Available online at www.sciencedirect.com

ScienceDirect

Green Energy & Environment 4 (2019) 146e153www.keaipublishing.com/gee

Research paper

Ionic liquid/tetraglyme hybrid Mg[TFSI]2 electrolytes for rechargeableMg batteries

Zheng Ma a, Maria Forsyth b, Douglas R. MacFarlane a, Mega Kar a,*a School of Chemistry, Faculty of Science, Monash University, VIC 3800, Australia

b ARC Centre of Excellence for Electromaterials Science, IFM – Institute for Frontier Materials, Deakin University, Burwood VIC 3125, Australia

Received 10 September 2018; accepted 10 October 2018

Available online 17 October 2018

Abstract

The electrochemical reversibility of Mg in hybrid electrolytes based on mixtures of ionic liquid and glyme based organic solvents wasinvestigated for applications in rechargeable magnesium batteries (RMBs). The electrolytes demonstrate reversible reduction and oxidation ofMg only after being pre-treated with the dehydrating agent, magnesium borohydride, Mg[BH4]2, highlighting the importance of removing waterin Mg based electrolytes. The addition magnesium di[bis(trifluoromethanesulfonyl)imide] (Mg[TFSI]2) (0.3 M) to N-butyl-n-methyl-pyrroli-dinium bis(trifluoromethanesulfonyl)imide [C4mpyr][TFSI]/tetraglyme at a mole ratio of 1:2 showed stable CV cycling over almost 300 cycleswhile scanning electron microscopy (SEM) and X-ray diffraction (XRD) confirmed Mg deposition, showing non-dendritic morphology and awell-aligned growth. Further thermogravimetric analysis (TGA) demonstrated a mass retention of 79% at 250 �C for this electrolyte suggestingthat the presence of the ionic liquid increases thermal stability substantially making these hybrid electrolytes compatible for RMBs.© 2018, Institute of Process Engineering, Chinese Academy of Sciences. Publishing services by Elsevier B.V. on behalf of KeAi Communi-cations Co., Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Mg[TFSI]2; Ionic liquids; Tetraglyme; Rechargeable Mg battery; Mg[BH4]2 dehydrate reagent

1. Introduction

Among battery devices, Li-ion batteries have been givenspecial attention since they can deliver a large amount ofenergy in a small volume, as required in smartphones, laptopsand other portable devices [1]. Rechargeable magnesium batte-ries (RMBs) are promising alternative candidates, since Mg ischeap, safe and also has high volumetric energy capacity(3833 mA h cm�3) [2].

A major obstacle to the development of RMBs is findingsuitable electrolytes that are compatible with the Mg metalanode and enable high coulombic efficiency and long cycle life.Ethereal based solvents have been widely explored as Mgelectrolytes since ether oxygens can coordinate to Mg2þ ions in

* Corresponding author.

E-mail address: mega.kar@monash.edu (M. Kar).

https://doi.org/10.1016/j.gee.2018.10.003

2468-0257/© 2018, Institute of Process Engineering, Chinese Academy of Sciences

Ltd. This is an open access article under the CC BY-NC-ND license (http://creativ

the presence of various Mg salts to support reversible Mg elec-trochemistry [3–9]. Examples include Mg organohaloaluminateelectrolytes [3], Mg–Al chloro-complex (MACC) electrolytes[9], Mg[BH4]2 electrolytes [8] and Mg carborane electrolytes[10] dissolved in ethers, such as tetrahydrofuran (THF), ordimethoxyethane (DME). However, some of these electrolyteseither contain very toxic components or are corrosive, have lowanodic stability due to the presence of [BH4]

� (E0 ¼ 1.7 V vs.Mg) [8], or can be costly to synthesise, making them not prac-tical for high voltage batteries.

Ionic liquids (ILs) generally show negligible vapour pressureand flammability, endowing them with high thermal stability[11].While many ILs can be expensive, their role as co-solventsin an organic medium can reduce the cost, increase the thermalstability, extend the electrochemical operating window andimprove the conductivity of the electrolyte [12]. By furtherchanging the functional groups of the cations or anions, ILs canbe endowed with specific physical and electrochemical

. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co.,

ecommons.org/licenses/by-nc-nd/4.0/).

147Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

properties [13]. In recent studies [14,15], Mg[BH4]2/IL elec-trolytes demonstrated that ether-functionalised ILs supportreversible Mg reduction and oxidation. Kar et al. demonstratedMg cycling from ILs based on multiple ethereal functionalgroups attached to a quaternary ammonium cation [15], whileWatkins and co-authors synthesised PEGylated pyrrolidiniumbased ILs to achieve reversible Mg electrochemistry with anefficiency ~ 90% [14]. In both these studies, ether oxygen co-ordination to Mg2þ ions was found to play a key role in thereversibility of Mg. However, in both cases, limited by theoxidative potential of Mg[BH4]2, these electrolytes offer onlylimited voltage range for a device.

