J. Electroanal. Chem., 238 (1987) 247-258 Elsevier Sequoia S.A., Lausanne - Printed in The Netherlands
REDUCTION OF CARBON DIOXIDE ON PARTIALLY-IMMERSED PLATE ELECTRODE AND Au-SPE ELECTRODE
Department of Apphed Nagoya 466 (Japan)
Nagoya Instriute of Technology.
(Received 19th May 1987)
The electrocatalytic reduction of carbon dioxide was carned out on two types of Au plate electrode; one was partially immersed in electrolyte solution containing CO, and the other was completely immersed. Their surface areas in contact with the solution were kept equal, so that the difference m Faradaic efficiency for the reduction of CO, to CO between the two types of electrodes was examined. The effects of surface treatments of the electrodes and electrolytes on the Faradaic efficiency were also investigated for each type of electrode. When the electrode surface was treated by electroplating Au on the Au plate, the partially-immersed electrode brought about a higher Faradaic efficiency for CO, reduction than the completely immersed one both in 1 mol dme3 KHCO, and in 0.1 mol dm-’ (C,H,),NClO, aqueous solution at room temperature. The Au-SPE (SPE: sohd polymer electrolyte) electrode was applied as a gas diffusion electrode to the reduction of CO,, and the Faradauz efficiency for CO production was compared with that obtained by the partially-immersed Au plate electrode. The Au-SPE electrode gave rise to a higher Faradaic efficiency than the partially-immersed electrode.
So far, a number of studies have been carried out on the electrocatalytic reduction of carbon dioxide at metal electrodes. They have been summarized in recent papers [l-5]. It has been found that CO, in aqueous solutions reduces to formic acid at high Faradaic efficiencies with some indication of further reduction to formaldehyde and methanol. Recently, Hori et al. [4,6] have carried out the electrochemical reduction of CO, in aqueous KHCO, solution at different temperatures by means of several kinds of metal electrodes; the principal products are formic acid on Cd, Sn, Pb and Zn, carbon monoxide on Ag and Au and methane on Cu. Summers et al.  have reported that the electrolysis using molybdenum metal electrodes in 0.2 mol dme3 Na,SO, carbon dioxide saturated aqueous solution at pH 4.2 produced methanol as the major carbon-containing product. We have found
that the metal electrodes, which catalyze the electrochemical reduction of CO*, are classified into three groups according to their main products from CO*, i.e., metallic In, Hg and Pb are selective for the formation of formic acid in aqueous solution; metallic Hg, Pb and Tl are selective for the production of oxalic acid; and metallic Zn, In, Sn and Au for carbon monoxide production in non-aqueous solutions . Carboxylic acids with larger molar masses, such as tartaric, malonic, glycolic, propionic and n-butylic acids, have also been formed at substantial concentrations in non-aqueous solutions . It has been known since the late nineteenth century that the exposure of part of a platinum electrode to hydrogen gas above the electrolyte solution results in a marked increase in current . Will [lo] studied the effect of the surface roughness of the platinum electrode on the current-potential curve during oxidation of hydrogen and found that if the electrode is partially immersed in electrolyte solution, the magnitude of the oxidation current increases with increasing surface roughness, while for a completely-submerged electrode there is no influence of surface roughness. Takahashi and Ito [ll] observed that with flat plate metal electrodes such as Au, Ag, Pt and Pd partially immersed in aqueous KOH the magnitudes of polarization for the reduction of oxygen are lowered, compared with those for the completely-immersed ones. Recently, Ogura and Watanabe  have employed a partially-immersed Pt plate electrode coated with Eve&t’s salt for the electrocatalytic reduction of carbon monoxide in aqueous methanol and have found that the Faradaic efficiency for the conversion of carbon monoxide into methanol is enhanced on the partially-immersed electrode, compared with that observed in the completely-immersed electrode system. It is understood that these results are caused by the rapid transport of the reactant gas to the active zone of the electrode through the three-phase (electrode/solution/gas) interface occurring on the partially-immersed flat plate electrode. This phenomenon has been utilized in gas diffusion electrodes, which are made up with a large number of cylindrical pores with each pore consisting of flat plates partially immersed in electrolyte solution. A Pt-SPE electrode, which is one of the gas diffusion electrodes, has been applied to the electrochemical hydrogenation of olefinic double bonds, the Kolbe reaction and the Brown-Walker reaction [13,14]. In the present work, first, the electrocatalytic reduction of carbon dioxide was carried out by means of flat plate metal electrodes partially and completely immersed in electrolyte solution in order to examine whether the enhancement of the Faradaic efficiency for CO, reduction was achieved in the partially-immersed electrode system. Gold was used as the electrode material and 1 mol dm-3 KHCO, and 0.1 mol dm- 3 (C2H,),NC104 aqueous solutions as electrolytes. The electrode surfaces were treated in two different ways, namely, by electroplating Au on the Au plate and by polishing the Au plate mechanically with abrasive paper. The Au-SPE electrode was employed as a gas diffusion electrode for the electrocatalytic redution of CO, in aqueous 1 mol dmp3 KHCO,. The Faradaic efficiency for CO, reduction was compared with that obtained by the partially-immersed Au plate electrode.
