Strategies to develop stable alkali metal anodes for rechargeable batteries

Intercalation-based materials were commercialized soon after due to their increased safety and stability. The theoretical gravimetric and volumetric capacities of graphite, which is the most commonly used anode in Li-ion batteries, are 372 mAh g⁻¹ and 735 mAh cm⁻³, respectively [16]. The capacity of intercalation-based anodes is approaching their theoretical value and cannot cope with the current energy demand [17–22]. Among the others, the alloying and conversion-based chalcogenide anode candidates provide a higher capacitance but are limited by their rate performance at high current rates and cycle life due to the problems caused by volume expansion (leading to pulverisation and low CE) [23]. Anodes for the Na ion batteries (SIBs) fall into many diverse regions with carbon, titanium, chalcogenides, alloying and organic materials acting as key players [24–30]. The most widely used anode in SIB is hard carbon that deliver a capacity of 300 mAh g⁻¹ at a lower potential and for potassium is graphite. A revisit of the alkali metal anodes has recently sparked huge interest in researchers to develop batteries with high energy density [9, 31, 32].

Though lots of work can be done on the cathode side towards developing high-energy-density batteries [33], the difficulties of increasing the alkali ion transfer per transition metal and the slow progress in extending the electrochemical window make it challenging. There is huge scope for modifying the anode towards increasing the energy density [23]. While the high theoretical specific capacity and high negative redox potential of lithium make it a dominant choice in today's power sources [34], Na and K, with their low cost and abundance in the earth's crust, show promise towards the development of large-scale energy systems [34–37].

The scarcity of the Li reserves in the earth's crust and the associated price hike of Li metal at this time of high energy demand have led us to look into alternate battery chemistries for energy storage [38, 39]. Na and K, due to their abundance and low cost, are considered promising candidates for futuristic anodes for aqueous and solid-state batteries [24, 40–46]. As Na and K do not alloy with the aluminium, replacing the copper current collector would further lead to reduced cost and improved gravimetric energy densities in these systems [47–52]. Li-ion, being a stronger Lewis acid (due to its smaller ionic size) than Na and K, has a direct effect on the de-solvation energy [47, 53]. More solvated polar molecules are needed for the stabilization of the Li ions, which increases the de-solvation energy along with the increase in the solvation shell radius. This increases the interfacial resistance during the ionic movement through the SEI, resulting in reduced ionic conductivity. In this regard, Na and K could offer better rate performances to meet high-power applications compared to Li [54, 55]. Due to the soft nature of Na and K, a smaller amount of mechanical pressure can have a huge impact on dendrite suppression. In this review, we attempt to project the main challenges associated with alkali metal anodes and discuss various strategies to mitigate the issues in-order to realize the faster implementation of solid-state alkali metal batteries.

The parasitic reduction reaction is initiated when the energy of the lowest unoccupied molecular orbital of the electrolyte is less than the electrochemical potential of the anode, leading the formation of the solid electrolyte interface [63]. The term solid electrolyte interface and the possible model for the same was introduced by Peled in 1979 [64]. The ideal property of the SEI includes perfect ionic conductance with quality of electronic insulation [23]. A stable SEI acts as a protective layer and prevents further decomposition of electrolytes and loss of active material. The mechanical stability and optimum thickness of the SEI can directly impact the cycle life and Coulombic efficiency.

Generally, two different models have been proposed for the SEI: Mosaic and multilayer models. In the former model, the reductive reactions are considered to simultaneously produce an interface with mixed components [65] in the multilayer model, the components are formed inhomogeneously concerning the depth of the interphase [66]. The inorganic species with low oxidation states (Li₂O, Li₃N, LiF, LiOH, and Li₂CO₃) are formed closer to the alkali metal and the organic species with higher oxidation states (ROCO₂Li, ROLi, and RCOO₂Li) are formed closer to the electrolyte contact [23].

While stable passivation acts as protection, its formation is more difficult in Na and K compared to that of Li due to their high reactivity. An uneven interface is formed for Na when it comes in contact with the electrolyte, far worse than the case of Li. The pits and heaves within these rough interphases formed during the battery assembly can act as further seeding sites for dendrites [67]. The higher reactivity of the K compared to the other two alkali metals makes it more difficult to form a stable passivation layer [68]. Moreover, the reactivity itself reduces the number of compatible and stable electrolytes for the K-based battery systems and a mere adoption of electrolytes from other systems is not possible [69]. The dissolution of the passivation layer can expose fresh alkali metals and lead to further electrolyte decomposition. These new sites can initiate dendrite formation as well. The Lewis acidity difference of the alkali metal ions can lead to differences in the solubility of these decomposition products [34]. The weak Coulombic interaction of Na⁺ with the solvate molecules has led to the dissolution of its decomposition products (sodium oxide, hydroxide and carbonates) in comparison to its Li counterparts [70].

