solution.

The influence of electrode composition on the electrochemical i-E curves recorded from the open circuit potential (ocp) towards negative potentials are shown in Fig. 1(a).

Fig. 1- i-E, (a) and iph-E, (b) curves recorded during triangular potential cycles scanned towards negative values from ocp on djurleite (-) and chalcocite (-----) electrodes in 0.05 M Na2B4O7. Scan rate : 0.03 Vs-1.

The first cathodic current contribution C1 is related to the electroreduction of compounds formed at the electrode surface in ocp conditions. According to the literature[28], one of the possible products is cupric oxide that can react with non stoichiometric copper sulphides phases trough the general reaction

The extent of this reaction is controlled by the amount of Cu_2-x-yS phases available in the electrode surface which explains the greater charge observed at C1 for djurleite as compared with chalcocite. This is supported by observing the differences found for both phases at more negative potentials: the electroreduction of djurleite to chalcocite [25-28] gives rise to cathodic peak C2 whereas only a current plateau, which extends over at least 0.40 V, is observed for the chalcocite phase. The current increasing at the end of the scan (C3) concerns the electroreduction to metallic copper of both, the chalcocite electrode as well as the chalcocite formed from djurleite. Considering the value of the negative switching potential probably the reduction be incomplete and some amounts of both phases remain at the electrode surface. The inverse of this reaction accounts for peak A3 during the positive scan. Further, the current increasing (A2) observed at the end of the scan for both phases corresponds to the formation of non stoichiometric copper sulphides trough the inverse of reaction (1).

The photoelectrochemical responses recorded simultaneously to each of the previous i-E curves are shown in Fig. 1(b). Whereas for djurleite no photocurrent is present during the potencial cycle, chalcocite electrode displays cathodic an anodic photocurrent peaks. The cathodic contributions are an small peak followed of a wide photocurrent front with a maximum at ca. -0.67 V. Afterwards the photocurrent diminishes , falls to zero at about -0.40 V, then undergoes a sign change and afterwards increases to reach a maximum which extends to the end of the scan. The first approach to explain the different behavior of both phases towards illumination is to consider the influence of composition regarding their conducting properties. In fact, as most of the mineral sulphides, chalcocite and djurleite exhibit semiconductor character. According to the literature[29], they display a p-type semiconductivity due to the presence of copper vacancies that may act as acceptor centers . The following values for the band gap energy have been reported[30]:

The absence of photocurrent in djurleite is probably due to the presence of an excess of acceptors centers within the gap, arising in a quasi-metallic behavior for this phase. Thus, any photogenerated electron-hole pair immediately recombines. Instead, photocathodic currents observed at chalcocite electrode means that the electric field in the space charge region rapidly separates photostimulated electron-hole pairs, with the holes migrating into the bulk of the semiconductor and the electrons to the surface for effecting reduction and then giving rise to a negative photocurrent. On the other hand, the photoanodic current might arise from the formation of an inversion layer that converts from p to n ­ type electrode surface.

In order to get deeper insights about the origin of photocurrents on chalcocite enriched electrode, we performed a set of experiments in a wider potential interval. The starting potential was chosen negative enough so as to ensure the reduction of any amount of remaining non stoichiometric phases. Further, by reducing the djurleite electrode to chalcocite in the same manner, a pure chalcocite electrode will be also obtained [28]. Figure 2 presents a set of voltammograms with the corresponding i_ph-E curves, scanned in the positive direction and at increasing anodic switching potential.

Fig. 2- Effect of the positive switching potential (1, 2, 3) on i-E, (a) and iph-E, (b) curves on a chalcocite electrode in 0.05 M Na2B4O7. Scan rate: 0.03 Vs-1.

Following the analysis of Fig. 2, it is evident that as the scans are extended towards more positive potentials, the photocathodic currents increase with respect to that observed previously in Fig. 1 (about ten fold for the main peak when comparing the current scale in both figures). It appears that the formation of a layer coming from chalcocite electro - oxidation would account of this behavior. As predicted by potential-pH diagrams [28] such layer would be composed of non stoichiometric copper sulphide and cupric oxide formed through the global reaction:

As the film is being formed patches of both phases cover partially the chalcocite surface explaining thus the small cathodic photocurrent peaks seen during the positive scan. Up to -0.10 V the electrode surface is completely covered by a duplex layer with a thin inner of non stoichiometric phase and a top film of cupric oxide. The overall charge under the anodic electrochemical peak is 2.31 mC cm^-2 corresponding to a thickness of about 3.4 nm on the basis of roughness factor of 2 and the molar volume of 12.428 cm³ mol^-1 for CuO. 1 mC cm^-2 = 1.289 nm of Cu0 and a monolayer corresponds to 84 mC cm^-2.

