- Email: [email protected]
Hydrogen peroxide generation in flow-mode capacitive deionization
Journal of Electroanalytical Chemistry 776 (2016) 101–104
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
Short communication
Hydrogen peroxide generation in flow-mode capacitive deionization Taeyoung Kim a, Jihyun Yu a, Choonsoo Kim a, Jeyong Yoon a,b,⁎ a School of Chemical and Biological Engineering, College of Engineering, Institute of Chemical Processes(ICP), Seoul National University (SNU), Gwanak-gu, Daehak-dong, Seoul 151-742, Republic of Korea b Asian Institute for Energy, Environment & Sustainability (AIEES), Seoul National University (SNU), Gwanak-gu, Daehak-dong, Seoul 151-742, Republic of Korea
a r t i c l e
i n f o
Article history: Received 28 March 2016 Received in revised form 28 June 2016 Accepted 1 July 2016 Available online 02 July 2016 Keywords: Capacitive deionization Carbon electrode Hydrogen peroxide Faradaic reaction
a b s t r a c t Capacitive deionization (CDI) is a process that uses the potential difference developed between two porous electrodes to remove dissolved salts from water known as desalination. In a typical CDI cell consisting of carbon electrodes, a non-Faradaic reaction usually controls the desalination process. However, undesirable Faradaic reactions have been reported which can lower the energy efficiency leading to unexpected side reactions. Among the various side reactions discussed so far, hydrogen peroxide (H2O2) generation has been suggested to occur on the cathode due to the pH fluctuation. The H2O2 generation was recently observed in a batchmode CDI operation, but little is known about the H2O2 generation in a flow-mode operation and the range of applied potentials to each electrode for producing H2O2. In this study, the generation of H2O2 was quantified during the charging step in a flow-mode CDI operation with supporting electrochemical characterizations on cathode and anode, which constantly produced approximately 0.1 mg of H2O2 for 50 cycles at applied voltage of 1.2 V. The applied potentials to cathode and anode were in the range for H2O2 production and carbon oxidation reactions. It can provide one of explanations to the pH increase in the effluent at the beginning of the CDI operation since oxygen and proton are consumed to produce H2O2 (O2 + 2H+ + 2e− = H2O2). Considering a typical applied voltage for CDI operation (b 1.2 V), the generation of hydrogen peroxide is inevitable on the cathode due to the thermodynamic redox potential under neutral pH condition. The quantification of H2O2 is significant finding in that it contributes to better understanding Faradaic reaction in CDI and providing foundation to further studies about detrimental effects in long-term operation or prevention of biofouling by employing hydrogen peroxide, which is a typical problem in most water treatment processes. © 2016 Published by Elsevier B.V.
1. Introduction Capacitive deionization (CDI) has received much attention as a desalination process lately [1–6]. CDI desalinates brackish water when passed between two porous electrodes, mostly made of activated carbons, by applying an electrical potential which is followed the formation of a double layer. CDI consists of two main processes, charging and discharging. In the charging process, ions are attracted and stored in the electrodes, and consequently, the solution passing between the electrodes can be desalinated. In the discharge process, the stored ions are released by reversing the potential of each electrode, and therefore the electrodes can be regenerated. CDI is an electrochemically driven process in which a non-Faradaic reaction intended process controls the desalination process. Therefore, the appropriate operation is required to achieve efficient desalination without any undesirable energy loss.
⁎ Corresponding author at: Daehak-dong, Gwankak-gu, Seoul 151-742, Republic of Korea. E-mail address: [email protected] (J. Yoon).
http://dx.doi.org/10.1016/j.jelechem.2016.07.001 1572-6657/© 2016 Published by Elsevier B.V.
To avoid various Faradaic reactions for an energy-efficient process, the applied potential has been typically limited to 1.2 V [7–10] because the electrolysis of water is negligible in this potential range. However, unexpected side reactions have been reported in an early CDI study [7] even when the applied potential was kept below 1.2 V. Incomplete decay of the current was proposed as evidence for parasitic electrochemical reactions together with current leakage. Afterwards, several Faradaic reactions were postulated by altering the applied potential and following the pH changes [11]. Reduction of dissolved oxygen (pH increase) and oxidation of the carbon electrodes (pH decrease) were suggested as possible explanations for the pH changes during the process. Another study revealed that dissolved oxygen could be involved in the side reactions by observing the cumulative charge and pH changes with and without nitrogen purging [12]. Changes of pH was also examined using a 3-electrode system by varying applied potentials [13]. Increasing the applied voltage above 1.2 V caused more side reactions, which decreased the effluent pH below 5 and induced discrepancy between effluent conductivities and ion concentrations [14]. So far, evidence for unexpected side reactions has been provided using various experimental approaches such as nitrogen sparging, pH changes, and potential adjustments. Previous studies have reported
102
T. Kim et al. / Journal of Electroanalytical Chemistry 776 (2016) 101–104
the oxidation of the carbon in the anode by using X-ray photoelectron spectroscopy (XPS) analysis and observing the change in potential of zero charge [12,13,15]. On the other hand, Faradaic reactions in the cathode have been poorly investigated; several electrochemical reactions related to the reduction of dissolved oxygen were suggested to explain the effluent pH fluctuations [11–13]. Recently, Faradaic reactions in a batch-mode CDI was examined, which reported produced H2O2 and decreased dissolved oxygen concentrations [16]. However, such measurements were performed in a batch-mode CDI operation and applied electrode potentials for driving Faradaic reactions remain unclear. Therefore, a suitable electrochemical analysis supporting the reactions that occur on the cathode could provide a better understanding of the CDI process and its mechanism. Herein, we investigated the generation of H2O2 as one of postulated Faradaic reactions in flow-mode CDI operation. As evidence of oxygen reduction, hydrogen peroxide was specifically detected together with pH changes in the effluent. In addition, the applied potential to each electrode was separately measured using a reference electrode, and current profiles in the corresponding potential region were observed to examine possible Faradaic reactions. 