Introduction

Technological developments associated with the needs of electronic devices with high performance and portability have stimulated research alongside the enhancements of new forms of energy storage and supply.1 The improvement of energy storage devices will allow the achieving of the most important current challenges, namely, the reduction of our dependence on fossil fuels through the development of advanced materials that can efficiently store electricity from renewable energy sources.2

In recent decades, several research groups have intensified the search for alternative energy sources, from solar, wind, hydraulic, nuclear, or chemical energies. Rechargeable batteries are based on chemical energy.3 Nowadays, lithium ion batteries (LIBs) are the most capable energy storage devices. However, lithium has limited abundance.4

On the other hand, the cost for the extraction and manufacture of lithium and sodium are more expensive than that for magnesium (Mg); magnesium is a sustainable alternative to stop the excessive consumption of lithium.5 Another advantage of the use of magnesium over lithium is the lack of dendrite growth, which has a positive impact on battery performance; nevertheless, one of the disadvantages corresponds to the short-term loss of storage capacity once the battery is partially discharged.6 Although magnesium offers a great potential for low-cost and safe energy storage devices, rechargeable magnesium ion batteries (MIBs) are currently at the research stage.7

Recently developed MIBs are still far from competitive with lithium batteries, in terms of output voltage and energy density.8 The design of novel electrolytic materials could offer the potential to provide high-purity magnesium anodes as well as good compatibility.9 Likewise, a highly conductive electrolyte is of main importance for a high-rate battery performance, with, furthermore, the need to avoid unstable flammable electrolytes such as those relying on ether solvents. One of the main challenges in MIBs is the inferior electrochemical performance compared to LIBs, because magnesium cyclability is difficult to achieve in most electrolytes.10,11,12

In this work, magnesium bis(oxalate)borate, Mg(BOB)2 was prepared by a low-temperature solid-state synthesis route. The Mg(BOB)2 was dissolved in various solvents and evaluated as an electrolyte. It was found that magnesium salt has a high solubility in tetrahydrofuran (THF), and that this mixture allowed Mg stripping/plating on the surface of a platinum working electrode when cycled electrochemically in three-electrode cells, showing electrochemical characteristics for a potential application in MIBs, as well as in other energy storage and conversion devices.

Experimental

Synthesis of Mg(BOB)2

Mg(BOB)2 was synthesized using magnesium hydroxide, Mg(OH)2, (reagent grade ≥ 95%; Sigma-Aldrich), boric acid, H3BO3, (ACS reagent ≥ 99.5%; Sigma-Aldrich), and oxalic acid, C2H2O4, (anhydrous ≥ 99%; Sigma-Aldrich) as starting materials (Table I). Stoichiometric amounts were weighed and mixed into the reaction pathway as:

$$ 4{\text{C}}_{2} {\text{H}}_{2} {\text{O}}_{4} + 2{\text{H}}_{3} {\text{BO}}_{3} + {\text{Mg}}\left( {{\text{OH}}} \right)_{2} \to {\text{Mg}}\left( {{\text{BOB}}} \right)_{2} + 8{\text{H}}_{2} {\text{O}} $$
(1)
Table I Raw materials as Mg(BOB)2 precursors
Full size table

The reaction was carried out by a low-temperature solid-state method, the first step consisting of drying all the precursors inside a vacuum oven at 60°C for 1 h, and subsequently mixing them for 15 min within an agate mortar and preparing dry-pressed pellets by applying a pressure of 2 MPa (gauge pressure). The homogeneous mixture was compressed into various cylindrical pieces with an area of 1.3 cm2 and a thickness of 0.5 cm. Afterwards, the pellets were initially heat-treated at 110 °C for 3 h, to remove residual water, then a subsequent heat treatment at 150 °C for 12 h was carried out, thus obtaining Mg(BOB)2 salt. The reaction product was transferred inside an argon-filled (O2 and H2O < 1 ppm) glove box (Omni-Lab 0210; VAC) as rapidly as possible, avoiding exposure to ambient atmosphere, then the pellets were ground before analysis.

Acetonitrile (ACN), propylene carbonate (PC), ethylene carbonate (EC), dimethyl carbonate (DMC), and tetrahydrofuran (THF) solvents were purchased from Sigma-Aldrich as battery-grade reagents and employed as received.

