Review—Interfaces: Key Issue to Be Solved for All Solid-State Lithium Battery Technologies

Sulfide-based solid electrolytes are advantageous to ASSLBs owing to their high ionic conductivity and deformability. Because sulfide ions are larger than oxide ions, they leave wider channels in the structure acting as conduction paths. In addition, sulfide-based SSEs is relatively soft and deformable materials. They are deformed and packed densely only by cold pressing, and in the cold-pressed form the electrolyte particles connect tightly to make grain boundary resistances low.²⁹ In fact, the highest ionic conductivities among sulfide-based SSEs had already reached 10⁻³ S cm⁻¹ in the early 1980s in Li₂S-P₂S₅-LiI glass.³⁰ Years development in this period had been improving the performance of sulfide-based SSEs,^31–33 the highest ionic conductivities have exceeded 10⁻² S cm⁻¹ in LGPS in 2011.¹⁴ The properties of typical sulfide-based solid electrolytes are listed in Table I.

The highest lithium ionic conductivity of the Li₂S–GeS₂–P₂S₅ crystalline system is 1.2 × 10⁻² S cm⁻¹ for Li₁₀GeP₂S₁₂, while that of the Li₂S–P₂S₅ glass-ceramic system remains in the order of 10⁻³ S cm⁻¹. Y. Seino et al. reported that a heat-treated 70Li₂S–30P₂S₅ glass-ceramic conductor has an extremely high ionic conductivity of 1.7 × 10⁻² S cm⁻¹ and the lowest conduction activation energy of 17 kJ mol⁻¹ (0.18 eV, 1 kJ mol⁻¹ = 1.036427 × 10⁻² eV) at room temperature. In their opinion, the heat treatment reduce the grain boundary resistance and influence of voids, resulting in increasing the Li⁺ ionic conductivity of the solid electrolyte.³³ The finding of LGPS with an outstanding ionic conductivity of 1.2 × 10⁻² S cm⁻¹ has made the use of sulfide-based SSEs an appealing option for ASSLBs. However, the high cost of Ge largely limits the practical use of LGPS. Creating isostructural analogues by substituting Ge with Si, Al, or Sn, has been suggested.^34,39,40

Another promising sulfide-based SSEs, halogen-doped Li₆PS₅X (X = Cl, Br, I) argyrodites, have attracted intensive attention due to their relatively high ionic conductivity and moderate electrochemical stability window. They are synthesized by solid-state sintering, mechanical alloying, or liquid-phase synthesis processes.^{35–38,41,42} However, the reported ionic conductivity values of Li₆PS₅X are still rather scattered in the range of 10⁻² to 10⁻⁵ S cm⁻¹ at room temperature. The highest ionic conductivity of 5.6 × 10⁻³ S cm⁻¹ is measured from Li₇P_2.9Mn_0.1S_10.7I_0.3 at room temperature with a low activation energy of 0.208 eV.³⁶

High Li-ion conductivity and excellent stability are essential properties for solid electrolytes used in ASSLBs. Various types of oxide-based solid electrolytes have been developed, including cubic garnet-type Li₇La₃Zr₂O₁₂ (LLZO),^43,44 lithium phosphorus oxynitride (LiPON),⁴⁵ NaSICON-type LiM₂(PO₄)₃ (M = Ti, Ge, Zr, Hf),^46,47 and perovskite-type Li_3xLa_2/3−xTiO₃ (LLTO)^48,49 etc. Among these solid electrolytes, cubic-phase LLZO has been widely studied because of its excellent stability against lithium metal anode and cathodes.^50,51 The garnet-type cubic LLZO ceramics and their doped variants like Li_6.25La₃Zr₂Al_0.25O₁₂ deliver high ionic conductivity of about 10⁻⁴ S cm⁻¹. Generally, supervalent doping of Al³⁺, Ga³⁺ and Ge^{4+ 52,53} at the Li sites as well as Ta⁵⁺ and Nb^{5+ 54} at the Zr sites has been conducted to stabilize the cubic garnet and increase the ionic conductivity. It has been reported that supervalent cation Zr⁴⁺ substitution on the site can increase the vacancy concentration, which can improve Li-ion conductivity by lowering the activation energy of Li ions, and that the substitution of an element with a larger ion diameter can enlarge the size of the Li-ion migration pathway and increase the mobility of Li-ions.^55,56 The ionic conductivity can be obtained with 4.78 × 10⁻⁴ S cm⁻¹ at 20 °C, while the activation energy was only about 0.29 eV when substitution of Zr⁴⁺ sites by the same value of Ge⁴⁺ in the Li₇La₃Zr₂O₁₂. The results indicate that doping of Ge⁴⁺ in the LLZO was able to increase the density of the pellets as well as reduce the sintering temperature.⁵³ J. Gai and coworkers substituted the Zr sites of a Li₇La₃Zr₂O₁₂ electrolyte with Nb and Y elements simultaneously to improve its Li-ion conductivity and air stability. The Li₇La₃ZrNb_0.5Y_0.5O₁₂ pellets treated at 1230 °C for 15 h had a total conductivity of 8.29 × 10⁻⁴ S cm⁻¹ at 30 °C, having much higher conductivities than undoped ones.⁵⁴ Dual substitution strategy can farther enhance Li⁺ ionic conductivity in LLZO solid electrolyte. Substituted the Li⁺ sites by Ga³⁺and replaced of Zr ⁴⁺ by Sc cation, a total ionic conductivity has been achieved 1.8 × 10⁻³ S cm⁻¹ for Li_6.65Ga_0.15La₃Zr_1.90Sc_0.10O₁₂ at 300 K, which is the highest ionic conductivity observed for the garnets so far.⁵⁷ The properties of typical oxide-based solid electrolytes are listed in Table II.

