Inorganic-Organic Ionic Liquid Electrolytes Enabling High Energy-Density Metal Electrodes for Energy Storage

Electrochimica Acta 220 (2016) 609–617

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Inorganic-Organic Ionic Liquid Electrolytes Enabling High Energy-Density Metal Electrodes for Energy Storage M. Forsytha,* , G.M.A. Girarda , A. Basilea , M. Hildera , D.R. MacFarlaneb , F. Chena , P.C. Howletta a b

Institute for Frontier Materials (IFM), Deakin University, Burwood, Victoria 3125, Australia School of Chemistry, Monash University, Victoria 3800, Australia

A R T I C L E I N F O

Article history: Received 31 July 2016 Received in revised form 17 October 2016 Accepted 20 October 2016 Available online 21 October 2016 Keywords: ionic liquids lithium battery sodium phosphonium electrolytes solid-electrolyte interphase

A B S T R A C T

It has recently been shown, in the case of the bis(fluorosulfonyl)amide (FSI) based ionic liquids, that as the concentration of the alkali metal salt (LiFSI or NaFSI) is increased, the alkali metal cation transference number increases, despite an increase in viscosity and decrease in conductivity. At the same time significant enhancements in electrochemical stability and rate performance of devices are also observed. Here we overview some of the recent findings already in the literature and in addition demonstrate the feasibility of stable, high rate room temperature lithium battery cycling in an electrolyte comprised of 60 mol% LiFSI in a trimethyl, isobutyl phosphonium FSI ionic liquid using a high voltage NMC cathode. We also demonstrate that the high rate cycling of lithium and sodium metal in these phosphonium FSI electrolytes leads to a nanostructured anode deposit and a lowering of the interfacial impedance, suggesting a stable SEI layer formation. Finally, we propose a hypothesis that may explain some of the observations thus made, by which the high alkali ion concentration in these mixed electrolyte systems leads to the effective elimination of the mass transport limitations that are chiefly responsible for the formation of dendrites in traditional electrolytes. This work suggests that a new type of ionic liquid consisting of a mixture of metal cations with organic cations can provide a solution to the instability of the reactive alkali metal anodes and hence enable higher energy density technologies. ã 2016 Elsevier Ltd. All rights reserved.

1. Introduction With our ever-increasing push to move away from fossil fuels to cleaner energy sources, there is a need to improve energy storage technologies. In particular, with respect to electrochemical energy storage, safer, longer cycle-life and cost effective technologies need to be developed. In some applications, such as transportation, improved energy density and higher capacity is also critical if EVs are to be a realistic alternative. Higher capacity can be achieved by moving from intercalation electrodes, such a graphite, towards metal electrodes, eg. lithium or sodium and these can be coupled with higher capacity cathodes such as sulphur or air systems. The use of metal anodes with traditional organic electrolytes, however, is not feasible from the point of view of safety, and also stable, long term cycling is not possible [1–4]. Room Temperature Ionic Liquids (RTILs) based on the bis(trifluoromethanesulfonyl) amide (TFSI) or the bis(fluorosulfonyl)amide (FSI) anions have

* Corresponding author. E-mail address: [email protected] (M. Forsyth). http://dx.doi.org/10.1016/j.electacta.2016.10.134 0013-4686/ã 2016 Elsevier Ltd. All rights reserved.

recently been of particular interest as alternative electrolytes for lithium rechargeable batteries due to their superior safety properties in opposition to commonly used organic carbonate electrolytes [5–10]. While such ILs have been shown to facilitate efficient cycling of lithium metal, poor cycling performance at higher rates, related to poor transport properties (high viscosity, low ionic conductivity) has mainly prevented them from being used as electrolytes in commercial devices. Thus, the search for lower viscosity, higher conductivity electrolytes has led to the investigation of new ionic liquids [11]. Novel FSI ionic liquids (ILs) based on a phosphonium cation have been reported with promising properties that make them candidates for alternative electrolytes [9,11–13]. These features include improved transport properties and wider electrochemical windows in comparison to the more commonly studied cyclic ammonium-based (eg pyrrolidinium) ILs. One of the critical factors dictating the cycling performance of a lithium battery is the lithium salt concentration present in the electrolyte. In general an increase in salt concentration leads to an increase in viscosity and decrease in conductivity, but it also

