All-Solid-State Anode-Free Sodium Batteries: Challenges and Prospects

Abstract

All-solid-state anode-free sodium batteries present a special and especially important kind of energy storage device. Unfortunately, the industrial production of such batteries has been absent up to now, although the prospects of their development seem to be rather optimistic. The present mini review considers the fundamental advantages of all-solid-state anode-free sodium batteries as well as challenges in their creation. The advantages of all-solid-state anode-free sodium batteries reveal themselves when comparing them with ordinary sodium-ion batteries, sodium metal batteries, sodium batteries with liquid electrolyte, and their lithium counterparts.

1. Introduction

So-called “anode-free” batteries are a specific but very promising kind of battery belonging to the post-lithium-ion era. Such batteries were first mentioned at the very end of the last century [1]. These were lithium metal batteries with solid electrolytes. Subsequently, various versions of anode-free lithium metal and sodium metal batteries with both solid and liquid electrolytes appeared. Surely, these batteries have two electrodes, and during discharge, one of them is the cathode and the other is the anode. Anode-free batteries are manufactured in a completely discharged state, and the negative electrode consists only of a current collector without active mass. During the first charge, lithium or sodium is deposited on this current collector at the expense of ions contained in the active material of the positive electrode. In the literature, one can find various synonyms for the term “anode-free battery”, such as “anode-less battery”, “zero-excess Li-metal battery”, “reservoir-free battery”, and so on.

The absence of active material on the negative electrode of a freshly assembled battery ensures a certain reduction in its weight and volume, i.e., an increase in energy density. An even more important advantage of anode-free sodium batteries over usual sodium metal ones is the significant simplification and reduction in cost of their production technology (at the assembly stage).

Unlike conventional batteries, the “electrolyte/negative electrode current collector” interface is of great importance for anode-free batteries, along with the “electrolyte/positive electrode active material” and “electrolyte/negative electrode active material” interfaces. There is currently extensive literature on anode-free lithium and sodium batteries with liquid electrolytes and on anode-free lithium batteries with solid electrolytes. Studies on the development of anode-free sodium batteries with solid electrolytes are still in their infancy. This very circumstance was the authors’ main motivation for writing this mini review. At the same time, many of the problems of such batteries are also inherent in anode-free sodium batteries with liquid electrolytes [2,3,4,5,6,7,8] and anode-free lithium batteries with solid electrolytes [7,9,10,11], as well as conventional (non anode-free) sodium batteries with solid electrolytes [12,13,14,15,16,17]. For the most part, these problems are related to sodium deposition during charging in the form of dendrites and their encapsulation, i.e., deposition of a solid electrolyte interphase (SEI) film on them.

The deposition of dendrites forming upon cathodic sodium is known to be connected with different factors. Firstly, in anode-free sodium metal batteries (no matter with liquid or solid electrolytes), sodium metal deposits directly onto foreign material of the current collector. The nucleation process in this case needs a rather high activation energy; therefore, it is non-uniform what is favorable for the growth of separate whiskers. Secondly, the above-mentioned SEIs, as a rule, contain a significant amount of flaws, which results in non-uniform electric fields and uncontrolled dendritic growth. The dendrite growth inevitably results in the breakdown of the separators and ultimately results in the short circuit of the battery. Sodium is a rather strong reducer, and in the course of its cathode deposition, a fresh, very active surface is formed. It is on this active surface that a passive film (SEI) grows, and since sodium is deposited in the form of dendrites, in many cases, the film completely envelops individual sodium particles, preventing their electronic contact with the current collector (encapsulation phenomenon).

The interface effect between electrode and solid electrolyte materials is the most critical issue for modern all-solid-state sodium batteries (both anode-free and ordinary). An ideal (or, at least, proper) interface between electrolyte and electrodes should have good compatibility and suitable mechanical strength as well as high ionic conductivity. In other words, this interface must reveal good sodiophility. Unfortunately, the contact of sodium metal and solid electrolytes faces interrelated challenges such as sluggish charge transfer kinetics, high interfacial resistance, uneven potential, and localized current density distribution, ultimately accelerating dendrite development.

The solid electrolytes for anode-free and ordinary sodium batteries must have high ionic conductivity along with a sufficiently high transference number of sodium ions, low electronic conductivity, a wide electrochemical window, chemical and mechanical stability, good mechanical properties, facile preparation, and environmental friendliness.

The solution to these problems usually comes down to modifying the current collector, electrolytes, and cycling mode (protocol).

2. The Interface Between a Solid Electrolyte and a Current Collector

Electrochemical processes at the interface between the solid electrolyte and the negative electrode current collector occur in anode-free batteries not only at the first charge, but also at further cycling during normal battery operation. In principle, such processes can also occur in a conventional (non-anode-free) sodium battery with a solid electrolyte during full discharge (or overdischarge) of the battery, but this is more likely to be related to improper operation (abuse) and should be avoided. The negative electrode current collector in a sodium battery can be made of different materials, and it is important to note that, unlike anode-free lithium batteries, in sodium batteries, this current collector can be made of both copper and aluminum, since sodium does not form alloys with aluminum and is not inserted into it.

The processes at the interface of the current collector with a solid electrolyte differ to a certain extent from the processes at the interface of the current collector with a liquid electrolyte. In the latter case, the entire surface of the current collector wetted with electrolytes is the contact surface. At the interface of the solid current collector with a solid electrolyte, contact is ensured only at individual points (spots), and the configuration of these spots may depend on the degree of compression of the electrolyte and the current collector (Figure 1).

A salient feature of such interfaces is the presence of voids (pores). As can be seen, the area of the electrochemically active surface is smaller than the total surface area, which leads to a certain decrease in the total current. This phenomenon, so-called current constriction, is well known in physics for contact between metals or semiconductors [19,20,21], but it has not been sufficiently studied in electrochemistry [18,22,23,24,25]. A natural consequence of current constriction is an increase in the total interfacial resistance [26,27], which also depends on the charge transfer resistance, the presence of SEI, and the presence of various contaminants. Of course, the picture shown in Figure 1 is not static, but it changes continuously when the current passes; the pore volume increases during the anodic dissolution of the metal and decreases during cathodic deposition. According to Faraday’s laws, the passage of 1 Ah is accompanied by a change in the volume of sodium by 0.88 cm³, which is 1.7 times greater than during the deposition/dissolution of lithium. In this regard, batteries with solid electrolytes compare favorably with batteries with liquid electrolytes, in which there is no free volume and the deposition of metallic sodium leads to the occurrence of internal stresses. Accordingly, the interfacial resistance increases in the anodic phase of cycling and decreases in the cathodic phase [24,25].