Although amorphous lithium phosphorus oxynitride (LiPON) shows the relatively low Li ionic conductivity of 1–3 × 10⁻⁶ S cm⁻¹ at 25 °C, it has the good chemical stability with the Li metallic anode, and the broad electrochemical potential window of 0–5.5 V (vs Li/Li⁺), making it a competitive electrolyte material.^45,61 There were many studies focusing on the optimization of LiPON electrolyte layer in the past years, especially on the improvement of Li ionic conductivity, including the investigation on the deposition condition and the optimization of thickness. D. L. Xiao et al. had deposited Li-compensated LiPON thin film solid electrolyte by sputtering a sintered Li-rich Li_3.3PO₄ target instead of the normally Li₃PO₄ target. The Li ionic conductivity in Li-LiPON was improved to 3.2 × 10⁻⁶ S cm⁻¹ from the 2.4 × 10⁻⁶ of LiPON.⁶²

The name of NaSICON was initiated from sodium super ion conductor with formula NaM₂(PO₄)₃, where M is a cation. NaSICON-type Li-ion solid electrolyte LiM₂(PO₄)₃ can be prepared by replacing Na-ion in NaM₂(PO₄)₃ by Li-ion. The NaSICON-type Li-ion solid state electrolyte LiM₂(PO₄)₃ (M = Ti, Ge, Zr, Hf) has some advantages, such as excellent stability in ambient atmosphere, mass production and low cost. By partially replacing M⁴⁺ ions by trivalent ions (Al³⁺,Sc³⁺), Li_1+xAl_xGe_2−x(PO₄)₃, Li_1+xAl_xTi_2−x(PO₄)₃ and Li_1+xSc_xTi_2–x(PO₄)₃ were obtained with low porosities and enhanced ionic conductivities.^63–65 Y. Liu et al. have prepared Li_1.5Al_0.5Ge_1.5(PO₄)₃ (LAGP) electrolytes by a facile sol-gel method in aqueous solution. The ionic conductivity of LAGP at 30 °C was 3.1 × 10⁻⁴ S cm⁻¹, with an activation energy value of 0.37 eV.⁴⁶ E. Zhao and coworkers have prepared Li_1.3Al_0.3Ti_1.7(PO₄)₃ (LATP)/Li_1.3Al_0.3Ge_1.7(PO₄)₃ (LAGP) bi-layer structured solid state electrolyte via a simple dry pressing and post-calcination method. This electrolyte sample exhibited a high ionic conductivity of 3.4 × 10⁻⁴ S cm⁻¹ and a negligible electronic conductivity of 9.6 × 10⁻⁹ S cm⁻¹ at room temperature.⁶³ R. Kahlaoui et al. have prepared Li_1+xSc_xTi_2–x(PO₄)₃ (0 ≤ x ≤ 0.5) using a conventional solid-state reaction. For x = 0.2, the Li_1.2Sc_0.2Ti_1.8(PO₄)₃ exhibits high bulk ionic conductivity (2.5 × 10^–3 S cm^–1) and low activation energy (E_a = 0.25 eV) at room temperature.⁶⁵ However, the LATP electrolyte tends to react with lithium anode due to the Ti⁴⁺/Ti³⁺ redox reaction. The problem can be solved by adopting a barrier layer introduced between electrolyte and lithium anode, which requires sufficient stability against lithium metal and block the contacts between them so as to prevent the undesired chemical reactions.

The perovskite-type (ABO₃) solid electrolyte Li_0.34(1)La_0.51(1)TiO_2.94(2) (LLTO), which showed bulk ionic conductivity of 1 × 10⁻³ S cm⁻¹ and total ionic conductivity of higher than 2 × 10⁻⁵ S cm⁻¹ at room temperature was firstly prepared by Y. Inaguma and coworkers.⁴⁸ Y. Inaguma et al. also synthesized a perovskite-type lanthanum lithium titanate (LLTO) with the nominal composition of Li_0.29La_0.57TiO₃ by a solid state reaction. The total lithium ion conductivity from Li_3xLa_2/3−xTiO₃ (3x = 0.29) was calculated to be 5.2 × 10⁻³ S cm⁻¹ at 300 K, with an activation energy of 0.45 eV.⁵⁹ A common drawback of LLTO solid electrolytes is that they contain Ti element. The Ti is in tetravalent state and can be reduced easily by contact with low-potential anodes, such as Li_3xLa_2/3−xTiO₃ with the reduction potentials of 1.6 V (vs Li⁺/Li). Thus to increase the electrochemical stability, it is necessary to substitute Ti by other elements. Substituted Ti⁴⁺ with Zr⁴⁺ and Ta⁵⁺, a new composition Li_2x−ySr_1−xTa_yZr_1−yO₃ were selected and synthesized. An optimized composition of Li_3/8Sr_7/16Ta_3/4Zr_1/4O₃ (LSTZ) showed the ionic conductivity with bulk and grain boundary ones of 2 × 10⁻⁴ S cm⁻¹ and 1.33 × 10⁻⁴ S cm⁻¹ at 30 °C, respectively. This solid electrolyte was found to be stable above 1.0 V against metallic lithium.⁶⁰