Materials Analytical grade potassium hydrogen carbonate was used without further purification. Tetraethylammonium perchlorate was prepared from analytical grade tetraethylammonium bromide and perchloric acid, recrystallized repeatedly from water and dried at 343 K under vacuum . The Au plates were used as the substrates for the Au plate electrodes. Analytical grade hydrogen tetrachloroaurate(II1) was used without further purification. Apparatus Au plates of two different sizes (0.5 x 0.9 cm and 0.5 X 1.8 cm) were used as the working electrodes. The smaller plate was employed as the completely-immersed electrode and the larger as the partially-immersed electrode so that the surface areas in contact with the electrolyte were kept equal. Electrical connection was made by spot-welding Au wire on one surface of the Au plate electrode, with epoxy resin being used to fix the Au plate to a Pyrex glass tube for the lead wire as well as to prevent the spot-welded surface from making contact with the electrolyte solutions. The surfaces of the Au plate electrodes were treated in two different ways, namely, by electroplating Au (5 C cmm2) in aqueous H[AuCl,] at a current density of 20 mA cm-2 and by polishing the surface mechanically with No. 1500 abrasive paper. The electrode areas submerged in the solution were measured with a cathetometer. The Au-SPE electrode was prepared as follows . A cation exchange membrane, [email protected]
315 (Du Pont), was used for the SPE. A sheet of SPE (3 cm in diameter) was first soaked in boiling water for 0.5 h and then mounted in a Teflon cell in such a way that one side of the SPE contacted with 3 x lop3 mol dmp3 H[AuCl,] aqueous solution (2 ml) and the other with 1.25 mol dmp3 NaOH aqueous solution (100 ml) containing the reducing agent, hydrazine. It took ca. 1 h to deposit the Au on the SPE at room temperature in the dark. The silver-silver chloride electrode saturated with KC1 was employed as the reference electrode for potential measurements. A Pyrex “H” type electrolytic cell consisting of two compartments was employed for the electrochemical measurements with the Au plate electrodes. The two compartments were separated with a cation exchange membrane, Nafion 315. In one compartment the working Au electrode and the reference electrode were immersed. The Pt plate electrode as a counter electrode was dipped in the other one. The electrochemical measurements with the Au-SPE electrode were carried out by using an electrolytic cell consisting of three compartments made of Teflon, which is illustrated schematically in Fig. 1. The cathode and middle compartments were separated by the Au-SPE electrode with the surface coated with Au facing the cathode side, and the anode and middle compartments by Nafion 315. The middle compartment was designed so that the electrolysis products were prevented from
Fig. 1. Schematic diagram of electrolytic cell with Au-SPE
oxidation at the anode. Ohmic contact of the Au of the Au-SPE electrode with the lead wire was achieved with a Pt gauze inserted between the Au-SPE electrode and the cathode compartment.