Some features of the interfaces of solid electrodes with a solid electrolyte associated with the occurrence of current constriction were studied using examples of different metals and different electrolytes. Thus, in [28,29,30,31], the interfaces of silver with AgI and AgBr and copper with CuBr, as well as platinum with CoO, were studied, and the occurrence of potential oscillations under galvanostatic conditions was discovered. The characteristic features of these oscillations depend on the current density, temperature, and compression force (pressure). Potential oscillations arise as a result of the periodic formation and disappearance of micropores and the corresponding changes in the contact surface area. Electrochemical dissolution occurs mainly at the points of dislocation exit to the surface, and it is in these places that micropores are formed. The walls of the micropores provide the rapid diffusion of metal adatoms.

Similar conclusions regarding the nature of current constriction were obtained in [32,33,34] using the example of lithium deposition from amorphous solid electrolyte LiPON (lithium phosphorus oxynitride) on substrates of different metals. In all cases, an increase in the current density of lithium deposition led to a decrease in the size of nuclei and to an increase in their number per unit surface area. The nucleation overpotential during lithium deposition on platinum (with which lithium forms an alloy) turned out to be significantly lower than the nucleation overpotential during lithium deposition on copper (when alloy formation is excluded). Potential oscillations were also observed during lithium deposition on platinum.

At the interface between copper and a crystalline lithium-conducting electrolyte with the garnet structure Li₇La₃Zr₂O₁₂ (LLZO), oscillations of even the open-circuit potential were observed, which, in turn, depended on the compression force [35] (Figure 2). An increase in the number of nuclei with increasing lithium deposition current density was also noted here, as well as preferential nucleation on surface defects [29,36].

The most interesting and important aspect from this article’s point of view is the work of [37], which studied the processes of sodium deposition and dissolution at the boundary of a copper current collector and a solid electrolyte NaSICON (Na SuperIonic CONductor). The electrolyte in this work had the composition Na_3.4Zr₂Si_2.4P_0.6O₁₂ (NZSP) (it should be noted that this is the electrolyte often used in studies of all-solid-state sodium-metal batteries [14,38,39,40,41,42,43,44,45,46]; one can find in the literature both notations “NaSICON” and “NASICON”). The processes of nucleation and deposit growth were studied at different current densities and different stack pressures using time-dependent impedance spectroscopy, confocal light microscopy, and electron microscopy. The studies were carried out using the electrochemical system Cu|NZSP|Na, where sodium served simultaneously as an auxiliary electrode and a reference one (quasi-reference electrode). The reliability of using such a reference electrode was confirmed in [38]. The galvanostatic curves of sodium deposition had an usual appearance, with a sharp shift of the potential to the negative direction at the moment of switching on the current, corresponding to the nucleation overpotential, and a subsequent gradual shift of the potential to the positive direction until reaching a steady-state value, corresponding to the current density [32,47,48]. Intriguingly, in this stationary section of the galvanostatic curve, sporadic fluctuations were recorded, the amplitude of which decreased somewhat with an increase in current density and stack pressure. Potential fluctuations were caused by local delamination of the copper current collector from the solid electrolyte and the corresponding increase in interphase resistance.

It is interesting that the nucleation overvoltage at a constant stack pressure was practically independent on the current density, but at a constant current density, it increased linearly with the increase in stack pressure (Figure 3). The latter fact is most intriguing, and it could be associated with a change in the pattern of contact between the electrode and the electrolyte (Figure 1). It should be emphasized here that it is the overvoltage of the nucleation stage that is considered, and not the deposit growth in the steady state (in the latter case, the overvoltage increases with increasing current density [37]). It is important to note that the influence of the stack pressure on the processes at the boundary of the current collector with the solid electrolyte has a clear physical nature. Some effects described for the boundary of the current collector with the liquid electrolyte [49,50] more likely than not are some artifacts.

At the same time, an increase in the current density was found to result in the formation of more uniform sodium deposits. Surprisingly, in this work, a practical absence of the effect of stack pressure on the morphology of sodium deposits was found; although, for lithium deposition under similar conditions, this effect is very noticeable [35,51]. This difference in the nature of lithium and sodium deposition is explained by the higher plasticity of sodium (the yield strength for lithium is 16 MPa [52], while for sodium it is only 0.2 MPa [53]).

In [36], the processes of sodium deposition and dissolution at the interface of a current collector made of aluminum with a carbon black coating and NZSP as an electrolyte were studied using electron backscatter diffraction (EBSD). It was found that sodium under these conditions was deposited in the form of fairly large crystallites (up to 150 μm wide), with the intercrystallite boundaries being strictly perpendicular to the plane of the current collector. During anodic dissolution of these deposits, voids (pores) were formed, and the interfacial resistance increased sharply.

To reduce the interfacial resistance at the interface between the solid electrolyte Na₃Zr₂Si₂PO₁₂ (NZSP(it is worth noting that the abbreviation NZSP is used for all stoichiometry of the compounds Na_xZr₂Si_yP_zO₁₂, and proper values of indexes x, y, z must be clear from context)) and the aluminum current collector, the work [54] proposes introducing an intermediate elastic ferroelectric layer with a thickness of about 3 μm, containing nanoparticles of barium titanate and carbon. Such a layer on the one hand dampens volumetric changes during the deposition and dissolution of sodium, and on the other hand, it makes the sodium ion flows more uniform. Accordingly, the sodium deposition is not in the form of dendrites but in the form of a smooth layer. In addition, the presence of this layer leads to a noticeable decrease in the nucleation overvoltage (from 53 to 4 mV at a current density of 0.1 mA/cm², and from 73 to 14 mV at a current density of 1 mA/cm²). A laboratory model of an anode-free battery with the described intermediate layer, NZSP as an electrolyte, and a positive electrode based on Na₃V₂(PO₄)₃ (NVP) withstood 300 cycles with a current density of 0.1 mA/cm² with an insignificant loss of capacity (4.6%).