Solid polymer electrolytes (SPEs) have several advantages, such as high flexibility, low-weight, safe, enhanced electrode/electrolyte interfacial compatibility, and low processing costs.⁶⁶ Poly(ethylene oxide) (PEO)-based polymer electrolytes have been studied widely due to their outstanding film-forming ability and relatively better compatibility towards electrodes. The PEO is a semicrystalline polymer at room temperature which has the capacity to dissolve high concentration of metal salts. The major drawbacks of PEO-based SPEs are that they exhibit low ionic conductivity (10⁻⁶ to 10⁻⁸ S cm⁻¹) at room temperature, poor mechanical strength, and inferior electrochemical window. Many strategies have been carried out to improve the performance of pristine PEO, like copolymerization,⁶⁷ cross-linking,⁶⁸ polymer blending^69,70 and incorporating fillers such as SiO₂, Al₂O₃, BaTiO₃, and garnet oxide into the polymer matrix.^71–75 Except for PEO-based system, various lithium-ion conductive polymer materials, such as (polymethacrylic acid) (PMAA),⁷⁶ polycarbonate-based polyurethane,⁷⁷ modified silyl-terminated polyether (MSTP),⁷⁸ poly(propylene carbonate) (PPC),⁷⁹ and Jeffamine^®-based^80,81 polymer have been exploited for polymer electrolyte matrix. P. Deng et al. have used lithium (fluorosulfonyl)(pentafluoroethylsulfonyl)imide (LiFPFSI) as a conducting salt for building LiFPFSI/PEO SPEs. The LiFPFSI/PEO electrolytes show a high ionic conductivity of 6.2 × 10⁻⁴ S cm⁻¹ at 80 °C and a lower glass transition temperature.⁸² The development of solid polymer electrolytes in recent years has proved that the SPEs are promising candidates to replace currently available organic liquid electrolytes for ASSLBs. The performance of some typical solid polymer electrolytes is listed in Table III.

To improve overall performance of SPEs, the polymers are being coordinated with other components to form composite polymer electrolytes (CPEs). The integration of polymer-based electrolytes and inorganic solid electrolyte filler to construct composite electrolytes to balance ionic conductivity, mechanical strength, flexibility, and interfacial stability is a promising strategy to improve electrolyte performances.

K. Karthik et al. had prepared lithium garnet composite polymer electrolyte membrane consisting of large molecular weight polyethylene oxide (PEO) complexed with lithium perchlorate (LiClO₄) and lithium garnet oxide Li_6.28Al_0.24La₃Zr₂O₁₂ (Al-LLZO) by solution-casting method. The maximized Li⁺ conductivity at 30 °C is 4.40 × 10⁻⁴ S cm⁻¹ and the electrochemical window is 4.5 V for PEO/LiClO₄ + 20 wt% Al-LLZO membrane.⁸³ L. Chen and B. J. Goodenough et al. have fabricated composite electrolyte consisting of polyethylene-oxide and garnet (Li_6.4La₃Zr_1.4Ta_0.6O₁₂) containing LiTFSI as the lithium salt. The composite electrolytes, from "ceramic-in-polymer" to "polymer-in-ceramic," are flexible and mechanically robust and have a Li⁺ conductivity σ_Li > 10⁻⁴ S cm⁻¹ at 55 °C.⁸⁴ X. Wang et al. have synthesized CPEs with a three-dimensional (3D) perovskite-type Li_0.33La_0.557TiO₃ (LLTO) network as a nano-backbone in poly(ethylene oxide) matrix. The LLTO nanofiber network was fabricated by electrospinning and subsequent calcination processes. For the 3D-CPE, a small mount of the PEO-LiTFSI solution was cast on the LLTO nanofiber network, after drying at RT in a glove box the composite membranes were hot-pressed at 75 °C and quenched in liquid nitrogen. These 3D-CPE membranes had much better thermal stability and enhanced mechanical strength in comparison with solid polymer electrolytes. And an ionic conductivity as high as 1.8 × 10⁻⁴ S cm⁻¹ was achieved at room temperature for PEO + 40 wt% LiTFSI. The electrochemical window of the 3D-CPEs was 4.5 V vs Li/Li⁺.⁸⁵

Zhang et al. used Li_6.75La₃Zr_1.75Ta_0.25O₁₂ (LLZTO) ceramics powders to modify poly (vinylidene fluoride) (PVDF) polymer electrolyte. The flexible PVDF/LLZTO composite electrolytes were prepared by a conventional solution-casting method in which LLZTO powders were dispersed into a PVDF matrix. The ionic conductivity of the PVDF/LLZTO CPE with 10 wt% LLZTO particles can reach as high as about 5 × 10⁻⁴ S cm⁻¹ at 25 °C with the activation energy of 0.20 eV.⁸⁶ Liu et al. fabricated a novel solid composite polymer electrolyte using the electrospinning method to disperse 15 wt% Li_0.33La_0.557TiO₃ nanowires into polyacrylonitrile (PAN)-LiClO₄ matrix. Because of the fast ion transport on the surfaces of ceramic nanowires acting as conductive network in the polymer matrix, an ionic conductivity of 2.4 × 10⁻⁴ S cm⁻¹ at room temperature was obtained. In addition, the LLTO nanowire filled composite polymer electrolyte shows an enlarged electrochemical stability window in comparison to the one without fillers.⁸⁷ The properties of some typical composite polymer electrolytes is listed in Table IV.

