Electrochemical reduction of CO2 using palladium modified boron-doped diamond electrodes: enhancing the production of CO

Prastika Krisma Jiwantia and Yasuaki Einaga*ab
aDepartment of Chemistry, Keio University, 3-14-1 Hiyoshi, Yokohama 223-8522, Japan. E-mail: einaga@chem.keio.ac.jp
bACCEL, Japan Science and Technology Agency, 5-3 Yonbancho, Chiyoda 102-8666, Japan

Received 12th March 2019 , Accepted 5th April 2019

First published on 6th April 2019

In recent years, boron-doped diamond (BDD) has been utilized as an electrode for the electrochemical reduction of CO2, and several reports have been published on this. The wide potential window of BDD enables the hydrogen evolution reaction, which competes with CO2 reduction, to be suppressed. On the other hand, the high overpotential is still a problem. We attempted to overcome this problem by depositing metal on the BDD electrode. Pd metal was chosen to modify the surface of the BDD electrode (PdBDD). Employing this electrode at a lower potential of −1.6 V vs. Ag/AgCl, we increased the production of CO (53.3% faradaic efficiency) from the reduction of CO2. We present various attempts made to improve the CO production.


The earth's atmosphere contains copious amounts of carbon dioxide (CO2), in which the concentration of CO2 gas increases every year and, unfortunately, this is tipping the balance of environmental stability towards a dangerous level.1,2 The conversion of this abundant CO2 gas by direct reduction,3–5 or utilizing in electrosynthesis reactions,6 has been studied by many research groups around the world, as one of the attempts to obtain value-added chemicals,7 or even convert it back to fuel, in which, it is also necessary in order to protect the environment.8 Of the many techniques that have been examined, electrochemical reduction is one of the most widely used. However, this technique has drawbacks, which are: the CO2 gas is thermodynamically stable and the CO2 molecules are chemically inert.9 A high overpotential is mostly needed for CO2 electrochemical reduction. This high overpotential causes hydrogen evolution to take place, and this interferes with the reduction of CO2.

Meanwhile, the BDD electrode has been widely researched for numerous applications including for synthesis, which shows exceptional and attractive results.10–13 BDD has a wide potential window, which can help suppress the production of H2. Moreover, the mechanical and chemical stability of BDD make it suitable for practical applications.14 In addition, recently the BDD electrode has also been utilized in the electrochemical reduction of CO2. Both modified and unmodified electrodes have been used. Our group has successfully utilized bare BDD electrodes in the reduction of CO2 and produced HCOOH with a high faradaic efficiency of 94.7% and a selectivity greater than 99% using a circulation flow cell.15 Since this initial study, many improvements have been made in the system we used and detailed scientific studies have been undertaken.16,17 However, a high overpotential was still needed to reduce CO2, and, moreover, only C1 products were successfully produced with these bare BDD electrodes.

Therefore, attempts were made to modify the surfaces of these BDD electrodes with metal nanoparticles, in order to decrease the overpotential or find new products with higher numbers of carbon atoms. With a copper modified BDD electrode (Cu-BDD), we were able to improve the production of C2/C3 species. Ethanol, acetaldehyde, and acetone were produced with faradaic efficiencies of 42.4%, 13.7%, and 7% respectively.18 After this study, modifying the surface of the BDD electrode with other promising metals was explored. There have not been many studies on metal modified BDD electrodes for CO2 reduction; however there are several reports19–22 that show that metal modified BDD electrodes could improve the efficiency, produce new products, and lower the overpotential for CO2 reduction. Spataru et al. deposited RuO2 on BDD to possibly produce methanol besides formic acid using a longtime polarization method.23 On the other hand, one of the metals that did not feature in these studies is palladium (Pd), despite it being a well-known and widely used metal catalyst. Pd is in group VIIIB of the periodic table (i.e. Pt, Ir, Ni, etc.) and is noted for its ability to electrochemically produce H2 gas.24,25 Therefore, in this study, we examined Pd modified BDD electrodes for the electrochemical reduction of CO2 to CO, which is a widely important material used in the chemical industry for the manufacture of, for example, alcohols, aldehydes, and liquid hydrocarbons.