In [55,56], the initial stages of sodium whisker deposition at the interface of a gold current collector and polycrystalline Na-β″-alumina as an electrolyte were studied using the elegant method of in situ transmission electron microscopy (TEM) in combination with crystal orientation analysis (Figure 4). The sodium whiskers were found to grow along grain boundaries, and nucleation occurs not only directly at the interface of the electrolyte with gold, but also in the bulk of the electrolyte at the grain boundaries.

Thus, reducing the intergranular volume (in other words, increasing the relative density of the solid electrolyte) is an effective means of combatting dendrite formation.

3. The Interface Between a Solid Electrolyte and Sodium

In anode-free sodium batteries, the interface “solid electrolyte/sodium” appears just after the start of charging a freshly assembled or completely discharged battery. Naturally, such a boundary is also present in conventional (not anode-free) sodium metal batteries, and in the literature, one can find a lot of works on the study of such a boundary (see, for example, the review [57]). Of course, the fundamental issue is the compatibility of the solid electrolyte and sodium [43,58]. The interface “solid electrolyte/sodium” has many features of the interface “solid electrolyte/current collector”, including the presence of voids and contact spots and, in general, the phenomenon of current constriction and, accordingly, increased interphase resistance. A significant problem is the risk of sodium deposition in the form of dendrites or whiskers, and the solution to this problem is associated with the modification of the solid electrolyte/sodium interface. It is the presence of voids and contact spots at such an interface that leads to uneven current density distribution and promotes the formation of dendrites [27,45,59]. It is worth noting that, since sodium has higher plasticity and a higher self-diffusion coefficient, the risk of dendrite formation at the solid electrolyte/sodium interface is less than at the solid electrolyte/lithium interface [60].

The penetration of sodium dendrites into a solid electrolyte depends on the presence of intercrystalline pores, i.e., on the deviation of the true density of the solid electrolyte from theoretical values, as well as on the ionic and electronic conductivity of the electrolyte (naturally, the former should be as large as possible, the latter as small as possible). In [61], it was shown that doping the solid electrolyte NASICON with ions of rare earth elements Pr³⁺, Eu³⁺, and Lu³⁺, which have a larger radius than the replaced ion Zr⁴⁺, leads to a noticeable increase in density, a decrease in electronic conductivity, and, accordingly, to a decrease in the risk of dendrite formation.

SEI on the sodium surface adjacent to a solid electrolyte compares favorably with SEI formed in a liquid electrolyte. In [39], the composition of SEI on a sodium electrode in a liquid electrolyte (1 M NaClO₄ in the mixture 95% propylene carbonate and 5% fluoroethylene carbonate) and in contact with a solid electrolyte (NASICON, Na₃Zr₂Si₂PO₁₂) was studied using time-of-flight secondary ion mass spectrometry (TOF-SIMS). Figure 5 shows the material distribution across the SEI thickness in both cases. In the SEI formed in the liquid electrolyte, the products of electrolyte decomposition (ions CH₂O⁻, C₂F⁻, NaO⁻, and ClO₂⁻) as well as the products transferred from the positive electrode (ions MnO⁻, FeO⁻, MnF₃⁻, and FeF₃⁻) were identified. In contact with the solid electrolyte, SEI was formed as a sodium oxide. It is important to emphasize that the SEI in contact with the solid electrolyte was much thinner than the SEI in the liquid electrolyte.

In [62] it is noted that natural SEIs at the sodium/NASICON interfaces have much higher protective properties than SEIs at the lithium/solid electrolyte interfaces.

High-quality SEI at the sodium/solid sulfide electrolyte (Na_2.895W_0.3Sb_0.7S₄) interface is obtained from a film of polyperfluoromethyl isopropyl ether [CF(CF₃)CF₂O]_x(CF₂O)_y. When in contact with sodium metal, this film is enriched by NaF and reliably prevents dendrite growth and reduces interfacial resistance [63].

Various strategies of modification of the solid electrolyte/sodium interface contribute to the improvement of the quality of this boundary, including the reduction of the interfacial resistance and the suppression of dendrite formation. Thus, in [40], it was shown that one of the reasons for the increased interfacial resistance at the NZSP-sodium interface is the presence of functional groups (in particular, hydroxyl and carbonate) on the NZSP surface. Annealing of NZSP in air at a temperature of 800 °C allows for cleaning the surface, and additional annealing at a temperature of 1000 °C leads to partial surface restoration and an increase in sodiophilicity and completely eliminates dendrite formation. It was found, in [41], that annealing of NZSP leads to the formation of a protective nanometric layer of Na₃PO₄. Annealing of NASICON to clean its surface from oxygen-containing groups is also recommended in [64]. In [65,66], it was shown that the tendency to form oxygen-containing functional groups on the NZSP surface depends on the annealing mode. Synthesis of NZSP in a two-stage mode (first stage 1300 °C, 15 min, second stage 1200 °C, 12 h) allows for it to reduce the number of surface groups, increase the ceramics density, and also increase its ionic conductivity.

A similar result is described in [67], where it is shown that surface treatment of NASICON with liquid sodium at a temperature of 380 °C leads to the formation of a reduced sodiophilic surface layer, which provides a tenfold decrease in interfacial resistance. In [14], it is proposed to treat the surface of NZSP with a suspension of SbF₃ in liquid sodium. With such treatment, the surface of the solid electrolyte is covered with a highly sodiophilic layer of Na-Sb alloy containing NaF particles, which provide ionic conductivity. An original method for creating a sodiophilic surface of a solid electrolyte is described in [68]. Here, a porous layer of the same electrolyte about 0.1 mm thick is applied to the surface of a monolithic electrolyte (NZSP doped with Ca²⁺). The porous layer is impregnated with a SnCl₂ solution and then annealed in air, converting all the tin into SnO₂. After this, the porous layer is impregnated with liquid sodium, whereby the pore walls are coated with metallic tin, and the entire porous layer acquires increased sodiophilicity.