610

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

greatly increases the amount of lithium cations available in the electrolyte. Indeed, highly concentrated electrolytes (>1 M salt concentration in solvent) have only been studied in the last few years [11,14–16]. In 2013, Yoon et al. reported for the first time the cycling performance at room temperature of a Li | LiCoO2 cell with a solution of up to 3.2 mol kg 1 of lithium salt in an IL electrolyte. The IL used was an ammonium-based IL, N-propyl-N-methylpyrrolidinium bis(fluorosulfonyl)imide (C3mpyrFSI). The cells with the highest salt content IL electrolyte showed better rate capability, even at high C rates (3C and 5C), than that of a cell with an organic electrolyte. More recently Na metal batteries have been considered as an alternative energy storage technology, with FSI based ionic liquids being demonstrated as suitable electrolytes for sodium metal cycling, in particular at higher concentrations of dissolved NaFSI. Indeed, Ding et al. reported 123 mAh g 1 for a NaCrO2 cathode using a 25 mol% NaFSI in C3mpyrFSI electrolyte (at 20 mAg 1 and 363 K). Thus, in both Na and Li cases, these high alkali salt concentration ionic liquid electrolytes present unique properties that lead to long term stable and efficient cycling. High concentration electrolytes such as these can be considered as a new dimension in ionic liquids research whereby the alkali salt concentrations can be more than 50 mol% and lead to speciation and structures which alter the basic ion transport mechanisms and perhaps decouple the alkali ion transport from vehicular transport. These mixed inorganic-organic cation ionic liquid electrolytes may also present a unique interface between the electrolyte and the reactive metal anode which influences its dissolution and deposition processes. In this paper, we summarise some of the transport and physicochemical properties of mixtures of ionic liquids with LiFSI and NaFSI salts and demonstrate effective cycling of alkali metal anodes in these metal cation/phosphonium cation based FSI ionic liquids. 2. Experimental Given that this article attempts to bring together already published work as well as highlighting some very recent, as yet unpublished data, it was considered appropriate to provide a detailed methodology section in the supplementary information. Here we briefly outline the nature of the experiments discussed. Ionic conductivity of the ionic liquid electrolytes is typically measured using a.c. impedance spectroscopy, as discussed in the references cited herein [11,15]. Vibrational spectroscopy is also used to probe ion speciation as a function of concentration in the LiFSI based ionic liquids using ATR-FTIR techniques. Symmetric cell cycling for both Li and Na cells, to determine the cycling stability in different electrolyte formulations, has been previously described [11,15,17] and is also presented in this paper. The plated and stripped surfaces of both Li and Na are examined using SEM techniques, following careful washing with a dimethyl carbonate (DMC) solvent. Full cell cycling of lithium metal batteries were carried out using a high voltage LiNi1/3Mn1/3Co1/3O2 (NMC, L&F Material Co. Ltd) using an Al current collector (Battery use, Hohsen). The charge-discharge tests of the Li cells discussed below were conducted with 4.2–2.75 V cut-off voltages at 0.5C, 1C, 2C and 4C current rates at 25  C. 3. Results and Discussion 3.1. Conductivity and transport The effect of LiFSI and NaFSI concentration on the conductivity in FSI ionic liquids is summarised in Fig. 1 from previous literature data [18–20]. This series of electrolytes includes a comparison of the C3mpyrFSI and trimethylisobutylphosphonium (P111i4FSI) ionic

Fig. 1. Comparison of conductivity in LiFSI and NaFSI containing ionic liquid electrolytes as a function of concentration; A+ = Li+ or Na+. Electrolytes include LiFSI in *-C3mpyrFSI [14], and P111i4FSI at ^- 25  C and ~- 50  C. NaFSI in C3mpyrFSI at &- 25  C and !- 50  C [20].

liquids. In all cases the addition of the alkali metal salt decreases the ionic conductivity, however, at the higher concentrations there appears to be little difference between the pyrrolidinium and phosphonium based systems (comparing the Yoon and Girard data at 25  C). For the most part, the decrease in conductivity reflects an increase in viscosity with increasing alkali metal ion concentration which results from strong coordination of the Li+ or Na+ ion to the anions and ultimately strong association and clustering, as has previously been reported and is discussed further below [18–20]. Interestingly, at room temperature the Na+ ion has a more dramatic effect on reducing conductivity (increasing viscosity) than the Li+ ion, likely due to the higher coordination number demanded by the larger Na+ cation [21]. At higher temperatures however, the differences appear less pronounced. Table 1 summarises the viscosity and alkali cation (Li+ or Na+) transference number measurement as determined using a modified Bruce-Vincent method [22,23] for a series of ionic liquid electrolytes in comparison with a typical organic electrolyte. Most significantly, t+ increases substantially with increasing concentration of the Li+ ion and approaches 0.5 at 50 mol% and above. In comparison, t+ for the organic electrolyte is 0.26. In the case of NaFSi in C3mpyFSI, tNa+ measured both by Yoon et al. and Hagiwara Table 1 Viscosity and Li/Na transference number of various electrolytes at 25  C obtained from various literature sources. The organic electrolyte (1 M LiPF6 in EC-DMC) values are included for comparison. Electrolyte

h/mPa s ( 10%) t(Li+/Na+)

1 M LiPF6-EC-DMC Neat P111i4FSI 0.5 mol kg 1 LiFSI in P111i4FSI 1.0 mol kg 1 LiFSI in P111i4FSI 2.0 mol kg 1 LiFSI in P111i4FSI 3.2 mol kg 1 LiFSI in P111i4FSI 3.8 mol kg 1 LiFSI in P111i4FSI 0.7 mol kg 1 LiFSI in P2225TFSI 0.7 mol kg 1 LiFSI in P222(2O1)TFSI Neat C3mpyrFSI