of cyclic voltammetry
The electrolytic cells were kept gas tight. CO, gas was supplied with the cyclic flow system, which was assembled from a mini-peristaltic pump, a manometer, a gas sampler and glass tubing. Prior to the electrolytic measurements with the Au plate electrodes which were either partially or completely immersed in the catholyte solution (1 mol dmp3 KHCO, or 0.1 mol dmp3 (CzH,),NCIO,), CO, gas was first supplied into the catholyte of the cell from a CO, gas cylinder and exhausted through the cyclic gas flow system for 30 min. After the CO, saturation was established both in the electrolyte and in the flow system, the line was closed to make it gas tight. The gas in the system was circulated by the mini-peristaltic pump at the rate of 10 ml min-’ at atmospheric pressure, while the electrolysis was being performed. For the electrolytic measurements with the Au-SPE electrode, the anode and middle compartments were charged with 1 mol dm-3 KHCO, aqueous solution and the cathode compartment with the CO, gas alone using the same cyclic gas flow system as described above. The gas in the system was circulated by the mini-peristaltic pump at the rate of 10 ml mm’ at atmospheric pressure, during the electrolysis. The cyclic voltammograms were obtained by cyclic linear sweep voltammetry, using a linear scanning unit (Hokuto Denko, LS-2D). The experiments on preparative CO, reduction for the analyses of the reduction products and for the evaluation of the Faradaic efficiencies were performed
potentiostatically at 298 K by using a potentiostat (Hokuto Denko, PS-303) and an electronic coulometer (Hokuto Denko, HF-201). The amounts of products in the gas and solution phases resulting from the electrolysis were determined chromatographically according to the same procedures as described previously . RESULTS AND DISCUSSION
Au plate electrode system In Figs. 2(a) and (b), the cyclic potential-current curve measured on the electroplated electrode partially immersed in 0.1 mol dme3 (C2H,),NC10, solution under a CO, atmosphere is compared with that measured on the completely-immersed electrode. The curves obtained under an N, atmosphere without CO, are also shown for comparison. It is apparent that the current densities of the CO,-dissolved solution in the cathodic potential range of - 1.0 to - 1.7 V are higher than those of the backgrounds (those of the N,-dissolved solutions), which results from the reduction of CO,. It seems that the cathodic current in this potential range is slightly more enhanced on the partially-immersed electrode than on the completelyimmersed one. This result suggests that the controlled potential electrolysis in this potential range by means of the electroplated electrode partially immersed in 0.1 aqueous solution may result in enhancement of the mol dmm3 (C,H,),NClO, Faradaic efficiency for the CO, reduction. As far as the voltammograms were concerned, no distinct difference in current between the partially-immersed electrode and the completely-immersed one was observed in the other systems. From the controlled potential coulometry experiments, it was found that the only reduction product obtained is carbon monoxide, irrespective of the types of electrodes and electrolytes and of the methods of the surface treatments. The present result is consistent with that reported previously in aqueous KHCO, . In Table 1 are given the Faradaic efficiencies for CO production and H, evolution on the polished electrodes partially and completely immersed in 1 mol dme3 KHCO, solution, and in Table 2 are those in 0.1 mol dmp3 (C2H,),NC104 solution. It is seen that the Faradaic efficiency for CO production in the (C2H,),NC10, solution is much better with the value as high as 68%. compared with that of ca. 16% in the KHCO, solution. By inspection of the current efficiencies for CO formation in Tables 1 and 2, it is concluded that there is no significant difference in Faradaic efficiency between the partially-immersed electrode and the completely-immersed one, when the electrodes are treated with abrasive paper. The lack of Faradaic balances, especially in Table 1, is probably caused by low accuracies in the quantitative determination of hydrogen. The Faradaic efficiencies for CO production and H, evolution on the electroplated electrodes partially and completely immersed in the KHCO, solution are shown graphically in Fig. 3, and those in the (C,H,),NClO, solution in Fig. 4. As is obvious from the figures, significant differences in Faradaic efficiency for CO formation between the partially-immersed
252 Potential -2.0
Fig. 2. Cyclic voltammograms of 0.1 mol dme3 (C2H,),NC104 solutions saturated with CO, and N,. (a) Partially-immersed electrode; (b) completely-immersed electrode. Electrode surfaces were treated by electroplating. Sweep rate: 50 mV s-‘.
on Au plate electrodes partially and Faradaic efficiencies for CO production and Hz evolution completely immersed in 1 mol dmp3 KHCO, solution. Electrode surfaces were treated by polishing mechanically with No. 1500 abrasive paper. Quantity of electricity passed: 30 C; potential: - 1.5 V vs. Ag-A&l Faradax
electrode and the completely-immersed one are observed in both KHCO, and (C,H,),NClO, solutions, when the electrodes are treated by electroplating of Au. The largest difference in Faradaic efficiency is 15% in the (C,H,),NClO, solution.
