Aluminum electrolytes for Al dual-ion batteries

In the search for sustainable energy storage systems, aluminum dual-ion batteries have recently attracted considerable attention due to their low cost, safety, high energy density (up to 70 kWh kg−1), energy efficiency (80–90%) and long cycling life (thousands of cycles and potentially more), which are needed attributes for grid-level stationary energy storage. Overall, such batteries are composed of aluminum foil as the anode and various types of carbonaceous and organic substances as the cathode, which are immersed in an aluminum electrolyte that supports efficient and dendrite-free aluminum electroplating/stripping upon cycling. Here, we review current research pursuits and present the limitations of aluminum electrolytes for aluminum dual-ion batteries. Particular emphasis is given to the aluminum plating/stripping mechanism in aluminum electrolytes, and its contribution to the total charge storage electrolyte capacity. To this end, we survey the prospects of these stationary storage systems, emphasizing the practical hurdles of aluminum electrolytes that remain to be addressed.

owing to their low cost, relatively high energy densities of up to ≈70 Wh kg −118 , and cyclic stability. In this review, we discuss recent developments in Al electrolytes for ADIBs covering the topics of charge storage capacity and the operating mechanism of ADIBs. In addition, we analyze in detail the impacts of acidity, ionic conductivity, the Al 3+ /Al redox potential, and the electrochemical voltage window of the Al electrolytes on the performance of ADIBs. Finally, with respect to the practical application of ADIBs, the compatibility of current collectors with Al electrolytes is covered in the last section of this review.
Historical aspects of the development of Al electrolytes for ADIBs Although the research on Al electrolytes for ADIBs may appear to be a new subject, this is a misconception. The employment of Al electrolytes in ADIBs based on molten salts was assessed for the first time in the 1970s by Fouletier et al. 19 Fouletier's ADIB was composed of metallic aluminum and graphite as negative and positive electrodes, respectively, which were immersed into molten LiCl/AlCl 3 salts (ca. 140°C). In 1988, Gifford et al. 20 extended research on ADIBs towards the use of roomtemperature ionic liquids (RTILs) based on imidazolium chemistry (AlCl 3 :1,2-dimethyl-3-propylimidazolium chloride). In 2015, Al electrolytes received a great deal of attention since the publication of Dai et al. 21 on ADIBs employing a metallic Al anode, two synthetic forms of graphite as the cathode (CVD-grown graphitic foam and pyrolytic graphite) and 1-ethylmethylimidazolium ionic liquid electrolyte (AlCl 3 :EMIMCl) 21 . Dai's ADIB showed a high reversibility over thousands of cycles and graphite cathodic capacities of up to 67 mAh g -1 . Those first publications have initiated the exploration of deep eutectic solvents (DESs) as Al electrolytes for ADIBs by Dai et al. 22,23 , Jiao et al. [24][25][26] , and others [27][28][29][30][31] . A schematic illustration of the different classes of Al electrolytes presently used in ADIBs is shown in Fig. 1a. A comparison of the various Al electrolytes with respect to their density, viscosity, ionic conductivity, theoretical charge storage capacity, and cost is given in Table 1.

Working principle of ADIBs and Al electrolytes based on RTILs
The basic configuration of ADIBs comprises a carbonaceous (graphite, zeolite-templated carbon (ZTC), or graphene) or organic positive electrode, chloroaluminate ionic liquid electrolyte and metallic aluminum anode as demonstrated in Fig. 1b. ADIBs operate as an electrochemical energy storage system employing reversible intercalation/insertion of the AlCl 4 − anion species into the positive electrode upon charge (oxidation). Concomitantly, the electroplating of aluminum occurs at the negative electrode of ADIBs. The working principle of ADIBs can be represented by the following cathodic and anodic half-reactions during charge: On the negative electrode : On the positive electrode : where C is graphite 32 , ZTC 33 , graphene 34 or organic active material 35,36 . Consequently, the mechanism of ADIBs is significantly different from the "rocking-chair" metal-ion batteries. There is no one-directional motion of Al 3+ ions from the positive to the negative electrodes. Al species are