Room Temperature Ionic Liquids as Electrolyte Additives for the HER in Alkaline Media

Three 1-ethyl-3-methylimidazolium ([Emim])-based room temperature ionic liquids (RTILs), namely [Emim][Ac] (acetate), [Emim][EtSO 4 ] (ethylsulfate) and [Emim][MeSO 3 ] (methanesulfonate), are tested as additives to alkaline solutions for hydrogen evolution reaction (HER). Electrochemical measurements are performed with a platinum foil working electrode in 8 M KOH solution and after adding 1–2 vol.% of each ionic liquid. Tafel plots are constructed from the polarization curves recorded in different electrolyte compositions and the main HER kinetic parameters are determined. The highest cathodic currents are found upon [Emim][MeSO 3 ] addition. The Volmer reaction is evidenced as the rate determining step by both Tafel analysis and electrochemical impedance spectroscopy (EIS) measurements. EIS data reveal a signiﬁcant decrease of the overall impedance in the RTILs added solutions. This effect can be due to surface pre-adsorption of the RTIL additives, which stabilize the intermediate hydrogen atoms formed in the Volmer step, thus modifying adsorption and charge transfer processes at the metal-electrolyte interface. This may lead to increased HER currents, as in the case of [Emim][MeSO 3 ] addition. The results support the beneﬁts of using small amounts of selected RTILs as electrolyte additives for the HER in alkaline media.

The efficient production of high purity hydrogen gas (H 2 ) has been considered of fundamental importance in addressing global energy issues while minimizing the environmental impact. 1,2 Currently, most of the H 2 production is achieved by the reforming of fossil or biofuels. An attractive alternative is the production of H 2 by water electrolysis. 3,4 This is an advantageous method for H 2 production because it is not dependent on fossil hydrocarbon sources and originates no carbon emissions. Also, the H 2 produced by this method is very pure, and can rely exclusively on renewable primary energy sources. 5,6 Industrial water electrolysis cells typically employ nickel (Ni) and other metal-based electrodes that operate in potassium (or sodium) hydroxide solution with concentration ranging from 6 to 9 M and in the 60-80 • C temperature range. 7,8 However, the overall energy efficiency of electrolysis, which is partly related to the hydrogen evolution reaction (HER) and to the ohmic resistivity of the electrolytic bath, is relatively limited. 9,10 Major technical problems with these cells are the low stability of the electrode materials and the low conductivity of the alkaline aqueous solutions under cell operating conditions when formation of gas bubbles is pronounced. 11 Thus, bubbles must be effectively eliminated from the solution by an optimized interelectrode distance in an appropriate cell design. 12 In this context, the use of room temperature ionic liquids (RTILs) promises to be a viable alternative, 11,13,14 because their use in replacement of aqueous electrolytes or in a mixture with conventional aqueous electrolytes has the potential to modify the electrode-electrolyte interfaces by affecting the intermolecular interactions. 15 RTILs may acceptably be defined as semi-organic salts composed entirely of organic cations and organic or inorganic anions at room temperature. Besides their wide range of fluidity, they possess high ionic conductivity, excellent thermal and chemical stability, high heat capacity and high cohesive energy density. 11,16 The use of 1-butyl-3-methylimidazolium [ 6 ], as electrolyte (additives) for HER has been the subject of research. 13,14,17 These imidazolium salt electrolytes have unique properties in terms of electrical conductance, and chemical and electrochemical stability. 18 It has been reported that the production of H 2 through water electrolysis using different electrodes, like stainless steel, Ni, platinum (Pt) and low-carbon steel, in presence of these RTILs led to higher H 2 production rates at lower operation costs. 13,14,17 The best HER performance was obtained with the lowcarbon steel electrode in an electrolyte of 10 vol.% of [Bmim][BF 4 ] in water. 