- Email: [email protected]
Beneficial effect of added water on sodium metal cycling in super concentrated ionic liquid sodium electrolytes
Journal of Power Sources 379 (2018) 344–349
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Short communication
Beneficial effect of added water on sodium metal cycling in super concentrated ionic liquid sodium electrolytes
T
Andrew Basile∗, Shammi A. Ferdousi, Faezeh Makhlooghiazad, Ruhamah Yunis, Matthias Hilder, Maria Forsyth, Patrick C. Howlett∗∗ Institute for Frontier Materials (IFM), Deakin University, Burwood, Victoria 3125, Australia
H I G H L I G H T S
G RA P H I C A L AB S T R A C T
Controlled [H O] benefits ionic liquid • electrolyte's physicochemical proper2
• •
ties. Demonstrated Na0/+ plating/stripping containing large amounts of H2O up to 500 ppm. H2O promotes [FSI]- breakdown forming a beneficial Solid-electrolyte interphase.
A R T I C L E I N F O
A B S T R A C T
Keywords: Room temperature ionic liquid Additives Sodium metal Superconcentrated Solid-electrolyte interphase
The plating and stripping performance of sodium metal in an ionic liquid electrolyte is improved when including water as an additive. Herein we report for the first time the trend of improved cycling behavior of Na0/+ in Nmethyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide with 500 ppm H2O. The addition of water to this ionic liquid electrolyte promotes the breakdown of the [FSI]- anion towards beneficial SEI formation. The benefits during plating and stripping of sodium is observed as lower total polarization during symmetrical cell cycling and decreased electrode/electrolyte interface impedance. Sodium metal surfaces after cycling with 500 ppm H2O are shown to be smooth in morphology in comparison to lower additive concentrations. The outcome of adventitious moisture benefiting Na0/+ cycling in an ionic liquid, contrary to conventional electrolytes, allows flexibility in ionic liquid electrolyte design to the benefit of battery manufacturers.
1. Introduction A global focus towards alternative energy storage technologies has recently seen the ‘beyond lithium’ research area drastically expand. This is due to several drawbacks expected for future lithium ion technologies mainly pertaining to high cost, and low natural abundance/ access of starting materials [1]. As sodium and lithium share similarities in their alkali metal electrochemistry, there is rationale for renewed interest in the area of sodium batteries. It is now envisioned that ∗
sodium secondary cells, amongst others, will be implemented in parallel to lithium based devices in the near future [2,3]. However, in order to be competitive with future Li-ion technologies, it is essential to enable the highest possible energy densities for sodium based devices (1677 Wh kg−1 Na-O2) [4]. In doing so, we may harness the sodium metal as an anode rather than hard carbon materials which are currently used (ca. 100–300 mAh g−1) [5]. The path towards enabling sodium metal anodes has proven difficult, mainly due to the reactivity of sodium with water and/or the flammable conventional electrolytes
Corresponding author. Corresponding author. E-mail addresses: [email protected] (A. Basile), [email protected] (P.C. Howlett).
∗∗
https://doi.org/10.1016/j.jpowsour.2018.01.044 Received 10 December 2017; Received in revised form 14 January 2018; Accepted 18 January 2018 0378-7753/ © 2018 Elsevier B.V. All rights reserved.
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
conductivity values decrease an order of magnitude when compared to the neat C3mpyrFSI. This is clearly attributed to the rise in viscosity shown in Fig. 1b. This electrolyte still maintains a liquid phase as shown previously to be possible within highly concentrated ionic liquid systems [8]. The increased viscosity and concomitant drop in conductivity of the system can be reversed through the addition of water to the electrolyte. Here, a concentration of 500 ppm water within the 50 mol% electrolyte exhibits conductivity values close to that of the neat C3mpyrFSI -albeit with greater viscosity. Practical applications of batteries cause internal heat generation which in turn raises the device temperature above ambient conditions. These intermediate temperatures can allow improved electrode kinetics in the full device, and it is only the fact that the traditional organic electrolytes lead to instabilities at elevated temperatures which has limited battery operation to room tempera ture [18]. Fig. 1a reveals that at a temperature of 50 °C the 500 ppm water containing electrolyte is within the same order of magnitude to the neat electrolyte. This 50 °C intermediate temperature then used as a controlled temperature for all cell cycling herein, consistent with our previous reports for the purpose of drawing comparison [8,19]. The symmetrical cell cycling for these three 50 mol% NaFSI in C3mPyrFSI electrolytes shown in Fig. 2a compares the polarization of the cell containing < 20 ppm H2O against those considered ‘wet’, comprising of 100 ppm and 500 ppm of water. The electrolyte containing less than 20 ppm is considered dry in this instance. The initial polarization for each cell begins at high values, larger than 250 mV, with the highest belonging to the cell with dry electrolyte at ca. 360 mV. As is consistently observed with ionic liquid electrolytes this overpotential relaxes upon continued cycling to values ca. 100 mV after 20 cycles. It is at this point after cycling that the effect of water as an additive can be observed clearly. The overpotentials upon completing the final polarization step are ca. 80 mV, 100 mV, 150 mV for the electrolyte systems comprising of ca. 500 ppm < dry < 100 ppm respectively, and are shown in Fig. 2b. It is this most ‘wet’ electrolyte which exhibits the highest stability in the voltage profile during the 20 plating and stripping cycles, suggesting that the (i) electrode-electrolyte interface formed has lowest resistance, and/or that (ii) the addition of water to the electrolyte promotes diverse speciation and an increased diffusion for the Na+ ion. Investigations to determine the strength of these proposed mechanisms for beneficial cycling are currently underway in our labs. Unlike the dry and 100 ppm electrolyte, this cell's voltage-time curve shows very little sign of tailing at the completion of each polarization step. This is indicative of a beneficial solid-electrolyte interphase (SEI) which permits the ingress/ egress of sodium without Coulombic inefficiency. If this were not the case, the tailing (which is seen predominantly in the dry cell) would be more apparent at the completion of cell cycling, and suggestive of pristine sodium being stripped from the counter electrode [20]. This
carried forward from Li-ion technologies [6]. In order to remove this safety concern from future sodium metal devices, a new class of electrolyte have shown promise, namely highly concentrated ionic liquid electrolytes [7–9]. Recent results have shown highly concentrated ionic liquid electrolytes allow a large amount of charge migration in both lithium and sodium plating and stripping [8]. In addition, the passing of large charges has been reported to prepare a surface with a stable electrodeelectrolyte interface [7]. These ionic liquids are typically non-flammable and exhibit negligible vapor pressure thus promoting their use as safe electrolytes. Nonetheless, ionic liquid electrolytes still falter in their high viscosity and lower conductivity, the latter being compounded with the addition of such high salt concentrations [10]. Various additives are included to alleviate viscosity (H2O) [11–13], flammability (e.g. ionic liquids) [6], or promote beneficial interface formation via sacrificial breakdown (fluoroethylene carbonate, vinylene carbonate) [14]. Importantly, for ionic liquid electrolytes, a beneficial solid-electrolyte interphase has been shown to form without the use of such additives [15]. The effect of water has been studied in non-fluorinated ionic liquid electrolytes for both Li and Na cells [16,17]. Both of these reports concluded that an increase above ca. 100–200 ppm H2O within the ionic liquid electrolyte was detrimental to cell cycling and that the water content should be kept to a controlled minimum. Herein we report for the first time the stable plating and stripping of sodium metal from N-methyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide comprising of 50 mol% sodium bis(fluorosulfonyl)imide (1:1 NaFSI:C3mpyrFSI) with an additional 500 ppm H2O. These results are compared with ionic liquids bearing the FSI- anion paired with an alkyl phosphonium or an alkoxy ammonium cation. With this higher concentration of H2O, the viscosity and conductivity concern may be alleviated, enabling cells exhibiting lower overpotentials required for the Na0/+ redox process than previously reported for implementation in Na metal cells. Additionally, at a high concentration of 50 mol% Na+ salt the lower cost of NaFSI salt, compared to the IL, effectively reduces the cost of these electrolytes. 2. Results & discussion The physicochemical properties of the neat C3mpyrFSI ionic liquid and electrolyte systems are shown in Fig. 1 and describe the trend of improved conductivity and viscosity with the inclusion of water as an additive. The conductivity curves measured at 10 °C intervals in Fig. 1a show that neat C3mpyrFSI maintains the highest conductivity values at all temperatures. As is common to this class of electrolyte, the conductivity for the system can be seen to steadily increase through three orders of magnitude in the temperature range −20 °C to 120 °C. It is upon the addition of a high 50 mol% concentration of NaFSI that these
Fig. 1. (a) Conductivity and (b) viscosity values for the C3mpyrFSI ionic liquid (shown at right) and the 50 mol% NaFSI containing electrolytes including increased amounts of water.