5.1 (<10) 41 55 79 163 323 571 – – 33

0.26  0.07 – – 0.28  0.05 0.30  0.06 0.40  0.01 0.46  0.04 0.44  0.05 0.54  0.04 –

[25] [18] [18] [18] [18] [18] [18] [12] [12] [26,27]

96 163 299 780 – 49.4 85.7 136 253

0.01  0.001 0.07  0.005 0.19  0.03 0.31  0.01 0.32  0.01 0.11  0.01 0.14  0.006 0.185  0.03 0.183  0.008

[17,20] [17,20] [17,20] [17,20] [17] [15,19] [15,19] [15,19] [15,19]

0.8 mol kg 1.4 mol kg 2.2 mol kg 3.2 mol kg 4.0 mol kg 0.8 mol kg 1.6 mol kg 2.4 mol kg 3.2 mol kg

1

NaFSI in C3mpyrFSI NaFSI in C3mpyrFSI 1 NaFSI in C3mpyrFSI 1 NaFSI in C3mpyrFSI 1 NaFSI in C3mpyrFSI 1 LiFSI in C3mpyrFSI 1 LiFSI in C3mpyrFSI 1 LiFSI in C3mpyrFSI 1 LiFSI in C3mpyrFSI 1

Reference

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

et al. [20,24] is 0.3 at 50 mol%, equivalent to tLi+ in an organic electrolyte systems. This difference in transport number in essentially equivalent systems possibly reflects the different coordination environment and larger size of the Na+ compared with Li+. 3.2. The effect of alkali metal salt concentration on speciation and transport Ion speciation has been investigated in FSI based electrolytes using FTIR spectroscopy. Yoon et al. showed [14] that increasing Li+ concentration in the C3mpyr system, led to a preference for the cis conformer, which has been shown to have lower binding energy than the trans conformer [14]. This was attributed to the depletion of available FSI for metal ion coordination. Thus, further characterisation of the electrolytes was conducted using Fourier transform infrared (FTIR) spectroscopy in order to investigate the interactions between the three ions present in the IL electrolytes: Li+ P111i4+ and FSI . FTIR spectra for the IL electrolytes and pure LiFSI salt are presented in Fig. 2(a) and (b). The majority of peaks were assigned based on the assignments for a solution of LiFSI in C3mpyrFSI reported by Yoon et al. [14]. Peak assignments for the FTIR spectra of neat P111i4FSI and P111i4FSI based electrolytes are presented in Table S1. Fig. 2(a) presents the IR spectra for the different salt concentrations. The results showed S-N-S stretching vibration peaks (asymmetric and symmetric stretching vibrations at 828 and 731 cm 1 respectively) shifting to a higher wavenumber with increasing Li salt concentration. The S-F asymmetrical stretching, previously assigned by Yoon et al. to the peak at 830 cm 1 [14], shifts closer to the peak position found for the pure LiFSI salt. Fig. 2(b) shows the FTIR spectra for the asymmetric SO2 bending vibrations located at 566 cm 1. The spectra reveal the appearance of a shoulder at 595 cm 1 with increasing intensity as the LiFSI concentration increases. These observations suggest changes in speciation between Li+ and FSI ions that were confirmed by diffusion NMR analysis [11]. These results are consistent with those previously reported for the C3mpyrFSI [14] ionic liquid electrolytes suggesting very little effect of the ionic liquid cation on lithium coordination at these high LiFSI concentrations. This previous work suggested that a change in coordination from Li(FSI)3 to Li(FSI)2 species concomitant with a change from predominantly a trans towards a cis conformation of the FSI anion occurred with increasing Li salt. Interestingly, with increasing salt towards a 1:1 by mole concentration, the available FSI anions to satisfy the coordination environment of the alkali metal cation (between 3 and 5) will become fewer and insufficient unless each anion begins to coordinate in a multidentate fashion or to more than one metal cation, for example by bridging two or even three Li+ ions. This would result in significant aggregation and structuring that may also explain the changes observed in the vibrational spectroscopy reported here and in previous publications [14] and may also impact on the transport mechanism of the alkali metal cation.