depleted from the chloroaluminate ionic liquid during the charge of ADIBs and are consumed by both electrodes. The Al electrolytes that are used in ADIBs play a double function: they support the Al plating/stripping process and act as the source of AlCl 4 − ions being needed for the intercalation/insertion into the positive electrode during charge. Consequently, ADIB term is applied for batteries, were electrolyte acts as a source of two types of ions (i.e., Al 2 Cl 7 − and AlCl 4 − ) required for electrochemical energy storage. Thus, the composition of the electrolyte is changing upon charge and discharge, contrary to "rocking-chair" Al-ion battery systems, where electrolyte acts exclusively as a transmitter of the ions. The most conventional example of an Al electrolyte is a mixture of aluminum chloride and other chlorides comprising an organic cation, for instance, 1-butyl-3-methylimidazolium chloride (BMIM) and 1-ethyl-3methylimidazolium chloride (EMIM). As a consequence of the acid-base interactions between AlCl 3 (Lewis acid) and Cl -(Lewis base), the salt mixture becomes a liquid at room temperature, forming an RTIL. The latter is composed of AlCl 4 − anions that are charge-balanced with organic cations. The RTIL with an excess of Lewis acid AlCl 3 over Lewis base EMIMCl is comprised of both AlCl 4 − and Al 2 Cl 7 − ions. Importantly, ADIBs are operational only in acidic formulations. Solely Al 2 Cl 7 − ions enable the electroplating of aluminum, which therefore, occurs only in chloroaluminate melts with an excess of AlCl 3 [37][38][39][40][41][42][43][44][45][46] . As a result, the charge storage capacity of the chloroaluminate melt is a function of the concentration of Al 2 Cl 7 − ions in the RTIL. Electroplating, and therefore, the charging process stops when no Al 2 Cl 7 − ions are left in the ionic liquid, which results in the formation of the neutral melt (AlCl 3 :EMIMCl = 1). The highest molar ratio (r) between AlCl 3 and EMIMCl that forms an RTIL is ca. 2:1. AlCl 3 does not dissolve at higher molar ratios. Apart from the ionic melts based on AlCl 3 / EMIMCl, other RTILs have also been recently reported to be composed of 1-methyl-3-propylimidazolium chloride (MPIMCl) 47 , benzyltriethylammonium chloride (TEBACl) 48 , and 1,2-dimethyl-3-propylimidazolium chloroaluminate (DMPIMCl) 49 . Figure 2 illustrates the impact of acidity r on the charge storage capacity of the chloroaluminate ionic liquid electrolytes, which can be, in fact, called anolytes. The theoretical gravimetric and volumetric capacities of the ionic liquid C an can be calculated as follows: where F = 26.8 × 10 3 mAh mol −1 (the Faraday constant), x ¼ 3 4 (number of electrons that are used to reduce 1 mol of the Al 2 Cl 7 − ions), M AlCl3 is the molar mass of AlCl 3 (g mol −1 ), M ACl is the molar mass of the Cl − salt (g mol −1 ), r is the AlCl 3 :ACl molar ratio, and ρ is density of the chloroaluminate melt (g mL −1 ). A detailed description of the derivation of Eqs. 3 and 4 can be found in ref. 50 . For instance, the gravimetric charge storage capacities of the AlCl 3 : EMIMCl ionic liquid are equal to 19 mAh g −1 and 48 mAh g −1 for r = 1.3 and r = 2, accordingly. Notably, these capacities define the overall energy density of ADIBs 14,18,[50][51][52][53][54][55][56][57] . Moreover, it should be pointed out that these theoretical capacities are not always achievable experimentally, i.e., they depend on practically relevant experimental conditions and on whether Al 2 Cl 7 − ions can be fully depleted for Al electroplating. To figure out this point, recently, Kravchyk et al. 58 assembled the anolyte-limited cell in a three-electrode configuration with a significant excess of graphite cathode (anolyte-limited cell). Upon these measurements (Fig. 3a), in addition to the voltage profile of the full cell, the profiles for both positive and negative electrodes were recorded. It was demonstrated that the voltage profile at the negative electrode (E CE ) remained relatively stable during charging for 15 min. Upon further charging, however, the voltage dropped  The theoretical capacities of RTILs and inorganic molten salts were calculated using expression Eq. 3. The theoretical capacities of deep eutectic solvents were computed from Eq. 5 (see Supplementary Note 1) using the concentration of Al2Cl7 − ions in AlCl3:urea, AlCl3:Me-urea and AlCl3:Et-urea electrolytes reported in the ref. 22 .