345
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Fig. 2. Symmetrical Na|Na cell cycling for (a) 20 cycles highlighting the final polarization steps (b). Nyquist plots for the Na|Na cells comprising of (c) < 20 ppm (d) 100 ppm and (e) 500 ppm of water additive.
At higher magnification inset the surface is observed to be covered in deep narrow pits. In areas where the pitting is more severe (highlighted in Fig. S1), the sodium metal has a large surface area with high roughness. This roughness is increased in the case of the cell cycled using 100 ppm water electrolyte (Fig. 3b). No evidence of dendrites is observed. It is in viewing the sodium surface from the cell cycled with 500 ppm of water that a stark contrast can be made. This surface which provided the lowest resistance and overpotential during cycling is smoother and has lower pitting density than the other electrodes (Fig. 3c). The surface is also partially covered with residual electrolyte, and requires more vigorous washing, otherwise the ionic liquid electrolyte adheres to the sodium metal as shown in Fig. S2. In fact, at higher magnification inset Fig. 3c it can be seen that, what may be pits harbor residual IL observed having darker contrast. Overall, the smoothest surface prepared over 20 cycles, shows the best cycling performance. This is particularly interesting as the two rougher surfaces have greater relative surface area, and inherently would disperse the applied current density. Nonetheless, these 20 ppm and 100 ppm cells still maintain highest overpotentials during Na|Na cycling (Fig. 2a). The elemental distribution across the surface can be seen in Fig. 3d and e, and reveals a large occurrence of carbon and oxygen on the larger, rough sodium electrodeposits regardless of cycling in dry or ‘wet’ electrolytes, whereas the fluorine and sulfur share the same regions of the underlying sodium surface. In fact when comparing these surfaces, it is apparent that the breakdown of the anion occurs more favorably when the [FSI]- anion is in the presence of the water additive. Listed in Table 1 are the weight% values collected for each from EDS spectra shown in Fig. S3 revealing the increase of fluorine and sulfur breakdown products in particular. Quite interestingly these species are cohabitating the rough plated sodium in the ‘wet’ electrolyte (Fig. 3e),
increased stability is also observed when measuring for longer polarization times (1 h) over larger numbers of plating/strpping cycles, currently underway in our labs. It is interesting to note that the lower amount of water additive at ca. 100 ppm invokes a more negative impact on plating/stripping performance than the dry electrolyte. Contrary to that described for a similar ionic liquid based on the dicyanamide anion for both lithium and sodium electrochemistry [16,17]. This is more clearly visualized in the impedance data collected pre/post cycling each of the Na|Na cells. Prior to cycling, after an initial resting period of 24 h at 50 °C, the interfacial resistance (Rint) reaches ca. 260 Ω. This is also the case for the dry electrolyte (< 20 ppm, 305 Ω). However after cycling with the electrolyte containing 100 ppm H2O Rint falls to ca. 90 Ω and lower still for the dry cell at ca. 50 Ω. The slight difference in water content for these two cells is still enough to significantly alter the interface and affect a change in the plating/stripping. With an increase to 500 ppm (ca. 0.05% water by mass in electrolyte) the initial resting interface resistance is much larger ca. 505 Ω. The additional water in the electrolyte initially prepares a highly resistive interface, however this is not detrimental to symmetrical cell cycling as the resistance falls to ca. 28 Ω post plating/stripping. This phenomenon of decreased resistance post cycling has been reported when plating and stripping from highly concentrated ionic liquids, however the beneficial effect of water is yet to be described [8]. It is interesting to note that it is still possible to cycle sodium from these FSI− anion based ionic liquid electrolytes even with such a high concentration of water as an additive, unlike in the dicynanamide systems reported in the literature [16,17]. This is important from the perspective of cell design and construction, where adventitious moisture across cell manufacture will allow greater flexibility in electrolyte formulation, cell preparation and process design. 346
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Fig. 3. Scanning electron micrographs of the sodium electrode surfaces after 20 cycles in symmetrical orientation using electrolyte comprising water additive at a concentration of a, < 20 ppm (left), b, 100 ppm (center), and c, 500 ppm (right). Corresponding EDS mapping comparing the d, dry and e, ‘wet’ surfaces.
previously reported by our group [22] and is attributed to strong interactions between the ions, through strong coordination between Na+ and the functionalised ammonium cation carrying ethoxy/methoxy groups. This is also reflected in the larger overpotential for symmetrical cell cycling with this electrolyte. Upon completion of cell cycling the final overpotential for the cell comprising of the phosphonium electrolyte remained constant with only a small decrease to values near 80 mV. However, the cell with ammonium IL reached a stable polarisation potential only after ten cycles (ca. 6 h). The overpotential for the ammonium electrolyte cell decreased substantially yet still remains quite high at ca. 220 mV. These overpotential values are larger than in the pyrrolidinium case which is shown using a dashed line for comparison. Furthermore, the electrode-electrolyte interface (prior to cycling) on the sodium metal electrode when using the ammonium electrolyte with 500 ppm water shows a higher impedance (ca. 3200 Ω, Fig. 4c). Whereas in the ammonium after cycling, the interface resistance (shown in Fig. 4c) decreases to a much lower value of 300 Ω. This phenomenon is also shown in the phosphonium cell, Fig. 4d, in which the impedance associated with the electrode interface decreases after cycling to 40 Ω and indicates the formation of a beneficial SEI layer at the surface of the electrode -which permits facile ingress and egress of Na+. The effect of water in this P111i4FSI based ionic liquid electrolytes follows the same trend as for the C3mpyrFSI, as an additive towards beneficial SEI formation, showing little difference in the V-t profile when compared to the dry P111i4FSI electrolyte. However, in the N2(20201)3FSI ionic liquid this benefit is not observed. When compared to the dry N2(20201)3FSI electrolyte shown in Fig. S4, with ca. 20 ppm water, the dry Rint is smaller after cycling. This suggests that ionic liquids including the FSI− anion are capable of including water as an additive for beneficial cycling performance but their behavior depends also upon the ionic liquid's cation chemistry. In the N2(20201)3FSI case, the ether oxygens may coordinate the sodium ion more strongly and thus the water may have a less disruptive effect on the speciation and hence the nature of the species presenting to the electrode and the resultant SEI layer. This will be studied in further in future work. The effect of cation, anion and mixed salt is also being studied with respect to adventitious moisture in further detail targeting cost reduction of this class of ionic liquid electrolyte.