611

Fig. 3 presents a snapshot of ionic structure in a mixed ionic liquid composed of 50 mol% LiFSI and 50 mol% C3mpyrFSI generated from a molecular dynamics simulation recently reported [21]. Significant aggregation is observed at these high concentrations. The left side of the image omits the pyrrolidinium cation for clarity and the right side includes the molecular surfaces of the pyrrolidinum cations. Extensive ion aggregation is apparent at the high concentrations of LiFSI which is completely absent at 10 mol% LiFSI [21] and is consistent with the changes indicated by the FTIR spectra; the more extensive association is only possible through the FSI anion coordinating with two or more Li+ cations such that Lix(FSI]y(x y)+ species exist as aggregates. Such aggregation has been shown, via MSD analysis of the MD trajectory files, to result in relatively higher diffusivities for the Li+ or Na+ ions in highly concentrated, mixed salt electrolytes, compared to the ionic liquid species, and ultimately results in a higher tLi+ consistent with the transference number measurements determined electrochemically [14,21]. 3.3. Lithium metal cycling and battery performance Cycling measurements using symmetric metal cells are usually conducted to investigate the ability of a given electrolyte to sustain efficient plating and stripping of the alkali metal and the effect of rate (ie. current density) on the stability of this process. The magnitude of the cell overpotential and the evolution of acquired impedance spectra reflect the contributions and stability of mass transport and surface impedances during cell cycling. As an example, we present the behavior in the phosphonium ionic liquid, P111i4FSI, with 1.0 and 3.8 mol kg 1 of LiFSI. These were compared to an organic carbonate reference electrolyte commonly used in Li batteries: 1 M LiPF6 in EC:DMC (1/1 vol/vol) (Fig. S1). Cells were cycled at different current densities (j) ranging from 0.05 mA cm 2 to 1.0 mA cm 2 (0.05, 0.1, 1.0 and 0.05 mA.cm 2 corresponding to the amounts of charge of Li reduced/oxidised of 6.4  10 3, 1.3  10 2 and 0.13 mg cm 2) for 30 min (one plating process for 30 min and one stripping process for 30 min each) at room temperature (25  C). The potential of the Li metal working electrode (WE) during Li plating/stripping cycles in a Li | Li cell and EIS spectra of the Li WE are shown in Fig. 4. The EIS impedance data were recorded each time the current density was varied. It can be see here that, for any given applied current density, the polarisation remains relatively stable over 10 plating/stripping processes for each current density and that the voltage profiles under these conditions are very similar in the mixed IL electrolyte and the organic electrolyte. However, the overall impedance in the case of the IL electrolyte is significantly reduced. Fig. 5 presents the same cycling experiment for cells cycled at a current density of 1.5 mA cm 2 at 2 h intervals (i.e., 3 mAh cm 2) for 50 cycles, comparing the organic electrolyte with the mixed electrolytes (3.8 mol kg 1) at 25 and 50  C. Under these more demanding conditions, the advantage of the mixed electrolyte is obvious; at 25  C the polarisation of the

Fig. 2. FTIR transmittance spectra: (a) from 900 to 680 cm

1

; (b) from 700 to 500 cm 1.

612

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

Fig. 3. MD simulation images for 50 mol% LiFSI in C3mpyrFSI showing explicitly the Li+ (dark pink) and FSI (stick) ions. The right side of the image includes the molecular surface of the C3mpyr cations. The extensive association of the Li+ and FSI is apparent.

Fig. 4. (a) Voltage time profiles for a Li | Li symmetrical cell with 1.0 mol kg 1 and (b) 3.8 mol kg 1 LiFSI in P111i4FSI over 10 plating/stripping processes at j = 0.05, 0.1, 1.0 mA cm 2 and EIS spectra of the cell with (c) 1.0 mol.kg 1 and (d) 3.8 mol.kg 1 LiFSI in P111i4FSI, before and after cycling. All cells were cycled at 25  C.

carbonate electrolyte based cell begins to gradually increase whereas the high concentration ionic liquid system steadily decreases. These observations suggest that a more resistive interface forms at the metal surface in contact with in the carbonate electrolyte and this behaviour has previously been shown to result from the build-up of isolated Li deposits (‘dead’ lithium) and SEI/electrolyte decomposition components [28–31]. Even more dramatically, at 50  C the cycling behaviour becomes

unstable, indicating that the build-up/decomposition process is exacerbated at higher temperature in the EC/DMC electrolyte. In contrast, the 3.8 mol kg 1 LiFSI in P111i4FSI electrolyte continues to cycle with high stability and remarkably low polarisation. The EIS data measured at 25  C after 10 and 50 cycles in these electrolytes suggest that the impedance is significantly lower for the ionic liquid electrolytes as compared with the carbonate electrolyte; this is consistent with the lower polarisation. The longer term cycling

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

613

Fig. 5. A comparison of voltage profiles during subsequent lithium plating/stripping processes in Li | Li symmetrical cells at 25  C (left) and 50  C (right) at j = 1.5 mA cm (3 mAh cm 2) for 50 cycles. (a) and (b) 1 M LiPF6 in EC:DMC (1/1 vol/vol); (c) and (d) 3.8 mol kg 1 LiFSI in P111i4FSI electrolyte.

at 3 mAh cm 2 shows that these mixed electrolytes can sustain a significant rate of charging and discharging without failing. This implies that the lithium metal can be plated reversibly without the formation of ‘dead’ lithium or shorting due to dendrite growth.