A somewhat similar technique was used in [69]. Here, a thin SnO₂ layer was applied to the surface of a solid NASICON-like electrolyte of the composition Na_3.2Hf_1.9Ca_0.1Si₂PO₁₂. Upon contact with sodium metal, an intermediate Na₂O/Na_xSn layer was formed, which provided low interfacial resistance. To create a similar sodiophilic layer on the NASICON surface, the authors of [46] recommend applying a saturated aqueous solution of lead acetate to the surface and annealing in an argon atmosphere at a temperature of 550 °C. This creates a Pb/C@NZSP composite layer that is perfectly wetted by sodium. The interfacial resistance at the boundary of sodium with the solid electrolyte upon application of such a layer decreased from 391 to 1.5 Ohm·cm².

The authors of [70] proposed applying a suspension of Super P carbon black in a liquid Na–K alloy to the surface of a solid electrolyte Na₃Hf₂Si₂PO₁₂ (NHSP) to form a quasi-liquid composite layer C@Na-K with a thickness of about 7 μm (such a suspension has the properties of a non-Newtonian liquid).

In [71], it is proposed to apply a thin layer of AlF₃ to the NASICON surface (from an aqueous solution with subsequent drying and annealing at a temperature of 450 °C in a nitrogen atmosphere). This 350 nm thick AlF₃ layer evens out the unevenness of the NASICON surface and therefore reduces the tendency to dendrite formation and interfacial resistance. When such a modified NASICON comes into contact with metallic sodium, Al, NaF, and also NamAlnF₃ are formed. The resulting layer reliably prevents the deposition of sodium dendrites.

Another method for reducing the interfacial resistance at the solid electrolyte/sodium interface is by introducing a conducting polymer layer. In the already cited work [67], cross-linked poly (ethylene glycol) methyl ether acrylate (CPMEA) is proposed as such a polymer. In [72], a similar layer with a thickness of about 20 μm is made of a nanoporous product of copolymerization of 5,5′,6,6′-tetrahydroxy-3,3,3′,3′-tetramethyl-1,1′-spirobisindane and tetrafluoroterephthalo-nitrile, the molecular structure of which is shown in Figure 6.

A sodiophilic interlayer made of sulfurized polyacrylonitrile (sulfurized PAN) on the surface of NASICON is proposed in [73].

In addition to polymers, the sodiophilic interlayer can be made simply of metals, in particular tin [74], oxides, e.g., TiO₂ [75], or carbon materials, specifically nanotubes [76] or graphene [77].

A relatively simple way to reduce the interfacial resistance at the solid electrolyte/sodium interface is proposed in [78]. It is shown here that doping Na₃Zr₂Si₂PO₁₂ with Zn²⁺ ions (replacing part of Zr with Zn) results in an increase in the electrolyte conductivity and, most importantly, in a decrease in the interfacial resistance. The electrolyte of the composition Na_3.20Zr_1.90Zn_0.10Si₂PO₁₂ has a specific conductivity of 1.58 mS/cm, which is four times greater than that of the undoped electrolyte, and the interfacial resistance at the boundary with sodium for such an electrolyte was 25 Ohm·cm², whereas, for the undoped electrolyte, this resistance was 785 Ohm·cm². It was shown that zinc segregates on the surface of the solid electrolyte, where a layer of sodium-zinc alloy is formed, ensuring the sodiophilicity of the interface. The beneficial effect of doping the NASICON electrolyte with zinc was also reported in [79].

An even simpler way to reduce the interfacial resistance at the solid electrolyte/sodium interface is described in [80]. It is shown here that ultrasonic treatment of such an interface at room temperature for only 25 s (practically ultrasonic welding) results in the formation of a very close contact between sodium and the solid electrolyte NASICON (Na₃Zr₂Si₂PO₁₂) (Figure 7).

A symmetrical Na|Na₃Zr₂Si₂PO₁₂|Na cell assembled using such ultrasound solid welding successfully withstood more than 200 cycles at a current density of 0.1 mA/cm² (1300 h) without any signs of dendrite formation or increase in overvoltage, which remained equal to 12 mV. A similar cell assembled without ultrasound solid welding did not withstand even 20 cycles.

The effect of stack pressure on the interfacial resistance at the solid electrolyte/current collector interface has already been noted above. Indicative data on the effect of uniaxial compression on the resistance at the solid electrolyte/sodium interface are presented in the study [81]. Here, the resistance of a NASICON pellet sandwiched between two sodium plates was measured at different compression pressures. The results are shown in Figure 8.

One can see that even a not too large compression pressure results in a decrease in interface resistance by an order of magnitude.

In the end of this section, it stands to right mentioning Refs. [82,83]. By using the density functional theory (DFT), it is shown here that monolayer coatings from some chalcogenides, namely ScTe₂ and V₂S₂, demonstrate the excellent sodiophilic properties.

4. The Interface Between a Solid Electrolyte and a Cathode

The interface between solid electrolyte and positive electrode in anode-free sodium batteries differs both from the interface between a solid electrolyte and negative electrode in such batteries and from the interface between a solid electrolyte and positive electrode in batteries with liquid electrolytes (although, of course, the interface between a solid electrolyte and cathode in anode-less sodium batteries is no different from the same interface in conventional, non-anode-free batteries). First of all, it is worth noting that there is a wide variety of possible active materials for positive electrodes. In addition to the most popular sodium vanadate-phosphate Na₃V₂(PO₄)₃ [80,84,85,86,87,88,89], the following oxides can be named: Na_xCoO₂ [90], NaCrO₂ [91], NaNi_0.4Fe_0.2Mn_0.4O₂ (NFM) [92], Na_0.66Ni_0.33Mn_0.67O₂ [93], P2–Na_0.68Li_0.10Ni_0.25Mn_0.63Ca_0.01Ti_0.01O₂ (NM-L10CN) [94], sodium titanophosphate NaTi₂(PO₄)₃ [67], double-cyano-perovskite Na₂MnFe(CN)₆ [39], perylene-3,4,9,10-tetracarboxylic dianhydride (PTCDA) [95], and pyrene-4,5,9,10-tetraone (PTO) [78].