13 The results were correlated with ionic conductivity of the media in a way that too low ionic liquid (IL) concentration would limit the ionic conductivity, while too high IL concentration would lead to the formation of aggregates of ionic pairs. The decrease of current density for high IL concentrations was also associated with the formation of H 2 bubbles on the surface of the electrode, reducing the available electrochemical area. 13 20 and attributed the improvement to the easier proton-donating properties on the IL pre-adsorbed surface with the decreased HER activation energy. 14, 21 3-triethylammonium-propanesulfonic acid tetrafluoroborate ([TEA-PS][BF 4 ]) was also used as an electrolyte for water electrolysis. 22,23 This ionic conductor led to high current densities, high efficiencies and low activation energy, which was related to the high conductivity of the system, associated with facilitated proton transport in the aqueous IL media.  24,25 Therefore, RTILs may be used as additives to improve the properties (e.g., ionic conductivity) of electrolyte solutions used for alkaline water electrolysis. Moreover, RTILs with 1-ethyl-3methylimidazolium ([Emim]) cations have been reported as the ionic compounds with the highest conductivity. 26 RTILs are also known to affect the electrode surface or electrode/electrolyte interface, with great interest arising regarding their application, among other areas, in electrochemical double-layer capacitors, 27-34 CO 2 electroreduction, [35][36][37][38][39][40] and synthesis and functionalization of MoS 2 41 or carbon-based catalysts for the HER. 42 The main objective of the present study is to systematically evaluate the effect of RTILs addition, namely [   Electrochemical measurements.-Electrochemical tests were performed at 25 • C in a single-compartment glass cell of 125 mL using a PAR 273A potentiostat/galvanostat (Princeton Applied Research, Inc.) controlled by PowerSuite software. A conventional three-electrode setup was used, with a commercial Pt foil electrode (Metrohm 6.0305.100, 1 cm 2 ) as working electrode, Pt mesh (Johnson Matthey, 50 cm 2 ) as the counter electrode and a saturated calomel electrode (SCE, Hanna instruments, HI 5412) as the reference. All potentials within this study are given relative to SCE reference.
The initial electrochemical characterization of the system was performed by LSV measurements from the OCP (ca. −1.14 V) to −1.5 V at 10 mV s −1 . CA measurements were also carried out at increasing overpotentials, η, and the respective current density, j, recorded after stabilization. EIS measurements were performed at −1.2 V, in the frequency range of 100 kHz to 0.01 Hz with an AC potential amplitude of 5 mV, with the potentiostat coupled to a frequency response analyzer (Schlumberger SI 1255). Experimental EIS data were modelled using an equivalent circuit with ZView software (Scribner Associates, Inc.).

Results and Discussion
Polarization measurements.-The effect of RTILs as additives to KOH electrolyte on the HER kinetics and efficiency was initially evaluated by LSV measurements. The obtained cathodic polarization curves, i.e., current density as a function of potential applied to Pt electrode, in the four studied electrolyte compositions are presented in Figure 1.
It can be seen from Figure 1 that for a given potential value, the highest current densities are obtained after addition of [ To analyze the obtained results, it must be taken into account that all the ILs used within the present study share the same cation. Generally, the cation plays a crucial role in the IL's cathodic stability and potential window. Mousavi et al. reported that ILs with saturated cations with quaternary ammonium substituents typically show higher cathodic stability in comparison with ILs with aromatic cations. 28,29 Experimental measurements followed by computational calculations, where parameters such as the alkyl chain length, the size of the quaternary ammonium cation and the nature of the alkyl substituent were varied, provided information on the role of the lowest unoccupied molecular orbital (LUMO) energy levels on the limiting cathodic potentials, confirming that the LUMO energy levels decrease in the order butylpyridinium < ethylmethylimidazolium (used in this study) < tributylmethylammonium, while their stability to reduction increases in the same order. 28,29 On the other hand, three different anions were used in this work, for which there is no/little data on their effect on the ILs electrochemical behavior. It is know that different anions lead to different wetting of the electrode surface. 