2

SEM was used to investigate the surfaces of the Li metal electrodes after 50 cycles at 50  C for current densities of 1.5 mA cm 2 and 12 mA cm 2 (3 mAh cm 2 and 6 mAh cm 2 respectively). The charged Li metal surfaces (i.e., the electrode

Fig. 6. SEM images (EHT = 2.00 and 5.00 kV) of the plated Li electrode after 50 cycles at j = 1.5 mA cm rinsing with dimethylcarbonate.

2

and q = 3 mAh cm

2

in 3.8 mol kg

1

of LiFSI in P111i4FSI at 50 C after

614

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

Fig. 7. SEM images (EHT = 2.00 kV) of the plated Li electrode after 450 cycles at j = 12 mA cm 2 and q = 6 mAh cm 2 in 3.8 mol kg 1 of LiFSI in P111i4FSI at 50 C after rinsing with dimethylcarbonate.

surface obtained after the final deposition polarisation) are shown in Figs. 6 and 7. The first dramatic observation is that there is remarkable uniformity across the entire electrode with no evidence of dendrite formation; higher current densities appear to lead to even more uniform, smooth Li metal deposits. These images also show the deposit cross-sections. These indicate that, for the lower current densities, an approximately 20 mm thick deposit of lithium metal is present and approximately 100 mm in the case of the higher current density cycling. The highest magnifications clearly show a compact, nano-structured deposit. Such a deposit must still be protected by an SEI layer, and indeed a combination of NMR, FTIR and XPS characterisation of these surfaces [18] revealed the formation of a chemically stable SEI, derived from reaction of the Li metal surface with the IL components, after the first cycle. The SEI composition and thickness were found to be stable with increasing number of cycles and seem to be composed of an outer layer of the FSI anion reduction products e.g., Li2S, Li2NSO2F as well as Li2O, LiOH, Li2CO3 and LiF, as has previously been reported for pyrrolidinium FSI based IL electrolytes [32,33]. This SEI layer clearly supports the stable cycling of Li metal even under very harsh conditions. A model of how this might occur is presented further below. Figs. 8 and 9 demonstrate the potential of the high salt concentration, mixed salt electrolytes in lithium metal devices using a typical NMC high voltage electrode formulation. The performance is compared with a 1 M LiPF6 in EC-DMC (LP30) electrolyte. Remarkably, despite a significantly higher viscosity, the mixed salt electrolyte performs comparably even at room

temperature. The evolution of the charge and discharge capacity as a function of cycle number at different discharge rates at 25  C is shown in Fig. 8. The performance of the highly concentrated IL electrolyte was superior to that of the organic carbonate electrolyte at moderate discharge rates (i.e. 0.2C), maintaining a higher charge/discharge capacity. It should be noted that these capacity values were reversibly maintained after increasing the discharge rate up to 4C and decreasing it again down to 0.2C. Both the standard carbonate electrolyte and the concentrated ionic liquid electrolyte exhibited a comparable discharge capacity at faster rates from 1C to 2C. Indeed, even though the 1 M LiPF6 in EC-DMC standard electrolyte has what one would think of as more appropriate transport properties (i.e. higher ionic conductivity and lower viscosity) for rapid charge/discharge, the concentrated phosphonium IL system exhibited comparable discharge capacities. However, on increasing the discharge rate to 4C, a more significant decrease in specific capacity was observed with the highly concentrated IL electrolyte compared with the more fluid organic electrolyte. Nevertheless, it should be noted that extensive cycling of the organic electrolyte cell would be limited by the Li metal electrode stability (as shown in Fig. 5), particularly in the case where higher more practical cathode loadings are applied. This difference is indeed highlighted for extended cell cycling experiments as shown in Fig. 9. Here we see the charge-discharge curves of lithium battery cells containing the highly concentrated P111i4FSI IL electrolyte, which can be compared with cells containing the LiPF6 in EC-DMC electrolyte. The cells exhibit typical charge-discharge behaviour, however, the capacities obtained with the phosphonium electrolyte over long-term cycling (200 cycles) are significantly more stable than those for the LiPF6 in EC-DMC electrolyte. As shown in Fig. 9(b), the cell containing the phosphonium P111i4FSI electrolyte with high Li salt content indicated high discharge capacities (over 150 mAh g 1 at the first cycle) with an average capacity retention of 90% after 100 cycles. The coulombic efficiencies also approached 100% (ca. 98–99%) after the third cycle and are maintained above 95% over the 200 cycles. This is significantly superior to the overall cycling performance of the cell containing the LiPF6 in EC-DMC electrolyte used in Fig. 9(a), showing that the P111i4FSI electrolyte leads to an improvement in cycling efficiency and capacity retention for this NMC cathode at 1C despite the higher fluidity and conductivity of the LP30 electrolyte. 3.4.

23

Na cycling in mixed salt electrolyte systems

The cycling behaviour observed for Li metal in high salt concentration LiFSI systems is not unique to Li. Below we present some early data for a 45 mol% NaFSI salt in P1i444FSI ionic liquid at 50  C. Fig. 10 shows that the first charging of the Na electrode in this mixed salt electrolyte leads to a relatively large polarisation potential, however this quickly decreases to stable polarisation

Fig. 8. Dependence of charge-discharge capacity on cycle number at 25  C at different C-rates (0.2C, 1C, 2C and 4C) for Li | NMC cells containing (a) 1 M LiPF6 in EC-DMC (LP30) and (b) solution of 3.8 mol kg 1 of LiFSI in P111i4FSI electrolytes.