Table 1 Elemental distribution of the Na surface after 20 cycles (J = 1 mA cm−2) with 15 min polarization at 50 C°. Element
Na C N O F S
dry IL
‘wet’ IL
Weight/%
Weight/%
41.6 16.1 1.3 31.1 3.7 6.2
29.2 15.7 4.1 24.6 12.9 13.6
which does not occur in the dry (Fig. 3d). From that which we know regarding [FSI]- anion breakdown [21] for beneficial SEI growth at Li0 anodes [15], it can be postulated that the thiazate anion product, NSO− is providing the required passivation of high surface area sodium deposits required to permit lower surface impedance and stable cycling when in a superconcentrated and wet environment. The EDS analysis clearly shows that the presence of water modifies the SEI formed in these systems, supporting reduced surface impedance and better Na metal plating/stripping. Further studies will seek to understand the detail of these processes. To determine whether other FSI− based ionic liquids would also function as effective electrolytes with 500 ppm water, sodium symmetrical cells containing 55 mol% NaFSI in trimethyl(isobutyl)phosphonium (P111i4FSI) and 42 mol% NaFSI in N-ethyl-2-(2-methoxyethoxy)-N,N-bis(2-(2-methoxyethoxy)ethyl)-ethan-1-aminium bis(fluorosulfonyl)imide (N2(20201)3FSI) were prepared. These cells were then cycled at 50 °C at the same current density of 0.1 mA cm−2 with 15 min interval for each polarisation. These concentrations are chosen as they compare to the solubility limit and highest Na+ concentrations for these ionic liquids. Fig. 4 shows the cycling results and EIS before and after cycling in these ionic liquids based on different cations. The polarisation potential of the symmetrical cell with phosphonium IL is lower than its ammonium counterpart (Figure 4a, 80 mV vs 220 mV). This can be ascribed to the higher viscosity, lower conductivity and higher resistivity of the ammonium ionic liquid, which have been 347
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Fig. 4. The [N2(20201)3]+ and [P111i4]+ cations (top) and their performance in symmetrical Na|Na cells when using P111i4FSI and N2(20201)3FSI with 500 ppm H2O for (a) 20 cycles highlighting (b) the final cycle. Nyquist plots for the (c) N2(20201)3FSI and (d) P111i4FSI cells pre/post cycling.
3. Conclusions
4. Materials and methods
The use of water as an additive to alleviate viscosity concerns and boost the conductivity of ionic liquids has been achieved with larger than previously reported values up to 500 ppm. The variation from the dicyanamide anion to the bis(fluorosulfonyl)imide anion provides this allowance suggesting it is, at least in part, the breakdown of the anion in a wet environment that causes this change in behavior. This is shown across two different cations for C3mpyrFSI and P111i4FSI, and detrimental at 500 ppm for the N2(20201)3FSI. The cycling behavior of symmetrical cells containing wetted C3mpyrFSI perform comparably with the dry electrolyte system. Indicating that this class of IL is not disturbed by up to 500 ppm of H2O. The surface of the sodium metal electrode is also benefitted from the largest water addition, which provides an ideally smooth morphology during cycling, and the lowest surface impedance of all cells. This suggests an ideal SEI may be formed through the use of water as an additive in FSI− based highly concentrated ionic liquid electrolytes. An amount of ca. 500 ppm is shown to perform well for both improving bulk physicochemical properties of the electrolyte, in addition to improving the sodium surface for cycling. Ongoing research aims to determine the optimum water addition within this new class of electrolyte to the benefit of both (i) efficient charge transport and (ii) interface formation in sodium metal secondary cells.
4.1. Cell fabrication Sodium metal (Sigma, purity under paraffin oil) was rolled flat, polished using a nylon brush under hexane (RCI Leachem, 99%, 6 ppm H2O) and then punched into disks (8 mm diameter, 100 μm thickness). Symmetrical Na|Na cells were constructed using coin cell configuration (Hohsen, CR2032 stainless steel using 1 mm spacer and 1.4 mm spring). A polyethylene separator (gratis, Lydall, 7P03A, 50 μm thickness, 85% 0.3 μm porosity) was used to prevent short circuit and soaked with a 150 μL aliquot of electrolyte. All cell components were dried, and assembly took place within an argon glovebox (< 1 ppm H2O, < 1 ppm O2). All cells were allowed to rest for 24 h at 50 °C prior to cycling as a formation step. 4.2. Materials and methods N-methyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide (C3mpyrFSI, Solvionic, 99.9%), potassium bis(fluorosulfonyl)imide (KFSI) (Suzhou Fluolyte China, > 99.9%), trimethyl(isobutyl)phosphonium dimethylphosphate (P111i4DMP) (Cytec Solvay Canada, mixture with dimethylphosphoric acid), acetonitrile (Sigma Aldrich Australia, > 99.8%, H2O = < 10 ppm) and dichloromethane (Sigma Aldrich Australia, > 99.8%) were used as received. Tris[2-(2-methoxyethoxy)ethyl]amine (Sigma Aldrich Australia, 95%) was freshly distilled over KOH and the colorless second fraction was used same-day to synthesize the ammonium salt. Milli348
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Q water (18.2 MΩ cm−1) was used during all this work. P111i4FSI was synthesized through anion exchange between P111i4DMP and KFSI by known procedures [23]. The purity of P111i4FSI was confirmed by 1H, 13C, 19F Nuclear Magnetic Resonance Spectroscopy (NMR), mass spectrometry, melting point and potassium content (20 ppm, Ion Selective Electrode). N-ethyl-2-(2-methoxyethoxy)-N,Nbis(2-(2-methoxyethoxy)ethyl)-ethan-1-aminium bis(fluorosulfonyl) imide (N2(20201)3FSI)was synthesised by already known procedures by anion metathesis between N2(20201)3 Br and KFSI [24]. The purity of N2(20201)3FSI was confirmed by 1H, 13C, 19F NMR, mass spectrometry, potassium (15 ppm) and bromide (150 ppm) contents [22]. All NMR data for the synthesized ILs is present in the supplementary information. All ionic liquids were dried using Schlenk manifold techniques prior to salt and water additions. Sodium bis(fluorosulfonyl)imide (NaFSI, Solvionic, 99.9%) was added to each respective ionic liquid before an additional water removing step using Schlenk manifold techniques under vacuum with sodium hydride (Sigma Aldrich (60% dispersion in mineral oil). Hexane was stirred in the presence of calcium hydride overnight, and then distilled and kept over lithium metal in the argon glovebox (RCI Leachem, 99%). Water content < 20 ppm was considered as dry within this study. Water additions (Milli-Q, 18 MΩ cm−1) were made dropwise to spike up to 100 ppm or 500 ppm. All electrolyte formulation and storage takes place within an Ar environment using a glovebox.