According to this model, the first cathodic photocurrent peak seen in the reverse scans is associated with the reduction of the non stoichiometric phases in chalcocite, followed by the reduction of cupric oxide to cuprous oxide according to reaction:

Once the formation of the Cu₂O layer starts, apparently it is the semiconductor where the photoelectrons are generated with CuO as the substrate for the photoreduction :

The cathodic photocurrents observed in the range comprised between -1.20 and -0.90 V present hysteresis during the reverse scan associated to changes in electrode composition as the potential made more positive. Up to -0.25 V there is no evidence of chalcocite oxidation in the voltammogram and also no photocurrent peaks appear meaning that between -1.20 and -0.25 V a single p- type photoactive surface prevails on chalcocite, that phase being chalcocite itself, with an electronic structure essentially unchanged over this potential interval. Figure 3 shows the photocurrent dependence on the photon energy of the incident light after holding the potential at -0.60 V. Replotting spectra in the coordinates (Q.E.* hn)^1/2, where Q.E. and hn are the photocurrent quantum efficiency (in arbitrary units) and photon energy, respectively, has shown that absorption of a photon in the energy band 1 ­ 1.5 eV causes the indirect transition of electrons. Determined in this way, the Eg value falls within the interval Eg = 1.2 ± 0.1 eV, in good agreement with the Eg values above mentioned. Similar spectral responses and Eg values are obtained at -0.80 and -1.10 V. The photocathodic current observed in this potential range is associated to the electrochemical reduction of chalcocite to copper. As a p ­ type semiconductor, chalcocite display photoactivity under depletion conditions which prevail at potential negative of E_fb whereupon the field causes migration of the minority charge-electrons-to the semiconductor/electrolyte interface.

Fig. 3- Spectral response of the chalcocite electrode recorded at -0.60 V in 0.05 M Na2B4O7

This assumption is supported by the measurement of the spectral response recorded at -0.20 V which is represented in Fig. 4.

Fig. 4- Spectral response of the chalcocite electrode recorded at -0.20 V in 0.05 M Na2B4O7.

As potential is increased cathodically, it is possible that Cu₂O reacts with some of remnant non stoichiometric sulphide according to:

Following the negative scan, electroreduction of chalcocite to copper takes place at the end of the scan and the electrochemical and photoelectrochemical events are similar to that describes when the positive scan limit was less than -0.20 V.

Behavior in sodium ethyl xanthate solution.

As was previously reported [25] the presence of ethyl xanthate in the electrolytic solution lead to significant changes in i-E curves that are dependent on copper sulfide composition. The presence of djurleite arises in a multiplicity of anodic peaks associated to xanthate chemisorption, followed by the formation of a cuprous xanthate layer. On the contrary, as it is shown in Fig. 5, no chemisorption peaks are evidenced when chalcocite phase prevails.

Fig. 5- i-E curves of djurleite (a) and chalcocite (b) electrodes recorded in the potential interval of xanthate adsorption. Scan rate : 0.03 Vs-1. 0.05 M Na2B4O7 + 1x10-5 M of sodium ethyl xanthate.

Figure 6 accounts for the influence of ethyl xanthate ions on the i-E and i_ph-E behavior of chalcocite where several distinctive features should be underlined. First, there is no evidence of electrochemical anodic current contributions in the potential interval where it is assumed that xanthate chemisorption takes place[16].

Fig. 6 - i-E (a) and iph-E (b) curves recorded for a chalcocite electrode in 0.05 M Na2B4O7 containing increasing ethyl xanthate concentrations: (---) 1x10-5 M; (--) 1x10-4 M; (-) 1x10-3 M. Scan rate: 0.03 Vs-1

The corresponding (Q.E.* hn)^1/2 vs. hn plot leads to a band gap value of 2.05 eV for the indirect transition, which is in agreement with that generally reported for cuprous oxide films [19].

Second, the potential at which the anodic currents start is shifted towards positive values as xanthate concentration increases, however the current peak heights remain practically the same as without xanthate in solution. To account for these results, it is assumed that this peak involves the participation of the first copper sulphide oxidation product of chalcocite in the presence of xanthate physisorbed at the electrode surface, accompanied of cuprous xanthate formation:

The extent of this reaction is controlled by the amount of X^­ adsorbed and not by its solution concentration, thus explaining the independence of current peak on this variable. It is also apparent that through this reaction, copper oxide formation is prevented as inferred by the absence of the corresponding cathodic peaks that were seen in Fig. 2.

Changes in the electrode surface are also reflected in the photocurrent responses of Fig. 6. In the positive scan appears an anodic photocurrent that is sensitive to xanthate concentration and whose potential onset shifts towards negative values as xanthate concentration is increased. A cathodic current plateau is also seen in the negative scan for the higher concentration. To account for the positive photocurrent it is necessary to assume that, due to the presence of xanthate ions, a negative to positive field reversal has been produced in the space charge region of the chalcocite . Then, the photogenerated holes are drained to the surface to assist oxidation reactions. The presence of hysteresis in the photocurrent ands its absence for the current in the dark is typical of electrochemical transformation changing the fiat-band potential of the surface by the creation or annihilation of surfaces states, as is shown for enargite [31] and p-type CuInSe₂ [32]. On the other hand, the presence of a cathodic photocurrent plateau in the negative scan means that the field in the space charge region reversals again but now in the direction that the bands are bending downwards, allowing electrons to migrate to the surface in order to accomplish a reduction reaction.