sharply indicating the end of the Al plating process caused by the depletion of Al 2 Cl 7 − ions at the negative electrode. Importantly, the voltage profile at the negative electrode (E CE ) for the graphite-limited cell was constant during the entire charge, with a small overpotential of <50 mV (Fig. 3b) pointing to the access of Al 2 Cl 7 − ions at the negative electrode. Using this approach, Kravchyk et al. 58 performed rate capability measurements of anolyte-limited full cells at different current densities ranging from 5 to 20 mA g −1 for the AlCl 3 : EMIMCl ionic liquid formulations with r = 1.3-2.0 (see Fig. 3c, d). These experiments revealed two main points. First, as expected, the higher capacities of chloroaluminate melts could be obtained only using highly acidic formulations. For instance, the charge storage capacity of the AlCl 3 :EMIMCl anolyte with r = 1.3 was measured to be ca. 21 mAh g −1 at a current density of 20 mA g −1 . In contrast, at r = 2, the capacity was ca. 46 mAh g −1 . These results indicate that the highest energy density of the ADIBs can be obtained using chloroaluminate ionic liquids with r = 2, and, therefore, future works on ADIBs should be focused on the most acidic formulations. Second, the applied current density affects the charge storage capacity of the anolyte. This is reflected in the significant deviation in the voltage profiles of the negative electrode at high currents (Fig. 3a). As a result, low charge storage capacities (ca. 10-14% from theoretical values) were obtained at very high current densities of 1 A g −1 . These results suggest that the frequent statements regarding the high power density of ADIBs are not fully correct. Specifically, at high current densities, a significant drop in the energy density of ADIBs is foreseen. The latter is caused by the rate capability limitations of both the chloroaluminate ionic liquid anolytes and the cathodes of ADIB. In fact, these observations show that the charge storage capacities of the anolyte significantly deviate from the theoretical value at charge current densities higher than 20 mA g −1 .
Apart from the charge storage capacity, the acidity of the chloroaluminate ionic liquid electrolytes strongly influences their ionic conductivity. For instance, as revealed by Ferrera et al. 59 for the AlCl 3 -EMIMCl chloroaluminate ionic liquid, its conductivity gradually decreases from 20 to 10 mS cm −1 at room temperature upon increasing the AlCl 3 /EMIMCl molar ratio from 1.1 to 1.7, respectively. Furthermore, it was shown by Wang et al. 60 , that the ionic conductivity increases with decreasing halide ionic radius (from iodide to bromide and chloride anions). Moreover, the acidity variations strongly influence the electrochemical stability window of the electrolyte. Specifically, it has been determined that the oxidation of the AlCl 3 : 1-ethyl-buthyl-imidazolium ionic liquid with mole ratios of 1, 1.1, 1.5 and 2 takes place at approximately 2.6 V vs. Al 3+ /Al, while for the 0.8 mole ratio, the oxidation starts at 1.75 V vs. Al 3+ /Al. The oxidation stability limit is associated with the chlorine evolution at the cathode side 41,61 .
It should be noted that Wang et al. 62 assessed the Al 3+ /Al redox potential in the AlCl 3 -EMIMCl ionic liquid vs. the standard hydrogen electrode (SHE) using a specially made cell composed of a β-alumina solid-state electrolyte and Na reference electrode. It was stated that plating/stripping of aluminum in AlCl 3 -EMIMCl ionic liquid occurs at −0.7 V vs. SHE and 2.3 V vs. Li + /Li. This Al 3+ /Al redox potential is ca. 1 V higher than that observed in aqueous solutions.

Inorganic molten salts
Another type of electrolyte capable of reversible aluminum plating/ stripping electrochemistry are low-cost molten salt eutectics. One of the first investigations of aluminum electrodeposition in molten salt electrolytes was undertaken by Del Duca in 1971 63 . This work on AlCl 3 -NaCl and AlCl 3 -(LiCl-KCl) mixtures elucidated the kinetics of aluminum electroplating. Gale et al. 64 examined the subvalent ion effect during aluminum anodization in molten AlCl 3 -NaCl.
The melting point of a chloroaluminate eutectic mixture is dependent on its precursors and composition. The typical binary AlCl 3 /NaCl and AlCl 3 /KCl systems melt at above 108 and 128°C, respectively 65 . The addition of a third/fourth salt to the binary electrolyte can further reduce the melting point. For example, a ternary AlCl 3 /NaCl/KCl electrolyte (molar ratio: 61/26/13) has been reported to melt below 100°C 65 , and the AlCl 3 /LiCl/KCl system with a molar ratio of 59/29/12 turns entirely liquid at approximately 95°C. Recently, the quaternary AlCl 3 /NaCl/LiCl/KCl inorganic molten salt with the lowest eutectic temperature of less than 75°C was also systematically studied as the electrolyte in an ADIB 66 . Apart from the low cost, among the advantages of inorganic molten salts as electrolytes for ADIBs are their low viscosity and high ionic conductivity that facilitate the kinetics of AlCl 4 − insertion/deinsertion reactions. Additionally, with respect to the charge storage capacity, the lower molar mass of alkali chlorides yields higher values of gravimetric capacity over imidazolium or pyridinium chlorides at the same acidity (Fig. 2).