In comparison, the bis(trifluoromethanesulfonyl)imideanion ([TFSI]�), with its strong charge delocalization, is stableat more anodic potentials than Mg[BH4]2 [16] and can in somecases, also be used in electrolytes with high Mg2þ concen-tration, thus having the potential to increase the charging ca-pacity in RMBs [17]. Some researchers have attempted todemonstrate cycling of Mg2þ/Mg in Mg[TFSI]2/IL systems;however results have been contradictory [18–21]. The strongionic interactions, for instance between Mg2þ ions and tetra-fluoroborate [BF4]

� or [TFSI]� [18,21], were speculated toaccount for the high over-potential of Mg reduction, whichpromotes the decomposition of the electrolyte, leading to apassivating layer formed on the Mg electrode. In contrast,when Mg[TFSI]2 salts were dissolved in a mixture of an ILand a “glyme” solvent such as tetraglyme, reversible Mgelectrochemistry was seen [22–24]. Since ether oxygens of theglyme can dominate in the coordination with the Mg2þ ions,the ion-pairing interactions between Mg2þ ions and the[TFSI]� ions are weakened. These effects can contribute toimprove the accessibility of Mg2þ ions and allow them toparticipate in Mg reduction at a relatively low over-potential[22]. Nevertheless, the electrochemical performance of theabovementioned IL/glyme-based Mg[TFSI]2 electrolytes stillsuffers from poor cycling efficiency, i.e. < 30% over 10 cycles[23].

It is anticipated that impurities such as water can inhibitefficient cycling of Mg. Mg[TFSI]2 salts can easily crystallizeas the hexahydrate (i.e. Mg[TFSI]2$6H2O) [25] and are diffi-cult to dry, the water being tightly bound to the Mg2þ ion [26].The presence of H2O in Mg[TFSI]2 electrolytes will readilyreact with Mg(0), forming inactive Mg[OH]2 and MgO on theelectrode, which is a likely the cause of the poor cycling ef-ficiency. Thus Mg[TFSI]2 requires a rigorous drying treatmentto achieve stable electrochemical cycling of Mg. Previousexamples of such treatments include removing H2O by theaddition of molecular sieves [26] or by reacting Cl� basedelectrolytes with tiny amounts of Bu2Mg [27]. Mainly itshowed that to achieve high efficiency and stable cycling inMg[TFSI]2 electrolytes, a significant amount of Cl� is pref-erable in different organic solvent based system [28–31].Apart from the design of the electrolytes, a recent studyshowed that it is possible to achieve charge–discharge in a Mg/V2O5 coin cell with a chemically engineered Mg anode in Mg[TFSI]2 electrolytes [32], although the thermal stability of thisorganic solvent based electrolytes still need to be improved for

practical use. Other than these studies, our recent study of Mg[TFSI]2 in tetraglyme demonstrated that a trace amount of Mg[BH4]2 as a dehydrating reagent is essential for reversible Mgcycling and achieving high coulombic efficiency [16]. Weshowed that difficult to remove trace amounts of water in theelectrolyte could successfully be removed by addition of suf-ficient Mg[BH4]2 to react with the amount of water and thatreversible cycling of Mg could then be achieved.

In an effort to move away from pure organic solvents such astetraglyme, the goal of this work is to initially investigate theelectrochemical properties ofMg fromMg[TFSI]2 in two ILs: a)(2-methoxyethoxy)-N,N-bis(2-(2-methoxyethoxy)ethyl)-ethan-1-aminium bis(trifluoromethylsulfonyl)imide [N2(20201)3][TFSI] as previously studied by us and b) N-methyl-n-butylpyrrolidinium bis(trifluoromethanesulfonyl)imide [C4mpyr][TFSI], a more conventional IL that is widely explored as abattery solvent [33–37]. In an effort to further improve theelectrochemical performance, the mixing of these ILs withcoordinating, ether-based organic solvents, to form hybridelectrolytes was also studied.

2. Experimental information

2.1. Preparation of chemicals

Tetraglyme (Sigma–Aldrich, � 99%) and dimethyl poly-ethylene glycol (DMPEG-250, Sigma–Aldrich, � 99%) werevacuum distilled (~0.8 mbar at 140–165 �C) over sodium(Sigma–Aldrich) with benzophenone (Sigma–Aldrich) as in-dicator, and additionally dried with 3 Å molecular sieves(Sigma–Aldrich) for 72 h to obtain a water content < 10 ppm(measured by Karl Fisher titration). Tri-alkoxy ammoniumbis(trifluoromethanesulfonyl)imide ([N2(20201)3][TFSI]) wassynthesised by the method according to the literature [15] andwas dried on a Schlenk line at 35 �C for 48 h before transferringinto an Ar glovebox (IL water content ~ 20 ppm according toKarl Fischer titration). N-methyl-n-butyl pyrrolidinium bis(-trifluoromethanesulfonyl)imide [C4mpyr][TFSI] was used asreceived (Solvionic, 99.5%) (water content ~ 40 ppm accordingto Karl Fischer titration). Mg[TFSI]2 (Solvionic, 99.5%) wasdried on a Schlenk line at 80 �C for 48 h. Mg[BH4]2 was used asreceived (Sigma–Aldrich, 95%).

All the chemicals were transferred into an Ar glovebox(H2O level < 1 ppm and O2 level < 1 ppm) for electrolytepreparation and electrochemical tests.