its financial support through Australian Laureate Fellowship FL110100013 (MF) and FL120100019 (DRM) and under Australian Research Council's Discovery Projects funding scheme (DP160101178). All authors thank Peter Newman for instrument design. Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx. doi.org/10.1016/j.jpowsour.2018.01.044. References [1] B.L. Ellis, L.F. Nazar, Sodium and sodium-ion energy storage batteries, Curr. Opin. Solid State Mater. Sci. 16 (2012) 168–177. [2] B. Dunn, H. Kamath, J.-M. Tarascon, Electrical energy storage for the grid: a battery of choices, Science 334 (2011) 928–935. [3] V. Palomares, M. Casas-Cabanas, E. Castillo-Martínez, M.H. Han, T. Rojo, Update on Na-based battery materials. A growing research path, Energy Environ. Sci. 6 (2013) 2312–2337. [4] C. Pozo-Gonzalo, P.C. Howlett, D.R. MacFarlane, M. Forsyth, Highly reversible oxygen to superoxide redox reaction in a sodium-containing ionic liquid, Electrochem. Commun. 74 (2017) 14–18. [5] X. Dou, et al., Pectin, hemicellulose, or Lignin? Impact of the biowaste source on the performance of hard carbons for sodium-ion batteries, ChemSusChem 10 (2017) 2668–2676. [6] D. Monti, A. Ponrouch, M.R. Palacín, P. Johansson, Towards safer sodium-ion batteries via organic solvent/ionic liquid based hybrid electrolytes, J. Power Sources 324 (2016) 712–721. [7] G.M.A. Girard, et al., The role of Li concentration and the SEI layer in enabling high performance Li metal electrodes using a phosphonium bis(fluorosulfonyl)imide ionic liquid, J. Phys. Chem. C2 (2017), http://dx.doi.org/10.1021/acs.jpcc. 7b01929 acs.jpcc.7b01929. [8] A. Basile, et al., Extensive sodium metal plating and stripping in a highly concentrated Inorganic−Organic ionic liquid electrolyte through surface pretreatment, ChemElectroChem 4 (2017) 986–991. [9] A. Basile, et al., Enabling Sodium metal anodes using concentrated salt in ionic liquid electrolytes, The 67th Annual Meeting of the International Society of Electrochemistry, International Society of Electrochemistry, 2016. [10] F. Makhlooghiazad, et al., Phosphonium plastic crystal salt alloyed with a sodium salt as a solid-state electrolyte for sodium devices: phase behaviour and electrochemical performance, J. Mater. Chem. A 5 (2017) 5770–5780. [11] L. Suo, et al., ‘Water-in-salt’ electrolyte enables high-voltage aqueous lithium-ion chemistries, Science 350 (2015) 938–943. [12] S.S. Zhang, A review on electrolyte additives for lithium-ion batteries, J. Power Sources 162 (2006) 1379–1394. [13] F. Mizuno, T.S. Arthur, K. Takechi, Water in ionic liquid for electrochemical Li cycling, ACS Energy Lett. 1 (2016) 542–547. [14] M. Dahbi, et al., Black phosphorus as a high-capacity, high-capability negative electrode for sodium-ion batteries: investigation of the electrode/electrolyte interface, Chem. Mater. 28 (2016) 1625–1635. [15] A. Basile, A.I. Bhatt, A.P. O'Mullane, Stabilizing lithium metal using ionic liquids for long-lived batteries, Nat. Commun. 7 (2016). [16] A. Basile, H. Yoon, D.R. MacFarlane, M. Forsyth, P.C. Howlett, Investigating nonfluorinated anions for sodium battery electrolytes based on ionic liquids, Electrochem. Commun. 71 (2016) 48–51. [17] H. Yoon, et al., Lithium electrochemistry and cycling behaviour of ionic liquids using cyano based anions, Energy Environ. Sci. 6 (2013) 979. [18] M. Hilder, et al., Small quaternary alkyl phosphonium bis(fluorosulfonyl)imide ionic liquid electrolytes for sodium-ion batteries with P2- and O3-Na 2/3 [Fe 2/3 Mn 1/3 ]O 2 cathode material, J. Power Sources 349 (2017) 45–51. [19] C. Ding, et al., Na[FSA]-[C 3 C 1 pyrr][FSA] ionic liquids as electrolytes for sodium secondary batteries: effects of Na ion concentration and operation temperature, J. Power Sources 269 (2014) 124–128. [20] K.N. Wood, et al., Dendrites and pits: untangling the complex behavior of lithium metal anodes through operando video microscopy, ACS Cent. Sci. 2 (2016) 790–801. [21] I.A. Shkrob, T.W. Marin, Y. Zhu, D.P. Abraham, Why bis(fluorosulfonyl)imide is a ‘magic anion’ for electrochemistry, J. Phys. Chem. C 118 (2014) 19661–19671. [22] M. Hilder, et al., Effect of mixed anions on the physicochemical properties of a sodium containing alkoxyammonium ionic liquid electrolyte, Phys. Chem. Chem. Phys. 19 (2017) 17461–17468. [23] M. Hilder, et al., Physicochemical characterization of a new family of small alkyl phosphonium imide ionic liquids, Electrochim. Acta 202 (2016) 100–109. [24] M. Kar, et al., Ionic liquid electrolytes for reversible magnesium electrochemistry, Chem. Commun. 52 (2016) 4033–4036.
4.3. Instrumentation Water contents were measured using Karl Fisher potentiometric titration using a Model 756/831 KF Coulometer (MEP Instruments). Bromide content was measured by using IJ-Br combinational bromide selective electrode (Ionode) after calibration with 10 and 100 ppm bromide standards. Similarly potassium content was measured by Potassium Half-Cell Ion Selective Electrode after calibration with 10 and 100 ppm potassium standards. Viscosity data was collected using a rolling ball viscometer (Anton Paar Lovis 2000ME). Conductivity measurements were made using a hermetically sealed dip cell with embedded platinum electrodes (in-house designed) controlled using a MTZ-35 frequency response analyser (BioLogic) through a range of temperatures maintained by thermocouple. Samples were quenched prior to measurement at 10 °C steps with resting times of 2 h. All symmetrical Na|Na stripping/plating cycling was collected using a VMP3 MultiPotentiostat within temperature controlled ovens held constant at 50 °C. Electrochemical impedance spectroscopy was used to collect impedance of the symmetrical cells pre/post cycling. This was collected through a frequency range of 1 MHz–10 mHz by applying an alternating 10 mV RMS perturbation to the cell. All cycling experiments include duplicates for reproducibility. Purity of the ionic liquids was confirmed by NMR carried out using a DPX Spectrometer (Bruker) And electrospray ionization using Micromass Platform I IAPI QMS Electrospray mass spectrometer (Waters Corporation). All micrographs were collected using a JSM-IT300LV scanning electron microscope (JEOL Ltd Japan) using an accelerating voltage of 2.0 kV under high vacuum with secondary electron detector. All elemental mapping was performed within the SEM using a 50 mm2 X-Max energy dispersive spectroscopy detector (Oxford Instruments). Notes The authors declare no competing financial interest. Acknowledgment The authors acknowledge the ARC (Australian Research Council) for
349
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Short communication
Beneficial effect of added water on sodium metal cycling in super concentrated ionic liquid sodium electrolytes
T
Andrew Basile∗, Shammi A. Ferdousi, Faezeh Makhlooghiazad, Ruhamah Yunis, Matthias Hilder, Maria Forsyth, Patrick C. Howlett∗∗ Institute for Frontier Materials (IFM), Deakin University, Burwood, Victoria 3125, Australia
H I G H L I G H T S
G RA P H I C A L AB S T R A C T
Controlled [H O] benefits ionic liquid • electrolyte's physicochemical proper2
• •
ties. Demonstrated Na0/+ plating/stripping containing large amounts of H2O up to 500 ppm. H2O promotes [FSI]- breakdown forming a beneficial Solid-electrolyte interphase.