Structural and Morphological Characterization

The crystalline structure as well as the phase identification of Mg(BOB)2 were attained through McMaille software, with the results from prior characterization by x-ray diffraction (XRD; D2-Phaser; Bruker) with a CuKα radiation source (λ = 1.5418 Å). Particle size and morphology were analyzed using scanning electron microscopy (SEM; Neoscope JCM-6000; JEOL). Fourier-transform infrared spectroscopy (FTIR; Interspec 200-X) was carried out within the wave number range from 4000 cm−1 to 400 cm−1. Potassium bromide (KBr) pellets were used as a calibration blank and holder for sample analysis.

Preparation of Electrolytes

Mg(BOB)2 powder was used without further purification. The electrolytes were prepared at 50 mM concentration in each of the following solvents: ACN, PC, EC:PC (1:1 vol%), EC:PC:DMC (1:1:3 vol%), and THF. Every resulting solution was stirred for 24 h before electrochemical characterization. All the electrolytes based on the Mg(BOB)2 salt were prepared under inert conditions in a glove box. The electrochemical cyclability was evaluated for each electrolyte-prepared targeting to select the best solvent performance.

Electrochemical Measurements

Cyclic voltammetry (CV). Analysis was performed in three-electrode cells with a VMP3 potentiostat/galvanostat (Biologic Science Instruments) at room temperature using a platinum (Pt) working electrode (2 mm diameter; CH Instruments), and Mg foils (99.9% purity; Sigma-Aldrich) were used as counter and reference electrodes.

Linear scanning voltammetry (LSV). The electrolyte selected for the LSV test was 250 mM Mg(BOB)2/THF, which required several working electrodes (WE), such as copper (Cu), silver (Ag) plated with gold (Ag/Au), silver, glassy carbon (GC), stainless steel 304 (SS304), nickel (Ni), stainless steel 316 (SS316), Pt, and aluminum (Al). Mg foils were used as counter and reference electrodes.

Electrochemical impedance spectroscopy (EIS). The ionic conductivity of various electrolytes was studied by EIS. Glass-fiber membranes were soaked with 120 µL of the electrolyte with different concentrations, C, of Mg(BOB)2/THF, where C = 50 mM, 100 mM, 150 mM, 200 mM, 250 mM, 300 mM, 350 mM, and 400 mM. The membranes were then placed between a pair of SS316 blocking electrodes inside a CR2032 coin cell. The cells were conditioned through heating at 40 ºC for 12 h to obtain good contacts between the electrolytes and electrodes, then the cells were slowly cooled to room temperature and rested for 24 h before being tested on electrochemical workstations. EIS measurements were carried out in the temperature range of 30–80 °C, and every cell was allowed to reach thermal equilibrium for at least 3 h prior to each measurement. EIS curves were recorded in the frequency range from 1000 Hz to 10 mHz, applying a sinusoidal wave with an amplitude of 10 mV, as published elsewhere.13

The ionic conductivity values were calculated from Eq.2, where σ is the ionic conductivity, \(l\) is the thickness of the electrolyte-soaked glass-fiber membrane, \(S\) is the cross-section area of the glass fiber membrane, and \(R\) represents the resistance determined by the Nyquist plot (see Figure S1 in the supplementary information):

$$ \sigma = \frac{l}{S R} $$
(2)

Results and Discussion

Synthesis and Characterization of Mg(BOB)2

Figure 1 shows the results obtained by XRD of Mg(BOB)2 (see Figure S2 that include the XRD pattern of the precursors and reaction product). Using McMaille software for power diffraction pattern indexing14 via Expo2014 [https://www.ba.ic.cnr.it/softwareic/expo/], we were able to determine the structure cell parameters of Mg(BOB)2 as a monoclinic structure belonging to the P21/c space group, with a = 11.63 Å, b = 10.16 Å, c = 8.80 Å, α = γ = 90°, and β = 94.02°, a cell volume of 1039.29 Å3 and Z = 2 formulae per unit cell. The average size of the crystallites was determined from the four highest intensity peaks, employing Scherrer's equation (Eq. 3), where D is the crystal grain size, m is the full width at half-maximum (FWHM), λ is the wavelength of the x-rays, θ is the diffraction angle, and K is the Scherrer constant, expressed in radians with K = 0.89.

$$ D = \frac{K\lambda }{{m \cos \theta }} $$
(3)
Fig. 1
figure 1

XRD pattern of Mg(BOB)2; inset SEM image of the solid.