2.2. Electrolyte preparation

Different Mg[TFSI]2 electrolytes were prepared by addingvarious concentrations of Mg[TFSI]2 to either IL, glymes or amixture of the IL and glymes, and stirring at room temperaturefor up to 12 h to obtain a homogeneous sample. A controlelectrolyte was prepared with Mg[TFSI]2 and Mg[BH4]2 in[C4mpyr][TFSI] and propylene carbonate (99.7%, Sigma–

Aldrich, used as received).0.3 M Mg[TFSI]2/tetraglyme and 0.3 M Mg[TFSI]2/

[C4mpyr][TFSI] electrolytes were prepared as comparison

148 Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

samples for physical property studies. No Mg[BH]4 was addedin these experiments.

According to our previous study [16], Mg[TFSI]2 electro-lytes need to be treated with traces of Mg[BH4]2 to removewater and other impurities, to achieve reversible Mg electro-chemistry. Thus, all Mg[BH4]2-optimized electrolytes in thiswork were prepared by adding trace amounts of Mg[BH4]2step by step, and the optimum amount was determined fromthe cyclic efficiency evaluated from cyclic voltammetry (asshown in Supplementary Information).

2.3. Electrochemical measurements

All the electrochemical experiments were conducted on aVMP Potentiostat with Bio-lab software with a 3-electrodesystem. A Pt disk electrode (1 mm diameter) was used as theworking electrode, and Mg ribbon, scraped until shiny, wasused as the counter and reference electrodes. The cyclic vol-tammetry (CV) was carried out at a scan rate of 25 mV s�1.

A galvanostatic electrodeposition study was carried at acurrent of 1.5 mA cm�2 for 24 h in the electrolytes by using aglassy carbon working electrode (with a removable head), Ptwire counter electrode and Mg ribbon reference electrode. Thedeposit product was washed with dimethoxyethane (Sigma–

Aldrich, � 99%, H2O < 10 ppm) in the glovebox beforeconducting SEM and XRD characterization under driedconditions.

2.4. Characterization

Thermogravimetric analysis (TGA) (Mettler Toledo,Australia) was conducted over a temperature range of 25–

550 �C in a Nitrogen gas atmosphere at a heating rate of5 �C min�1. All the electrolyte samples were tested withoutMg[BH4]2 addition, as Mg[BH4]2 is not stable during themeasurement. The lack of Mg[BH4]2 was hypothesised not togreatly change the thermal stability because it was only usedfor dehydration, and the optimized additions were small(0.6 wt% maximum).

Density measurements were performed using an AntonPaar DMA 5000 density meter for all samples, from 25 �C to90 �C.

The viscosity was measured using an Anton Paar AMVnfor all samples from 25 �C to 90 �C.

The ionic conductivity of the samples was evaluated usingAC impedance spectroscopy in a frequency range of 0.1 Hz to10 MHz using a dip cell. The measurements were performedwith a frequency response analyzer, Solartron 1296, driven bySolartron impedance measurement software version 3.2.0. Thetemperature range was 25–90 �C for all the samples.

Scanning electron microscopy (SEM) and energy disper-sive spectroscopy (EDS) were conducted using a field emis-sion scanning electron microscope (Magellan 400 FEGSEMinstrument). X-ray diffraction patterns (XRD) were obtainedon a Bruker D8 ADVANCE ECO powder Xray instrumentusing Cu Ka radiation (l ¼ 0.15418 nm) in the 2q range from25� to 60� with a scanning step size of 0.01�.

3. Results and discussion

In our previous study we demonstrated that reversible Mgreduction and oxidation can be observed in Mg[TFSI]2/tetra-glyme after being treated with Mg[BH4]2 [16]. Therefore, inthis work, we investigate whether the addition of Mg[BH4]

2 toan IL/Mg[TFSI]2 electrolyte in which Mg electrochemicalactivity is otherwise absent, can facilitate Mg reduction andoxidation, even in the absence of ether solvents [15]. However,as shown in Fig. 1a, the electrochemical reversibility of Mgwas barely observable in the [N2(20201)3][TFSI] electrolyte(black curve) even with the addition of excess Mg[BH4]2(0.15 M). The reduction peak observed at an over-potential of�1.2 V vs. Mg is speculated to be decomposition of electro-lyte. No reductive or oxidative currents were observed at all in[C4mpyr][TFSI] indicating an immediate passivation in thiscase. In contrast, the addition of tetraglyme made a significantimprovement to the electrochemistry of Mg in the treated Mg[TFSI]2/IL electrolytes (Fig. 1b). As depicted in the blackcurve, by using a mixture of [N2(20201)3][TFSI]/tetraglyme(mole ratio 1:1), the electrolyte showed reversible Mg reduc-tion and oxidation. Moreover, in the presence of the moreconductive [C4mpyr][TFSI] (2.8 mS cm�1 vs. 1.2 mS cm�1

for [N2(20201)3][TFSI] at 25 �C) higher current densitieswere achieved (Fig. 1b, red curve). Thus, the presence oftetraglyme in these electrolytes appears to be crucial inachieving reversible Mg cycling. It is worth noting that no Mgreduction and oxidation was observed when tetraglyme wasreplaced with propylene carbonate (Fig. S1), indicating thatthe role of tetraglyme is beyond that of a simple ‘diluent’.These results suggest that a coordinating co-solvent such astetraglyme is needed, in order to disrupt the Mg–TFSI inter-action, which is otherwise sufficiently strong to limit the ac-tivity of the Mg ion.