A similar set of experiments were conducted employing a djurleite enriched electrode surface in presence of ethyl xanthate ions. However the response towards illumination (not shown), was the same as in its absence i.e., no photocurrents were evidenced during the potential scan cycle. This is again due to the quasi-metallic behavior describes before that was attributed to the presence of a high density of acceptors centers within the band gap, preventing in this way the separation of any pair of photogenerated electron ­ holes. These results, together with the differences found in the voltammetric responses of both phases, put in evidence that also different mechanisms operate in both phases regarding the xanthate adsorption process.

4. CONCLUSIONS

In this work new results concerning the influence of composition on the behavior of mineral copper sulfides in conditions that are relevant to their industrial processing have been obtained by using photoelectrochemical measurements. The main conclusion is that composition plays an important role in determining the semiconducting or metallic properties of copper sulfides of varying stoichiometries. These, in turn, lead to different mechanism for the adsorption of reagents that confer hydrophobicity to the minerals surface. The importance of controlling the actual nature of the composition at the interfaces before each electrochemical characterization is also relevant to the interpretation of experimental results.

5. REFERENCES

1. A. Kowal and A. Pomianowski, Electroanal. Chem. Interfacial Electrochem., 46 (1973) 411.

2. G.W. Heyes and W.J.Trahar, Int. J. Miner. Process., 6 (1979) 229.

3. P.E. Richardson, J.V Stout III and C.L. Proctor, Int. J. Miner. Process., 12 (1984) 73.

4. C.S. O'Dell, G.W. Walker and P.E. Richardson, J. Appl. Electrochem., 12 (1984) 55.

5. C.S. O'Dell, G.W. Walker and P.E.Richardson, J. Appl. Electrochem., 16 (1986). 544.

6. G.W. Walker, C.P. Walker and P.E. Richardson, Int. J. Miner Process., 18 (1986) 137.

7. P.K. Ackerman, G.H. Harris and R.R. Klimpel, Int. J. Miner. Process., 21 (1987) 141.

8. J. S. Hanson and D.W. Fuerstenau, Colloid Surface, 26 (1987) 133.

9. A. Pomianowski and W. Barzyk, Bull. Pd. Acad. Sci. Chem., 35 (1987) 461.

10. J.A. Mielczarski, J. Colloid Interface Sci., 120 (1987) 201.

11. W. Barzyk and A. Pomianowski, Mater. Sci. Forum, 25 - 26 (1988) 569.

12. L. Holysz and E. Chibowski, Int. J. Miner. Process., 24 (1988) 1.

13. J.O. Leppinen, C.I. Basilio and R. H. Yoon, Int. J. Miner. Process., 26 (1989) 259.

14. P. Nowak, Fizykochem. Probl. Miner., 22 (1990) 13.

15. J.R. Ross, J.P. Celis and A.S. Sudarsono, Int. J. Miner. Process. 29 (1990) 17.

16. R.Woods, C.A Young and R.H. Yoon, Int. J. Miner. Process., 30 (1990) 17.

17. J.S. Hanson and D.W. Fuerstenau, Int. J. Miner. Process., 33 (1991) 33.

18. P. Persson and B. Maismenten , Colloid Surface, 59 (1991) 279.

19. C.A. Young, C.I. Basilio and R.H. Yoon, Int. J. Miner. Process., 31 (1991) 265.

20. A.N. BuckJey and R. Woods, J.Electroanal. Chem., 357 (1992) 387.

21. Y. Persson, J. Coord. Chem., 32 (1994) 261.

22. G. Contini, K. Laajalehto, E. Suoninen, J. Colloid Interface Sci., 171 (1995) 234.

23. R. Woods and R.H. Yoon , Langmuir 13 (1997) 876.

24. J.A. Mielczarski, Langmuir 13 (1997) 878.

25. H. Gómez, J. Vedel . R. Córdova, R. Schrebler and L. Basáez, J. Electroanal. Chem., 388 (1995) 81.

26. P.E. Richardson and C.S. O'Dell, J. Electrochem. Soc., 132 (1985) 27. S. Fletcher and M.D. Horn, Int. J. Miner. Process., 33 (1990) 145.

28. C.A. Young, R. Woods and R.H. Yoon, in P.E. Richardson and R. Woods (eds). Proc. Int. Symp. on Electrochem. in Mineral and Metal Process. II, Electrochemical Society Pennington, N.J. vol. 88 ­ 21 (1988) p. 3-16.

29. R.T. Shuey, in "Semiconducting Ore Minerals", p.304, Elsevier, Amsterdam (1975).

30. S.C. Basic Data, O. Madelung (ed.), Springer-Verlag, Berlín (1996).

Th. Pauporté and D. Schuhmann, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 111 (1996) 1.

31. Th.Pauporté and D. Shuhmann, colloids and Surfaces A: Physicochemical and Engineering Aspects, 111 (1996).

32. D. Lincot, H. Gómez, J.Kessler, J.Vedel, B.Dimmler and H.W. Schock, Solar Energy Materials, 20 (1990) 67.