NaCl (99.5%), Na2SO4 (99%), H2SO4 (98%), HCl (35%), and HNO3 (60%) were purchased from Wako Pure Chemical Industries. PdCl2 (99%) was purchased from Sigma Aldrich. All the reagents were used without any further purification. Ultra-pure water was obtained from a Simply-Lab water system (Direct-Q UV3, Millipore).

Preparation of the working electrodes

The electrodes were prepared by depositing BDD (B/C 1%) on the surface of Si(111) wafers using a micro-wave plasma-assisted chemical vapor deposition system (Model AX-5400, CORNES Technology Corp.). The details of this process are described elsewhere.26 Pd particles were electrodeposited on the BDD surface using a chronoamperometry technique in a one compartment cell containing 10 ml of 1 mM PdCl2/0.1 M HCl with Pt spiral and Ag/AgCl (3 M NaCl) as counter and reference electrodes respectively. The applied potential for electrodeposition was −0.15 V, and this was done for 30 s (PdBDD30), 100 s (PdBDD100), 300 s (PdBDD300), 500 s (PdBDD500), and 1000 s (PdBDD1000). After electrodeposition, the electrodes were taken off from the cell, rinsed with ultra-pure water, and dried with N2 gas afterward. The modified electrodes were used as it is for CO2 electrochemical reduction. All the PdBDD electrodes were recovered by immersing in aqua regia for 10 minutes, and then ultrasonicated twice in ultra-pure water for 5 minutes. This was followed by a pretreatment in 0.1 M H2SO4, in which cyclic voltammetry (CV) from −3.5 V to 3.5 V (20 cycles) and then from 0 V to 3.5 V (10 cycles) at scan rates of 1 V s−1 was carried out. The electrodes were characterized using scanning electron microscopy (SEM) (JCM-6000, JEOL) and X-ray photoelectron spectroscopy (XPS) (JPS-9010TR, JEOL).

CO2 reduction method

Electrolysis of CO2 was conducted in two cells (15 ml in each cell) separated by a Nafion membrane. 0.1 M Na2SO4 was used as the anolyte, and 0.1 M NaCl as the catholyte. A Pt mesh and the PdBDD electrodes (0.754 cm2) were used as the counter and working electrodes, respectively. All potentials were measured against Ag/AgCl (3 M NaCl) unless otherwise stated. Prior to the electrochemical reduction, N2 gas at a flow rate of 100 sccm was bubbled for 15 minutes into the catholyte to remove any remaining oxygen gas, and this was followed by CO2 gas purging for the same time and at the same flow rate. A CV scan from −0.5 V to 1.2 V was performed after bubbling the gas through the catholyte. Electrochemical reduction of CO2 at various applied potentials (−1.4 V to −1.9 V) was then performed for 1 hour while stirring. CO2 gas at a low flow rate was fed into the cell to maintain the CO2 concentration during reduction. All the electrochemical measurements were recorded using a potentiostat (Autolab PGSTAT204, Metrohm Autolab B.V.)

Product analysis method

One liter of the gas products was collected and a 1 ml sample of the liquid was taken from the cell. Both were subsequently analyzed after the reduction process. The gas products were analyzed by gas chromatography with a flame ionization detector and a thermal conductivity detector (GC-2014, Shimadzu Corp.). The liquid was analyzed using high performance liquid chromatography (HPLC) with an electroconductivity detector (Prominence, Shimadzu Corp.).