Fig. 1b illustrates that, when compared to [N2(20201)3][TFSI], in addition to producing higher current densities,lower over-potentials can be achieved in [C4mpyr][TFSI].This is also evident when varying the concentration of Mg[BH4]2 in the electrolyte (Figs. S2 and S3). The best result wasobtained at the highest concentration i.e. 19 mM Mg[BH4]2 inMg[TFSI]2/[C4mpyr][TFSI]/tetraglyme (mole ratio 1:1).Herein our work will focus on this [C4mpyr][TFSI] basedhybrid electrolyte, comparing the role of different glymes andstudying the effect of varying concentration of Mg[BH4]2 andMg[TFSI]2 salts in the electrolyte. A similar study for the[N2(20201)3][TFSI] hybrids is shown in the supplementaryinformation (Fig. S2).

Since the concentration of dissolved Mg2þ ions in theelectrolyte plays an essential role in the electrochemicalbehaviour of Mg2þ [38], different amounts of Mg[TFSI]2 wereadded into the [C4mpyr][TFSI]/tetraglyme (mole ratio 1:1)hybrid solvents and subsequently treated with Mg[BH4]2.According to the pictures of the samples shown in Fig. 2a(0.1–0.5 M Mg[TFSI]2 in [C4mpyr][TFSI]/tetraglyme (moleratio 1:1)), Mg[TFSI]2 showed a limited solubility of ~0.3 M,and beyond this concentration a white precipitate wasobserved. It is important to note that different dissolved

(a)

0.0

-0.5

-1.0

-1.5-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

J / m

A cm

-2

E / V vs. Mg

6

4

2

0

-2

-4

-6-1.0 -0.5 0.0 0.5 1.0 1.5

E (V vs. Mg)

0.15 M Mg[TFSI]2 / 0.15 M Mg [BH4]2

[N2(20201)3][TFSI][C4mpyr][TFSI]

0.15 M Mg[TFSI]2/0.15 M Mg [BH4]2

[N2(20201)3][TFSI] / Tetraglyme[C4mpyr][TFSI] / Tetraglyme

J (m

A cm

-2)

6

4

2

0

-2

-4

-6-1.0 -0.5 0.0 0.5 1.0 1.5

E / V vs. Mg

0.15 M Mg[TFSI]2 / 0.15 M Mg [BH4]2

[N2(20201)3][TFSI]/tetraglyme[C4mpyr][TFSI]/tetraglyme

J / m

A cm

-2

(b)

Fig. 1. Cyclic voltammetry (CV) performance in mixed electrolytes of 0.15 M Mg[TFSI]2/0.15 M Mg[BH4]2: (a) using [N2(20201)3][TFSI] or [C4mpyr][TFSI] as

the solvents and (b) using [N2(20201)3][TFSI]/tetraglyme (mole ratio 1:1) or [C4mpyr][TFSI]/tetraglyme (mole ratio 1:1) as the solvents. Pt WE, Pt CE, Mg RE,

25 mV s�1, Ar atmosphere.

(a)

(b) 8

6

4

2

0

-2

-4

-6

-8

-1 0 1 2 3

0.1 M Mg[TFSI]2, 8 mM

0.2 M Mg[TFSI]2, 8 mM

0.3 M Mg[TFSI]2, 19 mM

0.4 M Mg[TFSI]2, 18 mM

0.5 M Mg[TFSI]2, 14 mM

E / V vs. Mg

0.1 M 0.2 M 0.3 M 0.4 M 0.5 M

J / m

A cm

-2

Fig. 2. (a) Digital images of Mg[TFSI]2 electrolytes in [C4mpyr][TFSI]/tet-

raglyme (mole ratio 1:1) hybrid solvent at different Mg[TFSI]2 concentrations,

0.1�0.5 M. (b) Cyclic voltammetry of Mg[BH4]2-optimized electrolytes in

[C4mpyr][TFSI]/tetraglyme (mole ratio 1:1) at different Mg[TFSI]2 concen-

trations, 0.1�0.5 M. The concentration of Mg[BH4]2 added to the electrolytes

are indicated. Pt WE, Pt CE, Mg RE, 25 mV s�1, Ar atmosphere.

8

4

0

-4

-8

-12-1 0 1 2 3

E / V vs. Mg

[C4mpyr][TFSI]/tetraglyme,19 mM Mg[BH4]2

[C4mpyr][TFSI]/DMPEG-250,34 mM Mg[BH4]2

J / m

A cm

-2

Fig. 3. CV performance in 0.3 M Mg[TFSI]2 electrolytes using the addition of

various glymes to [C4mpyr][TFSI] (mole ratio 1:1). All the electrolytes were

treated with Mg[BH4]2 as shown. Pt WE, Pt CE, Mg RE, 25 mV s�1, Ar

atmosphere.

149Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

concentrations of Mg[TFSI]2 will release different amounts ofwater present in the solution. Thus, the minimum concentra-tion of Mg[BH4]2 required to treat these electrolytes will vary.According to the CV results shown in Fig. 2b and summarizedin Table S2, the 0.3 M Mg[TFSI]2 sample displayed the best

performance in terms of the cycling efficiency and activity inthe Mg reduction process, and thus is used as the optimisedsalt concentration (containing 19 mM Mg[BH4]2) for thefollowing study.