One of the main reasons for the failure of commercialization of the alkali metal anodes were the safety issues. Thermal runaways, electrolyte combustion and cell explosion can be induced when the battery is short-circuited by the internal unevenly grown 'dendrites'. Non-uniform deposition of the alkali metal anodes leads to the formation of dendrites [9]. The mechanism of nucleation and the growth of the dendrites are not fully understood, even though many models have been proposed [32, 71–74].

Dendrites can occur in different morphologies: needle-like, tree-like and moss-like morphologies. The needle-like dendrites are the main cause of the separator piercing as they are grown in length and diameter without any branches. In contrast, the tree-like morphologies are not common but have been modelled and studied. Both these morphologies are mainly observed during the initial cycles. Upon ageing, the moss-like morphologies are observed, which due to their high specific area and weak mechanical nature leads to the increased decomposition of electrolyte along with the formation of dead alkali metal. In most cases, the morphologies co-exist in a battery and transformation of morphologies also happens depending on the experimental conditions [23].

Different models have been proposed for forming the dendrites, and different in-situ imaging techniques have also been used to understand the process. Generally, a high local current density within the vicinity of high ionic concentration can lead to the formation of dendrites. When the concentration of the cations near the anode decays to zero, the surface inhomogeneities of the anode can lead to a variation in the local electric field, which might lead to dendritic formation. SEI, made of different components with diverse migration and surface formation energies, can form inhomogeneities at the surface, creating hotspots for dendritic formations. Different morphologies of the dendrites have been observed, among which the needle and whisker-shaped ones lead to the short circuit while the 3D moss-like structure leads to dead metals [23].

When it comes to the practical application of high areal capacity (>3 mAh cm⁻²) and high current density (>3 mA cm⁻²), a poor understanding of the dendrite formation can lead to poor cyclability [7]. High current density, high areal capacity and reduced pressure environment of the practical batteries can lead to the quick formation of dendrites. The low plasticity of the Li metal anode can in turn, produce microscopic defective sites on the surface that acts as seeds for the dendritic formations [16]. Short circuits created by the dendrites can easily happen at the high current densities, while the low current densities, due to the powdering of the Li, can lead to more formation of dead lithium [16]. The dead alkali metal formation also depends upon the dissolution of the dendrites in their corresponding electrolytes. Faster dissolution of the Na dendrites has been observed compared to the Li in LiPF₆ and NaPF₆-based solvents. The high chemical activity of the Na leads to an accelerated reaction with the trace amount of water within the solvents, leading to faster dissolution [67].

The major problem with Li metal anodes is the Coulombic efficiency rather than the dendrites [75–77]. The current advances in the high-concentration electrolytes have mitigated the failures like thermal runaways caused by the dendrites by aiding the formation of interwoven whisker-like structures which seldom pierce the separator before the failure caused by the CE. The main cause of the CE being less than 99% can be attributed to the inactive Li that can lost in the formation of continuous SEI formation or in the electrically insulating environment. Among them the electronically cut out Li has been identified as the main cause for the reduced CE which can be due to the whisker like structure of dendrites formed due to the tortuosity factor, lack of porosity and due to the heterogenous SEI [78].

Weak mechanical stability of the Na dendrites due to their larger ionic radius and reduced metallic bonding has been observed. The accelerated dissolution of the Na dendrites along with the weak mechanical stability leads to its breaking and formation of dead Na [67]. Among the alkali metals, K showcases the weakest mechanical feature with the highest reactivity [79].

The lack of a host for the alkali ions, as in the case of graphite and silicon, leads to an infinite volume change and 100% depth of discharge for the alkali metals. Loss of active material through dissolution followed by the pulverisation, increase in the resistance due to the inhomogeneous porous deposition, cracking of the SEI and associated side reactions, and the overall capacity fade and loss of Coulombic efficiency are some serious problems associated with the volume expansion [23].

The alkali metal's packing density during the deposition depends on numerous factors like the morphology, cell configuration and anode structure. Theoretically, the volume change per mole of the Na and K are calculated to be twice and four times that of Li. Thus, the problems associated with the volume expansion will be elevated for Na and K compared to Li.