Results and discussion

Characterization of the PdBDD electrodes

Fig. 1 shows the CV curves for PdBDD electrodes with different deposition times, in comparison with a bare BDD electrode. CV scans from −0.5 V to 1.2 V were done after bubbling CO2 through the 0.1 M NaCl solution, and before carrying out the CO2 reduction process. With the bare BDD electrode, no peaks can be seen in this potential range, showing the excellent behavior of the BDD electrode, that is, the low background current. The CV curves for the BDD electrodes with Pd particles deposited on the surface are typical of Pd electrodes. With the negative-going potential, H2 adsorption occurs at around −0.3 V, and hydrogen evolution at >−0.4 V. With the positive-going potential, we can see the H2 desorption peak at around −0.3 V, followed by the oxidation peak of Pd0 to Pd2+ at about 0.7 V. The reduction of Pd2+ back to Pd0 can be seen at a potential of around 0.2 V. The CV curves also show that as the deposition time increases, the amount of deposited Pd particles increases, and thus the height of the Pd oxidation peak at 0.7 V increases.
image file: c9cp01409h-f1.tif
Fig. 1 CV curves of PdBDD electrodes with different Pd deposition times in 0.1 M NaCl solution.

All the PdBDD electrodes after electrodeposition at −0.15 V were examined by SEM (Fig. 2(1)–(5)). As can be seen in the figures, the particles are deposited quite homogeneously over the active parts of the surfaces, with average particle sizes of around ∼50 nm, except on the PdBDD500 and PdBDD1000 electrodes, where the particles have agglomerated and formed particles of larger size (∼160 nm). In addition, the mass of the Pd particles is calculated based on the total charge consumed during electrodeposition, assuming that there was no contribution to hydrogen evolution. On PdBDD30, the areal density of the particles is 2.75 μg cm−2, and this increases as the deposition time becomes longer (Fig. 2(6)). Moreover, the Pd particles on the surfaces of the electrodes were also characterized using XPS, and this showed peaks for Pd 3d5/2 at 335.3 eV and Pd 3d3/2 at 340.5 eV, which are representative of the metallic form of Pd (Pd0). Deconvolution shows that the amount of PdOx on the surface is less than 10%, and thus the catalytic effect of Pd is considered to come from the Pd0 species.

image file: c9cp01409h-f2.tif
Fig. 2 SEM images of PdBDD electrodes prepared at a potential of −0.15 V with deposition times of 30 s (1), 100 s (2), 300 s (3), 500 s (4), and 1000 s (5), and the areal density of the Pd particles vs. deposition time (6).

Electrochemical reduction of CO2 using PdBDD electrodes

The electrochemical reduction of CO2 was evaluated from CV plots in which the potential was scanned from −1.8 V to 1.2 V after N2 and CO2 bubbling (Fig. 3). The scans were carried out starting from 0 V to an anodic potential of 1.2 V, then to a cathodic potential of −1.8 V and back to 0 V. CO2 reduction was considered to have taken place in this potential range, as can be observed from the reduction peak at around −1.0 V, which is indicative of the production of reduced-CO2 species as it cannot be seen in the CV curves after N2 bubbling.
image file: c9cp01409h-f3.tif
Fig. 3 CV of PdBDD500 after N2 and CO2 gas bubbling performed at potentials from −1.8 V to 1.2 V with a scan rate of 100 mV s−1.

This result is consistent with the results mentioned in ref. 27 and 28, where adsorbed-CO (COad) is formed on the surface of the electrode during CV in a bicarbonate solution containing CO2 gas, and this COad is then oxidized to form CO2 at anodic potentials. In addition, in the presence of CO2, the hydrogen evolution has shifted to a more negative potential, which can be explained as being due to CO2 reduction products covering the surface of the BDD electrode, and thus decreasing the H2 evolution. Fig. S1 (ESI) shows the CV scans for the other PdBDD electrodes fabricated in this study. As the deposition time for Pd increases, the height of the oxidation peak of Pd (at 0.7 V) increases, and the intensity of the peak at −1.0 V, which indicates that an increasing amount of reduced-CO2 species has become attached to the surface of the BDD electrode.