Unlike the negative electrode, the positive one (cathode during discharge) contains a porous (essentially granular) active layer of a certain thickness. In order for such a layer to be able to operate across its entire thickness, it consists of a mixture of active material particles and a solid electrolyte, with the volume of solid electrolyte particles exceeding the percolation threshold to maintain the connectivity of all these particles. It is worth recalling that a certain amount of conductive additive, often carbon black, is also introduced into the active mass of the positive electrode to ensure electronic conductivity across the entire depth of the active layer. In addition, the actual particles of the active material are often coated with a nanometer layer of carbon. For example, in work [88], the active mass of the positive electrode included active material (Na₃V₂(PO₄)₃ with a carbon coating), a solid electrolyte (Na_3.4Zr_1.8Mg_0.2Si₂PO₁₂), a conductive additive (super P), and binder (PVDF—polyvinylidene fluoride) in a weight ratio of 65:20:5:10.

If the liquid electrolyte permeates the entire porous space and enables the entire internal surface to participate in the current-producing process, then the mixture of particles of the active material and the solid electrolyte provides only point-to-point contacts. A more elegant method of organizing good contact between the active material and the electrolyte is to coat each particle of the active material with the electrolyte, for example, using pulse laser deposition.

As in the case of the negative electrode, cycling (sodium insertion/extraction) of the positive electrode is accompanied by volume changes, however, they are generally smaller than that at the negative electrode. Thus, when extracting two sodium ions from Na₃V₂(PO₄)₃ (during charging), the volume decrease is only 0.21 cm³/Ah, i.e., four times less than the change in the volume of sodium (with due account for the density of Na₃V₂(PO₄)₃ equal to 3.156 g/cm³ [87,96]). A significant difference between the interfaces of the solid electrolyte with the negative and positive electrodes also consists of the fact that, in the former case, the volume changes are one-dimensional and directed normally to the substrate surface, while, in the latter case, there is a three-dimensional change in the volume of each particle (granule) of the active material.

To improve the contact of the cathode material with the solid electrolyte (in essence, to reduce the interfacial resistance), in [85], it was proposed to wet the contact area of the NASICON solid electrolyte with the positive electrode based on sodium vanadate phosphate with a very small amount (0.5 μL/cm²) of liquid electrolyte (0.8 M NaPF₆ in a mixture of ethylene carbonate with dimethyl carbonate) (Figure 9). This approach allowed them to significantly improve the battery characteristics, but due to partial evaporation and decomposition of the electrolyte, the cyclic resource was limited. A noticeable improvement was achieved by replacing the carbonate electrolyte with the same amount of ionic liquid (N-methyl-N-propylpiperidinium-bis(fluorosulfonyl) imide (PP13FSI)). In this version, the battery withstood 10,000 full charge–discharge cycles in the 10 C rate.

A similar approach was used in [93]. Here, the active mass of the positive electrode is a paste (in the authors’ terminology, “toothpaste”) of Na_0.66Ni_0.33Mn_0.67O₂ particles with a carbon coating and N-butyl-N-methylpyrrolidinium bis(fluorosulfonyl)imide (PY14FSI) ionic liquid, the structural formula of which is shown in Figure 10. This paste is simply applied to the surface of Na-β″-Al₂O₃ solid electrolyte using a doctor blade.

The authors of [84] proposed impregnating the active mass of the positive electrode with a so-called plastic–crystal Na+-electrolyte. This electrolyte is a complex obtained by heating a mixture of succinonitrile (N≡C(CH₂)₂C≡N) and NaClO₄ with a molar ratio of 20:1 at 65 °C. When cooled to room temperature, it solidifies, remaining X-ray amorphous, and it has an ionic conductivity of 3.3 mS/cm, which exceeds the conductivity of the lithium analogue [97].

As already mentioned above, the authors of [67] proposed a CPMEA polymer interlayer to reduce the interfacial resistance at the solid electrolyte/sodium interface. The same work proposes the same interlayer to reduce the interfacial resistance at the solid electrolyte/cathode interface. In this case, the cathode was made of sodium fluorotitanate using CPMEA as a binder that also has the properties of an ionic electrolyte.

Another polymer, namely a copolymer of vinylidene fluoride with hexafluoropropylene (PVDF–HFP) for the same purposes, is proposed in [92]. Here, a mixture of PVDF–HFP with a suspension of the active material NFM in N-methylpyrrolidone was applied directly to the surface of the solid electrolyte (Na₃Zr₂S_i2.2P_0.8O₁₂).

In [94], a very sophisticated active material of the positive electrode, specifically Na_0.68Li_0.10Ni_0.25Mn_0.63Ca_0.01Ti_0.01O₂ (NM-L10CN), was reported. A battery with such an electrode and solid NZSP electrolyte, successfully withstood 500 cycles at 1 C rate. Intriguingly, this article says nothing about measures to reduce interphase resistance at the boundary of the solid electrolyte with the cathode.

It is known that materials such as Prussian blue, although attractive in terms of specific capacity, are unstable in contact with liquid organic electrolytes. In the above-mentioned work [39], the performance of such material, specifically double-cyano-perovskite with formula Na₂MnFe(CN)₆ (NMHCF), was assessed in contact with the Na₃Zr₂Si₂PO₁₂ solid electrolyte. The active mass of the positive electrode in this case was made from a mixture of NMHCF, carbon black as an electrically conductive additive, and a solid polymer electrolyte based on polyethylene oxide and NaClO₄. Such a battery successfully withstood 200 cycles at the C/2 rate with a loss of only 10% of capacity.