The CPEs, incorporating fast ionic conductor as fillers in SPEs, combine synergistically the beneficial properties of high ionic conductivity and strength from inorganic solid electrolytes, as well as good interfacial properties and flexibility from SPEs. The developments of polymer-based composite electrolytes in recent years have proved that the CPEs are promising electrolyte candidates for flexible solid-state lithium-ion batteries.

The Li₇P₃S₁₁ (composed of Li₃PS₄ and Li₄P₂S₇) is a Li₂S-P₂S₅ system compound with triclinic crystal structure of slightly distorted PS₄³⁻ tetrahedra and P₂S₇⁴⁻ di-tetrahedra. Among the different LPS materials, glass ceramics with the composition 70Li₂S–30P₂S₅ (Li₇P₃S₁₁) show the highest conductivity of up to 1.7 × 10⁻² S cm⁻¹.³³ The tetragonal structure Li₁₀GeP₂S₁₂ (composed of Li₃PS₄ and Li₄GeS₄), which is Li₂S-GeS₂-P₂S₅ system, consists of negatively charged PS₄³⁻ and GeS₄⁴⁻ tetrahedral, which are surrounded by Li ions for charge compensation.¹⁰³ It exhibits the highest lithium ionic conductivity of 1.2 × 10⁻² S cm⁻¹ at 27 °C. Argyrodites of the type Li₆PS₅X (X = Cl, Br, I), which exhibit high lithium ion conductivities in the range of 10⁻² to 10⁻³ S cm⁻¹ at room temperature, are considered as suitable candidates for the fabrication of ASSLBs facilitating Li metal anodes. The thermodynamic stability of the different argyrodites in contact with lithium metal is of practical interest for long-term stability of ASSLBs. When contact with Li-metal electrode, the sulfide solid electrolytes decomposed forming new chemical species on the interface that affect the interface resistance.

To clarify whether sulfidic ceramics are stable in contact with lithium metal and which products are formed at the interface, S. Wenzel et al. investigated the formation of an interphase between Li₇P₃S₁₁ and lithium metal by a combined analytical approach, comprising in situ photoelectron spectroscopy (in situ XPS) and time-dependent electrochemical impedance spectroscopy.¹⁰⁴ Mostly, the stability of solid electrolytes in contact with lithium metal is investigated by electrochemical impedance spectroscopy (EIS) or cyclic voltammetry (CV), both of which allow to monitor interphase formation in situ but do not provide chemical information. Combining impedance spectroscopy with in situ XPS allows the chemical identification of reaction products and gives information about the type of interphase. The measurements clearly revealed the growth of an interfacial reaction zone consisting of the decomposition products Li₂S and Li₃P at the Li/Li₇P₃S₁₁ interface.¹⁰⁵

S. Wenzel et al. also obtained the data on the stability of Li₆PS₅X (X = Cl, Br, I) in contact with Li metal from an in situ X-ray photoemission technique in combination with time-resolved impedance spectroscopy.¹⁰⁶ They reported that Li₆PS₅X decomposes into multiple phases composed of Li₂S, Li₃P, possibly LiP, and LiX. In situ XPS in combination with time-resolved electrochemical measurements offers detailed information on the chemical reactions at the Li/LGPS interface. Contacting with Li-metal electrode, the Li₁₀GeP₂S₁₂ decomposed to form Li₂S, Li₃P, and Li−Ge alloys at the Li/Li₁₀GeP₂S₁₂ interface.¹⁰¹

The interphase formation reactions between sulfide SSEs and Li-metal were suggested as the followings^{93,101,104,106}:

$$\begin{eqnarray}&&{{\rm{Li}}}_{3}{{\rm{PS}}}_{4}+8{\rm{Li}}\to 4{{\rm{Li}}}_{2}{\rm{S}}+{{\rm{Li}}}_{3}{\rm{P}}\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{{\rm{Li}}}_{7}{{\rm{P}}}_{3}{{\rm{S}}}_{11}+24{\rm{Li}}\to 11{{\rm{Li}}}_{2}{\rm{S}}+3{{\rm{Li}}}_{3}{\rm{P}}\end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&{{\rm{Li}}}_{10}{{\rm{GeP}}}_{2}{{\rm{S}}}_{12}+20{\rm{Li}}\to 12{{\rm{Li}}}_{2}{\rm{S}}+2{{\rm{Li}}}_{3}{\rm{P}}+{\rm{Ge}}\end{eqnarray} \tag{ 3 }$$

or

$$\begin{eqnarray*}&&{{\rm{Li}}}_{10}{{\rm{GeP}}}_{2}{{\rm{S}}}_{12}+23.75{\rm{Li}}\to 12{{\rm{Li}}}_{2}{\rm{S}}+2{{\rm{Li}}}_{3}{\rm{P}}+1/4{{\rm{Ge}}}_{4}{{\rm{Li}}}_{15}\end{eqnarray*}$$

$$\begin{eqnarray}&&{{\rm{Li}}}_{6}{{\rm{PS}}}_{5}{\rm{X}}+8{\rm{Li}}\to 5{{\rm{Li}}}_{2}{\rm{S}}+{{\rm{Li}}}_{3}{\rm{P}}+{\rm{LiX}}\left({\rm{with}}\,{\rm{X}}={\rm{Cl}},\,{\rm{Br}},\,{\rm{I}}\right)\end{eqnarray} \tag{ 4 }$$