Electrochemical reduction of CO2 was then carried out using all the PdBDD electrodes at a potential of −1.5 V for 1 hour (Fig. 4). The main product was CO, with HCOOH as a side product. No other products were detected, and the total faradaic efficiencies, including the H2 production, were more or less 100%. The faradaic efficiency for CO production is calculated based on 2e consumed in the reduction process, according to the reaction:

CO2(g) + H(aq)+ + 2e → CO(g) + OH(aq) (1)

image file: c9cp01409h-f4.tif
Fig. 4 Faradaic efficiencies for the products from the electrochemical reduction of CO2 at a potential of −1.5 V for 1 hour using PdBDD electrodes with different Pd deposition times.

The faradaic efficiency for the production of CO increases as the amount of Pd particles increases, but decreases slightly for the longest deposition time. This can be explained by the instability of the deposited Pd particles, where the Pd particles have agglomerated with longer deposition times, making it easier for them to become detached from the surface of the electrode, thereby decreasing the current density for CO2 reduction (Fig. S2, ESI).

In a more detailed study, the dependence of the faradaic efficiency on the electrode potential from −1.4 V to −1.9 V was evaluated for each PdBDD electrode (Fig. 5). The maximum production rate for CO is 53.3% (HCOOH 9.3%, H2 39.9%), obtained with the PdBDD300 electrode at a potential of −1.6 V. This rate of production for CO is approximately five times higher than that obtained with a bare BDD electrode and also that obtained with a Pd metal electrode under the same conditions, where the production of CO was around ∼9% (Table S1, ESI). Thus, the deposited Pd particles on the surface of the BDD electrode may catalyze the production of CO at a relatively lower potential than that previously obtained with CO2 reduction on a bare BDD electrode (>−2.0 V), producing higher amounts of CO2 reduction products.15

image file: c9cp01409h-f5.tif
Fig. 5 Faradaic efficiency for CO2 electrochemical reduction products vs. electrode potential with (A) PdBDD30, (B) PdBDD100, (C) PdBDD300, (D) PdBDD500, and (E) PdBDD1000, with each reduction process performed for 1 hour.

The faradaic efficiency for CO production at potentials more negative than −1.5 V with the PdBDD30 and PdBDD100 electrodes is ∼20%, which is twice that obtained with a bare BDD electrode. Moreover, with the PdBDD300, PdBDD500 and PdBDD1000 electrodes, the faradaic efficiency can be improved to 30–50%, which is three to five times higher than that with a bare BDD. This shows that there needs to be a sufficient amount of Pd particles to obtain the best results for CO production. Small amounts of Pd particles can double the faradaic efficiency, whereas the stability of electrodes with higher amounts of Pd particles during electrochemical reduction is problematic. The SEM images and measurements of the sizes of the Pd particles on the surfaces of the BDD electrodes show the differences before and after electrochemical reduction at −1.5 V for 1 hour (Fig. S3–S5, ESI). The size of the Pd particles on the PdBDD500 and PdBDD1000 electrodes has decreased, possibly due to the removal of particles from the surface of the electrode. Unlike the particle size dependence studies in the literature,29,30 however, it is suggested that the particle size dependence in this range (50–100 nm) does not seriously affect CO production since faradaic efficiencies of up to ∼40% with PdBDD500, which has larger particles than that of PdBDD300, were achieved. In addition, since the production of HCOOH is much the same (more or less 5–10%) in all cases, it may be that the catalytic effect of the Pd particles only influences the production of CO.

Since the electrolyte was stirred during the measurements, the effect of the stirring speed was studied. Mass transfer to the surface of the electrode can have a significant effect on CO2 reduction and selectivity.31 Stirring enables the dissolved CO2 in the bulk to come closer to the electrode,32 further affecting the faradaic efficiency of the CO2 products. Stirring speeds of ∼120 rpm, ∼240 rpm, and ∼480 rpm were applied with the PdBDD300 electrode at a potential of −1.6 V (Fig. S6, ESI). At a low stirring speed (∼120 rpm), the faradaic efficiency for CO was 39.9%. This increased to 53.3% for the moderate stirring speed (∼240 rpm), but then dropped to 39.8% at the fast stirring speed (∼480 rpm), whereas hydrogen production at this fast speed was higher. This can be explained by the increase in current density during electroreduction. As we can see, the current density gradually increases with the stirring speed. Taking account of the exchange current density as a function of the calculated free energy of adsorbed hydrogen, known as “the volcano plot”,33,34 metals on the left hand side of the volcano plot have large H coverage in contrast to metals that stand on the right hand side which have small H coverage. Pd is on the left hand side of the volcano plot, where the high current density may result in a higher faradaic efficiency for hydrogen production. Therefore, this limits the CO2 reduction. Finally, an appropriate stirring speed is necessary to control the mass transfer, while considering the catalytic activity of the metal toward hydrogen production.