I -1.3 / V(vs.
Fig. 3. Faradaic efficiencies for CO production and H, evolution on Au plate electrodes partially and completely immersed in 1 mol drnm3 KHCO, solution. (e) CO, (A) H, on partially-immersed electrode; (0) CO, (A) Hz on completely-immersed electrode. Electrode surfaces were treated by electroplating. Quantity of electricity passed: 30 C.
-1.7 / V(vs.
-1 i' AslAsCl)
Fig. 4. Faradax efficiencies for CO production and Hz evolutton on Au plate electrodes partially and completely immersed in 0.1 mol dme3 (C,H,),NClO, solution. (0) CO, (A) H2 on partially-nnmersed electrode; (0) CO, (A) H, on completely-immersed electrode. Electrode surfaces were treated by electroplating. Quantity of electricity passed: 30 C.
The (C,H,),NClO, system yields much higher Faradaic efficiencies for CO production over the potentials studied than the KHCO, system. This tendency is in accord with that observed in the polished electrode system. The Faradaic efficiencies for CO production remain fairly constant over the potentials studied in the (C,H,),NClO, system, while they decrease rapidly as the applied potential becomes more negative in the KHCO, system. Inspection of the Faradaic efficiencies for CO production in Tables 1 and 2 and in Figs. 3 and 4 reveals that when the electrodes are completely immersed, no significant difference in Faradaic efficiency is observed between the electroplated electrode and the polished one. Electron microscopic examination of the surfaces of the electroplated electrode and of the polished one revealed that finely divided particles are deposited on the surface of the former electrode, while scratches caused by polishing are present on the surface of the latter. From this observation, it may be said that the former electrode has a higher surface roughness than the latter. If Will’s conclusion that the magnitude of the oxidation current of hydrogen on the Pt electrode partially immersed in electrolyte solution increases with increasing surface roughness [lo] is
Faradaic efficiencies completely immersed polishing mechanically - 1.7 V vs. Ag/AgCl
on Au plate for CO production and H, evolution solution. Electrode in 0.1 mol dm -3 (C,H,),NCIO., with No. 1500 abrasive paper. Quantity of electricity
electrodes partially and surfaces were treated by passed: 30 C; potential:
taken into consideration with the present electron microscopic observation, the relative increase in the Faradaic efficiency for CO production, which occurs when the electropolated electrode is raised partly above the solution level, may be ascribed to higher surface roughness of the electroplated electrode compared with that of the polished one. There was no difference in size of the meniscus between the electroplated electrode and the polished one, as determined with a cathetometer. As was pointed out by Will [lo], this finding suggests that the surface roughness may affect the state of the microscopic areas, like the upper edge of the meniscus and the adjacent film close to it, in such a way that the diffusion of CO, molecules through the microscopic areas into the active zone of the electrode is promoted. Au-SPE
Figure 5 illustrates the cyclic potential-current curves on the Au of Au-SPE in contact with 1 mol dm- 3 KHCO, solution under CO, and N, atmospheres. The current density of the CO, atmosphere in the cathodic potential range is higher than that of the background. This result indicates that CO, reduction occurs. The current density is reduced to about one tenth of that obtained on the Au plate electrodes. The decrease is ascribed mainly to the fact that the current density has been estimated just on the basis of the surface area of the SPE covered with Au, i.e., since all the Au particles deposited on SPE are not in contact with each other with electrical conduction, it is impossible to collect electricity from all the particles by means of the platinum gauze used in the present experiments. The only product in the gas phase obtained from the controlled potential coulometry experiments was carbon monoxide. The Faradaic efficiencies for CO production and H, evolution are shown graphically in Fig. 6, together with those obtained on the Au plate electrode partially and completely submerged in 1 mol dmp3 KHCO, solution. It is apparent that the Au-SPE electrode has achieved the highest Faradaic efficiency for CO production over all the applied voltages. The fact that the Faradaic efficiency for CO formation on the Au-SPE electrode decreases as the applied potential becomes more negative is characteristic of the KHCO, solution systems, as is the case on the Au plate electrodes.