Full size image

Furthermore, the value of the FWHM should also be expressed in radians to ensure unit consistency throughout the calculation process.15 The crystallite sizes of the five maximum intensities are presented in Table II. Based on the FWHM value obtained from the XRD pattern, it was found that the average size of the crystallites was 44.89 nm, with a maximum size of 54.15 nm and a minimum size of 35.47 nm. The morphology of Mg(BOB)2, observed in the inset in Fig. 1, was obtained through SEM. Irregular microparticles with a size in the range of micrometers can be seen, similar to research reported in the literature.16,17,18,19,20,21,22,23

Table II Crystallite size calculation results by Scherrer's equation
Full size table

The FTIR spectrum of the synthesized Mg(BOB)2 is shown in Fig. 2, This technique allows the identification of functional groups with the powder technique24 (see Figure S3 showing the FTIR spectrum of the precursors and the reaction product). The main infrared peaks of Mg(BOB)2 and their functional group characteristic bands are shown in Table III. The vibration peaks at wavenumbers 3446 and 3215 cm−1 corresponding to the O-H bonds are attributed to water absorbed by the particles.25 In turn, the bands observed at 2673 and 2256 cm–1 can be associated to stretch vibrations of the atoms in the C–O bond26 from gaseous CO2 absorbed in the analyzed powders. Absorption bands at 1824 and 1791 cm−1 indicate the presence of functional groups C = O with oscillations in and out of phase from BOB anions,27 respectively. Asymmetric and symmetric stretch COO were ascribed to wavenumbers at 1650 and 1456 cm−1, respectively.28 Fingerprint bands of Mg(BOB)2 salt are present at 1310, 1221, 1089, and 990 cm−1, which represent the C-O-B-O-C stretch, C-O-C asymmetric stretch, and the O-B-O and C-O-C symmetric stretches, respectively. Further, other absorption bands of O-B-O symmetric and O-B-O asymmetric stretches show at wavenumbers 811 and 699 cm−1, as well as the BO4 functional group at 486 cm−1, comparably to the FTIR spectra reported in the literature.29 The wavenumbers 654 and 613 cm−1 are typical peaks of COO− and B-O deformations, respectively.20

Fig. 2
figure 2

FTIR spectrum of Mg(BOB)2; inset structure of the molecule.

Full size image
Table III Absorption bands in Mg(BOB)2
Full size table

Electrochemical Properties of Mg(BOB)2 in Different Solvents

The solubility of Mg(BOB)2 was evaluated in different solvents, CV tests were performed via a low concentration of about 50 mM Mg(BOB)2 in several solvents, at an experimental scan rate of 25 mV s−1 within the potential range − 1.0–3.0 V (vs. Mg2+/Mg). As can be observed in Fig. 3 (see also Figure S4 showing the CV curves using different solvents), the amount of magnesium salt dissolved in the PC (blue graph), EC:PC (1:1 vol%, yellow graph), and EC:PC:DMC (1:1:3 vol%, green graph) solvents showed null Mg stripping/plating. This behavior can be attributed to these solvents being of organic carbonate type, hence considered unsuitable for Mg electrolytes, since cathodic decomposition forming a passivation layer occurred30 as observed to occur in Pt WE. In addition, those solvents represents a serious risk for safety in batteries, being highly volatile and flammable, which prevent compliance with the highest safety standards for the immediate applications required.31 Moreover, the use of ACN as a solvent showed Mg plating but not Mg stripping. Although it can function as an electron donor ligand, studies based on Mg electrolytes using ACN reported contradictory results.

Fig. 3
figure 3

CV results of Mg stripping/plating containing 50 mM Mg(BOB)2 dissolved in different solvents collected at a scan rate of 25 mV s−1 within the potential range from − 1.0 to 3.0 V versus Mg2+/Mg; inset zoom at 0.9 V to observe Mg stripping, and oxidation of the electrolyte in the first cycle (dashed line).