2. Materials and methods 2.1. Capacitive deionization experiments CDI experiments were conducted in a custom-built polycarbonate CDI module, in which 50-mm-diameter electrodes can be installed. Two pieces of activated carbon fibers (CH900-20, BET specific surface area = 2183 m2 g−1, pore volume = 0.89 cm3 g− 1, thickness = 500 μm, Kuraray Chemical Co., LTD, Japan) with filter paper (Whatman, USA) in between were installed inside the module. A feed solution (10 mM NaCl) was supplied using a peristaltic pump at a flow rate of 10 ml/min. The solution was injected through a hole (4 mm in diameter) located in the center and then spread radially outside. The outlet conductivity and pH were continuously measured with a conductivity meter (3574-10C, HORIBA, Japan) and a pH meter (9618-10D, HORIBA, Japan). The solution was discarded after measuring pH instead of being recycled (single-pass mode). A constant voltage (1.2 V) was applied to the CDI module for 6 min using a cycler (WBSC3000, WonATech, Republic of Korea) and short-circuited for 6 min. 2.2. Determination of hydrogen peroxide The concentration of hydrogen peroxide was determined with the 2,9-dimethyl-1,10-phenanthroline (DMP) method [17]. During CDI operation, 1 ml of the effluent was collected every 2 min followed by mixing with 1 ml of DMP solution (1 g of neocuproine (Aldrich) dissolved in 100 ml of ethanol), 1 ml of 0.01 M copper (II) sulphate (Aldrich), and 7 ml of deionized water. Then, the absorbance of the mixed solution was measured at 454 nm (Agilent Technologies, USA). The concentration of H2O2 was calculated with the following equation: ½H2 O2 ¼ Asample ‐Ablank =1500 where [H2O2] is the concentration of the hydrogen peroxide (M); Asample is the absorbance of the sample, and Ablank is the absorbance of the blank. The blank was taken from the effluent before the CDI experiment was started. For a long-term experiment, the effluent during the charging step (6 min) was collected (60 ml) in a glass bottle, and 1 ml of the collected sample was used to determine the concentration of H2O2. 2.3. Electrochemical characterizations Electrochemical characterizations were carried out in a custom-built electrochemical cell consisting of a polytetrafluoroethylene (PTFE) housing and two graphite current collectors. Two pieces of round-
shaped activated carbon fibers (18 mm in diameter) were assembled with a separator (cellulose nitrate, Advanced Microdevices, India). After the installation, the cell was filled with an electrolyte (1 M NaCl) and tightly pressed. First, the cell was charged and discharged at a constant current (2.5 mA cm−2) within 1.2 V using a cycler, which is the typical working voltage for CDI operations. Meanwhile, the potential of each electrode was separately measured using a reference electrode (Ag/AgCl in a saturated KCl solution). Cyclic voltammetry (CV) was performed for each electrode; the potential range (− 0.7 to − 0.1 V and −0.1 to 0.5 V vs. reference electrode for the cathode and the anode, respectively) was determined based on the galvanostatic charge/discharge test. The current profile was recorded at a scan rate of 2 mV s−1 using a potentiostat (PARSTAT 2273, Princeton Applied Research, USA). 3. Results and discussion Fig. 1a shows the representative conductivity and pH profiles during CDI operation. The conductivity profile decreased and increased upon charging (1.2 V, 0–6 min) and discharging (0 V, 6–12 min) which represents the typical behavior of desalination and regeneration [2,9,18]. Together with the conductivity, the pH was simultaneously recorded during the operation. The pH increased up to 11 during the charging step, indicating either hydroxide production or proton consumption. Note that such a high pH level could contribute to an increase in conductivity, and thus the effluent conductivity was significantly affected by this pH effect (Fig. S1) [14]. The increase in pH is quite different compared to the previously reported pH response during CDI operation, in which a pH decrease was observed when charging [12]. However, the charging step of the previous study was much longer (240 min vs. 6 min) when obtaining the pH profile, and a sharp peak was also observed prior to the decrease. In addition, other studies reported an increase in pH at the beginning of the charging process [11,14,16], which is relevant to what we observed in this study. Obviously, the increase in pH suggests that unexpected Faradaic reactions did occur under the CDI working voltage (1.2 V), even though it is not expected to generate any other side reactions including water electrolysis. Among the various possible electrochemical reactions, H2O2 generation by oxygen reduction has been proposed in CDI [11], which is a wellknown reaction in other fields using a carbon electrode as a cathode [19, 20]. In order to confirm H2O2 generation during the charging step of the CDI operation, the amount of H2O2 in the effluent solution was measured by applying different cell voltages. As can be seen in Fig. 1b, the higher applied voltage produced more H2O2, for which 0.094 ± 0.015 mg of H2O2 was produced after 6 min of charging under the typical CDI operating voltage (1.2 V). When considering the charge efficiency loss due to this Faradaic reaction, the amount of charge consumed for H2O2 generation appeared to be insignificant (below approximately 5%, data not shown). Further increasing the voltage to 1.6 V produced a slightly more increased H2O2 of 0.097 ± 0.005 mg, which is relevant result to the previous study reporting rather decreased H2O2 generation at an elevated applied voltage (1.5 V) in a batch-mode CDI operation [16]. At an applied voltage of 0.6 V, a considerable amount of hydrogen peroxide (0.047 ± 0.009 mg) was still generated, indicating that this reaction is inevitable when using a CDI cell consisting of carbon electrodes. An insignificant amount of H2O2 was detected when the applied voltage was 0.2 V. Fig. 1c shows the long-term cycle performance of the conductivity change and the concentration of the hydrogen peroxide generated during the charging steps. The CDI performance gradually deteriorated up to 10 cycles and leveled off for the next 40 cycles. In particular, the conductivity profile for the first charging step was quite different from the subsequent cycles (higher than the influent conductivity at the end of the charging step), which might be due to the repulsion of impurity ions or the abrupt increase in ionic species generated by Faradaic reactions (e.g., OH−). The degradation could be attributed to the oxidation
T. Kim et al. / Journal of Electroanalytical Chemistry 776 (2016) 101–104
103
Fig. 1. (a) Representative effluent conductivity (circles) and pH (triangles) profiles while charging (1.2 V for 6 min) and discharging (0 V for 6 min) the CDI cell, (b) cumulative concentration of hydrogen peroxide generated in different applied voltages (0.2, 0.6, 1.2, and 1.6 V), and (c) long-term cycle performance of the effluent conductivity profile (solid line) and hydrogen peroxide generation (squares).