In summary, we successfully modified the surfaces of BDD electrodes by depositing Pd particles onto them. These electrodes were applied to the electrochemical reduction of CO2. Using an electrode with 16.39 μg cm−2 of Pd particles on the surface, the main CO2 reduction product at a potential of −1.6 V was CO. The deposited Pd particles remained stable after electrochemical reduction at −1.5 V on the surfaces of electrodes with smaller Pd particles (PdBDD30, PdBDD100, and PdBDD300). For those electrodes with bigger particles (PdBDD500 and PdBDD1000), the size of the particles decreased after reduction. Now we have an alternative metal, Pd, that can be deposited on the surface of BDD electrodes to increase the reduction of CO2, and possibly decrease the overpotential for CO2 reduction when using a BDD electrode. In particular, we can increase the production of CO using these modified BDD electrodes.

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. J. S. Sawyer, Nature, 1972, 239, 23–26 CrossRef CAS.
  2. S. Willie, L. B. Sallie, B. R. Arthur and W. R. Zachary, Clim. Res., 1999, 13, 149–164 CrossRef.
  3. M. Gattrell, N. Gupta and A. Co, J. Electroanal. Chem., 2006, 594, 1–19 CrossRef CAS.
  4. J. Qiao, Y. Liu, F. Hong and J. Zhang, Chem. Soc. Rev., 2014, 43, 631–675 RSC.
  5. W. Zhang, Y. Hu, L. Ma, G. Zhu, Y. Wang, X. Xue, R. Chen, S. Yang and Z. Jin, Adv. Sci., 2018, 5, 1700275 CrossRef PubMed.
  6. X. Yuan, B. Lu, J. Liu, X. You, J. Zhao and Q. Cai, J. Electrochem. Soc., 2012, 159, E183–E186 CrossRef CAS.
  7. S. Möhle, M. Zirbes, E. Rodrigo, T. Gieshoff, A. Wiebe and S. R. Waldvogel, Angew. Chem., Int. Ed., 2018, 57, 6018–6041 CrossRef PubMed.
  8. J. Wei, Q. Ge, R. Yao, Z. Wen, C. Fang, L. Guo, H. Xu and J. Sun, Nat. Commun., 2017, 8, 1–8 CrossRef PubMed.
  9. Q. Lu, J. Rosen, Y. Zhou, G. S. Hutchings, Y. C. Kimmel, J. G. Chen and F. Jiao, Nat. Commun., 2014, 5, 1–6 Search PubMed.
  10. T. A. Ivandini and Y. Einaga, Chem. Commun., 2017, 53, 1338–1347 RSC.
  11. S. Lips and S. R. Waldvogel, ChemElectroChem, 2019, 6, 1649–1660 CrossRef CAS.
  12. N. Yang, S. Yu, J. V. Macpherson, Y. Einaga, H. Zhao, G. Zhao, G. M. Swain and X. Jiang, Chem. Soc. Rev., 2019, 48, 157–204 RSC.
  13. N. Yang, S. R. Waldvogel and X. Jiang, ACS Appl. Mater. Interfaces, 2016, 8, 28357–28371 CrossRef CAS PubMed.
  14. Y. Einaga, J. Appl. Electrochem., 2010, 40, 1807–1816 CrossRef CAS.
  15. K. Natsui, H. Iwakawa, N. Ikemiya, K. Nakata and Y. Einaga, Angew. Chem., Int. Ed., 2018, 57, 2639–2643 CrossRef CAS PubMed.