28 28 This implies that the RTILs addition modifies the electrode/electrolyte double layer, which affects the Pt electrocatalytic activity toward the HER. In fact, it has been suggested that an electroactive layer of IL ([TEA-PS][BF 4 ]) is formed at the electrode, enhancing its catalytic performance for HER. Thus, the presence of the IL was found to facilitate the mass transfer of the reactants so that HER kinetics are governed by the adsorption and charge transfer processes. 23 The typical HER in alkaline solutions occurs as a multi-step process (Eqs. 1-3). 47,48 First, a primary discharge step, so-called Volmer step, occurs by covering the metal surface with adsorbed protons (Eq. 1). Volmer step is then followed by a catalytic recombination of the adsorbed intermediates (MH ads ) via Tafel step (Eq. 2) or by an electrodesorption of the adsorbed species via Heyrovsky step (Eq. 3). 47,48 The establishment of the rate determining step requires the kinetic study of the HER. Thus, additional E vs. j plots were obtained by carrying out CA measurements at several different applied potentials. The imposed potentials ranged from the open circuit potential (OCP, ca. −1.14 V) up to −1.50 V and the corresponding current density after stabilization was recorded. The resulting E vs. j plots allowed construction of Tafel plots (η vs. log j) shown in Figure 2 and the application of Tafel analysis considering Tafel expression (Eq. 4), where a is the intercept related with the exchange current density, j 0 , corresponding to the electron transfer intrinsic rate. The Tafel slope, b, reflects the rate of change of j with η, with α being the charge transfer coefficient, R the ideal gas constant (8.314 J mol −1 K −1 ), T the temperature in Kelvin and F is Faraday's constant (96485 C mol −1 ). Tafel plots show excellent adjustment to Eq. 4, with correlation coefficients higher than 0.99. The kinetic parameters calculated for the HER in the studied electrolyte solutions are shown in Table II. As expected, the modification of the electrolyte composition by adding each RTIL influenced the reaction kinetics and generally led to higher j 0 values. The exception was for [Emim][EtSO 4 ] where the increase in the RTIL amount from 1 to 2 vol.% decreased the j 0 value. Moreover, the j 0 values obtained were ca. 10 times higher when compared to reported values for HER at Pt or Ni electrodes in 8 M KOH electrolyte without the use of RTILs. 7,49,50 As referred above, the HER can proceed via two possible pathways, Volmer-Heyrovsky or Volmer-Tafel. 47 When the α value is 0.5 at 25 • C, the Tafel slope for Volmer, Heyrovsky and Tafel step being the rate determining step (RDS) is 120, 40 and 30 mV dec −1 , respectively. 7,47 As shown in Table II the Tafel slopes obtained were higher than 120 mV dec −1 , suggesting the Volmer step as the RDS of HER in the studied electrolytes. Indeed, the RTILs addition generally led to higher Tafel slopes and lower charge transfer coefficients, as well as to higher currents clearly observed after addition of [Emim][MeSO 3 ].  It should be noted that effects related with dependency on adsorbed intermediates may be the origin of the differences obtained between the experimental and the theoretical Tafel slopes. 51,52 Electrochemical impedance spectroscopy analysis.-EIS measurements were performed and the corresponding Nyquist and Bode plots are presented in Figure 3 (for all the electrolyte compositions at an applied potential of −1.2 V) and Figure 4 (for the 2   3 ] added and the IL-free KOH electrolytes, at potentials ranging from the OCP to −1.5 V). No inductive loop at low frequencies was observed, which is in agreement with the Volmer step being the RDS, 53 as suggested by the Tafel analysis.

vol.% [Emim][MeSO
The high frequency intersection with the Z axis (insets in Figures  3 and 4) represents the sum of the resistances of the wiring, the electrolyte and the electrode material. 54 Additionally, three semi-circles and therefore three other contributions to the impedance of the system are suggested by the Nyquist plots in Figures 3, 4A and 4B, as well as by the corresponding peaks in the Bode plots (Figures 3 and  4C and 4D). The impedance of the system may then be modeled by the equivalent circuit presented in Figure 5. It consists of a resistor, R s , accounting for the resistance of the wiring, electrolyte and electrode material, in series with three contributions, each consisting of a resistance, R, and a constant phase element (CPE) in parallel.