5. Solid Electrolytes

In principle, the solid electrolyte for an anode-free sodium battery is no different from the solid electrolyte for a conventional (not-anode-free) battery. This explains the fact that there are a huge number of publications in the literature on solid electrolytes for conventional sodium batteries (it is high time to mention review papers such as [13,15,43,98,99,100,101,102,103,104,105,106,107]) but practically nothing on anode-free versions. Indeed, we are unaware of research devoted to this very topic. The main requirements for solid electrolytes for anode-free sodium batteries are obvious. Such electrolytes must possess sufficiently high conductivity for sodium ions and a sufficiently high transference number of such ions, a wide electrochemical stability window, chemical and mechanical stability, low electronic conductivity, ability to inhibit dendrites, good mechanical properties, facile preparation, and environmental friendliness. Of particular importance is the requirement for a high sodium ion transference number, since a deviation of this quantity from unity leads to the occurrence of concentration polarization with a corresponding decrease in discharge voltage and energy density.

Solid polymer electrolytes are sometimes included in the category of solid electrolytes, but polymers should be considered supercooled liquids rather than true solids. In addition, solid polymer electrolytes, as a rule, possess rather low sodium ion transference numbers. They are not considered in this article. Among the numerous inorganic solid electrolytes with sodium conductivity, groups such as NASICON-type materials, sulfides [108], and complex hydrides [100,108] can be distinguished. The most popular are NASICON-type materials, which were proposed as an electrolyte in 1976 as an alternative to β-alumina [109,110].

The composition of NASICON-type material can be described by the generalized formula NaMM′(PO₄)₃, where the M and M’ positions can contain various transition metal ions in the divalent, trivalent, tetravalent, or pentavalent states, and phosphorus can be partially replaced by silicon or arsenic [105]. Standard ceramic methods are used to synthesize NASICON-like electrolytes, i.e., sintering of precursor’s mixtures, which are obtained by simple dry mixing, sol-gel route, hydrothermal method, or ion exchange. Depending on the composition and synthesis method, NASICON can crystallize in the following structures: a rhombohedral (prevalent in NASICON system), monoclinic, triclinic, orthorhombic, corundum-like, etc. NASICON-like electrolytes have a specific ionic conductivity of about 0.1 mS/cm at room temperature, which depends on the specific composition and crystal structure [111]. For example, in [112], it was found that the introduction of a superstoichiometric amount of sodium into standard NASICON (transition from Na₃Zr₂Si₂PO₁₂ to Na_3.3Zr₂Si₂PO₁₂) leads to an increase in the room-temperature specific conductivity from 0.2 to 0.9 mS/cm. A similar result (2.4 mS/cm for Na_3.3Zr₂Si₂PO₁₂) was reported in [113]. In [114], materials of the compositions Na_3.4Zr₂Si_2.5P_0.5O₁₂ and Na_3.4Zr₂Si_2.4P_0.6O₁₂ with a specific conductivity of about 5 mS/cm were reported.

In [115], it was shown that doping NASICON with yttrium (by replacing the ZrO₂ precursor with (ZrO₂)_0.97(Y₂O₃)_0.03) leads to a fivefold increase in ionic conductivity. In general, doping with cations is an effective means of increasing the ionic conductivity of NASICON [106,116,117,118,119,120,121], with hafnium, yttrium, niobium, cobalt, and zinc being the most effective dopants. Among the NASICON-like materials with the highest ionic conductivity at room temperature, the following should be mentioned: Na_3.2Hf₂(SiO₄)_2.2(PO₄)_0.8 (specific conductivity 2.3 mS/cm) [122], Na_3.4Zr_1.8Ni_0.2Si₂PO₁₂ (2.28 mS/cm) [116], Na_3.3Hf₂(SiO₄)₂(PO₄) (1.1 mS/cm) [122], Na_3.36Zn_0.08Zr_1.92(SiO₄)_2.2(PO₄)_0.8 (1.1 mS/cm) [123], Na₃Zr_1.8Zn_0.2Si₂PO_11.8 (1.44 mS/cm) [120], Na_3.232Co_0.016Zr_1.984(SiO₄)_2.2(PO₄)_0.8 (1.41 mS/cm) [124], Na_2.1Y_0.1Zr_1.9(SiO₄)₂(PO₄) (1.98 mS/cm) [117], Na_3.5Zr_1.75Mg_0.25Si₂O₁₂ (2.5 mS/cm) [113], Na_3.4Sc_0.4Zr_1.6(SiO₄)₂(PO₄) (4.0 mS/cm) [125], Na_2.96Nb_0.04Zr_1.96(SiO₄)₂(PO₄) (4.95 mS/cm) [117], and Na_3.4Zr_1.9Zn_0.1Si_2.2P_0.8O₁₂ (5.27 mS/cm) [121].

As was indicated in Section 2, a decrease in the intergranular volume (an increase in the relative density of NASICON) is important from the point of view of combatting dendrite formation. Doping NASICON with cations in some cases leads to a noticeable increase in the relative density [68,85,121,126]. In [61], it was shown that large ions of rare earth elements such as Pr³⁺, Eu³⁺, and Lu³⁺ are particularly effective in this regard. Thus, if the relative density of undoped Na₃Zr₂Si₂PO₁₂ is 87.8%, then for Na_3.3Zr_1.7Pr_0.3Si₂PO₁₂, Na_3.3Zr_1.7Eu_0.3Si₂PO₁₂, and Na_3.3Zr_1.7Lu_0.3Si₂PO₁₂, it increases to 94.2%, 94.5%, and 92.3%.

Besides doping, various technological methods are an important means of increasing ionic conductivity. For example, in [127], a NASICON-like electrolyte was synthesized by sintering a mixture of canonical Na₃Zr₂Si₂PO₁₂ with the addition of 5% Na₂SiO₃ at a temperature exceeding the melting point of Na₂SiO₃. The resulting material had an ionic conductivity at room temperature as high as 1.45 mS/cm, which was associated with the formation of a non-stoichiometrically silicon-enriched phase at the intercrystalline boundaries.