Fig. 3 compares CVs of the treated 0.3 M Mg[TFSI]2/[C4mpyr][TFSI] in two different glyme based solvents. Forcomparison dimethyl polyethylene glycol (DMPEG, Mw

average ~250) was chosen as an alternative to tetraglyme, due tothe former having a higher thermal stability (Fig. S5). However,as shown in Fig. 3 and Fig. S4, while 0.3 M Mg[TFSI]2/([C4mpyr][TFSI]/DMPEG-250, 1:1 mol ratio) showed similarreversible Mg reduction and oxidation behaviour to that oftetraglyme, lower currents and a lower coulombic efficiency(summarized in Table S1) were observed. The difference maybe due to the large ether chain in DMPEG-250 more readilychelating to the metal ion. A similar observation was made by

150 Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

Kar et al. [39] in their work on electrochemical cycling of zincin ether based ionic liquids.

Although the boiling point of tetraglyme is relatively high, itsthermal stability is not ideal for safe and stable batteries. Theaddition of ionic liquid to tetraglyme is expected to increase thethermal stability of the electrolyte, but at the cost of somereduction in the conductivity of the Mg2þ ions. Thus an opti-mum ratio of IL to solvent will achieve a balance between thetwo properties [40]. Thermal, transport and electrochemicalproperties are compared in Fig. 4 as a function of mol fraction ofIL in the [C4mpyr][TFSI]:tetraglyme þ 0.3 M Mg[TFSI]2hybrid electrolyte. The viscosity of the various ratios as afunction of temperature is shown in Fig. 4a. The addition of Mg[TFSI]2 to neat [C4mpyr][TFSI] increases the ion interactionsand thus the viscosity. Upon the addition of various fractions oftetraglyme the viscosity decreases as a result of a simple diluenteffect. The conductivity as a function of temperature of thedifferent samples is compared in Fig. 4b. Generally, an increasein viscosity corresponds to a decrease in conductivity and maybe caused by enhanced ion–ion interactions in ILs [41,42].Interestingly, the conductivity values of some of the electrolytesin Fig. 4b do not correlate to their viscosity observed in Fig. 4a.For instance, the neat IL, [C4mpyr][TFSI] has the highestconductivity above 50 �C despite its high viscosity. Additionally

1000/T / K-1

T / °C

(a)

(c)

2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4

Log

(Vis

cosi

ty /

Poi

se)

0.5

0.0

-0.5

-1.0

-1.5

-2.0

-2.5

-3.0

-3.5

100

80

60

40

20

0

0 100 200 300 400 500

90 °C

50 °C25 °C

1:01:11:21:5[C4mpyr][TFSI] (No salt)

Mg [TFSI]2

[C4mpyr][TFSI] : Tetraglymeat

0:11:51:21:1

Wei

ght /

%

(

Fig. 4. (a) Viscosity and (b) Conductivity of 0.3 M Mg[TFSI]2 electrolytes with di

[TFSI] (no salt) in the temperature range: 25–90 �C. (No Mg[BH4]2 was added for

[C4mpyr][TFSI]/tetraglyme mole ratios, compared with pure Mg[TFSI]2 salt. (d) C

different solvent ratios; the optimized Mg[BH4]2 concentration is indicated in the

0.3 M Mg[TFSI]2 in [C4mpyr][TFSI]:tetraglyme (1:5) appearsto be the most fluid, whilst its conductivity is one of the lowestamongst all the electrolytes. This is especially evident at highertemperatures. A slight decrease in conductivity is also observedat elevated temperatures for electrolytes comprising of 0.3 MMg[TFSI]2 in [C4mpyr][TFSI]:tetraglyme (1:2) and (1:1). Adecrease in conductivity in the presence of tetraglyme may be aresult of some chelation by the ether oxygens to Mg2þ ions,reducing the mobility of the metal ions. In other cases, theaddition of certain molecular solvents to ILs may induce ionaggregation, reducing the number of charge carriers present,thereby reducing the ionicity [43]. Further investigation of thisby spectroscopic methods including Raman and NMR willenable us to determine the coordination and diffusivity of Mg2þ

ions and will be the focus of a future work.Fig. 4c shows the TGA traces of the various 0.3MMg[TFSI]2

hybrid electrolytes. The TGA curve of Mg[TFSI]2 showed themain decomposition starting at 345 �C. According toWatanabe'swork, in Mg[TFSI]2/tetraglyme solution (< 2.7 M), tetraglymeexists both as the component of coordinated [Mg(tetraglyme)][TFSI]2 complex and as free solvent molecules [26]. Althoughtetraglyme has a boiling point of ~275 �C, its vapour pressure is60.5 Pa at 95 �C [44]. Thus, 0.3 M Mg[TFSI]2/tetraglymeelectrolytes without any [C4mpyr][TFSI] (indicated as 0:1 in

1000 ( T / K-1)

E (V vs. Mg)

2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4

Log

(con

duct

ivity

/ S

cm

-1) -1.8

-2.0

-2.2

-2.4

-2.8

-3.0

-2.6

-1 0 1 2 3

9

6

3

0

-3

-6

-9

90 °C50 °C

25 °C1:01:11:21:5[C4mpyr][TFSI] (no salt)