256 Potential -1.5
/ V(vs. -1.0
Fig. 5. Cyclic voltammograms of CO, and Nz atmospheres KHCO, solution system. Sweep rate: 50 mV s- ‘.
in 1 mol dmm3
It was found from the electron microscopic examination of the Au-SPE electrode that the surface of the electrode has an analogous appearance to that of the electroplated Au plate electrode. Thus, it may be said that the Au-SPE electrode employed in the present work has high surface roughness, which has led to the effective formation of the three-phase interfaces and to the highest Faradaic efficiency for CO production in the Au-SPE electrode system among the systems studied.
The present experiments gave the following information: (1) It may be possible to enhance the Faradaic efficiency for CO, reduction on the partially-immersed Au plate electrode, if the electrode surface is treated so as to have as high a surface roughness as possible. (2) For the completely-submerged Au plate electrode, there is no significant influence of surface roughness on the Faradaic efficiency for CO, reduction.
Fig. 6. Faradaic efficiencies for CO production and Hz evolution on Au-SPE, partially-immersed and completely-immersed electrodes in 1 mol dm -3 KHC03 solution system. (o- - -0) CO, (A- - -A) H, on partially-immersed electrode; co, (A -----A) H, on Au-SPE electrode; (0 -0) A) H, on completely-immersed electrode. Partially- and completely-im0) CO, (A(Omersed electrode surfaces were treated by electroplating. Quantity of electricity passed: 30 C.
(3) With the Au plate electrodes, the (C,H,),NClO, solution system yields a much higher Faradaic efficiency for CO production than the KHCO, solution system. This result emphasizes that the choice of electrolytes is of crucial importance for the achievement of high Faradaic efficiency for CO1 reduction. (4) It is feasible for the Au-SPE electrode to be used as a gas diffusion electrode for the electrocatalytic reduction of CO,. (5) The lead on the Au-SPE electrode should be designed in such a way that the efficiency for electricity collection is raised. ACKNOWLEDGEMENT
This work was supported in part by a Grant-in-Aid for Scientific Research No. 62603525 from the Ministry of Education, Science and Culture, of Japan.
258 REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
P.G. Russell, N. Kovac, S. Srinivasan and M. Steinberg, J. Electrochem. Sot., 124 (1977) 1329. K. Ito, S. Ikeda, T. Iida and A. Nomura, J. Electrochem. Sot. Jpn. (Denki Kagaku), 50 (1982) 463. K. Kapusta and N. Hackerman, J. Electrochem. Sot., 130 (1983) 607. Y. Hori, K. Kikuchi and S. Suzuki, Chem. Lett., (1985) 1695. Y.B. Vassihev, VS. Bagotzky, N.V. Osetrova, O.A. Khazova and N.A. Mayorava, J. Electroanal Chem., 189 (1985) 271. Y. Hori, K. Kikuchi, A. Murata and S. Suzuki, Chem. Lett., (1986) 897. D.P. Summers, S. Leach and K.W. Frese Jr., J. Electroanal. Chem., 205 (1986) 219. S. Ikeda, T. Takagi and K. Ito, Bull. Chem. Sot. Jpn., 10 (1987) 2517. W.R. Grove, Philos. Mag., 14 (1893) 127. F.G. Will, J. Electrochem. Sot., 110 (1963) 145. T. Takahashi and K. Ito, J. Electrochem. Sot. Jpn. (Denki Kagaku), 30 (1962) 431. K. Ogura and H. Watanabe, J. Chem. Sot. Faraday Trans., 81 (1985) 1569. Z. Ogumi, K. Nishino and S. Yoshizawa, Electrochim. Acta, 26 (1981) 1779. Z. Ogumi, H. Yamashita, K. Nishino, Z. Takehara and S. Yoshizawa, Electrochim. Acta, 28 (1983) 1687. D.D. Pen-in, W.L.F. Aramarego and D.R. Perrin, Purification of Laboratory Chemicals, Pergamon. Oxford, 1980. Z. Ogumi, S. Nishio and S. Yoshizawa, J. Electrochem. Sot. Jpn. (De&i Kagaku), 52 (1981) 212.