of the anode, leading to carbon oxidation followed by capacity reduction (also see Fig. S2) [12,13,15]. Interestingly, H2O2 generation also showed a similar trend; its generation was the highest at the beginning and gradually decreased. After 10 cycles, the generation of hydrogen peroxide was kept constant at the amount of approximately 0.1 mg. To further investigate H2O2 generation in CDI operation, an electrochemical analysis was performed. Fig. 2a shows the potential applied to each electrode which was separately measured using a reference electrode while applying a cell voltage from 0 to 1.2 V. The potential ranges were − 0.1 to 0.5 V for the anode and − 0.1 to − 0.7 V for the cathode (coulombic efficiency = 79%). Note that applied potentials for the anode and cathode were similar regardless of electrolyte concentrations (see Fig. S3). Both profiles were non-linear, which is indicative of Faradaic reactions as opposed to a linear profile for non-Faradaic reactions. In the cell voltage profile, the line is no longer linear above 0.4 V; therefore, the Faradaic reaction began around this applied voltage. This is in good agreement with the H2O2 detection (Fig. 1b) even at the low applied voltage (0.6 V). Based on the measured potential range of each electrode, CV was performed separately with the anode and cathode (Fig. 2b). In the cathode CV profile, a cathodic current from − 0.4 to − 0.7 V was observed indicating Faradaic reactions other than double layer charging. Three cathodic reactions could occur in this region at pH 7 as follows [21]: O2 þ 2Hþ þ 2e− →H2 O2 ; E0 ¼ 0:087Vvs:Ag=AgClKClðsat0 dÞ
ð1Þ
2Hþ þ 2e− →H2 ; E0 ¼ −0:61Vvs:Ag=AgClKClðsat0 dÞ
ð2Þ
þ
−
0
H2 O2 þ 2H þ 2e →2H2 O; E ¼ 1:153Vvs:Ag=AgClKClðsat dÞ 0
ð3Þ
Generation of H2O2 on a carbon electrode is a well-known electrochemical process through the reduction of oxygen [19–24]. Therefore,
Fig. 2. (a) Representative voltage profile of the cell voltage and potential of the anode and cathode. The cell was charged and discharged at a constant current (±2.5 mA cm−2) in the range of 0 and 1.2 V while open-circuit electrode potentials of each electrode were recorded using a reference electrode. Dashed lines indicate redox potentials (V vs. Ag/ AgCl) for O2 and H2 evolutions at pH 7. (b) Cyclic voltammograms of the cathode (−0.7 to −0.1 V) and anode (−0.1 to 0.5 V) at a scan rate of 2 mV s−1.
104
T. Kim et al. / Journal of Electroanalytical Chemistry 776 (2016) 101–104
the cathodic current can be mainly explained by reaction 1. Proton reduction (reaction 2) and H2O2 reduction (reaction 3) are possible reactions in this potential region. However, proton reduction (reaction 2) would be insignificant because oxygen reduction (reaction 1) first consumes the proton as observed in the pH change (Fig. 1a). In the positive potential region (−0.1 to 0.5 V), we also observed a Faradaic current indicating the oxidation of the carbon [12]. This reaction would be coupled with oxygen reduction in the cathode. The abrupt pH increase during the charging step (Fig. 1a) can be explained in relation to the Faradaic reaction described above. According to reaction 1, protons (2 mol) are consumed to produce the hydrogen peroxide (1 mol). The concentration of the hydrogen peroxide under 1.2 V was approximately 0.05 mM such that 0.1 mM of protons could be consumed in the cathode. If we assume that the cathode reaction explained above is the major Faradaic reaction at the beginning, the estimated resulting pH based on the proton consumption (0.1 mM) is 10, which is in good agreement with the pH of the effluent (Fig. 1a). Therefore, the pH change of the initial charging step can be partly explained by H2O2 generation. However, the reaction related to carbon oxidation and its impact on ionic species remain unclear; therefore, further studies are required on the energy efficiency, the use of H2O2 as an oxidizing agent, and the detrimental effects of H2O2 on CDI performance. 4. Conclusions In this study, we demonstrated the generation of H2O2 as a result of the Faradaic reaction, which produced approximately 0.1 mg of H2O2 for 50 cycles during the charging step. The electrochemical analysis confirmed that H2O2 generation is inevitable on the cathode based on its thermodynamic redox and measured applied potentials. The initial increase in effluent pH could be partly explained by H2O2 generation because this reaction consumes both oxygens and protons. Because this study revealed that H2O2 is produced in a flow-mode CDI operation, H2O2 could be used as an oxidizing agent to overcome possible biofouling issues in CDI. Our finding could also lead to fundamental studies to elucidate unproven phenomena such as pH fluctuation and performance decay. Acknowledgements This research is supported by Korea Ministry of Environment as “Converging technology project (2014001640002)” and by a grant (code 16IFIP-B065893-04) from Industrial Facilities & Infrastructure Research Program funded by Ministry of Land, Infrastructure and Transport of Korean. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jelechem.2016.07.001. References [1] M. Suss, S. Porada, X. Sun, P. Biesheuvel, J. Yoon, V. Presser, Water desalination via capacitive deionization: what is it and what can we expect from it? Energy Environ. Sci. 8 (8) (2015) 2296–2319.