The formation of Li₂S as major product, which exhibits a relatively low ionic conductivity (∼10⁻¹³ S cm⁻¹ for bulk Li₂S), the reduction interphase layer (Li₂S, Li₃P, LiX or Li-Ge alloy ) has much lower ionic conductivity, increasing the Li/SSEs interfacial resistance. The stability of lithium binary compounds vs lithium anode is calculated from experimental thermodynamic data. The stability window (V vs Li metal) is in the sequence: Li₃P < Li₂S < LiI < Li₂O < LiBr < LiCl < LiF.¹⁰⁷ The formation of Li₂S and the instability of Li₃P at the interface degrade the interfacial properties between solid electrolyte and metal anode. Thus the interface modification will be required to form a proper low-impedance interface to apply sulfide SSEs in ASSLBs.

A. Kato et al. revealed that the Au thin film between Li metal and Li₃PS₄ electrolytes prevented void generation after the initial Li dissolution and increased the sites for Li deposition. The more uniform morphology changes of Li metal contributed to the better reversibility of Li utilization.⁹³ D. Xie et al. also prepared Sb₂O₅-doped β-Li₃PS₄ solid electrolytes with good stability against Li metal. The glass-ceramic Li₃P_0.98Sb_0.02S_3.95O_0.05 exhibited an excellent stability, could be due to existence of the oxygen ions in the structure, suppressing the side reactions between sulfide electrolyte and lithium.¹⁰⁸

To suppress the Li dendrite growth, R. Xu et al. used LiF (or LiI) layer at the interface between Li and sulfide electrolyte and penetrated methoxyperfluorobutane (HFE) (or I solution) inside of sulfide electrolyte. The effect of LiF and LiI interphase layer on the interface resistance and Li dendrite suppression were evaluated in a symmetrical Li/electrolyte/Li cells using LiF (or LiI) coated Li, and HFE (or I) coated/infiltrated Li₇P₃S₁₁ electrolyte. The interfacial resistance was evaluated by electrochemical impedance spectroscopy (EIS). Their results indicated that the LiF and LiI interphase between Li and electrolyte effectively suppressed the side reactions and dramatically reduces the interfacial resistance. In comparison, LiF was easier to form dense/uniform layer on lithium metal and it had advantages over LiI in maintaining interface stability and inhibiting the growth of lithium dendrite.¹¹

L. Cong et al. used perfluoropolyethers (PFPEs) interfacial stabilizers coating between Li anode and Li₁₀GeP₂S₁₂ composite solid-state electrolyte, dramatically improving the interfacial compatibility with Li anode. The SEM and XPS analyses revealed that the PFPEs coating in situ form a LiF-rich solid electrolyte interphase layer, not only facilitating solid-solid interfacial contacts, but also promoting the fast Li⁺ mobility in electrode/electrolyte interfaces.¹⁰⁹

J. Zhang et al. fabricated Li₆PS₅Cl/poly(ethylene oxide) composite solid electrolytes by liquid-phase process. The Li₆PS₅Cl/PEO composite SSEs could effectively inhibit the interfacial reactions and lithium dendrite formation when the PEO content reached the optimal value. With Li₆PS₅Cl/5 wt% PEO SSE the as-assembled sandwich-type LiNi_0·8Co_0·1Mn_0·1O₂/composite SSE/Li battery exhibited enhanced cycling performance with a capacity retention rate of 91% over 200 cycles at 0.05 °C and 30 °C.¹¹⁰

Recently, research efforts have focused on the study and design of materials with high ionic conductivity and the understanding of ion transport mechanisms by computational methods. The quantum mechanics-based methods, density functional theory (DFT), and experimental tools have enabled the study of Li-metal/SSE interfaces.^90,111,112 Camacho-Forero and coworkers used density functional theory (DFT) and ab initio molecular dynamics (AIMD) simulations to learn more about the Li-metal/sulfide-type SSEs (LGPS, Li₂P₂S₆, β-Li₃PS₄, and Li₇P₃S₁₁) interfacial chemistry.¹¹³ They found that, in general, the SSE/Li interface was very unstable and induced the formation of multiple solid phases such as Li₂S, Li₃P, Li₁₇Ge₄, etc. The anions that were close to the interface quickly decomposed during the DFT optimization, S–P bond breaking and Li₃P and Li₂S forming when Li₂P₂S₆ and Li₇P₃S₁₁ reacted with Li-metal. In case of LGPS, S–P and S–Ge bonds cleavaging, Li₁₇Ge₄, Li₃P, and Li₂S formed from LGPS reduction reaction with the Li-metal. These results are in good agreement with in situ XPS observations of sulfide-type SSEs instability in contact with Li metal.^101,105 The compounds that are formed in the interface are likely to increase the ionic transport resistance. They also found that artificial Li₂S thin-film on Li-metal had a strong passivation effect at the interface. However, ionic transport kinetics studies of this and other thin film layers are required in order to gain a more complete understanding of their effect in the overall performance of the batteries.