Sulfides are the second most important group of inorganic solid electrolytes with sodium ion conductivity [128]. The most important are electrolytes based on Na₃PS₄ [129], Na₃SbS₄ [130,131], and Na₁₁Sn₂PS₁₂ [132]. Sulfide electrolytes are characterized by a wide electrochemical stability window (over 5 V) and stability to sodium [133,134,135,136]. There are amorphous (glassy), crystalline, and glass-ceramic samples. Plain amorphous examples of the Na₂S–P₂S₅ system have rather modest ionic conductivity at room temperature (about 0.01 mS/cm). The transition to crystalline forms and to glass ceramics makes it possible to increase this conductivity to fractions of mS/cm [129,133,135,137,138,139]. An even greater increase in ionic conductivity can be achieved through doping, mainly anionic [140,141,142,143,144,145,146].

The positive effect of replacing part of the phosphorus in Na₃PS₄ with arsenic is described in [147]. The material with the composition Na₃P_0.62As_0.38S₄ has a room-temperature conductivity of about 1.46 mS/cm. The most effective cationic dopants are silicon, germanium, and tin. According to [145], the ionic conductivity at room temperature for materials with the compositions Na_3.125Si_0.125P_0.875S₄ and Na_3.0625Sn_0.0625P_0.9375S₄ amounts to 2.99 and 10.7 mS/cm. The ionic conductivity of Na_3.1Ge_0.1Sb_0.9S₄ at room temperature is 5.1 mS/cm [146].

In [142,144,148], it is shown that the substitution of sulfur in Na₃PS₄ with selenium leads to a noticeable increase in ionic conductivity. Thus, according to [144], Na₃PSe₄ has a specific conductivity of 1.16 mS/cm at room temperature. More importantly, the transference number for the Na⁺ cation in Na₃PSe₄ is practically equal to unity.

Probably, the compound Na_2.88Sb_0.88W_0.12S₄ has the highest ionic conductivity among doped sulfides; the literature gives room-temperature ionic conductivity values of 2.5 [149], 19 [150], and 32 mS/cm [151]. This discrepancy is due to different heat treatment conditions of the ceramics; the value of 32 mS/cm was obtained after 12 h annealing at 275 °C, which resulted in higher crystallinity and density. Figure 11 and Figure 12 compare the crystal structure of Na_2.88Sb_0.88W_0.12S₄ and undoped Na₃SbS₄. The inclusion of tungsten in crystalline Na₃SbS₄ results in the appearance of Na vacancies in the lattice, which promotes the formation of a cubic phase, which can facilitate isotropic three-dimensional conductivity of fast ions.

The transference number of the Na⁺ cation in this case was also close to unity. Interestingly, the Na_2.88Sb_0.88W_0.12S₄ composites with sodium iodide have an even higher ionic conductivity, which depends on the NaI content. The maximum conductivity of 36 mS/cm is exhibited by the Na_2.88Sb_0.88W_0.12S₄·0.50NaI composite [149] (Figure 13). At the same time, the introduction of NaI into the composite with Na₃PS₄ does not lead to a noticeable increase in conductivity [152].

The properties of the Na_3–xP_1–xW_xS₄ electrolyte depend on both the composition (degree of tungsten addition) and the synthesis method [153]. With the same mechanochemical synthesis, compositions with x = 0.09, 0.15, 0.18, and 0.21 had specific conductivities of 1.84, 8.8, 2.47, and 1.69 mS/cm, respectively. At the same time, according to data from [154], the Na_2.9P_0.9W_0.1S₄ electrolyte synthesized by solid-phase synthesis had a conductivity of 13 mS/cm, while Na_2.9Sb_0.9W_0.1S₄ had a conductivity of 41 mS/cm. It is worth noting that such properties of Na₁₁Sn₂PS₁₂ are not too sensitive to small changes in composition. According to [155], electrolytes with compositions Na₁₁Sn₂PS₁₂, Na_10.875Sn₂PS₁₂, and Na_11.125Sn₂PS₁₂ have conductivities of 2.4, 2.3, and 2.4 mS/cm, respectively. At the same time, the conductivity of Na₁₀SnP₂S₁₂ is only 0.4 mS/cm [156].

Important dopants for sulfide electrolytes are halides. Thus, in [140], it was shown that Na_2.9375PS_3.9375Cl_0.0625 has a conductivity of 1.38 mS/cm at room temperature. In a later work of the same team [157], for Na₃PS_3.8Cl_0.2, Na_2.875PS_3.875Cl_0.125, and Na_2.9375PS_3.875Cl_0.125, the ionic conductivity values of 1.6, 6.38, and 9.03 mS/cm are reported. In [158], a series of halogen-substituted sulfides of the general formula Na_3−xSbS_4−xHal_x (0 ≤ x ≤ 0.3, Hal = Cl, Br, and I) was studied, and it was found that Na_2.85SbS_3.85Br_0.15 has the maximum room-temperature ionic conductivity of 2.87 mS/cm.

Of interest is the effect of doping sulfide electrolytes with oxygen on their conductivity and stability in contact with sodium. The authors of [159] investigated the introduction of small amounts of phosphorus and oxygen (in a ratio of 4 P:10 O) into the Na₃SbS₄ glass ceramic electrolyte and found that the highest ionic conductivity of 3.82 mS/cm was possessed by the composition Na₃SbP_0.16S₄O_0.4, pre-annealed in a vacuum at a temperature of 500 °C for 10 h (annealing led to an increase in crystallinity). At the same time, in [150], it was discovered that the introduction of oxygen into Na_2.88Sb_0.88W_0.12S₄ due to the substitution of sulfur led to a slight decrease in ionic conductivity. Moreover, based on first-principle calculations in [160], it was shown that the ionic conductivity of Na₇P₃O₁₁ at room temperature is only 3 μS/cm, whereas for Na₇P₃S₁₁ and Na₇P₃Se₁₁, values of 10.97 and 12.56 mS/cm, respectively, were obtained.

Of other types of solid electrolytes, mention should be made of hydride (and borohydride) systems [161,162], thioborates [163], and the chloride electrolyte (Na_1−xZr_xLa_1−xCl₄) [164], although they are of much less importance than NASICON and sulfides.

6. Forecast of Possible Characteristics of Future All-Solid-State Anode-Free Sodium Batteries

As is clear from the Section 1, we are unaware of any literature data on the characteristics of specific all-solid-state anode-free sodium batteries. However, published data on the electrochemical properties of ordinary all-solid-state sodium battery systems provide an idea of the possible performance of future anode-free counterparts.

