Letter Cite This: J. Phys. Chem. Lett. 2019, 10, 3296−3300
pubs.acs.org/JPCL
Potassium Metal as Reliable Reference Electrodes of Nonaqueous Potassium Cells Tomooki Hosaka,† Shotaro Muratsubaki,† Kei Kubota,†,‡ Hiroo Onuma,† and Shinichi Komaba*,†,‡ †
Department of Applied Chemistry, Tokyo University of Science, Shinjuku, Tokyo 162-8601, Japan Elements Strategy Initiative for Catalysts and Batteries (ESICB), Kyoto University, 1-30 Goryo-Ohara, Nishikyo-ku, Kyoto 615-8245, Japan
‡
Downloaded via NOTTINGHAM TRENT UNIV on July 29, 2019 at 12:29:09 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: Potassium metal electrochemical cells are widely utilized to examine potassium insertion materials for nonaqueous potassium-ion batteries. However, large polarization during K plating−stripping and unstable rest potential are found at the potassium electrodes, which leads to an underestimation of the electrochemical performance of insertion materials. In this study, the electrochemical behavior of K-metal electrodes is systematically investigated. Electrolyte salts, solvents, and additives influence the polarization of K metals. Although a highly concentrated electrolyte of 3.9 M KN(SO2F)2/1,2-dimethoxyethane realizes the smallest polarization of 25 mV among all the electrolytes investigated in this study, the polarization of K metals is still larger than those of Li and Na metals. The issue of inaccurate rest potential is solved by pretreating the K electrodes with a plating−stripping process, which is essential in evaluating the intrinsic electrode performance of potassium insertion materials.
P
Moreover, a large deviation of >100 mV from the average value (represented as error bars in Figure 1a) raises a concern regarding the reproducibility (see detailed OCV data in the Supporting Information, Figure S1). Similar phenomena of unstable OCV were previously reported for Mg//Mg and Ca// Ca cells, which deviated by ca. 40 mV from 0 V.7 From the OCV measurement, K metal is not suitable to be used as a reference electrode in KPF6/EC:DEC without any treatments. Figure 1b shows the voltage profiles of the symmetry cells during successive cycles of 10 h of plating−stripping at ±25 μA cm−2. The Li//Li and Na//Na cells showed a small voltage drop of less than 5 and 10 mV, respectively. These values are consistent with those in the literature.2,8 In contrast, the K//K cell showed a significant polarization of 100 mV or larger. In addition, voltage spikes of nearly 1 V were frequently observed in the initial few cycles. During the subsequent cycles, the polarization gradually decreased to 100−200 mV. A stark difference in their electrochemical impedance data was also found before and after the cycles, as shown in Figure 1c−e. Li//Li and Na//Na cells showed two semicircles in the highfrequency region of 0.1−0.9 kHz and middle-frequency region of 0.7−2 Hz (see Figure S2) which can be ascribed to passivation and surface resistive species like alkaline oxide, respectively.9 The total interfacial resistance (Rint) ranges between 100−500 Ω cm2 in Li//Li and Na//Na cells, consistent with a previous report.8 On the other hand, the K//K cell showed a huge semicircle in the frequency region
otassium-ion batteries (KIBs) have recently attracted much attention because of the abundance of potassium resources, lower standard electrode potential of K/K+,1 and weaker Lewis acidity of K+ ion compared to lithium.2 Because these features are advantageous for high-voltage/power operation of KIBs, researchers have shed light on development of new potassium insertion materials.3−5 To investigate their electrochemical potassium insertion, electrochemical cells equipped with metallic potassium as a counter electrode (CE) are widely utilized, similar to Li and Na cells.6 Recently, we reported a polarization issue of potassium plating−stripping process unlike Li and Na metals.2 The undesired polarization results in inaccurate evaluation of the electrochemical properties of the electrode materials when K metal is used as a CE. In this Letter, we investigate the polarization and rest potential of the potassium electrode to confirm its applicability to counter and reference electrodes. Moreover, we find that a galvanostatic pretreatment to refresh the K metal surface is highly effective to ensure an accurate equilibrium potential of the K/ K+ couple, enabling us to demonstrate a reproducible and reliable evaluation of potassium insertion materials in K cells. Prior to evaluation of the polarization behavior of alkalimetal electrodes, the open-circuit voltage (OCV) of symmetric A//A cells (A = Li, Na, and K) was examined using an APF6/ ethylene carbonate (EC):diethyl carbonate (DEC) solution as an electrolyte. The OCV measurements were started after 20 min of the cell fabrication. The Li//Li and Na//Na cells showed quite small OCVs of less than 5 mV, as shown in Figure 1a. In contrast, the K//K cell represented a significantly larger OCV of >500 mV in an average of 10 cells at the beginning and 50 mV even after a 24 h relaxation period. © 2019 American Chemical Society
Received: March 12, 2019 Accepted: May 1, 2019 Published: May 1, 2019 3296
DOI: 10.1021/acs.jpclett.9b00711 J. Phys. Chem. Lett. 2019, 10, 3296−3300
Letter
The Journal of Physical Chemistry Letters
Figure 1. (a) OCV curves of two-electrode Li//Li, Na//Na, and K//K cells filled with 1.0 M LiPF6/EC:DEC, 1.0 M NaPF6/EC:DEC, and 0.8 M KPF6 EC:DEC, respectively. Averaged OCV of 10 cells was plotted for the K//K cell. (b) Voltage profiles of the symmetric cells during continual plating−stripping process. Nyquist plots of (c) Li//Li, (d) Na//Na, and (e) K//K cells before and after 5 and 20 cycles of plating−stripping.