A R T I C L E I N F O
A B S T R A C T
Keywords: Room temperature ionic liquid Additives Sodium metal Superconcentrated Solid-electrolyte interphase
The plating and stripping performance of sodium metal in an ionic liquid electrolyte is improved when including water as an additive. Herein we report for the first time the trend of improved cycling behavior of Na0/+ in Nmethyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide with 500 ppm H2O. The addition of water to this ionic liquid electrolyte promotes the breakdown of the [FSI]- anion towards beneficial SEI formation. The benefits during plating and stripping of sodium is observed as lower total polarization during symmetrical cell cycling and decreased electrode/electrolyte interface impedance. Sodium metal surfaces after cycling with 500 ppm H2O are shown to be smooth in morphology in comparison to lower additive concentrations. The outcome of adventitious moisture benefiting Na0/+ cycling in an ionic liquid, contrary to conventional electrolytes, allows flexibility in ionic liquid electrolyte design to the benefit of battery manufacturers.
1. Introduction A global focus towards alternative energy storage technologies has recently seen the ‘beyond lithium’ research area drastically expand. This is due to several drawbacks expected for future lithium ion technologies mainly pertaining to high cost, and low natural abundance/ access of starting materials [1]. As sodium and lithium share similarities in their alkali metal electrochemistry, there is rationale for renewed interest in the area of sodium batteries. It is now envisioned that ∗
sodium secondary cells, amongst others, will be implemented in parallel to lithium based devices in the near future [2,3]. However, in order to be competitive with future Li-ion technologies, it is essential to enable the highest possible energy densities for sodium based devices (1677 Wh kg−1 Na-O2) [4]. In doing so, we may harness the sodium metal as an anode rather than hard carbon materials which are currently used (ca. 100–300 mAh g−1) [5]. The path towards enabling sodium metal anodes has proven difficult, mainly due to the reactivity of sodium with water and/or the flammable conventional electrolytes
Corresponding author. Corresponding author. E-mail addresses: [email protected] (A. Basile), [email protected] (P.C. Howlett).
∗∗
https://doi.org/10.1016/j.jpowsour.2018.01.044 Received 10 December 2017; Received in revised form 14 January 2018; Accepted 18 January 2018 0378-7753/ © 2018 Elsevier B.V. All rights reserved.
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
conductivity values decrease an order of magnitude when compared to the neat C3mpyrFSI. This is clearly attributed to the rise in viscosity shown in Fig. 1b. This electrolyte still maintains a liquid phase as shown previously to be possible within highly concentrated ionic liquid systems [8]. The increased viscosity and concomitant drop in conductivity of the system can be reversed through the addition of water to the electrolyte. Here, a concentration of 500 ppm water within the 50 mol% electrolyte exhibits conductivity values close to that of the neat C3mpyrFSI -albeit with greater viscosity. Practical applications of batteries cause internal heat generation which in turn raises the device temperature above ambient conditions. These intermediate temperatures can allow improved electrode kinetics in the full device, and it is only the fact that the traditional organic electrolytes lead to instabilities at elevated temperatures which has limited battery operation to room tempera ture [18]. Fig. 1a reveals that at a temperature of 50 °C the 500 ppm water containing electrolyte is within the same order of magnitude to the neat electrolyte. This 50 °C intermediate temperature then used as a controlled temperature for all cell cycling herein, consistent with our previous reports for the purpose of drawing comparison [8,19]. The symmetrical cell cycling for these three 50 mol% NaFSI in C3mPyrFSI electrolytes shown in Fig. 2a compares the polarization of the cell containing < 20 ppm H2O against those considered ‘wet’, comprising of 100 ppm and 500 ppm of water. The electrolyte containing less than 20 ppm is considered dry in this instance. The initial polarization for each cell begins at high values, larger than 250 mV, with the highest belonging to the cell with dry electrolyte at ca. 360 mV. As is consistently observed with ionic liquid electrolytes this overpotential relaxes upon continued cycling to values ca. 100 mV after 20 cycles. It is at this point after cycling that the effect of water as an additive can be observed clearly. The overpotentials upon completing the final polarization step are ca. 80 mV, 100 mV, 150 mV for the electrolyte systems comprising of ca. 500 ppm < dry < 100 ppm respectively, and are shown in Fig. 2b. It is this most ‘wet’ electrolyte which exhibits the highest stability in the voltage profile during the 20 plating and stripping cycles, suggesting that the (i) electrode-electrolyte interface formed has lowest resistance, and/or that (ii) the addition of water to the electrolyte promotes diverse speciation and an increased diffusion for the Na+ ion. Investigations to determine the strength of these proposed mechanisms for beneficial cycling are currently underway in our labs. Unlike the dry and 100 ppm electrolyte, this cell's voltage-time curve shows very little sign of tailing at the completion of each polarization step. This is indicative of a beneficial solid-electrolyte interphase (SEI) which permits the ingress/ egress of sodium without Coulombic inefficiency. If this were not the case, the tailing (which is seen predominantly in the dry cell) would be more apparent at the completion of cell cycling, and suggestive of pristine sodium being stripped from the counter electrode [20]. This
carried forward from Li-ion technologies [6]. In order to remove this safety concern from future sodium metal devices, a new class of electrolyte have shown promise, namely highly concentrated ionic liquid electrolytes [7–9]. Recent results have shown highly concentrated ionic liquid electrolytes allow a large amount of charge migration in both lithium and sodium plating and stripping [8]. In addition, the passing of large charges has been reported to prepare a surface with a stable electrodeelectrolyte interface [7]. These ionic liquids are typically non-flammable and exhibit negligible vapor pressure thus promoting their use as safe electrolytes. Nonetheless, ionic liquid electrolytes still falter in their high viscosity and lower conductivity, the latter being compounded with the addition of such high salt concentrations [10]. Various additives are included to alleviate viscosity (H2O) [11–13], flammability (e.g. ionic liquids) [6], or promote beneficial interface formation via sacrificial breakdown (fluoroethylene carbonate, vinylene carbonate) [14]. Importantly, for ionic liquid electrolytes, a beneficial solid-electrolyte interphase has been shown to form without the use of such additives [15]. The effect of water has been studied in non-fluorinated ionic liquid electrolytes for both Li and Na cells [16,17]. Both of these reports concluded that an increase above ca. 100–200 ppm H2O within the ionic liquid electrolyte was detrimental to cell cycling and that the water content should be kept to a controlled minimum. Herein we report for the first time the stable plating and stripping of sodium metal from N-methyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide comprising of 50 mol% sodium bis(fluorosulfonyl)imide (1:1 NaFSI:C3mpyrFSI) with an additional 500 ppm H2O. These results are compared with ionic liquids bearing the FSI- anion paired with an alkyl phosphonium or an alkoxy ammonium cation. With this higher concentration of H2O, the viscosity and conductivity concern may be alleviated, enabling cells exhibiting lower overpotentials required for the Na0/+ redox process than previously reported for implementation in Na metal cells. Additionally, at a high concentration of 50 mol% Na+ salt the lower cost of NaFSI salt, compared to the IL, effectively reduces the cost of these electrolytes. 2. Results & discussion The physicochemical properties of the neat C3mpyrFSI ionic liquid and electrolyte systems are shown in Fig. 1 and describe the trend of improved conductivity and viscosity with the inclusion of water as an additive. The conductivity curves measured at 10 °C intervals in Fig. 1a show that neat C3mpyrFSI maintains the highest conductivity values at all temperatures. As is common to this class of electrolyte, the conductivity for the system can be seen to steadily increase through three orders of magnitude in the temperature range −20 °C to 120 °C. It is upon the addition of a high 50 mol% concentration of NaFSI that these
Fig. 1. (a) Conductivity and (b) viscosity values for the C3mpyrFSI ionic liquid (shown at right) and the 50 mol% NaFSI containing electrolytes including increased amounts of water.