CPE elements are frequently used in the analysis of impedance data to account for real systems and generally associated with a distribution of relaxation times on the electrode surface. 54,55 In fact, equivalent circuits known as 2CPE or 3CPE models have been applied to the study of HER on porous electrodes in alkaline media. 53,[56][57][58] The impedance of a CPE can be expressed by Eq. 5, where Q 0 is a pseudocapacitance, ω is the angular frequency (equal to 2πf, f is the linear frequency), i is equal to √ −1, and n is an empirical constant that can take values between 0 and 1. 59 The true capacitance, C, can be estimated from Eq. 6. 59 In the case of n = 1, the CPE is a pure capacitor.
The high-frequency relaxation, visible in the insets in Figures 3 and  4, roughly above 1 kHz, may be related to the resistance of the charge transfer process, R 1 , and the capacitance of the double layer, C 1 . At intermediate frequencies, between ca. 0.1 Hz and 1 kHz, R 2 and C 2 are probably associated with the superficial mass transfer of adsorbed H. R 3 and C 3 , regarding the low frequency contribution, detected in some cases below 0.1 Hz, may be assigned to the oscillation of the H 2 concentration at the electrode/electrolyte interface. 60 For applied potentials above −1.2 V, the low frequency relaxation could not be correctly registered and the semi-circles were not well resolved due to high instability of the impedance measurements for frequencies below 0.1 Hz. This occurrence may be due to the increase of the hydrogen evolution at those potentials and the consequent high amount of gas bubbles in the vicinity of the electrode.
The parameters related with the several resistance and capacitance contributions are obtained from the fittings in Figures 3 and 4 and are presented, respectively, in Tables III and IV. The resistance val-ues were obtained directly as the fit parameters, whereas the true capacitances were estimated from CPE parameters using Eq. 6.
In spite of the small shift of the high frequency interception with the Z axis with the addition of RTILs and with the variation of their concentration, it may be seen in Table III 3 ], a decrease of 23% in R s was observed. As all measurements were performed using the same experimental setup, only changing the electrolyte composition, it may be concluded that the resistance of the 8 M KOH electrolyte was considerably decreased by the RTILs addition.
The resistance of the high frequency contribution, R 1 , visible in the insets of Figures 3 and 4, has a magnitude of around 2.34 cm 2 in the case of the IL-free KOH electrolyte solution. This contribution is considerably decreased upon the RTILs addition, with the exception of the addition of 1 vol.% of [Emim][EtSO 4 ], which revealed a more moderate reduction. A clear decrease was however observed when 2 vol.% of this RTIL was added. The capacitance values for this contribution, C 1 , are in the 54-150 μF cm −2 range, which are close to previously reported values for double layer capacitance. 61,62 It should be mentioned that the experimental capacitance-potential curves of ILs contradict the Gouy-Chapman double layer model so that the potential of zero charge obtained from an electrocapillary maximum (via a dropping mercury electrode) does not correspond to the capacitance minimum. 63 Thus, alternative double layer models for ILs have been suggested. 27 Double layer capacitance of an IL is determined by the IL's structure and the electrode material. 28 Cation size was shown to have strong effect on the electrolyte viscosity and conductivity, as well as the capacitance. [30][31][32][33] Comparison of ILs with cations of different structure, including imidazolium, ammonium, pyridinium, piperidinium, and pyrrolidinium, showed that imidazolium-and pyridinium-based ionic liquids provide the highest capacitance. [30][31][32][33] Studies of the influence of anion on the double layer capacitance are somewhat rarer. Still, double layer capacitance of several ionic liquids with structurally diverse anions, namely tetrafluoroborate, trifluoromethanesulfonate and trifluoromethanesulfonimide, has been recently explored. 28 It was shown that double layer capacitance of a quaternary ammonium salt with a methoxyalkyl group on the nitrogen atom depends more on the nature of the anion than on the cation structure. 34 Regarding the intermediate frequency contribution, with a resistance of 158.10 cm 2 in the case of the IL-free KOH electrolyte, it is observed a remarkable three-fold decrease in all the electrolytes upon RTIL addition, nearly four-fold in the case of the addition of 2 vol.% [Emim][MeSO 3 ]. It should be noted that the increase in [Emim][EtSO 4 ] concentration from 1 to 2 vol.% does not lead to a decrease of this resistance, showing the highest resistance of the RTIL-added electrolytes, thus demonstrating the worst performance among all studied RTILs, in agreement with previous HER results (Figures 1 and 2). The capacitances obtained for this contribution are of the order of 10 −4 F cm −2 .