[2] S. Porada, R. Zhao, A. Van Der Wal, V. Presser, P. Biesheuvel, Review on the science and technology of water desalination by capacitive deionization, Prog. Mater. Sci. 58 (8) (2013) 1388–1442. [3] Z.Y. Yang, L.J. Jin, G.Q. Lu, Q.Q. Xiao, Y.X. Zhang, L. Jing, X.X. Zhang, Y.M. Yan, K.N. Sun, Sponge‐templated preparation of high surface area graphene with ultrahigh capacitive deionization performance, Adv. Funct. Mater. 24 (25) (2014) 3917–3925. [4] S.J. Yang, T. Kim, K. Lee, Y.S. Kim, J. Yoon, C.R. Park, Solvent evaporation mediated preparation of hierarchically porous metal organic framework-derived carbon with controllable and accessible large-scale porosity, Carbon 71 (2014) 294–302. [5] R. Zhao, O. Satpradit, H. Rijnaarts, P. Biesheuvel, A. van der Wal, Optimization of salt adsorption rate in membrane capacitive deionization, Water Res. 47 (5) (2013) 1941–1952. [6] S. Porada, P. Biesheuvel, V. Presser, Comment on sponge‐templated preparation of high surface area graphene with ultrahigh capacitive deionization performance, Adv. Funct. Mater. 25 (2) (2015) 179–181. [7] J.C. Farmer, D.V. Fix, G.V. Mack, R.W. Pekala, J.F. Poco, Capacitive deionization of NaCl and NaNO3 solutions with carbon aerogel electrodes, J. Electrochem. Soc. 143 (1) (1996) 159–169. [8] S. Porada, L. Borchardt, M. Oschatz, M. Bryjak, J. Atchison, K. Keesman, S. Kaskel, P. Biesheuvel, V. Presser, Direct prediction of the desalination performance of porous carbon electrodes for capacitive deionization, Energy Environ. Sci. 6 (12) (2013) 3700–3712. [9] T. Kim, J. Yoon, Relationship between capacitance of activated carbon composite electrodes measured at a low electrolyte concentration and their desalination performance in capacitive deionization, J. Electroanal. Chem. 704 (2013) 169–174. [10] H. Yin, S. Zhao, J. Wan, H. Tang, L. Chang, L. He, H. Zhao, Y. Gao, Z. Tang, Three-dimensional graphene/metal oxide nanoparticle hybrids for high-performance capacitive deionization of saline water, Adv. Mater. 25 (43) (2013) 6270–6276. [11] J.-H. Lee, W.-S. Bae, J.-H. Choi, Electrode reactions and adsorption/desorption performance related to the applied potential in a capacitive deionization process, Desalination 258 (1) (2010) 159–163. [12] Y. Bouhadana, E. Avraham, M. Noked, M. Ben-Tzion, A. Soffer, D. Aurbach, Capacitive deionization of NaCl solutions at non-steady-state conditions: inversion functionality of the carbon electrodes, J. Phys. Chem. C 115 (33) (2011) 16567–16573. [13] I. Cohen, E. Avraham, Y. Bouhadana, A. Soffer, D. Aurbach, Long term stability of capacitive de-ionization processes for water desalination: the challenge of positive electrodes corrosion, Electrochim. Acta 106 (2013) 91–100. [14] J.J. Lado, R.E. Pérez-Roa, J.J. Wouters, M.I. Tejedor-Tejedor, M.A. Anderson, Evaluation of operational parameters for a capacitive deionization reactor employing asymmetric electrodes, Sep. Purif. Technol. 133 (2014) 236–245. [15] J.J. Lado, R.E. Pérez-Roa, J.J. Wouters, M.I. Tejedor-Tejedor, C. Federspill, M.A. Anderson, Continuous cycling of an asymmetric capacitive deionization system: an evaluation of the electrode performance and stability, J. Environ. Chem. Eng. 3 (4) (2015) 2358–2367. [16] D. He, C.E. Wong, W. Tang, P. Kovalsky, T.D. Waite, Faradaic reactions in water desalination by batch-mode capacitive deionization, Environ. Sci. Technol. Lett. 3 (5) (2016) 222–226. [17] A.N. Baga, G. Johnson, N.B. Nazhat, R.A. Saadalla-Nazhat, A simple spectrophotometric determination of hydrogen peroxide at low concentrations in aqueous solution, Anal. Chim. Acta 204 (1988) 349–353. [18] S. Porada, L. Weinstein, R. Dash, A. Van Der Wal, M. Bryjak, Y. Gogotsi, P. Biesheuvel, Water desalination using capacitive deionization with microporous carbon electrodes, ACS Appl. Mater. Interfaces 4 (3) (2012) 1194–1199. [19] R.A. Rozendal, E. Leone, J. Keller, K. Rabaey, Efficient hydrogen peroxide generation from organic matter in a bioelectrochemical system, Electrochem. Commun. 11 (9) (2009) 1752–1755. [20] T. Matsue, M. Fujihira, T. Osa, Oxidation of alkylbenzenes by electrogenerated hydroxyl radical, J. Electrochem. Soc. 128 (12) (1981) 2565–2569. [21] Z. Qiang, J.-H. Chang, C.-P. Huang, Electrochemical generation of hydrogen peroxide from dissolved oxygen in acidic solutions, Water Res. 36 (1) (2002) 85–94. [22] A. Da Pozzo, L. Di Palma, C. Merli, E. Petrucci, An experimental comparison of a graphite electrode and a gas diffusion electrode for the cathodic production of hydrogen peroxide, J. Appl. Electrochem. 35 (4) (2005) 413–419. [23] P. Tatapudi, J.M. Fenton, Synthesis of hydrogen peroxide in a proton exchange membrane electrochemical reactor, J. Electrochem. Soc. 140 (4) (1993) L55–L57. [24] J. Forti, R. Rocha, M. Lanza, R. Bertazzoli, Electrochemical synthesis of hydrogen peroxide on oxygen-fed graphite/PTFE electrodes modified by 2-ethylanthraquinone, J. Electroanal. Chem. 601 (1) (2007) 63–67.