345
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Fig. 2. Symmetrical Na|Na cell cycling for (a) 20 cycles highlighting the final polarization steps (b). Nyquist plots for the Na|Na cells comprising of (c) < 20 ppm (d) 100 ppm and (e) 500 ppm of water additive.
At higher magnification inset the surface is observed to be covered in deep narrow pits. In areas where the pitting is more severe (highlighted in Fig. S1), the sodium metal has a large surface area with high roughness. This roughness is increased in the case of the cell cycled using 100 ppm water electrolyte (Fig. 3b). No evidence of dendrites is observed. It is in viewing the sodium surface from the cell cycled with 500 ppm of water that a stark contrast can be made. This surface which provided the lowest resistance and overpotential during cycling is smoother and has lower pitting density than the other electrodes (Fig. 3c). The surface is also partially covered with residual electrolyte, and requires more vigorous washing, otherwise the ionic liquid electrolyte adheres to the sodium metal as shown in Fig. S2. In fact, at higher magnification inset Fig. 3c it can be seen that, what may be pits harbor residual IL observed having darker contrast. Overall, the smoothest surface prepared over 20 cycles, shows the best cycling performance. This is particularly interesting as the two rougher surfaces have greater relative surface area, and inherently would disperse the applied current density. Nonetheless, these 20 ppm and 100 ppm cells still maintain highest overpotentials during Na|Na cycling (Fig. 2a). The elemental distribution across the surface can be seen in Fig. 3d and e, and reveals a large occurrence of carbon and oxygen on the larger, rough sodium electrodeposits regardless of cycling in dry or ‘wet’ electrolytes, whereas the fluorine and sulfur share the same regions of the underlying sodium surface. In fact when comparing these surfaces, it is apparent that the breakdown of the anion occurs more favorably when the [FSI]- anion is in the presence of the water additive. Listed in Table 1 are the weight% values collected for each from EDS spectra shown in Fig. S3 revealing the increase of fluorine and sulfur breakdown products in particular. Quite interestingly these species are cohabitating the rough plated sodium in the ‘wet’ electrolyte (Fig. 3e),
increased stability is also observed when measuring for longer polarization times (1 h) over larger numbers of plating/strpping cycles, currently underway in our labs. It is interesting to note that the lower amount of water additive at ca. 100 ppm invokes a more negative impact on plating/stripping performance than the dry electrolyte. Contrary to that described for a similar ionic liquid based on the dicyanamide anion for both lithium and sodium electrochemistry [16,17]. This is more clearly visualized in the impedance data collected pre/post cycling each of the Na|Na cells. Prior to cycling, after an initial resting period of 24 h at 50 °C, the interfacial resistance (Rint) reaches ca. 260 Ω. This is also the case for the dry electrolyte (< 20 ppm, 305 Ω). However after cycling with the electrolyte containing 100 ppm H2O Rint falls to ca. 90 Ω and lower still for the dry cell at ca. 50 Ω. The slight difference in water content for these two cells is still enough to significantly alter the interface and affect a change in the plating/stripping. With an increase to 500 ppm (ca. 0.05% water by mass in electrolyte) the initial resting interface resistance is much larger ca. 505 Ω. The additional water in the electrolyte initially prepares a highly resistive interface, however this is not detrimental to symmetrical cell cycling as the resistance falls to ca. 28 Ω post plating/stripping. This phenomenon of decreased resistance post cycling has been reported when plating and stripping from highly concentrated ionic liquids, however the beneficial effect of water is yet to be described [8]. It is interesting to note that it is still possible to cycle sodium from these FSI− anion based ionic liquid electrolytes even with such a high concentration of water as an additive, unlike in the dicynanamide systems reported in the literature [16,17]. This is important from the perspective of cell design and construction, where adventitious moisture across cell manufacture will allow greater flexibility in electrolyte formulation, cell preparation and process design. 346
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Fig. 3. Scanning electron micrographs of the sodium electrode surfaces after 20 cycles in symmetrical orientation using electrolyte comprising water additive at a concentration of a, < 20 ppm (left), b, 100 ppm (center), and c, 500 ppm (right). Corresponding EDS mapping comparing the d, dry and e, ‘wet’ surfaces.
previously reported by our group [22] and is attributed to strong interactions between the ions, through strong coordination between Na+ and the functionalised ammonium cation carrying ethoxy/methoxy groups. This is also reflected in the larger overpotential for symmetrical cell cycling with this electrolyte. Upon completion of cell cycling the final overpotential for the cell comprising of the phosphonium electrolyte remained constant with only a small decrease to values near 80 mV. However, the cell with ammonium IL reached a stable polarisation potential only after ten cycles (ca. 6 h). The overpotential for the ammonium electrolyte cell decreased substantially yet still remains quite high at ca. 220 mV. These overpotential values are larger than in the pyrrolidinium case which is shown using a dashed line for comparison. Furthermore, the electrode-electrolyte interface (prior to cycling) on the sodium metal electrode when using the ammonium electrolyte with 500 ppm water shows a higher impedance (ca. 3200 Ω, Fig. 4c). Whereas in the ammonium after cycling, the interface resistance (shown in Fig. 4c) decreases to a much lower value of 300 Ω. This phenomenon is also shown in the phosphonium cell, Fig. 4d, in which the impedance associated with the electrode interface decreases after cycling to 40 Ω and indicates the formation of a beneficial SEI layer at the surface of the electrode -which permits facile ingress and egress of Na+. The effect of water in this P111i4FSI based ionic liquid electrolytes follows the same trend as for the C3mpyrFSI, as an additive towards beneficial SEI formation, showing little difference in the V-t profile when compared to the dry P111i4FSI electrolyte. However, in the N2(20201)3FSI ionic liquid this benefit is not observed. When compared to the dry N2(20201)3FSI electrolyte shown in Fig. S4, with ca. 20 ppm water, the dry Rint is smaller after cycling. This suggests that ionic liquids including the FSI− anion are capable of including water as an additive for beneficial cycling performance but their behavior depends also upon the ionic liquid's cation chemistry. In the N2(20201)3FSI case, the ether oxygens may coordinate the sodium ion more strongly and thus the water may have a less disruptive effect on the speciation and hence the nature of the species presenting to the electrode and the resultant SEI layer. This will be studied in further in future work. The effect of cation, anion and mixed salt is also being studied with respect to adventitious moisture in further detail targeting cost reduction of this class of ionic liquid electrolyte.