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

Fig. 9. Charge-discharge curves of Li | NMC cells at 25  C at 1C (CC mode, 0.25 mA cm electrolytes.

2

): (a) 1 M LiPF6 in EC-DMC (LP30) and (b) solution of 3.8 mol kg

615

1

of LiFSI in P111i4FSI

metal surface, a relatively compact, dendrite free layer is observed with nano-structuring evident at higher magnifications. 3.5. Morphology development and influence of high alkali metal ion concentration

Fig. 10. Voltage time profiles for a Na | Na symmetrical cell with 45 mol% NaFSI P1i444FSI over 12 plating/stripping processes at 0.5 mA cm 2 for a polarisation duration of 1 h.

potentials of less than 100 mV with stable cycling and shows no sign of failure at least for 12 cycles. This initial large polarisation potential is directly related to the current density applied to the sodium metal electrodes, with a lower current density requiring more plating/stripping processes to reach a stable polarisation potential (i.e. 100 mV). Fig. 10 shows that a current density relating to a charge of 0.5 mAh cm 2 is effective to reach the required degree of morphology change to permit efficient sodium cycling in successive cycles. The surface of the cycled Na electrode after washing with DMC is presented in Fig. 11. As with the charged Li

The morphology of both the Na surfaces and the Li surfaces shown above suggest a nucleation and growth mechanism of the electrodeposit that leads to nanostructuring and compact layer formation, rather than dendrite formation. This may arise from the fact that there is such a high concentration of the alkali ions present at the interface to begin with. Such a high Li+/Na+ concentration, coupled with a transport mechanism which is less prone to mass transport limitations as a result of faster diffusion through a structural rearrangement mechanism rather than vehicular diffusion, would lead to a high nucleation rate of the metal deposit. The nanostructured nature of the deposit may be a result of the slower overall diffusion of the non electrochemically active ions (i.e. the phosphonium ions) and/or the possibility of rapid SEI layer formation upon nucleation of the Li metal particle due to the potential for greater reactivity of a high surface area Li with neighbouring FSI anions. Furthermore, the first discharge (stripping) cycle will result in yet a further increase of the alkali metal ion concentration adjacent to the electrode, that, given the slower overall dynamics of the electrolyte, may lead to slow diffusion away from the interface. When the current is reversed, these ions are then able to rapidly deposit as Li or Na metal. A schematic model of the Li (or likewise Na) metal/electrolyte interface and the formation of the plated metal surface is shown in Fig. 12. These notions will be further explored using molecular

Fig. 11. Plated Na metal surfaces post cycling with 0.5 mAh cm

2

of charge during plating/stripping for 12 cycles.

616

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

Fig. 12. Schematic model of the metal electrolyte interface with a mixed salt electrolyte.

dynamics simulation and in-situ spectroscopy including in-situ NMR in future work.

program. We are also indebted to Cytec for supply of ionic liquids and their precursors.

4. Conclusions

Appendix A. Supplementary data

We have presented the physicochemical properties and electrochemical performance of a series of FSI based ionic liquid electrolytes in which the concentration of NaFSI or LiFSI salt mixed with the organic FSI salt approaches or exceeded 50 mol% and thus the metal cation becomes the majority cation in the mixture. In these mixed cation ionic liquid electrolytes, we observed high metal ion transference numbers, despite the lower conductivity and the higher viscosity. Interestingly the phosphonium systems seems to result in higher transference numbers for equivalent concentrations compared to the pyrrolidinium based ILs, although we need to acknowledge the difficulty in obtaining accurate values from the traditional Bruce-Vincent-Evans technique commonly used. MD simulations support a change in diffusion mechanism at these high concentrations resulting from structural rearrangement and hopping of the alkali metal cation, rather than vehicular transport. The symmetric cell cycling for both Li and Na in these mixed cation electrolytes showed high rate performance and the reversible deposition and stripping of more than 100 mm layers of Li metal in these electrolytes. The SEM images demonstrated smooth metal surfaces with cross-sections showing a nanostructured metal deposit on the charged anode surface. Full Li metal cells using NMC cathodes were shown to cycle efficiently with capacities of 130 mAh g 1 for 200 cycles. These materials deserve further investigation from both a fundamental understanding of the structure and speciation of the ions in the mixture, as well as in application in metal anode based devices.

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. electacta.2016.10.134.