  16. M. Tomisaki, K. Natsui, N. Ikemiya, K. Nakata and Y. Einaga, ChemistrySelect, 2018, 3, 10209–10213 CrossRef CAS.
  17. J. Xu, K. Natsui, S. Naoi, K. Nakata and Y. Einaga, Diamond Relat. Mater., 2018, 86, 167–172 CrossRef CAS.
  18. P. K. Jiwanti, K. Natsui, K. Nakata and Y. Einaga, Electrochim. Acta, 2018, 266, 414–419 CrossRef CAS.
  19. N. Roy, Y. Hirano, H. Kuriyama, P. Sudhagar, N. Suzuki, K. Katsumata, K. Nakata, T. Kondo, M. Yuasa, I. Serizawa, T. Takayama, A. Kudo, A. Fujishima and C. Terashima, Sci. Rep., 2016, 6, 38010 CrossRef CAS PubMed.
  20. N. S. R. Cuellar, K. Wiesner-Fliescher, O. Hinrichsen and M. Fleischer, ChemistrySelect, 2018, 3, 3591–3595 CrossRef.
  21. N. Roy, N. Suzuki, Y. Nakabayashi, Y. Hirano, K. Katsumata, K. Nakata, A. Fujishima, C. Terashima and H. Ikari, ChemElectroChem, 2018, 5, 2542–2550 CrossRef CAS.
  22. Y. Nakabayashi, Y. Hirano, Y. Sakurai, A. Okazaki, H. Kuriyama, N. Roy, N. Suzuki, K. Nakata, K. Katsumata, A. Fujishima and C. Terashima, J. Appl. Electrochem., 2018, 48, 61–73 CrossRef CAS.
  23. N. Spataru, K. Tokuhiro, C. Terashima, T. N. Rao and A. Fujishima, J. Appl. Electrochem., 2003, 33, 1205–1210 CrossRef CAS.
  24. W. Zheng, H. W. Man, L. Ye and S. C. E. Tsang, Energy Technol., 2017, 5, 937–944 CrossRef CAS.
  25. R. Kortlever, I. Peters, C. Balemans, R. Kas, Y. Kwon, G. Mul and M. T. M. Koper, Chem. Commun., 2016, 52, 10229–10232 RSC.
  26. T. Yano, D. A. Tryk, K. Hashimoto and A. Fujishima, J. Electrochem. Soc., 1998, 145, 1870–1876 CrossRef CAS.
  27. D. Kolbe and W. Vielstich, Electrochim. Acta, 1996, 41, 2457–2460 CrossRef CAS.
  28. S. Pérez-Rodríguez, E. Pastor and M. J. Lázaro, J. CO2 Util., 2017, 18, 41–52 CrossRef.
  29. R. Reske, H. Mistry, F. Behafarid, B. R. Cuenya and P. Strasser, J. Am. Chem. Soc., 2014, 136, 6978–6986 CrossRef CAS PubMed.
  30. D. Gao, H. Zhou, J. Wang, S. Miao, F. Yang, G. Wang, J. Wang and X. Bao, J. Am. Chem. Soc., 2015, 137, 4288–4291 CrossRef CAS PubMed.
  31. D. Raciti, M. Mao, J. H. Park and C. Wang, Catal. Sci. Technol., 2018, 8, 2364–2369 RSC.
  32. C. F. C. Lim, D. A. Harrington and A. T. Marshall, Electrochim. Acta, 2017, 238, 56–63 CrossRef CAS.
  33. S. Trasatti, J. Electroanal. Chem. Interfacial Electrochem., 1972, 39, 163–184 CrossRef CAS.
  34. Y. J. Zhang, V. Sethuraman, R. Michalsky and A. A. Peterson, ACS Catal., 2014, 4, 3742–3748 CrossRef CAS.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9cp01409h

This journal is © the Owner Societies 2019