Table 1. Comparison of the characteristics of all-solid-state sodium batteries based on different electrochemical systems.

Electrolyte	Cathode	Anode	Voltage Range, V	Ionic Conductivity of Electrolyte, RT, S/cm	Cathode Specific Capacity, mAh/g	Cycles	Ref.
Na3Zr2Si2PO12	Na2MnFe(CN)6	Na	4.0–2.5	1.2·10−3 at 60 °C	120 at 0.5 C	200	[39]
Na3.4Zr2(SiO4)2.4(PO4)0.6	Na3V2(PO4)3	Na	3.8–2.0	5·10−3	0.6 mAh/cm2 at 0.5 mA/cm2	70	[45]
Na3Zr2Si2PO12	Na3V2(PO4)3	Na	3.8–2.0	6·10−4	88 at 0.2 C	100	[46]
Na3Zr2Si2PO12	NaCu1/9Ni2/9Fe1/3Mn1/3O2	BaTiO3/C@Al	3.8–2.0	6·10−4	93.1 at 0.1 mA/cm2	300	[54]
Na3Zr2Si2PO12	Na3.5V0.5Mn0.5Fe0.5 Ti0.5(PO4)3	Na	4.4–2.0	6·10−3	152 at 1 C	1000	[65]
Na3Zr2Si2PO12	Na3V1.5Cr0.5(PO4)3	Na	4.3–2.5	0.85·10−3	103 at 1 C	400	[66]
Na3.2Zr1.9Ca0.1Si2PO12	Na3V2(PO4)3	Na	3.8–2.7	1.67·10−3	94.9 at 1 C	450	[68]
Na3.2Hf1.9Ca0.1Si2PO12	Na3V2(PO4)3	Na/SnO2	3.8–2.7	1.07·10−3	103.1 at 0.5 C	300	[69]
β″-Al2O3	Na3V2(PO4)3	Na	3.8–2.5	1.17·10−4	100 at 0.2 C	100	[76]
Na3Zr2Si2PO12	Na3V2(PO4)3	Na	4.0–2.6	4.3·10−4	110 at 0.1 mA/cm2	900	[80]
Na3Zr2Si2PO12	Na3V2(PO4)3	Na	4.0–2.5	3.3·10−4	90 at 0.1 C	100	[84]
Na3.4Zr1.6Sc0.4Si2PO12	Na3V2(PO4)3	Na	3.8–2.3	1.77·10−4	100 at 1 C	300	[86]
Na-β″-Al2O3	Na0.66Ni0.33Mn0.67O2	Na	3.8–2.5	1.0·10−3	79 at 6 C	10,000	[93]
Na3.4Zr1.9Zn0.1Si2.2P0.8O12	FeS2	Na	2.5–0.7	5.27·10−3	236.5 at 0.1 C 133.1 at 0.5 C	100 300	[121]
Na3.1Zr1.95Mg0.05Si2PO12	Na0.9Cu0.22Fe0.3Mn0.48O2	Na	3.8–2.5	1.33·10−3	57.9 at 0.5 C	100	[126]
Na3PS4	NaCrO2	Na15Sn4	4.0–1.0	4.6·10−4	90 at 0.013 mA/cm2	20	[129]
Na3PS4	NaCrO2	Na15Sn4	3.0–1.0	4.0·10−4	250	10	[131]
Na3SbS4	NaCrO2	Na	4.0–1.5	1.1·10−3	108 at 0.05 mA/cm2	1	[134]
t-Na2.9375PS3.9375Cl0.0625	TiS2	Na	2.4–1.2	1.14·10−3	80 at 0.1 C	10	[140]
Na3P0.62As0.38S4	TiS2	Na-Sn	2.4–1.2	1.46·10−3	150 at 0.02 C	10	[147]
Na2.88Sb0.88W0.12S4·0.5NaI	TiS2	Na-Sn	2.4–1.2	10−2	100 at 0.057 C	5	[149]
Na2.88Sb0.88W0.12S3.7O0.3	TiS2	Na-Sn	2.4–1.2	2.2·10−3	130 at 0.057 C	5	[150]
Na2.85P0.85W0.15S4	TiS2	Na-Sn	2.4–1.2	8.8·10−3	140 at 0.057 C	5	[153]
Na3.0PS3.8Cl0.2	Na3V2(PO4)3	Na	3.8–2.5	1.96·10−3	100 at 0.1 C	10	[157]
Na0.7Zr0.3La0.7Cl4	NaCrO2	Na2Sn	3.4–2.0	2.9·10−4	114 at 0.3 C	70	[164]

provides literature data on various electrochemical systems of all-solid-state sodium batteries.

7. Conclusions

All-solid-state anode-free sodium batteries present a special category of energy storage devices. Up to date, there are neither industrially produced batteries, nor laboratory ones like these such batteries, but the prospects for their creation are quite optimistic. This optimism is explained by the fundamental advantages of solid-state anode-free sodium batteries over other categories.

The advantages of anode-free sodium batteries over sodium-ion batteries are greater specific capacity and energy density. The advantages of anode-free sodium batteries over conventional sodium metal batteries are the absence of operations with metallic sodium during production and assembly, that is, in the radical simplification and reduction of the cost of technology. The advantages of batteries with a solid electrolyte over batteries with a liquid electrolyte are greater chemical stability, good mechanical properties, and increased safety. In addition, solid electrolytes usually have a sodium cation transference number close to one, which ensures an increase in energy density. The advantages of solid-state anode-free sodium batteries over their lithium counterparts include the much lower cost and greater availability of sodium, as well as the ability to use aluminum rather than copper as a current collector.

The above-listed advantages of all-solid-state anode-free sodium batteries have certain economic and environmental aspects. The already mentioned simplification and reduction in cost of the production technology of such batteries, associated with the exclusion of operations with sodium metal, also contribute to a significant scaling of production, which is especially important for energy storage systems. The exclusion of liquid electrolytes provides certain environmental advantages both in the production and in the recovering of batteries.