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
Short communication
Hydrogen peroxide generation in flow-mode capacitive deionization Taeyoung Kim a, Jihyun Yu a, Choonsoo Kim a, Jeyong Yoon a,b,⁎ a School of Chemical and Biological Engineering, College of Engineering, Institute of Chemical Processes(ICP), Seoul National University (SNU), Gwanak-gu, Daehak-dong, Seoul 151-742, Republic of Korea b Asian Institute for Energy, Environment & Sustainability (AIEES), Seoul National University (SNU), Gwanak-gu, Daehak-dong, Seoul 151-742, Republic of Korea
a r t i c l e
i n f o
Article history: Received 28 March 2016 Received in revised form 28 June 2016 Accepted 1 July 2016 Available online 02 July 2016 Keywords: Capacitive deionization Carbon electrode Hydrogen peroxide Faradaic reaction
a b s t r a c t Capacitive deionization (CDI) is a process that uses the potential difference developed between two porous electrodes to remove dissolved salts from water known as desalination. In a typical CDI cell consisting of carbon electrodes, a non-Faradaic reaction usually controls the desalination process. However, undesirable Faradaic reactions have been reported which can lower the energy efficiency leading to unexpected side reactions. Among the various side reactions discussed so far, hydrogen peroxide (H2O2) generation has been suggested to occur on the cathode due to the pH fluctuation. The H2O2 generation was recently observed in a batchmode CDI operation, but little is known about the H2O2 generation in a flow-mode operation and the range of applied potentials to each electrode for producing H2O2. In this study, the generation of H2O2 was quantified during the charging step in a flow-mode CDI operation with supporting electrochemical characterizations on cathode and anode, which constantly produced approximately 0.1 mg of H2O2 for 50 cycles at applied voltage of 1.2 V. The applied potentials to cathode and anode were in the range for H2O2 production and carbon oxidation reactions. It can provide one of explanations to the pH increase in the effluent at the beginning of the CDI operation since oxygen and proton are consumed to produce H2O2 (O2 + 2H+ + 2e− = H2O2). Considering a typical applied voltage for CDI operation (b 1.2 V), the generation of hydrogen peroxide is inevitable on the cathode due to the thermodynamic redox potential under neutral pH condition. The quantification of H2O2 is significant finding in that it contributes to better understanding Faradaic reaction in CDI and providing foundation to further studies about detrimental effects in long-term operation or prevention of biofouling by employing hydrogen peroxide, which is a typical problem in most water treatment processes. © 2016 Published by Elsevier B.V.
1. Introduction Capacitive deionization (CDI) has received much attention as a desalination process lately [1–6]. CDI desalinates brackish water when passed between two porous electrodes, mostly made of activated carbons, by applying an electrical potential which is followed the formation of a double layer. CDI consists of two main processes, charging and discharging. In the charging process, ions are attracted and stored in the electrodes, and consequently, the solution passing between the electrodes can be desalinated. In the discharge process, the stored ions are released by reversing the potential of each electrode, and therefore the electrodes can be regenerated. CDI is an electrochemically driven process in which a non-Faradaic reaction intended process controls the desalination process. Therefore, the appropriate operation is required to achieve efficient desalination without any undesirable energy loss.
⁎ Corresponding author at: Daehak-dong, Gwankak-gu, Seoul 151-742, Republic of Korea. E-mail address: [email protected] (J. Yoon).
http://dx.doi.org/10.1016/j.jelechem.2016.07.001 1572-6657/© 2016 Published by Elsevier B.V.
To avoid various Faradaic reactions for an energy-efficient process, the applied potential has been typically limited to 1.2 V [7–10] because the electrolysis of water is negligible in this potential range. However, unexpected side reactions have been reported in an early CDI study [7] even when the applied potential was kept below 1.2 V. Incomplete decay of the current was proposed as evidence for parasitic electrochemical reactions together with current leakage. Afterwards, several Faradaic reactions were postulated by altering the applied potential and following the pH changes [11]. Reduction of dissolved oxygen (pH increase) and oxidation of the carbon electrodes (pH decrease) were suggested as possible explanations for the pH changes during the process. Another study revealed that dissolved oxygen could be involved in the side reactions by observing the cumulative charge and pH changes with and without nitrogen purging [12]. Changes of pH was also examined using a 3-electrode system by varying applied potentials [13]. Increasing the applied voltage above 1.2 V caused more side reactions, which decreased the effluent pH below 5 and induced discrepancy between effluent conductivities and ion concentrations [14]. So far, evidence for unexpected side reactions has been provided using various experimental approaches such as nitrogen sparging, pH changes, and potential adjustments. Previous studies have reported
102
T. Kim et al. / Journal of Electroanalytical Chemistry 776 (2016) 101–104
the oxidation of the carbon in the anode by using X-ray photoelectron spectroscopy (XPS) analysis and observing the change in potential of zero charge [12,13,15]. On the other hand, Faradaic reactions in the cathode have been poorly investigated; several electrochemical reactions related to the reduction of dissolved oxygen were suggested to explain the effluent pH fluctuations [11–13]. Recently, Faradaic reactions in a batch-mode CDI was examined, which reported produced H2O2 and decreased dissolved oxygen concentrations [16]. However, such measurements were performed in a batch-mode CDI operation and applied electrode potentials for driving Faradaic reactions remain unclear. Therefore, a suitable electrochemical analysis supporting the reactions that occur on the cathode could provide a better understanding of the CDI process and its mechanism. Herein, we investigated the generation of H2O2 as one of postulated Faradaic reactions in flow-mode CDI operation. As evidence of oxygen reduction, hydrogen peroxide was specifically detected together with pH changes in the effluent. In addition, the applied potential to each electrode was separately measured using a reference electrode, and current profiles in the corresponding potential region were observed to examine possible Faradaic reactions. 