including ca. 8 Hz, resulting in Rint = ca. 15 kΩ cm2 before plating−stripping. In addition, linear plots inclined at 45° corresponding to Warburg impedance appeared in the lowfrequency region, which should be attributed to formation of a resistive surface layer of K+ conduction. Notably, the Warburg impedance disappeared after 5 and 20 cycles, and the Rint also decreased to less than 10 kΩ cm2 after 20 cycles, which agrees with the reduction of polarization after the cycles in Figure 1b. As studied in Li- and Na-metal batteries,8 electrode resistance depends on the surface layer formed by electrolyte decomposition. Therefore, we further examined the polarization behavior of K//K cells with different solvents, salts, and additives. Figure 2a shows voltage profiles of K//K cells filled with different KPF6 solutions: 1 M KPF6/propylene carbonate (PC), 1 M KPF6/EC:PC, 0.8 M KPF6/EC:dimethyl carbonate (DMC), and 0.8 M KPF6/EC:DEC. Similar to the KPF6/ EC:DEC, use of KPF6/EC:PC and KPF6/EC:DMC leads to large polarization of >500 mV at initial plating−stripping cycles, and it converged to approximately 100 mV after several cycles. Oppositely, the KPF6/PC cell showed continuous increase in polarization during cycles, which is possibly due to continuous electrolyte decomposition and deposition of the decomposition products on K metal with simultaneous electrolyte degradation. Indeed, discoloration of KPF6/PC to yellow was observed when a piece of K metal was just soaked in the electrolyte for 14 days (Figure 2d) while no visual change was confirmed in KPF6/EC:DEC (Figure 2e). In contrast to the KPF 6 /PC system, a potassium bis(fluorosulfonyl)amide (KFSA)/PC cell showed large but constant polarization, as shown in Figure 2b. Furthermore, no voltage spikes were observed with all of the KFSA solutions including 3.9 M KFSA/1,2-dimethoxyethane (DME), which is the highly concentrated and promising electrolyte for demonstrating KIBs10 and K-metal batteries.11 Actually, the
Figure 2. Voltage profiles of repeated K plating−stripping in K//K cells filled with (a) 1.0 M KPF6/PC, 1.0 M KPF6/EC:PC, 0.8 M KPF6/EC:DMC, 0.8 M KPF6/EC:DEC; (b) 1.0 M KFSA/PC, 1.0 M KFSA/EC:DEC, 3.9 M KFSA/DME; (c) 0.8 M KPF6/EC:DEC with or without 1 vol % of FEC, DFEC, VC, or ES. Photographs of K metal discs and electrolytes of the (d) KPF6/PC, (e) KPF6/EC:DEC, (f) KFSA/PC, (g) KFSA/EC:DEC, and KPF6/EC:DEC with (h) FEC and (i) ES after 14 d.
3.9 M KFSA/DME cell undergoes the least polarization of ca. 50 mV among all the electrolytes used in this study. The suppressed polarization can be attributed to the lower resistive and durable SEI on the surface.11 Moreover, minor and no 3297
DOI: 10.1021/acs.jpclett.9b00711 J. Phys. Chem. Lett. 2019, 10, 3296−3300
Letter
The Journal of Physical Chemistry Letters
Figure 3. (a) Voltage profile of a K//K cell filled with 0.8 M KPF6/EC:DEC before, during, and after a K plating−stripping cycle. (b) OCV curves of K//K cells filled with 1.0 and 2.0 M KFSA/EC:DEC, 3.9 M KFSA/DME, and 0.8 M KPF6/EC:DEC after a K plating−stripping cycle. (c) Schematic illustrations explaining the change of K-metal surface and Rint before, during, and after the K plating−stripping.
a reference electrode (RE) is poorly reproducible and significantly different from 0 V, as shown in Figure 1a. To solve this issue, an effect of electrochemical pretreatment of K metal is applied to realize the reliable potential of K/K+. Because the reduction of polarization and Rint was successful as described in Figures 1 and 2, the adapted pretreatment is plating−stripping of K metals between CE and RE inside three-electrode cells. Figure 3a displays a voltage curve of a K//K cell with 0.8 M KPF6/EC:DEC. As discussed above, the initial OCV was 0.87 V, far from 0 V. After the plating− stripping, the OCV of the K//K cell is immediately equilibrated to