The interface modification at the Li anode/solid electrolyte interface was shown to be a useful technique for improving the cycle stability and the rate of a charge-discharge reaction of ASSLBs. Strategies for improving the solid-solid interfacial issues include forming an intermediate Li-metal alloy that changes the wettability of the garnet surface to develop a good contact between the Li metal anode and garnet electrolyte with Au,¹¹⁴ or Ge¹¹⁵ coatings on LLZO substrate, and decreasing the interfacial impedance using aluminum oxide (Al₂O₃).¹¹⁶ Moreover, it has been shown the effect that mixing LLZO with a polymer to improve its flexibility,¹¹⁷ introducing Li₃PO₄ as an additive in garnet-type Li_6.5La₃Zr_1.5Ta_0.5O₁₂ could improve the interfacial compatibility and suppress Li-dendrite formation.¹¹⁸

C.-L. Tsai et al. synthesized Ta-substituted LLZO (Li_6.6La₃Zr_1.6Ta_0.4O₁₂) via solid-state reaction. Gold (Au) buffer layers (∼20 nm) were sputtered onto both sides of the LLZO pellets for Li dendrite prevention. The interface resistance was dramatically reduced by using thin layers of Au buffer to improve the contact between LLZO and Li electrodes, which prevented the Li dendrite growth due to a more homogeneous current distribution.¹¹⁴ Depositing a thin germanium (Ge) (20 nm) layer on the garnet LLZO (Li_6.85La_2.9Ca_0.1Zr_1.75Nb_0.25O₁₂) could reduce the garnet/Li-metal interfacial resistance. By applying this approach, the garnet/Li-metal interfacial resistance decreased from ∼900 Ω cm² to ∼115 Ω cm² due to an alloying reaction between the Li metal and the Ge.¹¹⁵

X. Han et al. effectively addressed the large interfacial impedance between a lithium metal anode and the garnet Li₇La_2.75Ca_0.25Zr_1.75Nb_0.25O₁₂ electrolyte using ultra thin aluminium oxide (Al₂O₃) by atomic layer deposition (ALD) technique. As in Fig. 2a, a schematic of the wetting behavior between garnet and Li metal illustrated that without surface modification, garnet had insufficient physical contact with Li metal. With the ALD coating, the ultra thin Al₂O₃ layer helped the molten Li metal to coat the garnet surface with no interfacial void space. The SEM images in Fig. 2b clearly demonstrated the enhancement of interfacial contact by applying the ALD-Al₂O₃ ultra thin layer on garnet surface. A significant decrease of interfacial impedance, from 1710 Ω cm² down to 1 Ω cm², was observed at room temperature, effectively negating the lithium metal/garnet interfacial impedance. The reduction revealed that the oxide coating enabled wetting of metallic lithium in contact with the garnet electrolyte surface and the lithiated-alumina interface allowed effective lithium ion transport between the lithium metal anode and garnet electrolyte.¹¹⁶

Zoom In Zoom Out Reset image size

J. B. Goodenough and co-workers prepared a solid polymer electrolyte membrane composed of lithium poly(acrylamide-2-methyl-1-propane-sulfonate) (PAS) and polyethylene oxide (PEO). The PEO-PAS was coated on garnet electrolyte of Li_6.5La₃Zr_1.5Ta_0.5O₁₂ (LLZTO) to make a polymer-ceramic single-ion conducting solid electrolyte. The interfacial impedance of a Li/PEO-PAS/LLZTO/PEO-PAS/Li cell demonstrated a dramatically lower overall resistance, from around 5000 Ω for Li/LLZTO/Li to less than 400 Ω. The impedance curves of Li/PEO-PAS/LLZTO/PEO-PAS/Li symmetric cells during the lithium plating-stripping cycles at a constant current didn't exhibit any obvious change. After cycling, the lithium-metal anode was carefully peeled off and observed with SEM. The lithium-metal surface showed a uniform layer of deposited lithium without obvious dendrite formation, which indicates that the PEO-PAS coating layer not only effectively lowered the interfacial impedance, but also successfully suppressed dendrite formation owing to a more uniform interface and Li⁺ homogeneous flux across the interface.¹¹⁷

B. Xu et al. had introduced Li₃PO₄ as a second phase additive for Li_6.5La₃Zr_1.5Ta_0.5O₁₂ garnet electrolyte and resulted in decreased interfacial resistance and Li/garnet/Li cell successfully cycled at a current density of 0.1 mA cm⁻² at 60 °C for 60 h.¹¹⁸ The outstanding interface stability can be attributed to the in situ reaction of the Li flux with Li₃PO₄ to form a self-limiting and ion-conducting interphase Li₃P near the grain boundaries.