High-temperature Na–S batteries, zebra batteries and thin film Li batteries are some of the commercialized battery systems that employ alkali metal anodes. While the HT Na–S batteries have been utilized for grid storage, the lithium thin film batteries found its application in wearable electronics. With the success of the Li-ion batteries, Li metal anode is expected to be the earliest to commercialize.

Mechanical stress can also cause the formation of cracks in the lithium deposit, which can serve as nucleation sites for dendrite growth and increase the likelihood of short circuits. In addition, stress can cause the growth of dendrites to become more uneven and result in the formation of thicker and more irregular dendrites.

Acid-base theory in oxygen doping: aO-GNR (carboxylic group containing graphene nanoribbons) with an extra pair of electrons in the carboxylic functional group site is expected as an electron-rich donor that naturally acts as Lewis base sites to absorb Lewis acidic Li ions through acid-base interactions strongly. Most electronegative F doping is not good: F, O & N doping forms sigma-bonds with adjacent carbon and withdraws electrons through inductive effects. O and N atoms participate in the delocalized pi system of GNR that contributes to their negative charge states. In contrast, the filled p orbitals of the F atom form a p–pi conjugation with the carbon plane that feedbacks electrons from the F atom to carbon. As a result, the F atom is less electronegative than the N and O atoms. S, Cl, Br, due to their increased size, leads to weak interaction with Li [83].

On the other hand, in the case of boron doping, boron is electropositive, and the adjacent carbon atoms are electronegative due to the smaller electronegativity of the B atom (2.04) against the C atom (2.55).

Local geometrical and electronic structures of the dopant also play a vital role. Herein, the 'local dipole', is defined as the dipole formed by the doping atom and its adjacent carbon atom. A strong local dipole not only renders a strong ion-dipole force toward Li ions, but also delivers an induced dipole to the absorbed Li ions. Binding energy is found to be proportional to 'log (0.5 × electronegativity + local dipole)'. Co-doping of O and B increases the local dipole [83].

Bader charge and charge density difference analyses. The correlation between binding energy and charge transfer is further probed. When charge transfer increases up to 0.9 e−, a surge of the binding energy is sparked. Therefore, a critical charge transfer around 0.9 e− is determined.

In general, the growth rate of dendrites can be reduced by decreasing the voltage applied to the battery, increasing the lithium-ion concentration in the electrolyte, or modifying the electrolyte. Ongoing research aims to understand the kinetics of lithium dendrite growth and develop strategies to control or prevent dendrite formation in lithium metal anodes. Stress is one of the factors that can affect the growth of lithium dendrites in lithium metal anodes. Dendrites are highly branched and brittle structures prone to breakage and cracking under mechanical stress. In addition, the growth of dendrites can be influenced by the formation of mechanical stress within the lithium deposit, which can result from the mismatch of thermal expansion coefficients between the lithium and the substrate.

High shear modulus and ionic conductivity enable the inorganic coating to suppress dendrite growth and allow its electrochemical reactions at the electrodes to occur [89, 92].

• The fabrication methods of inorganic coating can readily achieve low thickness and high homogeneity to improve interfacial stability without severe damage to the electrode effectively.
• Inadequate thickness or compositional control of the coatings may lead to compromised SEI homogeneity. More importantly, given the limited Li-ion conductivity or poor flexibility of the coating materials, the cracking of the artificial SEI layers during cycling will induce even greater inhomogeneity on the Li metal surface, exacerbating dendrite growth and side reactions. Electrolyte degradation owing to the highly reactive Li metal has also been a serious concern [93–96].

Various coatings of alumina [97], fluorinated graphene [98–100], carbon [101–103], nanodiamonds [104], polyacetylene [105], tetraethoxysilane [106], lithium phosphorus oxynitride [107] have been reported as an effective anode interface layer. The thickness of amorphous carbon coating affects the electrochemical performances from two aspects; the thick coating can prevent the formation of dendritic lithium very efficiently but lead to a large resistance of Li transfer [108, 109]. Achieving good performance of Li/C electrodes needs a moderate thickness of a-C coating [110].

Belov et al [105] reported Cu₃N-polymer composite coating, with Cu₃N nanoparticles dispersed in tetrahydrofuran with styrene butadiene rubber (SBR). The protective film on the surface of the lithium anode was prepared by drop casting. Cu₃N provided mechanical strength to suppress Li dendrite propagation while the SBR maintained the film's integrity without cracking during battery cycling. The film showed good mechanical strength and flexibility accounted for by the synergistic action of Cu₃N and SBR.