Acknowledgements This work was funded at various stages by the Australian Research Council under the discovery schemes (DP130101652 and DP160101178) and the Linkage project LP120200181. MF and DRM are grateful to the ARC for funding through the Laureate Fellowship

References [1] D. Aurbach, A. Zaban, A. Schechter, Y. Eineli, E. Zinigrad, B. Markovsky, The Study of Electrolyte-Solutions Based on Ethylene and Diethyl Carbonates for Rechargeable Li Batteries. 1. Li Metal Anodes, J Electrochem Soc 142 (9) (1995) 2873–2882. [2] J.B. Goodenough, Evolution of Strategies for Modern Rechargeable Batteries, Accounts Chem Res 46 (5) (2013) 1053–1061. [3] W. Xu, J.L. Wang, F. Ding, X.L. Chen, E. Nasybutin, Y.H. Zhang, J.G. Zhang, Lithium metal anodes for rechargeable batteries, Energ Environ Sci 7 (2) (2014) 513–537. [4] D.P. Lv, Y.Y. Shao, T. Lozano, W.D. Bennett, G.L. Graff, B. Polzin, J.G. Zhang, M.H. Engelhard, N.T. Saenz, W.A. Henderson, P. Bhattacharya, J. Liu, J. Xiao, Failure Mechanism for Fast-Charged Lithium Metal Batteries with Liquid Electrolytes, Adv Energy Mater 5 (3) (2015). [5] K. Tsunashima, F. Yonekawa, M. Kikuchi, M. Sugiya, Effect of Quaternary Phosphonium Salts in Organic Electrolyte for Lithium Secondary Batteries, Electrochemistry 79 (6) (2011) 453–457. [6] D.R. MacFarlane, N. Tachikawa, M. Forsyth, J.M. Pringle, P.C. Howlett, G.D. Elliott, J.H. Davis, M. Watanabe, P. Simon, C.A. Angell, Energy applications of ionic liquids, Energ Environ Sci 7 (1) (2014) 232–250. [7] H.-B. Han, S.-S. Zhou, D.-J. Zhang, S.-W. Feng, L.-F. Li, K. Liu, W.-F. Feng, J. Nie, H. Li, X.-J. Huang, M. Armand, Z.-B. Zhou, Lithium bis(fluorosulfonyl)imide (LiFSI) as conducting salt for nonaqueous liquid electrolytes for lithium-ion batteries: Physicochemical and electrochemical properties, J Power Sources 196 (7) (2011) 3623–3632. [8] M.A. Navarra, Ionic liquids as safe electrolyte components for Li-metal and Liion batteries, Mrs Bull 38 (7) (2013) 548–553. [9] K. Tsunashima, Y. Sakai, M. Matsumiya, Physical and electrochemical properties of phosphonium ionic liquids derived from trimethylphosphine, Electrochem Commun 39 (0) (2014) 30–33. [10] D.R. MacFarlane, M. Forsyth, P.C. Howlett, M. Kar, S. Passerini, J.M. Pringle, H. Ohno, M. Watanabe, F. Yan, W. Zheng, S. Zhang, J. Zhang, Ionic liquids and their solid-state analogues as materials for energy generation and storage, Nature Reviews Materials 1 (2016) 15005. [11] G.M.A. Girard, M. Hilder, H. Zhu, D. Nucciarone, K. Whitbread, S. Zavorine, M. Moser, M. Forsyth, D.R. MacFarlane, P.C. Howlett, Electrochemical and physicochemical properties of small phosphonium cation ionic liquid

M. Forsyth et al. / Electrochimica Acta 220 (2016) 609–617

[12]

[13]

[14]

[15]

[16]

[17]

[18] [19] [20]

[21]

[22] [23]