2. Materials and methods 2.1. Capacitive deionization experiments CDI experiments were conducted in a custom-built polycarbonate CDI module, in which 50-mm-diameter electrodes can be installed. Two pieces of activated carbon fibers (CH900-20, BET specific surface area = 2183 m2 g−1, pore volume = 0.89 cm3 g− 1, thickness = 500 μm, Kuraray Chemical Co., LTD, Japan) with filter paper (Whatman, USA) in between were installed inside the module. A feed solution (10 mM NaCl) was supplied using a peristaltic pump at a flow rate of 10 ml/min. The solution was injected through a hole (4 mm in diameter) located in the center and then spread radially outside. The outlet conductivity and pH were continuously measured with a conductivity meter (3574-10C, HORIBA, Japan) and a pH meter (9618-10D, HORIBA, Japan). The solution was discarded after measuring pH instead of being recycled (single-pass mode). A constant voltage (1.2 V) was applied to the CDI module for 6 min using a cycler (WBSC3000, WonATech, Republic of Korea) and short-circuited for 6 min. 2.2. Determination of hydrogen peroxide The concentration of hydrogen peroxide was determined with the 2,9-dimethyl-1,10-phenanthroline (DMP) method [17]. During CDI operation, 1 ml of the effluent was collected every 2 min followed by mixing with 1 ml of DMP solution (1 g of neocuproine (Aldrich) dissolved in 100 ml of ethanol), 1 ml of 0.01 M copper (II) sulphate (Aldrich), and 7 ml of deionized water. Then, the absorbance of the mixed solution was measured at 454 nm (Agilent Technologies, USA). The concentration of H2O2 was calculated with the following equation: ½H2 O2 ¼ Asample ‐Ablank =1500 where [H2O2] is the concentration of the hydrogen peroxide (M); Asample is the absorbance of the sample, and Ablank is the absorbance of the blank. The blank was taken from the effluent before the CDI experiment was started. For a long-term experiment, the effluent during the charging step (6 min) was collected (60 ml) in a glass bottle, and 1 ml of the collected sample was used to determine the concentration of H2O2. 2.3. Electrochemical characterizations Electrochemical characterizations were carried out in a custom-built electrochemical cell consisting of a polytetrafluoroethylene (PTFE) housing and two graphite current collectors. Two pieces of round-
shaped activated carbon fibers (18 mm in diameter) were assembled with a separator (cellulose nitrate, Advanced Microdevices, India). After the installation, the cell was filled with an electrolyte (1 M NaCl) and tightly pressed. First, the cell was charged and discharged at a constant current (2.5 mA cm−2) within 1.2 V using a cycler, which is the typical working voltage for CDI operations. Meanwhile, the potential of each electrode was separately measured using a reference electrode (Ag/AgCl in a saturated KCl solution). Cyclic voltammetry (CV) was performed for each electrode; the potential range (− 0.7 to − 0.1 V and −0.1 to 0.5 V vs. reference electrode for the cathode and the anode, respectively) was determined based on the galvanostatic charge/discharge test. The current profile was recorded at a scan rate of 2 mV s−1 using a potentiostat (PARSTAT 2273, Princeton Applied Research, USA). 3. Results and discussion Fig. 1a shows the representative conductivity and pH profiles during CDI operation. The conductivity profile decreased and increased upon charging (1.2 V, 0–6 min) and discharging (0 V, 6–12 min) which represents the typical behavior of desalination and regeneration [2,9,18]. Together with the conductivity, the pH was simultaneously recorded during the operation. The pH increased up to 11 during the charging step, indicating either hydroxide production or proton consumption. Note that such a high pH level could contribute to an increase in conductivity, and thus the effluent conductivity was significantly affected by this pH effect (Fig. S1) [14]. The increase in pH is quite different compared to the previously reported pH response during CDI operation, in which a pH decrease was observed when charging [12]. However, the charging step of the previous study was much longer (240 min vs. 6 min) when obtaining the pH profile, and a sharp peak was also observed prior to the decrease. In addition, other studies reported an increase in pH at the beginning of the charging process [11,14,16], which is relevant to what we observed in this study. Obviously, the increase in pH suggests that unexpected Faradaic reactions did occur under the CDI working voltage (1.2 V), even though it is not expected to generate any other side reactions including water electrolysis. Among the various possible electrochemical reactions, H2O2 generation by oxygen reduction has been proposed in CDI [11], which is a wellknown reaction in other fields using a carbon electrode as a cathode [19, 20]. In order to confirm H2O2 generation during the charging step of the CDI operation, the amount of H2O2 in the effluent solution was measured by applying different cell voltages. As can be seen in Fig. 1b, the higher applied voltage produced more H2O2, for which 0.094 ± 0.015 mg of H2O2 was produced after 6 min of charging under the typical CDI operating voltage (1.2 V). When considering the charge efficiency loss due to this Faradaic reaction, the amount of charge consumed for H2O2 generation appeared to be insignificant (below approximately 5%, data not shown). Further increasing the voltage to 1.6 V produced a slightly more increased H2O2 of 0.097 ± 0.005 mg, which is relevant result to the previous study reporting rather decreased H2O2 generation at an elevated applied voltage (1.5 V) in a batch-mode CDI operation [16]. At an applied voltage of 0.6 V, a considerable amount of hydrogen peroxide (0.047 ± 0.009 mg) was still generated, indicating that this reaction is inevitable when using a CDI cell consisting of carbon electrodes. An insignificant amount of H2O2 was detected when the applied voltage was 0.2 V. Fig. 1c shows the long-term cycle performance of the conductivity change and the concentration of the hydrogen peroxide generated during the charging steps. The CDI performance gradually deteriorated up to 10 cycles and leveled off for the next 40 cycles. In particular, the conductivity profile for the first charging step was quite different from the subsequent cycles (higher than the influent conductivity at the end of the charging step), which might be due to the repulsion of impurity ions or the abrupt increase in ionic species generated by Faradaic reactions (e.g., OH−). The degradation could be attributed to the oxidation
T. Kim et al. / Journal of Electroanalytical Chemistry 776 (2016) 101–104
103
Fig. 1. (a) Representative effluent conductivity (circles) and pH (triangles) profiles while charging (1.2 V for 6 min) and discharging (0 V for 6 min) the CDI cell, (b) cumulative concentration of hydrogen peroxide generated in different applied voltages (0.2, 0.6, 1.2, and 1.6 V), and (c) long-term cycle performance of the effluent conductivity profile (solid line) and hydrogen peroxide generation (squares).