Zheng et al reported a flexible, interconnected, hollow amorphous carbon nanosphere coating as an interfacial layer, which, due to its chemical stability when in contact with lithium, resulted in a Young's modulus of ∼200 GPa, and a minimal contribution to the impedance to charge transfer [111]. The hollow nanosphere layer was fabricated using a templated-assisted approach. Initially, polystyrene nanoparticles are deposited onto the Cu substrate followed by the deposition of amorphous carbon via flash-evaporation of carbon, and finally the polystyrene template by thermal decomposition. Coulombic efficiency of ∼99% was achieved for more than 150 cycles. Umeda et al demonstrated the formation of an electrochemically stable mesoporous SiO₂ film over lithium electrodes by reacting with lithium, stabilizing the surface [106]. The electrode was prepared by dipping freshly polished lithium in tetraethoxysilane for 5 min, which was wiped and directly used in test cells. It was demonstrated that after some 100 cycles of lithium plating and stripping, the impedance of the surface remains constant. Liu et al [104] have developed a double-layer nanodiamond thin film interface by microwave-plasma chemical vapour deposition which exhibits excellent stability against lithium, high mechanical strength (Young's modulus over 200 GPa) and low electrical conductivity. A Coulombic efficiency of >99.4% at 1 mA cm⁻² was achieved.

A spin-coated porous Al₂O₃ layer as a protective layer for a lithium–sulphur battery has been shown to improve the capacity retention of Li–S battery from 50% to 70% by spin coating slurry of Al₂O₃:PVDF (70:30) mixture over lithium metal anode. This technique is relatively simpler and cheaper [97]. Kozen et al reported atomic layer deposition (ALD) of Al₂O₃ with 14 nm thickness that shielded the Li surface from corrosion owing to exposure to environment, sulphur, and electrolyte. Using Li–S battery cells as a test system, thus modified anodes were shown to retain more capacity for up to 100 cycles compared to the cells assembled with bare Li metal anodes [112]. Li₃PO₄ is another material reported as an interphase layer for Li metal anode with excellent stability (0–4.7 V vs. Li/Li⁺) and low electrical conductivity (10⁻¹⁰ S cm⁻¹). The amorphous coating generated by magnetic sputtering stabilizes the interface between the lithium metal and the electrolyte. In an alternate approach, polypyrrole was used as the coating material for the Li powder anode for lithium-sulphur batteries. This technique was user-friendly, had a high preparation efficiency, and prevented the chemical interaction between the lithium anode and the lithium polysulfide dissolved in the liquid electrolyte [113].

A biomimetic design of artificial SEI films for Li batteries with high energy density and capacity has been developed by Jiang et al It involves rationally integrating ClO⁴⁻ decorated UiO-66 MOF and flexible lithiated Nafion binder (Li-Nafion) on the surface of Li metal anodes using the drop-casting method, as shown in figures 3(a) and (b). The film demonstrates excellent single-ion conductivity and remarkable mechanical strength [114].

Zoom In Zoom Out Reset image size

Li et al developed a robust all-organic artificial interphase as seen in figures 3(c) and (d) by incorporating porous polymer-based molecular brushes poly(oligo(ethylene glycol) methyl ether methacrylate)-grafted, hyper crosslinked poly(4‐chloromethylstyrene) nanospheres into a lithiated Nafion polymer matrix, and were able to achieve stable cycling at a current density of 10 mA cm⁻² for over 100 h [115].

In another work, a solid electrolyte interfacial layer is introduced as a gradient thin film consisting of perovskite framework on the top surface to isolate the liquid electrolyte and underneath Li–M alloy layer to induce the homogeneous deposition of Li metal (figures 3(e)–(g)). Based on the intrinsic feature of metal chloride perovskite, a fast Li ion transport gradient layer model was proposed to illustrate the shielding mechanism of perovskite thin film for the dense deposition of Li metal.

Wang et al [117] reported an artificial SEI layer formed of crosslinked and self-reinforced [LiNBH]_n chains fabricated in situ on the surfaces of Li metal anodes by the self-polymerisation of lithium amido borane, which was synthesized in situ dehydrogenation reaction between Li metal and ammonia borane as shown in figures 3(h)–(j). With the [LiNBH]_n layer, the Li metal anode exhibited stable cycling at a 3 mA cm⁻² for more than 700 h. Another group describes a dendrite-free Li composite anode with little volume change made of mechanically rolled and folded lithiophilic MoN nanosheets and 3D interpenetrating Li nanosheets. The planar Li plating caused by the conductive 2D MoN nanosheets' tiny lattice misfit with Li and excellent lithiophilic affinity inhibits dendritic formation in the perpendicular direction. In order to reduce volume change during cycling, the MoN nanosheets interlaced framework provides an abundance of Li accommodation sites, while the horizontally plated Li fills the nanoscale gaps between the MoN nanosheets [118].