Full size image

One study has reported the formation of a thick passivation layer on a Mg-coated electrode when working at high overpotential with electrolytes containing Mg salt/ACN.32 Nevertheless, another study has revealed ACN decomposition at a cathodic potential, with a simultaneous reduction in the current of Mg plating, owing to a reduction potential in the ACN that is closer to Mg2+/Mg (2.5 V and 2.3 V, respectively).33

In general, the use of ACN as a solvent for Mg electrolytes remains a challenge to understand, as well as the rest of the solvents. We conclude that the high insolubility of magnesium salts34,35,36 cannot be improved, even by increasing the solvent amounts. Therefore, the coordination of specific anions is more strongly controlled by the solvent structure instead of by the salt concentration in the range from 100 to 500 mM. Meanwhile, the solvation tendency of multivalent cations in these types of solvents provides information on the structural relationships, promoting the design and research to develop better electrolytes.37 Regarding the THF solvent, it was the only one displaying Mg plating at − 0.03 mA, and Mg was stripping at 0.005 mA of current. In this sense, THF was selected as the best solvent due to its high compatibility with Mg(BOB)2, similar to the type of salts in previous reports.38,39,40 When analyzing the electrolyte containing 50 mM Mg(BOB)2/THF, a decomposition was observed above 2 V, as can be seen in the inset in Fig. 3. This finding allowed us to modify the working potential window at this voltage for the subsequent cyclability tests with different concentrations of the Mg(BOB)2/THF electrolyte.

Electrochemical Behavior of Mg(BOB)2 in THF at Different Cycles

The electrolyte with a concentration of 250 mM Mg(BOB)2/THF was selected for its high current densities (J), and evaluated in three-electrode cells by means of CV curves up to 250 cycles, as can be seen in Fig. 4 (see Figure S5 and Table S1 in supplementary information). Starting from an initial open circuit voltage (OCV) of 1.10 V (inset a in Fig. 4) until reaching a final OCV ≈ 1.3 V after 250 cycles. The cyclability started in the cathodic zone, and, for the first cycle, the highest amount of cathodic charge current density (Mg plating) of about J ≈ -1.74 mA cm−2 was obtained (inset b in Fig. 4), as well as an anodic charge current density of J ≈ 0.40 mA cm−2 (Mg stripping) related to other concentrations (see Table S1). For the tenth cycle, the same amount of J was obtained in the Mg plating; however, this was not the case for the J in the Mg stripping process, since its original value decreased by up to 50%.

Fig. 4
figure 4

CV results of the Mg stripping/plating process containing 250 mM Mg(BOB)2 dissolved in THF, collected at a scan rate of 25 mV s−1 within the potential range of − 1.0 to 2.0 V versus Mg2+/Mg; insets a zoom at 0.9 V to observe Mg stripping, and b zoom at 0.3 V to observe Mg plating in the first cycle.

Full size image

For cycle number 50 (J ≈ − 1.46 mA cm−2 in Mg plating) and cycle number 100 (J ≈ − 1.34 mA cm−2 in Mg plating), the change in the amount of anodic charge current density was not greater than 0.02 mA cm−2, the value of J was about 0.11 mA cm−2 and remained constant for cycles 150, 200, and 250, whereas the values of J were only reflected in the Mg plating. There was a gradual drop to J ≈ − 1.14 mA cm−2 in the Mg plating after 250 cycles, resulting in highly competitive values, compared with those reported in the literature for Mg salts based on borates with cathodic charge currents of around ≈ 0.03 mA,13 offering a broad opening to a field of research to explore the performance of these novel electrolytes at high temperatures and with different solvents.

The three-electrode cells cycling up to 250 cycles (250 mM/THF) in the present work displayed a good cyclability performance from start to finish, compared to the other concentrations studied (see Figure S5 in supplementary information), showing excellent Mg stripping/plating, denoting a high reversibility, the principal characteristic of which is fundamental in electrolytes for secondary magnesium batteries.41

Anodic Stability of Mg(BOB)2 in THF

The LSV test results can be seen in Fig. 5. In a new electrolyte at the same molar concentration (250 mM/THF), different WE were compared with the objective of evaluating their anodic stability. It was found that Cu WE generated lower electrochemical stability (1.05 V vs. Mg2+/Mg), and that Al WE offered the highest electrochemical stability (3.5 V vs. Mg2+/Mg), typically the most used materials as current collectors for commercial batteries.42

Fig. 5
figure 5

LSV curves of the 250 mM Mg(BOB)2/THF electrolyte collected at a scan rate of 25 mV s−1. Cu (black), Ag/Au (green), Ag (blue), Pt (orange), GC (turquoise), SS304 (purple), SS316 (fuchsia), Ni (yellow), and Al (red) WE and Mg foils were used as counter/reference electrodes (Color figure online).