of the anode, leading to carbon oxidation followed by capacity reduction (also see Fig. S2) [12,13,15]. Interestingly, H2O2 generation also showed a similar trend; its generation was the highest at the beginning and gradually decreased. After 10 cycles, the generation of hydrogen peroxide was kept constant at the amount of approximately 0.1 mg. To further investigate H2O2 generation in CDI operation, an electrochemical analysis was performed. Fig. 2a shows the potential applied to each electrode which was separately measured using a reference electrode while applying a cell voltage from 0 to 1.2 V. The potential ranges were − 0.1 to 0.5 V for the anode and − 0.1 to − 0.7 V for the cathode (coulombic efficiency = 79%). Note that applied potentials for the anode and cathode were similar regardless of electrolyte concentrations (see Fig. S3). Both profiles were non-linear, which is indicative of Faradaic reactions as opposed to a linear profile for non-Faradaic reactions. In the cell voltage profile, the line is no longer linear above 0.4 V; therefore, the Faradaic reaction began around this applied voltage. This is in good agreement with the H2O2 detection (Fig. 1b) even at the low applied voltage (0.6 V). Based on the measured potential range of each electrode, CV was performed separately with the anode and cathode (Fig. 2b). In the cathode CV profile, a cathodic current from − 0.4 to − 0.7 V was observed indicating Faradaic reactions other than double layer charging. Three cathodic reactions could occur in this region at pH 7 as follows [21]: O2 þ 2Hþ þ 2e− →H2 O2 ; E0 ¼ 0:087Vvs:Ag=AgClKClðsat0 dÞ
ð1Þ
2Hþ þ 2e− →H2 ; E0 ¼ −0:61Vvs:Ag=AgClKClðsat0 dÞ
ð2Þ
þ
−
0
H2 O2 þ 2H þ 2e →2H2 O; E ¼ 1:153Vvs:Ag=AgClKClðsat dÞ 0
ð3Þ
Generation of H2O2 on a carbon electrode is a well-known electrochemical process through the reduction of oxygen [19–24]. Therefore,
Fig. 2. (a) Representative voltage profile of the cell voltage and potential of the anode and cathode. The cell was charged and discharged at a constant current (±2.5 mA cm−2) in the range of 0 and 1.2 V while open-circuit electrode potentials of each electrode were recorded using a reference electrode. Dashed lines indicate redox potentials (V vs. Ag/ AgCl) for O2 and H2 evolutions at pH 7. (b) Cyclic voltammograms of the cathode (−0.7 to −0.1 V) and anode (−0.1 to 0.5 V) at a scan rate of 2 mV s−1.
104
T. Kim et al. / Journal of Electroanalytical Chemistry 776 (2016) 101–104
the cathodic current can be mainly explained by reaction 1. Proton reduction (reaction 2) and H2O2 reduction (reaction 3) are possible reactions in this potential region. However, proton reduction (reaction 2) would be insignificant because oxygen reduction (reaction 1) first consumes the proton as observed in the pH change (Fig. 1a). In the positive potential region (−0.1 to 0.5 V), we also observed a Faradaic current indicating the oxidation of the carbon [12]. This reaction would be coupled with oxygen reduction in the cathode. The abrupt pH increase during the charging step (Fig. 1a) can be explained in relation to the Faradaic reaction described above. According to reaction 1, protons (2 mol) are consumed to produce the hydrogen peroxide (1 mol). The concentration of the hydrogen peroxide under 1.2 V was approximately 0.05 mM such that 0.1 mM of protons could be consumed in the cathode. If we assume that the cathode reaction explained above is the major Faradaic reaction at the beginning, the estimated resulting pH based on the proton consumption (0.1 mM) is 10, which is in good agreement with the pH of the effluent (Fig. 1a). Therefore, the pH change of the initial charging step can be partly explained by H2O2 generation. However, the reaction related to carbon oxidation and its impact on ionic species remain unclear; therefore, further studies are required on the energy efficiency, the use of H2O2 as an oxidizing agent, and the detrimental effects of H2O2 on CDI performance. 4. Conclusions In this study, we demonstrated the generation of H2O2 as a result of the Faradaic reaction, which produced approximately 0.1 mg of H2O2 for 50 cycles during the charging step. The electrochemical analysis confirmed that H2O2 generation is inevitable on the cathode based on its thermodynamic redox and measured applied potentials. The initial increase in effluent pH could be partly explained by H2O2 generation because this reaction consumes both oxygens and protons. Because this study revealed that H2O2 is produced in a flow-mode CDI operation, H2O2 could be used as an oxidizing agent to overcome possible biofouling issues in CDI. Our finding could also lead to fundamental studies to elucidate unproven phenomena such as pH fluctuation and performance decay. Acknowledgements This research is supported by Korea Ministry of Environment as “Converging technology project (2014001640002)” and by a grant (code 16IFIP-B065893-04) from Industrial Facilities & Infrastructure Research Program funded by Ministry of Land, Infrastructure and Transport of Korean. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jelechem.2016.07.001. References [1] M. Suss, S. Porada, X. Sun, P. Biesheuvel, J. Yoon, V. Presser, Water desalination via capacitive deionization: what is it and what can we expect from it? Energy Environ. Sci. 8 (8) (2015) 2296–2319.