In addition to the classifications cited, we have many approaches that use additives like F, S and N-containing solvents in the electrolyte that can modify the SEI of the metal battery and promote the stability [119–125]. In an unconventional approach, very recently, a salt-philic, solvent-phobic polymer coating for Li metal electrodes was employed as a metal protection strategy [126]. The coating selectively transported the salt over solvent and thus promoted a salt-derived SEI formation thereby offering long-term cycling stability.

This section details the use of 3D frameworks or scaffolds to accommodate and distribute alkali metal ions uniformly in the empty volume of the porous framework in order to reduce the local current density and minimizes the formation of dendrite. It is well known that additives such as LiNO₃, FEC etc, have been introduced as additives in LIB electrolytes to reduce the interface issues associated with lithium metal overlooking the issues associated with volume expansion.

Further, the continuous consumption of liquid electrolytes upon cycling hampers the use of additives as a sustainable solution for batteries that require long-term cycling. On the other hand, solid-state electrolyte is a possible solution, but the limited ionic conductivity together with issues such as large interfacial impedance impedes their practical use in batteries. A consolidated summary of different materials used as inorganic modifiers and their fabrication route is detailed in figure 4.

Zoom In Zoom Out Reset image size

In order to ensure the efficient exchange of Li ions across the electrolyte and anode, the ion flux and the sluggish kinetics of the current collector surface play a crucial role. Current collector modification is, therefore, an area of interest for metal anode research to date [145–151]. A high crystalline mismatch between the metallic anode and the chosen 3D substrate leads to an increased overpotential.

In order to reduce the nucleation overpotential, the common practice is to opt for metallic substrates with BCC lattice structure similar to that of alkali metal anode. In the case of lithium metal batteries, interestingly, lithium also deposits on lithiophilic substrates, which tend to dissolve lithium partially or form an alloy with Li, like gold, even though they have an FCC lattice. Cu is among the least favourable substrates due to its reduced Li solubility and FCC structure [127, 138, 152]. Dealloying from Cu–Zn alloy has been reported using various metals, including chemical treatment [133], and vacuum distillation (due to the reduced vapour pressure of Zn) [132]. Chemical treatment of copper foil with ammonia converts them to hydroxide and is further reduced to bundles of copper fibres of submicron diameter roughly perpendicular to the foil [129]. The 3D Cu porous skeleton is obtained by electrodeposition with control over the surface pore size and thickness by tuning the solution conc, time, and surfactants added [138].

The pore size of the 3D current collector is another important point of consideration [129, 153].

When considering lithium as an example, the smaller pores cannot provide enough Li nucleation sites and fail in confining the lithium dendrite growth, subsequently leading to more deposition in the outer surface. On the other hand, very large pores, as in commercial copper foam, lead to poor electrical contact and insufficient support for the deposited lithium, leading to the increased formation of dead lithium [133]. Porous metal current collectors (figure 5) are still under consideration for large-scale applications [150]. In another work, 35 µm thick commercial copper foil is processed by a pulsed laser micro-processing system, leading to an in-situ decomposition of bulk copper into nanoparticles with a dense thin oxide coating which acts as lithiophilic nucleation sites. By precisely tuning the pore volume and depth by adjusting the laser parameters, the lithium volume expansion can be accommodated [137]. Atomic layer deposition is also a common surface modification strategy for the fabrication of 3D dendrite-ee metal anodes [154, 155]. These current collectors play a pivotal role in anode-less batteries as well.

Zoom In Zoom Out Reset image size

When looking at the feasibility of carbon-based current collectors, materials such as carbon nanotubes (CNTs) and carbon felt and cloth are widely explored. One easy, scalable way to synthesize CNT over Ni foam is by exposing it to an alcohol lamp where the non-fully burned carbon species acts as the source and exposure to the air oxidizes the surface of freshly prepared CNT. The ketonic species in the CNT with lone pair act as electron rich donors with filled p-orbital acting as strong active sites for Li adsorption and act as current collectors [131].