The R 3 resistance detected in some cases, particularly at −1.2 V, and with a magnitude of around 30 cm 2 for the KOH-only electrolyte, is again substantially decreased in the RTIL-added electrolyte solutions. This contribution showed capacitance values of the  order of 10 −2 F cm −2 in the RTIL-added electrolytes and a higher value of the order of 10 −1 F cm −2 in the case of the IL-free KOH solution. Figure 4 and Table IV compare the variation with the applied potential of the Nyquist and Bode plots of the 2 vol.% [Emim][MeSO 3 ]added and the IL-free KOH electrolytes. The impedance is drastically decreased with the increasing applied potential and, as previously discussed, the RTIL-added electrolyte shows lower resistance at a given potential, regarding practically all the contributions. Above −1.3 V, both C 1 and C 2 are of 10 −4 F cm −2 order of magnitude, with the capacitance values slightly decreasing with the increasing potential.
The sum of the R 1 and R 2 resistances at a given applied potential is related with the HER kinetics and its reciprocal value is directly related to the faradaic current density. The Tafel plots in Figure 2 show that HER is charge transfer-controlled in the studied potential range, thus η vs. log (R 1 + R 2 ) −1 plot should be linear and its slope equal to the Tafel slope. 61,62 This is in fact the case and may be confirmed in Figure 6, which shows that the Tafel plots (from Figure 2B) and the η vs. log (R 1 + R 2 ) −1 plots concerning the 8 M KOH + 2 vol.% of [Emim][MeSO 3 ] electrolyte have equivalent slopes of 180 and 176 mV dec −1 , respectively. The separation between the two lines is 1.24, consistent with a Langmuir-type monolayer adsorption of the H ads to the surface of the electrode, which has a theoretical value of 1.29. 61 It should be noted that the equivalent slopes of the experimental and simulated plots in Figure 6, as well as the agreement with the predicted separation between the two lines, 61 support the applicability of the model used in the present analysis of EIS experimental data for the HER in alkaline media. Generally, the addition of RTILs led to a pronounced decrease of the overall impedance of the system. As cathodic polarization was applied, the similarities observed in the impedance response of the Pt electrode in RTIL-added electrolytes may be related with the presence of the same cation, [Emim]. Cation adsorption to the electrode and/or cation-metal complex adsorbed intermediate species are affecting the reduction processes, as it was reported by Rosen et al. 35 65 Indeed, the important role of the adsorption of ions onto electrode surfaces and the formation of intermediate complexes during electrocatalytic processes in ILs is acknowledged 37,39,40,66,67 and may have an effect on the present HER results. In this way, the reported results may be originated from surface pre-adsorption of the IL additives, stabilizing the intermediate hydrogen atoms formed in Volmer step and modifying adsorption and charge transfer processes at the metal/electrolyte interface.
Future prospects are focused on the effect of the addition of RTILs with different cations and anions to the electrolyte. Additionally, attention will be devoted to the RTILs behavior at higher temperatures, at which a decrease of their viscosity and a consequent enhancement of the electrolyte conductivity are expected, with the corresponding improvement of the electrolysis cell efficiency. The effect of adding RTILs to the KOH electrolyte solution while using Ni electrodes will also be envisaged, as the impact of this behavior on practical electrolysis cells, which use Ni-based cathodes instead of Pt and other noble metals, can lead to considerable advances in commercial alkaline electrolyzers.  8.08 × 10 −3 , 8.87 × 10 −3 , 1.35 × 10 −2 and 2.10 × 10 −2 mA cm −2 , respectively. Tafel analysis suggests the Volmer step as the RDS, which is supported by the impedance measurement results. EIS experimental data were modelled with an equivalent circuit and showed a remarkable decrease in the overall impedance of the system after addition of the RTILs, with generally all contributions being decreased. A Langmuir-type monolayer adsorption of H ads to the electrode surface was also observed. These environmental-friendly electrolyte additives were shown to improve