[2] S. Porada, R. Zhao, A. Van Der Wal, V. Presser, P. Biesheuvel, Review on the science and technology of water desalination by capacitive deionization, Prog. Mater. Sci. 58 (8) (2013) 1388–1442. [3] Z.Y. Yang, L.J. Jin, G.Q. Lu, Q.Q. Xiao, Y.X. Zhang, L. Jing, X.X. Zhang, Y.M. Yan, K.N. Sun, Sponge‐templated preparation of high surface area graphene with ultrahigh capacitive deionization performance, Adv. Funct. Mater. 24 (25) (2014) 3917–3925. [4] S.J. Yang, T. Kim, K. Lee, Y.S. Kim, J. Yoon, C.R. Park, Solvent evaporation mediated preparation of hierarchically porous metal organic framework-derived carbon with controllable and accessible large-scale porosity, Carbon 71 (2014) 294–302. [5] R. Zhao, O. Satpradit, H. Rijnaarts, P. Biesheuvel, A. van der Wal, Optimization of salt adsorption rate in membrane capacitive deionization, Water Res. 47 (5) (2013) 1941–1952. [6] S. Porada, P. Biesheuvel, V. Presser, Comment on sponge‐templated preparation of high surface area graphene with ultrahigh capacitive deionization performance, Adv. Funct. Mater. 25 (2) (2015) 179–181. [7] J.C. Farmer, D.V. Fix, G.V. Mack, R.W. Pekala, J.F. Poco, Capacitive deionization of NaCl and NaNO3 solutions with carbon aerogel electrodes, J. Electrochem. Soc. 143 (1) (1996) 159–169. [8] S. Porada, L. Borchardt, M. Oschatz, M. Bryjak, J. Atchison, K. Keesman, S. Kaskel, P. Biesheuvel, V. Presser, Direct prediction of the desalination performance of porous carbon electrodes for capacitive deionization, Energy Environ. Sci. 6 (12) (2013) 3700–3712. [9] T. Kim, J. Yoon, Relationship between capacitance of activated carbon composite electrodes measured at a low electrolyte concentration and their desalination performance in capacitive deionization, J. Electroanal. Chem. 704 (2013) 169–174. [10] H. Yin, S. Zhao, J. Wan, H. Tang, L. Chang, L. He, H. Zhao, Y. Gao, Z. Tang, Three-dimensional graphene/metal oxide nanoparticle hybrids for high-performance capacitive deionization of saline water, Adv. Mater. 25 (43) (2013) 6270–6276. [11] J.-H. Lee, W.-S. Bae, J.-H. Choi, Electrode reactions and adsorption/desorption performance related to the applied potential in a capacitive deionization process, Desalination 258 (1) (2010) 159–163. [12] Y. Bouhadana, E. Avraham, M. Noked, M. Ben-Tzion, A. Soffer, D. Aurbach, Capacitive deionization of NaCl solutions at non-steady-state conditions: inversion functionality of the carbon electrodes, J. Phys. Chem. C 115 (33) (2011) 16567–16573. [13] I. Cohen, E. Avraham, Y. Bouhadana, A. Soffer, D. Aurbach, Long term stability of capacitive de-ionization processes for water desalination: the challenge of positive electrodes corrosion, Electrochim. Acta 106 (2013) 91–100. [14] J.J. Lado, R.E. Pérez-Roa, J.J. Wouters, M.I. Tejedor-Tejedor, M.A. Anderson, Evaluation of operational parameters for a capacitive deionization reactor employing asymmetric electrodes, Sep. Purif. Technol. 133 (2014) 236–245. [15] J.J. Lado, R.E. Pérez-Roa, J.J. Wouters, M.I. Tejedor-Tejedor, C. Federspill, M.A. Anderson, Continuous cycling of an asymmetric capacitive deionization system: an evaluation of the electrode performance and stability, J. Environ. Chem. Eng. 3 (4) (2015) 2358–2367. [16] D. He, C.E. Wong, W. Tang, P. Kovalsky, T.D. Waite, Faradaic reactions in water desalination by batch-mode capacitive deionization, Environ. Sci. Technol. Lett. 3 (5) (2016) 222–226. [17] A.N. Baga, G. Johnson, N.B. Nazhat, R.A. Saadalla-Nazhat, A simple spectrophotometric determination of hydrogen peroxide at low concentrations in aqueous solution, Anal. Chim. Acta 204 (1988) 349–353. [18] S. Porada, L. Weinstein, R. Dash, A. Van Der Wal, M. Bryjak, Y. Gogotsi, P. Biesheuvel, Water desalination using capacitive deionization with microporous carbon electrodes, ACS Appl. Mater. Interfaces 4 (3) (2012) 1194–1199. [19] R.A. Rozendal, E. Leone, J. Keller, K. Rabaey, Efficient hydrogen peroxide generation from organic matter in a bioelectrochemical system, Electrochem. Commun. 11 (9) (2009) 1752–1755. [20] T. Matsue, M. Fujihira, T. Osa, Oxidation of alkylbenzenes by electrogenerated hydroxyl radical, J. Electrochem. Soc. 128 (12) (1981) 2565–2569. [21] Z. Qiang, J.-H. Chang, C.-P. Huang, Electrochemical generation of hydrogen peroxide from dissolved oxygen in acidic solutions, Water Res. 36 (1) (2002) 85–94. [22] A. Da Pozzo, L. Di Palma, C. Merli, E. Petrucci, An experimental comparison of a graphite electrode and a gas diffusion electrode for the cathodic production of hydrogen peroxide, J. Appl. Electrochem. 35 (4) (2005) 413–419. [23] P. Tatapudi, J.M. Fenton, Synthesis of hydrogen peroxide in a proton exchange membrane electrochemical reactor, J. Electrochem. Soc. 140 (4) (1993) L55–L57. [24] J. Forti, R. Rocha, M. Lanza, R. Bertazzoli, Electrochemical synthesis of hydrogen peroxide on oxygen-fed graphite/PTFE electrodes modified by 2-ethylanthraquinone, J. Electroanal. Chem. 601 (1) (2007) 63–67.