electrolytes with high lithium salt content, Phys Chem Chem Phys 17 (2015) 8706–8713, doi:http://dx.doi.org/10.1039/C5CP00205B. V.L. Martins, N. Sanchez-Ramirez, M.C.C. Ribeiro, R.M. Torresi, Two phosphonium ionic liquids with high Li+ transport number, Phys Chem Chem Phys 17 (35) (2015) 23041–23051. M. Hilder, G.M.A. Girard, K. Whitbread, S. Zavorine, M. Moser, D. Nucciarone, M. Forsyth, D.R. MacFarlane, P.C. Howlett, Physicochemical characterization of a new family of small alkyl phosphonium imide ionic liquids, Electrochim Acta 202 (2016) 100–109. H. Yoon, A.S. Best, M. Forsyth, D.R. MacFarlane, P.C. Howlett, Physical properties of high Li-ion content N-propyl-N-methylpyrrolidinium bis (fluorosulfonyl)imide based ionic liquid electrolytes, Phys Chem Chem Phys 17 (6) (2015) 4656–4663. H. Yoon, P.C. Howlett, A.S. Best, M. Forsyth, D.R. MacFarlane, Fast Charge/ Discharge of Li Metal Batteries Using an Ionic Liquid Electrolyte, J Electrochem Soc 160 (10) (2013) A1629–A1637. J. Qian, W.A. Henderson, W. Xu, P. Bhattacharya, M. Engelhard, O. Borodin, J.-G. Zhang, High rate and stable cycling of lithium metal anode, Nat Commun 6 (2015). H. Yoon, H. Zhu, A. Hervault, M. Armand, D.R. MacFarlane, M. Forsyth, Physicochemical properties of N-propyl-N-methylpyrrolidinium bis (fluorosulfonyl)imide for sodium metal battery applications, Phys Chem Chem Phys 16 (24) (2014) 12350–12355. G.M.A. Girard, Investigation of phosphonium bis(fluorosulfonyl)imide-based ionic liquid electrolytes for lithium batteries, Deakin University, 2016. H.G. Yoon, A study of pyrrolidinium bis(fluorosulfonyl)imide based ionic liquids for lithium metal batteries, Monash University, 2013. C. Ding, T. Nohira, R. Hagiwara, K. Matsumoto, Y. Okamoto, A. Fukunaga, S. Sakai, K. Nitta, S. Inazawa, Na[FSA-C3C1pyrr][FSA] ionic liquids as electrolytes for sodium secondary batteries: Effects of Na ion concentration and operation temperature, J Power Sources 269 (2014) 124–128. F. Chen, M. Forsyth, Elucidation of transport mechanism and enhanced alkali ion transference numbers in mixed alkali metal-organic ionic molten salts, Phys Chem Chem Phys 18 (28) (2016) 19336–19344. P.G. Bruce, C.A. Vincent, Polymer Electrolytes, J Chem Soc Faraday T 89 (17) (1993) 3187–3203. J. Evans, C.A. Vincent, P.G. Bruce, Electrochemical Measurement of Transference Numbers in Polymer Electrolytes, Polymer 28 (13) (1987) 2324–2328.

617

[24] K. Matsumoto, Y. Okamoto, T. Nohira, R. Hagiwara, Thermal and Transport Properties of Na[N(SO2F)2-N-Methyl-N-propylpyrrolidinium][N(SO2F)2] Ionic Liquids for Na Secondary Batteries, The Journal of Physical Chemistry C 119 (14) (2015) 7648–7655. [25] M. Dahbi, F. Ghamouss, F. Tran-Van, D. Lemordant, M. Anouti, Comparative study of EC/DMC LiTFSI and LiPF6 electrolytes for electrochemical storage, J Power Sources 196 (22) (2011) 9743–9750. [26] A.I. Bhatt, A.S. Best, J. Huang, A.F. Hollenkamp, Application of the N-propyl-Nmethyl-pyrrolidinium Bis(fluorosulfonyl)imide RTIL Containing Lithium Bis (fluorosulfonyl)imide in Ionic Liquid Based Lithium Batteries, J. Electrochem. Soc. 157 (2010) A66–A74. [27] A. Nazet, S. Sokolov, T. Sonnleitner, T. Makino, M. Kanakubo, R. Buchner, Densities, Viscosities, and Conductivities of the Imidazolium Ionic Liquids [Emim][Ac], [Emim][FAP], [Bmim][BETI], [Bmim][FSI], [Hmim][TFSI], and [Omim][TFSI], Journal of Chemical & Engineering Data 60 (8) (2015) 2400– 2411. [28] X.-B. Cheng, R. Zhang, C.-Z. Zhao, F. Wei, J.-G. Zhang, Q. Zhang, A Review of Solid Electrolyte Interphases on Lithium Metal Anode, Advanced Science 3 (3) (2016) n/a-n/a. [29] G. Bieker, M. Winter, P. Bieker, Electrochemical in situ investigations of SEI and dendrite formation on the lithium metal anode, Phys Chem Chem Phys 17 (2015) 8670–8679, doi:http://dx.doi.org/10.1039/C4CP05865H. [30] P.C. Howlett, D.R. MacFarlane, A.F. Hollenkamp, High lithium metal cycling efficiency in a room-temperature ionic liquid, Electrochemical and Solid-State Letters 7 (5) (2004) A97–A101. [31] P.C. Howlett, N. Brack, A.F. Hollenkamp, M. Forsyth, D.R. MacFarlane, Characterization of the lithium surface in N-methyl-N-alkylpyrrolidinium bis(trifluoromethanesulfonyl) amide room-temperature ionic liquid electrolytes, J Electrochem Soc 153 (3) (2006) A595–A606. [32] A. Budi, A. Basile, G. Opletal, A.F. Hollenkamp, A.S. Best, R.J. Rees, A.I. Bhatt, A.P. O'Mullane, S.P. Russo, Study of the Initial Stage of Solid Electrolyte Interphase Formation upon Chemical Reaction of Lithium Metal and N-Methyl-N-PropylPyrrolidinium-Bis(Fluorosulfonyl)Imide, J Phys Chem C 116 (37) (2012) 19789– 19797. [33] A. Basile, A.I. Bhatt, A.P. O’Mullane, A Combined Scanning Electron Micrograph and Electrochemical Study of the Effect of Chemical Interaction on the Cyclability of Lithium Electrodes in an Ionic Liquid Electrolyte, Aust J Chem 65 (11) (2012) 1534–1541.