Table 1 Elemental distribution of the Na surface after 20 cycles (J = 1 mA cm−2) with 15 min polarization at 50 C°. Element
Na C N O F S
dry IL
‘wet’ IL
Weight/%
Weight/%
41.6 16.1 1.3 31.1 3.7 6.2
29.2 15.7 4.1 24.6 12.9 13.6
which does not occur in the dry (Fig. 3d). From that which we know regarding [FSI]- anion breakdown [21] for beneficial SEI growth at Li0 anodes [15], it can be postulated that the thiazate anion product, NSO− is providing the required passivation of high surface area sodium deposits required to permit lower surface impedance and stable cycling when in a superconcentrated and wet environment. The EDS analysis clearly shows that the presence of water modifies the SEI formed in these systems, supporting reduced surface impedance and better Na metal plating/stripping. Further studies will seek to understand the detail of these processes. To determine whether other FSI− based ionic liquids would also function as effective electrolytes with 500 ppm water, sodium symmetrical cells containing 55 mol% NaFSI in trimethyl(isobutyl)phosphonium (P111i4FSI) and 42 mol% NaFSI in N-ethyl-2-(2-methoxyethoxy)-N,N-bis(2-(2-methoxyethoxy)ethyl)-ethan-1-aminium bis(fluorosulfonyl)imide (N2(20201)3FSI) were prepared. These cells were then cycled at 50 °C at the same current density of 0.1 mA cm−2 with 15 min interval for each polarisation. These concentrations are chosen as they compare to the solubility limit and highest Na+ concentrations for these ionic liquids. Fig. 4 shows the cycling results and EIS before and after cycling in these ionic liquids based on different cations. The polarisation potential of the symmetrical cell with phosphonium IL is lower than its ammonium counterpart (Figure 4a, 80 mV vs 220 mV). This can be ascribed to the higher viscosity, lower conductivity and higher resistivity of the ammonium ionic liquid, which have been 347
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Fig. 4. The [N2(20201)3]+ and [P111i4]+ cations (top) and their performance in symmetrical Na|Na cells when using P111i4FSI and N2(20201)3FSI with 500 ppm H2O for (a) 20 cycles highlighting (b) the final cycle. Nyquist plots for the (c) N2(20201)3FSI and (d) P111i4FSI cells pre/post cycling.
3. Conclusions
4. Materials and methods
The use of water as an additive to alleviate viscosity concerns and boost the conductivity of ionic liquids has been achieved with larger than previously reported values up to 500 ppm. The variation from the dicyanamide anion to the bis(fluorosulfonyl)imide anion provides this allowance suggesting it is, at least in part, the breakdown of the anion in a wet environment that causes this change in behavior. This is shown across two different cations for C3mpyrFSI and P111i4FSI, and detrimental at 500 ppm for the N2(20201)3FSI. The cycling behavior of symmetrical cells containing wetted C3mpyrFSI perform comparably with the dry electrolyte system. Indicating that this class of IL is not disturbed by up to 500 ppm of H2O. The surface of the sodium metal electrode is also benefitted from the largest water addition, which provides an ideally smooth morphology during cycling, and the lowest surface impedance of all cells. This suggests an ideal SEI may be formed through the use of water as an additive in FSI− based highly concentrated ionic liquid electrolytes. An amount of ca. 500 ppm is shown to perform well for both improving bulk physicochemical properties of the electrolyte, in addition to improving the sodium surface for cycling. Ongoing research aims to determine the optimum water addition within this new class of electrolyte to the benefit of both (i) efficient charge transport and (ii) interface formation in sodium metal secondary cells.
4.1. Cell fabrication Sodium metal (Sigma, purity under paraffin oil) was rolled flat, polished using a nylon brush under hexane (RCI Leachem, 99%, 6 ppm H2O) and then punched into disks (8 mm diameter, 100 μm thickness). Symmetrical Na|Na cells were constructed using coin cell configuration (Hohsen, CR2032 stainless steel using 1 mm spacer and 1.4 mm spring). A polyethylene separator (gratis, Lydall, 7P03A, 50 μm thickness, 85% 0.3 μm porosity) was used to prevent short circuit and soaked with a 150 μL aliquot of electrolyte. All cell components were dried, and assembly took place within an argon glovebox (< 1 ppm H2O, < 1 ppm O2). All cells were allowed to rest for 24 h at 50 °C prior to cycling as a formation step. 4.2. Materials and methods N-methyl-N-propylpyrrolidinium bis(fluorosulfonyl)imide (C3mpyrFSI, Solvionic, 99.9%), potassium bis(fluorosulfonyl)imide (KFSI) (Suzhou Fluolyte China, > 99.9%), trimethyl(isobutyl)phosphonium dimethylphosphate (P111i4DMP) (Cytec Solvay Canada, mixture with dimethylphosphoric acid), acetonitrile (Sigma Aldrich Australia, > 99.8%, H2O = < 10 ppm) and dichloromethane (Sigma Aldrich Australia, > 99.8%) were used as received. Tris[2-(2-methoxyethoxy)ethyl]amine (Sigma Aldrich Australia, 95%) was freshly distilled over KOH and the colorless second fraction was used same-day to synthesize the ammonium salt. Milli348
Journal of Power Sources 379 (2018) 344–349
A. Basile et al.
Q water (18.2 MΩ cm−1) was used during all this work. P111i4FSI was synthesized through anion exchange between P111i4DMP and KFSI by known procedures [23]. The purity of P111i4FSI was confirmed by 1H, 13C, 19F Nuclear Magnetic Resonance Spectroscopy (NMR), mass spectrometry, melting point and potassium content (20 ppm, Ion Selective Electrode). N-ethyl-2-(2-methoxyethoxy)-N,Nbis(2-(2-methoxyethoxy)ethyl)-ethan-1-aminium bis(fluorosulfonyl) imide (N2(20201)3FSI)was synthesised by already known procedures by anion metathesis between N2(20201)3 Br and KFSI [24]. The purity of N2(20201)3FSI was confirmed by 1H, 13C, 19F NMR, mass spectrometry, potassium (15 ppm) and bromide (150 ppm) contents [22]. All NMR data for the synthesized ILs is present in the supplementary information. All ionic liquids were dried using Schlenk manifold techniques prior to salt and water additions. Sodium bis(fluorosulfonyl)imide (NaFSI, Solvionic, 99.9%) was added to each respective ionic liquid before an additional water removing step using Schlenk manifold techniques under vacuum with sodium hydride (Sigma Aldrich (60% dispersion in mineral oil). Hexane was stirred in the presence of calcium hydride overnight, and then distilled and kept over lithium metal in the argon glovebox (RCI Leachem, 99%). Water content < 20 ppm was considered as dry within this study. Water additions (Milli-Q, 18 MΩ cm−1) were made dropwise to spike up to 100 ppm or 500 ppm. All electrolyte formulation and storage takes place within an Ar environment using a glovebox.