Deep eutectic solvents
In the search for a more economical and environmentally sustainable alternative to imidazolium-based RTILs, a new class of ionic liquids, namely, DESs, were recently employed as Al electrolytes for ADIBs 43,44,67 . They are also known as ionic liquid analogs (ILAs) 67 . DESs can be defined as a mixture of a strongly Lewis acidic metal halide and an oxygen donor amide ligand, such as urea, acting as a Lewis base. Similar to the AlCl 3 -EMIMCl system 41 , AlCl 3 -urea forms through the exothermic reaction between AlCl 3 and urea according to the following equation: The comprehensive characterization of DESs by both Raman 22,24,27,68 and NMR 22,23,68 spectroscopies have revealed that only AlCl 4 − are present in the AlCl 3 -urea DESs at an AlCl 3 / urea ratio of 1.0 (neutral). However, as the acidity of the melt increases through the addition of AlCl 3 , the concentration of the Al 2 Cl 7 − species gradually increases when compared to the concentration of the AlCl 4 − species (Fig. 4a). The highest molar ratio (r) between the AlCl 3 and urea that forms an ionic liquid is ca. 1.5:1 22 . Importantly, as revealed by Ng et al. 27 , Al electroplating takes place only in the acidic melts (AlCl 3 /urea > 1.1). Similar observations have also been reported in other AlCl 3 -amide systems 69 . Aiming to identify which species are responsible for Al electrodeposition, Dai et al. 22 performed operando Raman spectroscopy measurements of AlCl 3 -urea DESs (AlCl 3 /urea = 1.4) during Al deposition. Upon Al plating (0 to −0.5 V vs. Al 3+ /Al), a large decrease in the intensity of the Al 2 Cl 7 − peak (313 cm −1 ) and a corresponding increase in the AlCl 4 − peaks (350, 445 cm −1 ) were observed (Fig. 4b). Upon Al stripping (0-0.5 V vs. Al 3+ /Al), the exact opposite processes were observed, namely, there was a large intensity increase and decrease in the Al 2 Cl 7 − and the AlCl 4 − peaks, respectively; no changes to any other peaks were detected. According to Dai's observations, the Al electrodeposition/stripping processes in AlCl 3 -urea DESs can be described as follows: 4Al 2 Cl À 7 þ 3e À $ 7AlCl À 4 þ Al. We note that previously 23,26 , it had been assumed that the cationic aluminum species ½AlCl 2 area ð Þ 2 þ were responsible for the Al electroplating reaction. This conclusion is in line with the work of Chu et al. 70 , whereby the Al plating pathways were investigated by density functional theory (DFT) calculations (Fig. 4c, d). Considering the difference in the dissociation energy barriers of Al 2 Cl 7and ½AlCl 2 urea ð Þ 2 þ is 0.28 eV, it has been assumed that the dissociation of ½AlCl 2 urea ð Þ 2 þ is less kinetically preferred, therefore, favoring the Al plating through the Al 2 Cl 7reduction. From this perspective, taking into consideration the concentration of Al 2 Cl 7 − in the acidic AlCl 3 -urea DESs, the charge storage capacity equals ca. 21 mAh kg −1 (for AlCl 3 /urea = 1.4). This value can be derived from Eq. 5 (see Supplementary Note 1) using the concentration of Al 2 Cl 7 − ions in AlCl 3 :Urea DES reported in the ref. 22 . For AlCl 3 :Me-urea and AlCl 3 :Et-urea anolytes, one can obtain similar capacities of ca. 16 mAh g −1 and 18 mAh g −1 , respectively.
In addition to AlCl 3 -urea DESs, several articles have been published reporting the ability of acetamide (AcAm) to form eutectic mixtures when mixed with AlCl 3 , yielding the reversible electrodeposition of aluminum 28,71 . As stated in ref. 28 , the AlCl 3 -AcAm DES comprises the anionic AlCl 4 − and Al 2 Cl 7 − species as well as complexed aluminum cationic species, such as [AlCl 2 (AcAm) 2 ] + . Another eutectic AlCl 3 -organic solvent system with a coordination behavior similar to that of AlCl 3 -amide is the AlCl 3 -diglyme electrolyte, which also supports reversible Al deposition/dissolution 72,73 . The active species in this electrolyte are AlCl 2 (diglyme) 2 + cations, which are paired with the AlCl 4 − anions.