In an alternate approach, vertically aligned CNTs over graphene (GCNTs) anchored by PVDF was synthesized using the CVD technique. The low density of the GCNTs, aligned CNTs (non-tortuosity) and a uniform conductive skeleton (due to anchoring of graphene and CNTs) leads to Li host without conductive additive providing a 3351 mAh g⁻¹ capacity with negligible contribution from the host. The high surface area leads to a reduced local current density which in turn leads to a low rate of Li-ion reduction to Li metal over GCNTs without exceeding the diffusion limit of Li-ion from the electrolyte into the pores [130]. In another work, CNTs were decorated with Al₂O₃ via ALD and employed as 3D current collector as shown in figure 6 [156]. The modified 3D electrode display a dendrite-free morphology and exhibited stable voltage profiles in an organic carbonate electrolyte, thus demonstrating electrochemical stability superior to that of planar copper current collectors.

Zoom In Zoom Out Reset image size

Like lithium metal protection, inorganic coating is a viable option for preventing dendrites and increasing the cycle life of sodium metal anode. The low melting point and extra sensitivity of sodium metal make it difficult to form protective coatings on them. Even within these restrictions, different inorganic coatings on sodium have been reported with materials like Na₃P, NaBr, NaI, NaF, etc. Sodium-based alloys come in handy with high ionic conductivity attributed to rich open ionic channels, ensuring a uniform deposition.

Additives like SbF₃ in the electrolyte have been shown to form Na–Sb alloy through an in-situ reaction with the sodium metal. Fang et al studied the formation of an in-situ bilayer on the sodium metal, with adamant Na–Sb as the inner layer and a compact NaF-rich SEI on the outer layer. While the former reduced the diffusion barrier of sodium ions, maintaining a smooth surface, the latter enhanced the stability of the interface. The synergetic effect of the bilayer leads to enhanced cycling stability, superior rate capability and high-capacity retention [157].

Functional groups within the protective layer can also be tuned to improve the anode performance. A sodium benzenedithiolate (PhS₂Na₂) rich protective layer on sodium metal, formed through a two-step process, (figure 7) was studied by Zhu et al [158]. The facile and scalable synthesis approach starts with a chemical reaction of S₈ with para-dichlorobenzene and sodium as catalysts to form polyphenylene sulphides. The intermediate product converts to stable PhS₂Na₂ coating during the first few electrochemical cycles. The protective layer exhibits reduced interfacial transport resistance compared to pristine sodium and boosted Na⁺ kinetics attributed to the Ph–S–Na functional groups present in the layer. A stable cycling without dendrite formation and low overpotential were observed for symmetric and full cells with the coated sodium anode [158].

Zoom In Zoom Out Reset image size

Due to the low surface diffusion barrier, large band gap and robust mechanical properties, NaBr has been explored as a protective coating for sodium metal [159]. The coating was shown to be effective in preventing dendrites and unwanted electrochemical reactions. While NaBr has a low surface diffusion barrier, it suffers from low diffusion kinetics due to the bulk diffusion barrier. Luo et al combined both properties through a hybrid artificial SEI composed of nano-crystalline NaBr and Na₃P composites. The coating was formed through a spontaneous reaction of phosphorous tribromide with sodium metal, and subsequent sodium plating was observed to form a smooth columnar structure without dendrite formations. Enhanced cycling stability and capacity retention were observed for full and symmetric cell studies [161].

While in-situ formation might have a drawback of structural reconstruction, ex-situ-controlled formation of Na₃P layer was explored by Shi et al on both sodium and Potassium metal anode. A simple rolling of ground red phosphorous on the metal leads to the formation of mechanically stable and ionically conducting coatings [160]. NaF offers a high Young's modulus (31.4 GPa), large surface energy and good electrochemical stability, making them a potential coating candidate for preventing dendrites in sodium metal anode. A heterogenous coating layer with NaF and cobalt nanoparticles was formed through a conversion reaction between sodium and CoF₂. The embedded cobalt nanoparticles with defective carbon coating enhanced the sodiophilicity and improved the Na⁺ conductivity. Interestingly, the NaF provided three times higher Young's modulus (7.1 GPa) compared to the bare sodium metal preventing dendrites [162].

Al₂O₃ is also explored as an efficient protective coating to limit the dendrite formation in sodium, similar to lithium. A glovebox-integrated ALD technique and plasma-enhanced ALD by Luo et al for Al₂O₃ coatings [163]. Considering the low melting point of Na, the ALD technique has a low reaction temperature, and a high ionic conducting NaAlO_x formation also occurs during the electrochemical cycling with the Al₂O₃ coating. The coated sodium exhibits a reduced lower initial overpotential, which remains stable during the cycling for symmetric cells. An optimized thickness of 25 cycles was studied to minimize moss-like dendrites and form smooth islands of sodium even during high current densities [164]. Molecular layer deposition has also been explored for the formation of inorganic-organic alucone protective layer [165]. Other coating techniques include the use of chemical vapour deposition for the formation of multilayer graphene and transferring over the sodium metal [166]. The cycling stability of sodium metal anode was shown to have greatly improved with an optimized graphene coating of 5 nm thickness.