its financial support through Australian Laureate Fellowship FL110100013 (MF) and FL120100019 (DRM) and under Australian Research Council's Discovery Projects funding scheme (DP160101178). All authors thank Peter Newman for instrument design. Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx. doi.org/10.1016/j.jpowsour.2018.01.044. References [1] B.L. Ellis, L.F. Nazar, Sodium and sodium-ion energy storage batteries, Curr. Opin. Solid State Mater. Sci. 16 (2012) 168–177. [2] B. Dunn, H. Kamath, J.-M. Tarascon, Electrical energy storage for the grid: a battery of choices, Science 334 (2011) 928–935. [3] V. Palomares, M. Casas-Cabanas, E. Castillo-Martínez, M.H. Han, T. Rojo, Update on Na-based battery materials. A growing research path, Energy Environ. Sci. 6 (2013) 2312–2337. [4] C. Pozo-Gonzalo, P.C. Howlett, D.R. MacFarlane, M. Forsyth, Highly reversible oxygen to superoxide redox reaction in a sodium-containing ionic liquid, Electrochem. Commun. 74 (2017) 14–18. [5] X. Dou, et al., Pectin, hemicellulose, or Lignin? Impact of the biowaste source on the performance of hard carbons for sodium-ion batteries, ChemSusChem 10 (2017) 2668–2676. [6] D. Monti, A. Ponrouch, M.R. Palacín, P. Johansson, Towards safer sodium-ion batteries via organic solvent/ionic liquid based hybrid electrolytes, J. Power Sources 324 (2016) 712–721. [7] G.M.A. Girard, et al., The role of Li concentration and the SEI layer in enabling high performance Li metal electrodes using a phosphonium bis(fluorosulfonyl)imide ionic liquid, J. Phys. Chem. C2 (2017), http://dx.doi.org/10.1021/acs.jpcc. 7b01929 acs.jpcc.7b01929. [8] A. Basile, et al., Extensive sodium metal plating and stripping in a highly concentrated Inorganic−Organic ionic liquid electrolyte through surface pretreatment, ChemElectroChem 4 (2017) 986–991. [9] A. Basile, et al., Enabling Sodium metal anodes using concentrated salt in ionic liquid electrolytes, The 67th Annual Meeting of the International Society of Electrochemistry, International Society of Electrochemistry, 2016. [10] F. Makhlooghiazad, et al., Phosphonium plastic crystal salt alloyed with a sodium salt as a solid-state electrolyte for sodium devices: phase behaviour and electrochemical performance, J. Mater. Chem. A 5 (2017) 5770–5780. [11] L. Suo, et al., ‘Water-in-salt’ electrolyte enables high-voltage aqueous lithium-ion chemistries, Science 350 (2015) 938–943. [12] S.S. Zhang, A review on electrolyte additives for lithium-ion batteries, J. Power Sources 162 (2006) 1379–1394. [13] F. Mizuno, T.S. Arthur, K. Takechi, Water in ionic liquid for electrochemical Li cycling, ACS Energy Lett. 1 (2016) 542–547. [14] M. Dahbi, et al., Black phosphorus as a high-capacity, high-capability negative electrode for sodium-ion batteries: investigation of the electrode/electrolyte interface, Chem. Mater. 28 (2016) 1625–1635. [15] A. Basile, A.I. Bhatt, A.P. O'Mullane, Stabilizing lithium metal using ionic liquids for long-lived batteries, Nat. Commun. 7 (2016). [16] A. Basile, H. Yoon, D.R. MacFarlane, M. Forsyth, P.C. Howlett, Investigating nonfluorinated anions for sodium battery electrolytes based on ionic liquids, Electrochem. Commun. 71 (2016) 48–51. [17] H. Yoon, et al., Lithium electrochemistry and cycling behaviour of ionic liquids using cyano based anions, Energy Environ. Sci. 6 (2013) 979. [18] M. Hilder, et al., Small quaternary alkyl phosphonium bis(fluorosulfonyl)imide ionic liquid electrolytes for sodium-ion batteries with P2- and O3-Na 2/3 [Fe 2/3 Mn 1/3 ]O 2 cathode material, J. Power Sources 349 (2017) 45–51. [19] C. Ding, et al., Na[FSA]-[C 3 C 1 pyrr][FSA] ionic liquids as electrolytes for sodium secondary batteries: effects of Na ion concentration and operation temperature, J. Power Sources 269 (2014) 124–128. [20] K.N. Wood, et al., Dendrites and pits: untangling the complex behavior of lithium metal anodes through operando video microscopy, ACS Cent. Sci. 2 (2016) 790–801. [21] I.A. Shkrob, T.W. Marin, Y. Zhu, D.P. Abraham, Why bis(fluorosulfonyl)imide is a ‘magic anion’ for electrochemistry, J. Phys. Chem. C 118 (2014) 19661–19671. [22] M. Hilder, et al., Effect of mixed anions on the physicochemical properties of a sodium containing alkoxyammonium ionic liquid electrolyte, Phys. Chem. Chem. Phys. 19 (2017) 17461–17468. [23] M. Hilder, et al., Physicochemical characterization of a new family of small alkyl phosphonium imide ionic liquids, Electrochim. Acta 202 (2016) 100–109. [24] M. Kar, et al., Ionic liquid electrolytes for reversible magnesium electrochemistry, Chem. Commun. 52 (2016) 4033–4036.
4.3. Instrumentation Water contents were measured using Karl Fisher potentiometric titration using a Model 756/831 KF Coulometer (MEP Instruments). Bromide content was measured by using IJ-Br combinational bromide selective electrode (Ionode) after calibration with 10 and 100 ppm bromide standards. Similarly potassium content was measured by Potassium Half-Cell Ion Selective Electrode after calibration with 10 and 100 ppm potassium standards. Viscosity data was collected using a rolling ball viscometer (Anton Paar Lovis 2000ME). Conductivity measurements were made using a hermetically sealed dip cell with embedded platinum electrodes (in-house designed) controlled using a MTZ-35 frequency response analyser (BioLogic) through a range of temperatures maintained by thermocouple. Samples were quenched prior to measurement at 10 °C steps with resting times of 2 h. All symmetrical Na|Na stripping/plating cycling was collected using a VMP3 MultiPotentiostat within temperature controlled ovens held constant at 50 °C. Electrochemical impedance spectroscopy was used to collect impedance of the symmetrical cells pre/post cycling. This was collected through a frequency range of 1 MHz–10 mHz by applying an alternating 10 mV RMS perturbation to the cell. All cycling experiments include duplicates for reproducibility. Purity of the ionic liquids was confirmed by NMR carried out using a DPX Spectrometer (Bruker) And electrospray ionization using Micromass Platform I IAPI QMS Electrospray mass spectrometer (Waters Corporation). All micrographs were collected using a JSM-IT300LV scanning electron microscope (JEOL Ltd Japan) using an accelerating voltage of 2.0 kV under high vacuum with secondary electron detector. All elemental mapping was performed within the SEM using a 50 mm2 X-Max energy dispersive spectroscopy detector (Oxford Instruments). Notes The authors declare no competing financial interest. Acknowledgment The authors acknowledge the ARC (Australian Research Council) for
349