Considering the low cost of DESs and their environmental friendliness, energy storage systems that utilize DES electrolytes will have significant economic and environmental cost advantages over conventional RTILs. However, such systems have a relatively low electrochemical stability window of 2.2 V vs. Al 3+ /Al in comparison with that of dialkylimidazolium chloridebased RTILs 22,23,27 . The latter are stable up to ∼2.45 V vs. Al 3+ /Al 62 . Apart from this, the conductivity of the AlCl 3 -amide systems is rather poor at room temperature (see Table 1) 25,43,74 . The lower conductivity and the sluggish kinetics of the DES electrolytes have been ascribed to their relatively low concentration of active ionic species as well as strong coordination interactions. Atomistic simulations demonstrated that aluminum ions in the RTILs have a weaker coordination and form Al-Cl complexes with a low stability compared to those in some molecular solvents, and this contributes to the facile ion transport and dissociation 75 .

Corrosion and current collectors
One of the main practical issues to address for the realization of cell prototypes of ADIBs is the high reactivity and corrosivity of aluminum anolytes. For instance, the coin-type cells composed of stainless steel corrode in chloroaluminate melts, requiring the employment of corrosion-free battery cases. Furthermore, contrary to the LIBs, where the Al foil is considered as an established current collector, the current collector for ADIBs at the positive electrode is still under development. Earth-abundant metals, such as aluminum and iron, are easily oxidized in aluminum anolytes at the high voltages of 4.5-5.25 V vs. Li + /Li being employed during positive electrode operation [76][77][78] . Therefore, oxidatively stable conductive materials such as tungsten, molybdenum, and glassy carbon are typically utilized in ADIBs 79,80 . Of note, it has recently been suggested to use titanium nitride (TiN) as a compelling current collector for ADIBs. As reported by Wang et al. 62 , TiN coated on a stainless steel or flexible polyimide substrate can be fabricated by a low-cost and scalable method, such as magnetron sputtering. The TiN oxidative stability in the AlCl 3 : EMIMCl ionic liquid is on par with that of W and Mo current collectors. Notably, graphitic electrodes without a current collector were also proposed by Di-Yan Wang et al. 81 as an alternative approach to address the oxidation issues. So-called freestanding natural graphite films were fabricated by graphite slurry tape casting on the Cu foil following its etching in an iron chloride (FeCl 3 ) solution.
Moreover, the research on the chemical stability of Al foil on the negative side of the AGDIB in acidic RTILs was performed recently by Tak et al. 82,83 . It has been demonstrated that Al surface corrodes in chloroaluminate melts. Furthermore, the corrosion enhances significantly upon increasing their acidity. Importantly, the works of Tak et al. 82,83 , pointed to the fact that additional research efforts should be carried out to mitigate or entirely suppress the reaction of Al with chloroaluminate electrolytes. The quest for non-corrosive electrolyte formulations, therefore, continues.

Outlook
Although significant advances were accomplished recently on Al electrolytes for ADIBs, substantial room remains for the improvement in their gravimetric/volumetric charge storage capacity and the efficiency of Al plating/stripping. We note that ionic melts used in ADIBs are not just electrolytes (ion-conductors), but represent an electrochemically active, capacity-and rate-limiting battery component. In this context, further research should be focused on finding the practical amounts of the electrolyte needed for the operation of ADIBs. Notably, most of the studies presented in the literature have employed a 2-10-fold excess of the electrolyte required to match the charge storage capacity of the positive electrode (cathode-limited cell). Such tests are acceptable for research purposes, but they do not provide correct and practically relevant information on achievable energy/ power densities or the cycling stability of ADIBs. It is also apparent that future research should focus on the decrease in the redox potential of the Al plating/stripping being ca. 1 V higher (− 0.7 V vs. SHE) than that in aqueous solutions (−1.66 V vs. SHE). Additionally, we suggest that subsequent work should stress the other issues associated with ADIB technology, one being the incompatibility of most metallic current collectors with the corrosive Al electrolytes. For instance, aluminum and stainless steel slowly oxidize in the AlCl 3 :EMIMCl ionic liquid when electrochemically polarized up to 2.5 V vs. Al 3+ /Al. Thus far, only tungsten, molybdenum, glassy carbon, chromium, and titanium nitrides have been identified as electrochemically stable current collectors in such batteries. In this framework, we note that any oxidation stability issues in the current collectors that cause a severe decrease in the coulombic efficiency of ADIBs will have a profound effect on their electrochemical performance, which is almost ignored in scientific publications. Towards this end, we state that all these advancements on Al electrolytes should be achieved at a low cost to keep the overall cost-competitiveness of ADIBs.