Inorganic coatings are shown to be more efficient and better strategy compared to current collector modifications to achieve prolonged cycling life of sodium metal anodes. While most of the coatings are made through in-situ reactions, ex-situ formation of uniform coatings is still expensive and less explored. Incorporating mechanical stability and, at the same time, sodiophilic properties to the coating layers are the key to future stable metal anodes.

The volume change of sodium during the deposition is twice that of lithium, and 3D host with a high surface area can be an effective method for longer cycling. The large surface area and uniform surface structures ensure an even electric field for better sodium metal anode depositions. The 3D metallic current collectors have been adopted for sodium, and as discussed in the previous section, copper has been chosen as a prospective material. One step dealloying of Cu–Zn alloy (Cu_0.7Zn_0.3) leads to forming a porous copper skeleton with high surface area, resulting in uniform deposition of sodium during plating [167]. Copper nanowire formation on 3D copper foam results in further increase in surface area and nucleation sites [168].

Unlike lithium, sodium does not react with aluminium, and the possibility of utilizing the latter as a lightweight and cheap alternative to copper current collectors has also been studied. The formation of 3D zinc nanofiber on an aluminium current collector through simple magnetron sputtering was studied. Low nucleation overpotential and small polarization voltage were achieved due to the high surface area with the in-situ formed sodiophilic NaZn₁₃ [169].

The morphology of sodium metal deposition is strongly related to the initial nucleation and its uniformity. Sodiophilic metal inclusion in the 3D host has been explored as an advantageous technique for homogeneous nucleation, and thereby ensuring a high Coulombic efficiency. The 3D nanostructured porous carbon particle containing carbon-shell-coated Fe nanoparticles was formed though carbonization of colloidal block polymer solution. A 3D hierarchical structure with ordered open channels leads to a high CE of 99.6% at 10 mA cm⁻² and 10 mAh cm⁻² [170].

Carbonization of electrospun PMMA over antimony-containing carbon fibre has been performed by Li et al and studied as 3D host for sodium metal anode [171]. The PMMA generates lotus root-like channels within the structure enabling uniform deposition, and also induces N and O functional groups in the carbon structure. A horizontal growth of sodium was observed during deposition, which is attributed to the gradient sodiophilicty within the 3D framework.

The 3D printing technique has recently been explored to synthesize hosts with embedded sodiophilic metals. Wang et al produced 3D Au/rGO using 3D printing followed by freeze drying and subsequent annealing. The host showcased remarkable cycle life (1200 h at 5 mAh cm⁻², 5 mA cm⁻²) at high current density and capacity. The technique is reported to have several advantages, such as being lightweight, the flexibility in tuning the thickness, periodicity in the microlattices [172].

Sophisticated top-down synthesis is required for the metal-embedded 3D sodium hosts, which acts as a barrier for large-scale production. Commercially viable materials have also been studied as candidates for sodium metal hosts. For instance, 3D carbon felt is used as a host where the sodium is inserted using a commercially scalable melt infusion strategy [143]. This flexibility of the felt makes it versatile, and can be used in different battery systems (cylindrical cells, pouch cells etc). Several factors including additional room to accommodate the sodium, a uniform Na⁺ flux, and associated lowered current density, lead to better performance in terms of high current density (5 mA cm⁻²) cycling. The low cost of the host further favours future commercialization. An alternate method through electrochemical plating inside carbon felt has also been explored [173].

While more materials are studied as potential hosts for sodium metal anode, the cost, availability and commercialization aspect should be considered. Incorporating sodiophilic metals in the 3D host is an interesting strategy, however, scalable methods with cheaper alternatives should be the priority for future research in this area.