Full size image

In order to evaluate the difference between most used current collectors for MIBs, two alloys, SS316 and SS304, were included in the LSV measurements. It was determined that SS316 achieved a better electrochemical stability. In the same way, this material has been reported to display a lower steel corrosion,43 whereas in this work, the material managed to reach a slightly higher value of ≈ 0.5 V above SS304, therefore, for the application in MIBs, contributing positively to employing a broad potential window, as this improvement will currently support the development of new materials.44

Otherwise, Ag and Ag/Au WE did not reach a potential beyond 2 V, since the electrolyte rapidly decomposed, so these results were not beneficial for its application in MIBs. Thus, GC, Ni, and Pt WE were able to withstand a voltage similar to an electroactive magnesium salt, as reported.45

For the preparation of prototype half-cells, we recommend using SS316 current collectors. We were able to verify in this work that three-electrode cells showed a good anodic stability of 2.8 V versus Mg2+/Mg, in accordance with previously reported research, employing CR2032 coin cell type SS316 in Mg2+/Li+ hybrid-ion batteries, with a suitable performance.46

Ionic Conductivity of Mg(BOB)2 in THF

Finally, the electrolytes were evaluated with the blocking electrode technique, using a CR2032 coin cell SS316 to measure its conductivity at different temperatures. As shown in Fig. 6, the ionic conductivities exhibited a linear behavior according to the Arrhenius equation (Eq. 2). In addition to this, the electrolyte with the concentration of 250 mM/THF again presented the best results at room temperature, as well as in the cyclability tests, achieving an ionic conductivity value of 4.9 × 10–5 S cm−1, remaining above a superior performance compared with the other combinations.

Fig. 6
figure 6

Arrhenius plot of the ionic conductivity in different electrolyte of C Mg(BOB)2/THF concentration, where C = 50 mM, 100 mM, 150 mM, 200 mM, 250 mM, 300 mM, 350 mM, and 400 mM; lines are just guides to the eye; the THF boiling point is 339 K.

Full size image

Likewise, the electrolyte with the concentration of 400 mM Mg(BOB)2/THF, which did not present good electrochemical properties in the cyclability measurements, showed an increase in conductivity with a value of about 7.9 × 10–5 S cm−1 at 80 °C, these results confirming that Mg electrolytes can withstand higher operating temperatures (see Figure S6 for results in half-cells at room temperature), thus improving its experimental specific capacities,47 similar to solid polymer electrolytes, which demonstrates a high dependency with temperature to improve efficiency in the batteries.41 Consequently, the 250-mM Mg(BOB)2 electrolyte has the potential to be further explored, as, in ionic conductivity values, it only displays one order of magnitude lower than its analog.48 The change in the curve slopes in Fig. 6 can be associated with the boiling point of the THF solvent.

Conclusions

A magnesium salt Mg(BOB)2 was obtained using oxalic acid, boric acid, and magnesium hydroxide as a low-cost and high-purity anhydrous raw material by applying solid-state synthesis. A Mg-based electrolyte prepared with 250 mM Mg(BOB)2/THF has a highly reversible Mg cycle behavior (> 250 cycles), a wide anodic stability (2.8 V vs. Mg2+/Mg), and a good ionic conductivity at room temperature (4.9 × 10–5 S cm−1). The electrolyte had good chemical and electrochemical properties, such as low toxicity, low volatility, once the devices were assembled, and high reversibility, when being evaluated in three-electrode cells and in half-cells, assuming the challenges of the complex interactions of the electrolyte Mg(BOB)2/THF with the different solvents studied, as well as with the prepared commercial electrodes. Therefore, more work is required to prove its full potential as a high-performance electrolyte for MIBs.