[C4mpyr][TFSI] :Tetraglyme1:5, 15 mM Mg [BH4]2

1:2, 18 mM Mg [BH4]2

1:1, 19 mM Mg [BH4]2

J(m

A cm

-2)

1000/T / K-1

E / V vs. Mg

2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4

Log

(Con

duct

ivity

/ S

cm

-1) -1.8

-2.0

-2.2

-2.4

-2.8

-3.0

-2.6

-1 0 1 2 3

9

6

3

0

-3

-6

-9

90 °C50 °C

25 °C1:01:11:21:5[C4mpyr][TFSI] (No salt)

[C4mpyr][TFSI] :Tetraglyme1:5, 15 mM Mg [BH4]2

1:2, 18 mM Mg [BH4]2

1:1, 19 mM Mg [BH4]2

J/ m

A cm

-2

b)

(d)

fferent [C4mpyr][TFSI]:tetraglyme mole ratios, compared with neat [C4mpyr]

these measurements). (c) TGA of 0.3 M Mg[TFSI]2 electrolytes with different

yclic voltammetry of Mg[BH4]2-optimized 0.3 M Mg[TFSI]2 electrolytes with

figure for each sample. Pt WE, Pt CE, Mg RE, 25 mV s�1, Ar atmosphere.

-0.6 0.0 0.6 1.2 1.8 2.4

16

8

0

-8

-16

-24

1st

41st

81st

121st

161st

201st

241st

281st

J / m

A cm

-2

E / V vs. Mg

Fig. 5. Long term CV cycling in the electrolytes of 0.3 M Mg[TFSI]2 in

[C4mpyr][TFSI]/tetraglyme (mole ratio 1:2) for over 280 cycles. Pt WE, Pt

CE, Mg RE, 25 mV s�1, Ar atmosphere.

151Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

Fig. 4c) shows two steps including (1) from 95 �C to 278 �Ccorresponding to the evaporation of uncoordinated tetraglyme,and (2) from 278 �C to 390 �C, being the evaporation of tetra-glyme from the coordinated [Mg(tetraglyme)][TFSI]2 complexand some decomposition of Mg[TFSI]2. An increase in thermalstability is observedwith an increasing ratio of [C4mpyr][TFSI]/tetraglyme. Beyond 410 �C, the weight loss is assigned to thedecomposition of [C4mpyr][TFSI] [11]. Overall, the massretention is significantly influenced by the [C4mpyr][TFSI]content in the electrolytes, as is essential for batteries thatmay be

Fig. 6. (a) SEM/EDS, (b) SEM, (c) SEM and (d) XRD of electrodeposit product

raglyme (mole ratio 1:2) containing 18 mM Mg[BH4]2. Galvanostatic deposition w

atmosphere.

exposed to elevated operating temperatures, for example in solarinstallations. The mass proportion of each component ofdifferent Mg[TFSI]2 electrolytes and their mass retention as250 �C is listed in Table S3.

Fig. 4d depicts the CV of treated electrolytes of the hybridmixtures. At all ratios, a reductive peak corresponding to theMg2þ þ 2e� / Mg process was observed at approximately�0.5 V vs. Mg. A subsequent oxidation peak was observed at0 V vs. Mg. The highest oxidative currents were observed for0.3 M Mg[TFSI]2/([C4mpyr][TFSI]:tetraglyme) (mole ratio1:2). A cyclic efficiency of approximately 87% was calculatedfor this treated hybrid electrolyte. Furthermore both mole ra-tios of (1:1) and (1:2) ([C4mpyr][TFSI]:tetraglyme) showanodic stability > 3.0 V vs. Mg, making these electrolytespromising candidates for high voltage cathodes.

Fig. 4 demonstrates that all four hybrid electrolytes arepromising candidates for RMBs. Therefore, the treated elec-trolyte 0.3 M Mg[TFSI]2 in [C4mpyr][TFSI]:tetraglyme (1:2)was chosen as the optimised system for further work herein.However future work will also focus on the other hybrid elec-trolytes with varying concentrations of tetraglyme. As shown inFig. 5, the electrolyte can support stable Mg cycling within thepotential range of�0.6 V to 2.5 V vs. Mg. Although there was adrop in efficiency over the first 40 cycles, stable Mg reductionand oxidation were observed overall for 280 cycles. Further-more, no [BH4]

� oxidation was observed throughout the wholeprocess.

As illustrated in Fig. 6a the electrodeposit from treatedelectrolyte of 0.3 M Mg[TFSI]2 in [C4mpyr][TFSI]:[Te-tragylme] [1:2] shows the trend to grow into a similar par-ticulate morphology of Mg deposits as described by Matsui

from the electrolyte of 0.3 M Mg[TFSI]2 electrolytes in [C4mpyr][TFSI]/tet-

as carried at 1.5 mA cm�2 on a glassy carbon working electrode for 24 h, Ar

152 Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

[45]. The EDS spectrum confirms the high purity of the Mgdeposit. Upon closer observation of the morphology in Fig. 6band c, we see a well-aligned growth of the Mg deposit, whichis in good agreement with the XRD result shown in Fig. 6d.The XRD spectrum in Fig. 6d also depicts the existence of apure Mg deposit. Compared to the standard hexagonal Mg(JCPDS 04-0770) as referred in the figure, the Mg depositshows a well grown plane of (100) at 32.3�, and another twopeaks at 36.7� and 57.6� assigned to (101) and (110),respectively. This dominant (100) plane of the Mg deposit is ingood accordance with the results obtained in previous workwhere it was deduced that the presence of [C4mpyr][TFSI]assisted the growth of the substrate-parallel (100) plane [23].However, in contrast to that work, the morphology of the Mgdeposit in this study is more uniform, which can be ascribed tothe removal of H2O and thereby avoiding significant MgO/Mg[OH]2 formation [46]. In the literature computational studieshave also indicated different energy barriers for atom stackingin different crystallographic orientations [45,47].