Although a routine XPS depth profiling could identify the constituent elements, and qualitative depth distribution of the passivation layer on the lithium samples, the lateral and depth resolutions of XPS analyses are rather low [190–192]. In addition, the sensitivity towards lithium is low and hydrogen cannot be detected at all. ToF-SIMS depth profiling can complement the XPS results with higher lateral resolution and quantitative depth information, even if ToF-SIMS results are not inherently compound-specific. Therefore, a group led by Henss applied ToF-SIMS to take advantage of its higher sensitivity, as well as of its superior lateral and depth resolution, to get more information about the three-dimensional distribution of the different compounds [193, 194]. Due to matrix effects, the ToF-SIMS results are only semiquantitative and not inherently compound-specific. Machine learning assisted logistic regression approaches are employed to ease the analyses process [195]. Through this approach, researchers were able to trace and identify different lithium compounds that are expected to appear on lithium metal anodes, namely, Li₂CO₃, Li₂O, Li₃N, LiH, and LiOH and qualitatively identify compounds on the lithium metal.

The evolution of electrode morphology is observed through operando high-resolution video capture and is directly correlated to the voltage traces. Further, in order to get an in-depth understanding of the electrochemical processes occurring on the electrode surfaces, to monitor the lithium platting and dissolution associated voltage polarization, a continuum-scale numerical model is developed. This technique is used to relate electrode morphology and compare electrochemical kinetics to the cell voltage [196]. When carrying out platting dissolution measurements and correlating the voltage profile shapes and EIS, there have been many instances wherein it is difficult to differentiate between soft shorts [197]. A clear picture of the electrode during cycling is evident from this operando video microscopy. This helps give insights on the dendrite evolution. Figures 10(a)–(e) shows a quartz cell design used for monitoring the dendrites [198]. The technique helps avoid the process of disassembling and drying, leading to significant changes in the surface morphology and damage to the anode.

Zoom In Zoom Out Reset image size

In an unconventional approach, researchers from MIT propose a method to quantify the impact of stack pressure and in-plane stresses on dendrite trajectory using fracture mechanics and also propose design approaches to achieve such stress [200]. The researchers found that the dendrite propagation direction could be deflected with applied stress in the direction perpendicular to the ion transport direction. Operando optical measurements were recorded (figure 11) to capture the propagation visually.

Zoom In Zoom Out Reset image size

Terahertz time-domain spectroscopy has been used in many research arenas [201–204]. It was used for the first time to understand the state of SEI in a battery electrode while cycling the cell. The group monitored the cell's impedance at different cycling states to verify that the THz transients are due to the SEI formation (figures 10(g) and (h)). The presented technique shows that terahertz spectroscopy can be a promising tool for operando electrochemical characterisation [199]. THz-TDS can detect alterations of the elusive SEI layer under cycling conditions, and more importantly, during rest periods. Through this study, it is possible to identify the SEI formation on the surface of the Si electrode during the lithiation process, and its partial dissolution during prolonged delithiation.

Operando x-ray studies of anodes are a powerful tool for investigating the electrochemical behaviour of anodes during battery operation [205]. These studies use x-ray techniques, such as x-ray absorption spectroscopy or x-ray diffraction (XRD), to probe the chemical and structural changes in the anode material in real-time during battery operation. Operando x-ray studies have provided valuable insights into the behaviour of metal anodes, including lithium metal, sodium, and potassium, in batteries and have helped to improve our understanding of the electrochemical behaviour of these materials. These studies have the potential to aid in the development of safer and more stable metal anode batteries for various applications, including electric vehicles and energy storage systems.

The benefits of operando x-ray studies of anodes include the ability to:

• (1)  
  Observe the evolution of the electrode structure and chemical composition during battery cycling.
• (2)  
  Study the distribution and kinetics of lithium or other metal ions within the anode during battery operation.
• (3)  
  Investigate the reaction mechanisms and kinetics of the electrode/electrolyte interface.

It is difficult to monitor the formation and growth of dendrites and other structures of sodium in real time due to their reactivity and gas evolution.

Techniques such as TEM cannot be used because the electron beam can possibly damage the material under observation. Even a small exposure time of one second may damage the lithium dendrite [206]. With the micro-electromechanical systems technology in in-situ liquid cell TEM, it has become feasible to observe the deposition and dissolution of dendrites in the charge-discharge cycles. There are a few successful cases of observing metallic dendrites in in situ electrochemical liquid cell experiments, such as lithium dendrite and silver dendrite [207].

By virtue of numerical modelling (e.g. finite element analysis, Monte Carlo calculation, nonlinear phase-field model, data-driven model, and molecular dynamics simulation), great research progress has been made in understanding protective SEI films [208–212]. These computational studies have played a pivotal role in understanding electrochemical battery kinetics owing to the complexity of obtaining real-time interfacial visualisation and quantitative characterisation through experiments. A very strong understanding of dendrite growth is emerged in recent times [213–219] and is beyond the scope of this review.