The interfacial layers, like alumina (Al₂O₃) and gold (Au), which can increase the wettability of lithium metal anode with LLZO and provide a good contact between the garnet-type electrolyte and lithium metal anode, were prepared by ALD or sputter which are too expensive and complex to be adopted for large scale applications. Y. Tian et al. prepared a composite electrolyte composed of Li_6.75La₃Zr_1.75Ta_0.25O₁₂ (LLZTO) particles embedded in the amorphous Li₃OCl (2 wt%) by grounding LLZTO and Li₃OCl powders, pressing the solid-state mixture into the pellet, reheating at 350 °C for 12 h and then quenching in a glove box to room temperature. As shown in Fig. 3, during the charge-discharge process of Li/LLZTO-2 wt% Li₃OCl/Li cell, the Li₃OCl in composite in situ reacts with the lithium metal to form a stable and dense interfacial layer, which not only greatly decreases the interfacial resistance between the electrolyte and lithium metal electrode, but also effectively suppresses the growth of lithium dendrite on the surface of lithium metal anode during lithium plating-striping.¹¹⁹ R. Sudo et al. investigated the effect of the Al₂O₃ content in LLZO on the interface behavior with lithium metal. Garnet-type lithium conducting solid electrolytes with the nominal Li₇La₃Zr₂O₁₂ composition and with various Al₂O₃ contents were prepared by solid state reaction at 1180 °C. The interface resistance between lithium metal and Al₂O₃-doped LLZO was significantly dependent on the Al₂O₃ content. The lowest grain boundary resistance and interface resistance were obtained for 0.5 wt% Al₂O₃-doped LLZO, with the highest relative density of 93.7%.¹²⁰

Zoom In Zoom Out Reset image size

The environment effect on interface resistance between lithium metal and LLZO are important for ASSLBs assemble and their usage. Recently, the Li/LLZO interface stability was observed under different environmental conditions, like temperature and current density,¹²¹ air exposure,¹²² moisture (relative humidity) and carbon dioxide.¹²³ It was demonstrated that by heating to 175 °C, the Li/LLZO interface resistance decreased dramatically (from 5822 Ω cm² (room temperature) down to 514 Ω cm² (175 °C)) and the maximum sustainable current density of the Li/LLZO interface was increased evidently from 30 °C to 175 °C.¹²¹ It was demonstrated that LLZO reacted with humid air，the most favorable pathway involving protonation of LLZO and formation of an LiOH intermediate, subsequent exposure to CO₂ converting the LiOH to Li₂CO₃. Moreover, the Li/LLZO interfacial resistance was correlated to LLZO surface chemistry. It was shown that the formation of contamination layers at the LLZO surface significantly increases in the Li/LLZO interfacial impedance. These results suggest that the LLZO can be synthesized, and densified in ambient air, measures must be taken to prevent subtle surface contamination to enable relatively low Li/LLZO interfacial resistance.

A. Schwöbel and coworkers studied interface layer formation between LiPON and metallic lithium using an in situ X-ray photoemission spectroscopy (in situ XPS) surface science approach. Because of the buried nature of the interface and the XPS surface sensitivity, they had used a surface science approach to analyze the interface with in situ XPS by exposing LiPON film stepwise to lithium in ultrahigh vacuum. Therefore at each exposure XPS measurements were conducted. They concluded that the chemical reactions lead to the decomposition into smaller units like Li₃PO₄, Li₃P, Li₃N and Li₂O, and that the interface reactions were not continuous, but quickly vanished due to the formation of a suitable passivation layer, which meant that the chemical reactions were finished after a certain period of time and not all the LiPON was destroyed. And that the disrupted interface region must be thin enough to allow Li-ion transfer. Meanwhile, they believed that in situ experiments combined with the highly surface sensitive XPS yield a valuable amount of information about the structure and the chemical composition directly at the interface.^103,107

C. Vinado and coworkers studied the interfacial behavior of Li₁₀SnP₂S₁₂ (LSPS) SSE with Li₃NbO₄-coated LiCoO₂ cathode, which the Li₃NbO₄ layer was coated by atomic layer deposition (ALD).¹³² Their study showed that the ALD coating largely improved the electrochemical performance of ASSLBs with the LSPS SSE. It is obvious that the ultrathin Li₃NbO₄ coating layer largely impedes the atomic inter-diffusion, resulting in much thinner interphase layer and hence lower interfacial resistance compared with uncoated particles (as shown in Fig. 4). It is demonstrated that surface protection of the cathode materials is necessary, at least for the layered cathodes (i.e. LCO, NCA (LiNi_0.8Co_0.15Al_0.05O₂), NMC(LiNi_1−x−yMn_xCo_yO₂)) and that the ALD coating with ultrathin oxide-based Li-ion conductors is effective in not only protecting the active-material/SE interface, but also reducing interfacial resistance.¹³³ Using LiNbO₃ coated LiCoO₂ as cathode, methoxyperfluorobutane (HFE) coated/infiltrated Li₇P₃S₁₁ glass-ceramic as electrolyte, and LiF coated Li metal as anode, the electrochemical performances of the LiNbO₃@LiCoO₂∣Li₇P₃S₁₁∣Li ASSLB showed a high reversible discharge capacity of 118.9 mAh g⁻¹ at 0.1 mA cm⁻² and retained 96.8 mAh g⁻¹ after 100 cycles.¹¹ Other coated cathodes, like Li₂O–SiO₂ glasses coated LiCoO₂,¹³⁴ Li₄Ti₅O₁₂ coated LiNi_0.8Co_0.15Al_0.05O₂,¹³⁵ LiAlO₂ coated Li(Ni_1/3Mn_1/3Co_1/3)O₂,¹³⁶ Li₃PO₄ coated LiNi_0.5Mn_1.5O₄ electrode,¹³⁷ Li₂MoO₄ coated Li[Ni_0.8Co_0.15Al_0.05]O₂,¹³⁸ LiTaO₃ coated on the LiNi_1/3Co_1/3Mn_1/3O₂ (NMC) cathode,¹³⁹ were shown that the coatings were effective in suppressing the side reaction between the cathode and the sulfide electrolyte and improving interfacial stability.