4. Conclusions

In conclusion, we have developed thermally-stable Mgelectrolytes based on Mg[TFSI]2 in a hybrid mixture of IL/tetraglyme. Treatment with Mg[BH4]2 enabled reversible Mgelectrochemistry in these hybrid electrolytes for over 280 cy-cles. After a series of optimization and comparison studies ofthe various treated electrolytes, 0.3 M Mg[TFSI]2 in [C4mpyr][TFSI]/tetraglyme at a mole ratio of 1:2 showed high thermalstability, ionic conductivity and enhanced electrochemicalperformance. With an anodic stability > 3.0 V vs. Mg, theseelectrolytes have potential to be used with high voltage cath-odes. By lowering the H2O content and thereby avoiding MgOformation on the Mg electrodeposit, a highly uniform andaligned deposit is obtained.

Conflict of interest

There are no conflicts of interest to declare.

Acknowledgement

The authors acknowledge Mr. Changlong Xiao fromMonash University for conducting the SEM & EDX and XRDmeasurements. DRM is grateful for support from the Austra-lian Research Council for his Australian Laureate Fellowship.

Appendix A. Supplementary data

Supplementary data to this article can be found online athttps://doi.org/10.1016/j.gee.2018.10.003.

References

[1] J.M. Tarascon, M. Armand, Nature 414 (2001) 359–367.

[2] R. Zhang, C. Ling, MRS Energy Sustain. 3 (2016) E1.

[3] D. Aurbach, Z. Lu, A. Schechter, Y. Gofer, H. Gizbar, R. Turgeman,

Y. Cohen, M. Moshkovich, E. Levi, Nature 407 (2000) 724–727.

[4] Y. Viestfrid, M.D. Levi, Y. Gofer, D. Aurbach, J. Electroanal. Chem. 576

(2005) 183–195.

[5] O. Mizrahi, N. Amir, E. Pollak, O. Chusid, V. Marks, H. Gottlieb,

L. Larush, E. Zinigrad, D. Aurbach, J. Electrochem. Soc. 155 (2008)

A103–A109.

[6] H.S. Kim, T.S. Arthur, G.D. Allred, J. Zajicek, J.G. Newman,

A.E. Rodnyansky, A.G. Oliver, W.C. Boggess, J. Muldoon, Nat. Com-

mun. 2 (2011) 427.

[7] Q.S. Zhao, Y.N. Nuli, Y.S. Guo, J. Yang, J.L. Wang, Electrochim. Acta

56 (2011) 6530–6535.

[8] R. Mohtadi, M. Matsui, T.S. Arthur, S.-J. Hwang, Angew. Chem. Int. Ed.

51 (2012) 9780–9783.

[9] C.J. Barile, E.C. Barile, K.R. Zavadil, R.G. Nuzzo, A.A. Gewirth,

J. Phys. Chem. C 118 (2014) 27623–27630.

[10] O. Tutusaus, R. Mohtadi, T.S. Arthur, F. Mizuno, E.G. Nelson,

Y.V. Sevryugina, Angew. Chem. Int. Ed. 54 (2015) 7900–7904.

[11] Z.-B. Zhou, H. Matsumoto, K. Tatsumi, Chem. Eur. J. 12 (2006)

2196–2212.

[12] G.B. Appetecchi, M. Montanino, A. Balducci, S.F. Lux, M. Winter,

S. Passerini, J. Power Sources 192 (2009) 599–605.

[13] D.R. Macfarlane, M. Forsyth, P. Howlett, M. Kar, S. Passerini, J. Pringle,

H. Ohno, M. Watanabe, F. Yan, W. Zheng, S. Zhang, J. Zhang, Nat. Rev.

Mater. 1 (2016) 15005.

[14] T. Watkins, A. Kumar, D.A. Buttry, J. Am. Chem. Soc. 38 (2015)

641–650.

[15] M. Kar, Z. Ma, L.M. Azofra, K. Chen, M. Forsyth, D.R. Macfarlane,

Chem. Commun. 52 (2016) 4033–4036.

[16] Z. Ma, M. Kar, C. Xiao, M. Forsyth, D.R. Macfarlane, Electrochem.

Commun. 78 (2017) 29–32.

[17] I. Shterenberg, M. Salama, Y. Gofer, E. Levi, D. Aurbach, MRS Bull. 39

(2014) 453–460.

[18] G.T. Cheek, W.E. O'grady, S.Z. El Abedin, E.M. Moustafa, F. Endres,

J. Electrochem. Soc. 155 (2008) D91–D95.

[19] Y. Nuli, J. Yang, R. Wu, Electrochem. Commun. 7 (2005) 1105–1110.

[20] N. Amir, Y. Vestfrid, O. Chusid, Y. Gofer, D. Aurbach, J. Power Sources

174 (2007) 1234–1240.

[21] G. Vardar, A.E.S. Sleightholme, J. Naruse, H. Hiramatsu, D.J. Siegel,

C.W. Monroe, ACS Appl. Mater. Interf. 6 (2014) 18033–18039.

[22] A. Kitada, Y. Kang, Y. Uchimoto, K. Murase, J. Electrochem. Soc. 161

(2014) D102–D106.

[23] A. Kitada, Y. Kang, K. Matsumoto, K. Fukami, R. Hagiwara, K. Murase,

J. Electrochem. Soc. 162 (2015) D389–D396.

[24] M.M. Huie, C.A. Cama, P.F. Smith, J. Yin, A.C. Marschilok,

K.J. Takeuchi, E.S. Takeuchi, Electrochim. Acta 219 (2016) 267–276.

[25] L. Xue, D.D. Desmarteau, W.T. Pennington, Solid State Sci. 7 (2005)

311–318.

[26] S. Terada, T. Mandai, S. Suzuki, S. Tsuzuki, K. Watanabe, Y. Kamei,

K. Ueno, K. Dokko, M. Watanabe, J. Phys. Chem. C 120 (2016) 1353–

1365.

[27] I. Shterenberg, M. Salama, H.D. Yoo, Y. Gofer, J.-B. Park, Y.-K. Sun,

D. Aurbach, J. Electrochem. Soc. 162 (2015) A7118–A7128.

[28] K. Shimokawa, H. Matsumoto, T. Ichitsubo, J. Phys. Chem. Lett. 9

(2018) 4732–4737.

[29] S. Hebi�e, F. Alloin, C. Iojoiu, R. Berthelot, J.-C. Lepretre, ACS Appl.

Mater. Interf. 10 (2018) 5527–5533.

[30] B. Lee, J.-H. Cho, H.R. Seo, S.B. Na, J.H. Kim, B.W. Cho, T. Yim,

S.H. Oh, J. Mater. Chem. A 6 (2018) 3126–3133.

[31] L. Zhou, Q. Liu, Z. Zhang, K. Zhang, F. Xiong, S. Tan, Q. An, Y.-

M. Kang, Z. Zhou, L. Mai, Adv. Mater. 30 (2018) 1801984.

[32] S.-B. Son, T. Gao, S.P. Harvey, K.X. Steirer, A. Stokes, A. Norman,

C. Wang, A. Cresce, K. Xu, C. Ban, Nat. Chem. 10 (2018) 532–

539.

[33] H. Yoon, H. Zhu, A. Hervault, M. Armand, D.R. Macfarlane, M. Forsyth,

Phys. Chem. Chem. Phys. 16 (2014) 12350–12355.

[34] S. Theivaprakasam, D.R. Macfarlane, S. Mitra, Electrochim. Acta 180

(2015) 737–745.

153Z. Ma et al. / Green Energy & Environment 4 (2019) 146–153

[35] D.R. Macfarlane, P. Meakin, J. Sun, N. Amini, M. Forsyth, J. Phys.

Chem. B 103 (1999) 4164–4170.

[36] S. Randstr€om, G.B. Appetecchi, C. Lagergren, A. Moreno, S. Passerini,

Electrochim. Acta 53 (2007) 1837–1842.

[37] A.I. Bhatt, A.S. Best, J. Huang, A.F. Hollenkamp, J. Electrochem. Soc.

157 (2010) A66–A74.

[38] C. Liao, N. Sa, B. Key, A.K. Burrell, L. Cheng, L.A. Curtiss,

J.T. Vaughey, J.-J. Woo, L. Hu, B. Pan, Z. Zhang, J. Mater. Chem. A 3

(2015) 6082–6087.

[39] M. Kar, B. Winther-Jensen, M. Forsyth, D.R. Macfarlane, Phys. Chem.

Chem. Phys. 15 (2013) 7191–7197.

[40] J.J. Fillion, J.E. Bennett, J.F. Brennecke, J. Chem. Eng. 62 (2017) 608–

622.

[41] D.R. Macfarlane, M. Kar, J.M. Pringle, Fundamentals of Ionic Liquids:

From Chemistry to Applications, Wiley-VCH Verlag GmbH & Co.

KGaA, Weinheim, 2017.

[42] J. Wang, Y. Yamada, K. Sodeyama, C.H. Chiang, Y. Tateyama,

A. Yamada, Nat. Commun. 7 (2016) 12032.

[43] D. Macfarlane, A.L. Chong, M. Forsyth, M. Kar, V. Ranganathan,

A. Somers, J.M. Pringle, Faraday Discuss. 206 (2018) 9–28.

[44] O. Aschenbrenner, S. Supasitmongkol, M. Taylor, P. Styring, Green

Chem. 11 (2009) 1217–1221.

[45] M. Matsui, J. Power Sources 196 (2011) 7048–7055.

[46] J.G. Connell, B. Genorio, P.P. Lopes, D. Strmcnik, V.R. Stamenkovic,

N.M. Markovic, Chem. Mater. 28 (2016) 8268–8277.

[47] C. Ling, D. Banerjee, M. Matsui, Electrochim. Acta 76 (2012) 270–274.