Zoom In Zoom Out Reset image size

The composite cathodes for ASSLBs based on sulfide solid electrolytes have been proven to be effective to improve the battery performance.¹⁴⁰ The composite cathodes are typically prepared by mixing the active material, the sulfide-based solid electrolytes and conductive additive. This is because sulfide materials could provide of good intimate contact with composite electrodes through a simple cold pressing due to its low Young's modulus. Composite cathode with high content of an active material could be prepared by a liquid phase process assisted by a dispersant agent to produce a better electrode and electrolyte interface. A composite cathode was prepared by the dispersion of LiNi_1/3Mn_1/3Co_1/3O₂ particles and carbon additive in the Li₆PS₅Cl-solution containing dispersant agent and subsequent drying at 180 °C. Bulk-type ASSB fabricated with the composite cathode containing 89 wt% of the active material showed an initial discharge capacity of 110 mAh g⁻¹ at 25 °C and maintained 95% discharge capacity after 15 cycles.¹⁴¹ Other composite electrodes using sulfide-based SSE coated on active materials by solution processes, for example, cathode materials of Li₆PS₅Cl-coated LiCoO₂,¹⁴² Li₆PS₅Br-coated LiCo_1/3Ni_1/3Mn_1/3O₂,⁴² Li₆PS₅Cl-coated Li₂S,¹⁴³ Li₆PS₅Cl-coated MoS₂,¹⁴⁴ and Li₆PS₅Cl-coated LiNi_0.8Co_0.1Mn_0.1O₂,¹⁴⁵ showed good electrode/electrolyte interface and improved electrochemical performance. All-solid state cells with these composite electrodes have been proved to achieve better electrochemical performance compared to non-coated active materials.

The key for realization of ASSLBs is to find an ideal SSE with high conductivity, wide electrochemical window and excellent compatibility with both cathode and anode. To solve interfacial problems of sulfide solid electrolytes like LGPS, the bi-layer construction of sulfide electrolytes was reported recently. J. E. Trevey et al. firstly revealed that ASSLBs constructed with the bi-layer sulfide electrolyte (Li∣77.5Li₂S-22.5P₂S₅/Li_3.3Ge_0.3P_0.7S₄∣LiCoO₂) exhibited a better electrochemical performance than those with the single layer electrolyte. They revealed that the different electrolyte configurations (Li∣(Li₂S-P₂S₅)∣LiCoO₂, Li∣(Li₂S-P₂S₅/Li₂S-GeS₂-P₂S₅)∣LiCoO₂) resulted in different electrochemical performance for ASSLBs.¹⁴⁶ The Li₁₀GeP₂S₁₂ (LGPS) and glass-ceramic Li₃PS₄ (LPS) bi-layer construction was reported by B. R. Shin et al. The combined use of the TiS₂-LGPS cathode and the LGPS-LPS SSE bi-layer where LPS forms an interface with the Li-In anode ((TiS₂-LGPS)∣(LGPS-LPS)∣Li-In cell) resulted in the best overall performance at 30 °C, exhibiting a capacity of ∼60 mAh g⁻¹ at 20 °C.¹⁴⁷

Based on the bi-layer electrolyte several ASSLBs with different electrode materials were reported.^148–150 The Li∣75% Li₂S-24% P₂S₅-1% P₂O₅/Li₁₀GeP₂S₁₂∣Fe₃S₄ @Li₇P₃S₁₁ ASSLB exhibited higher discharge capacity and better rate capability, the discharge capacity remained at a high value of 1001 mAh g⁻¹ at a current density of 0.1 A g⁻¹ after 200 cycles.¹⁴⁸ The Li∣70%Li₂S-29%P₂S₅−1%P₂O₅/Li₁₀GeP₂S₁₂∣FeS ASSB showed reversible discharge capacity of 550 mAh g⁻¹ after 50 cycles at a current density of 0.1 A g⁻¹ and exhibited superior rate performances.¹⁴⁹ D. Xie et al. prepared Sb₂O₅ doped β-Li₃PS₄ glass-ceramic electrolytes, and the sample of Li₃P_0.98Sb_0.02S_3.95O_0.05 exhibited the highest ionic conductivity of 1.08 × 10⁻³ S cm⁻¹ at room temperature. The constructed LiCoO₂∣LGPS/Li₃P_0.98Sb_0.02S_3.95O_0.05∣Li ASSB showed high discharge capacity of 133 mAh g⁻¹ at 0.1 C (1 C = 120 mA g⁻¹) in the range of 3.0–4.3 V vs Li/Li⁺) at 25 °C. Moreover, the ASSLBs with the doped glass-ceramic bi-layer electrolyte were still workable at −10 °C.¹⁰⁸

S. Hao et al. developed a new LLZO-based composite electrolyte by LiBr incorporation into Li_6.25La₃Zr₂Al_0.25O₁₂ (LLZO). The LiBr, which has much lower melting point than LLZO, would serve as a binder for interaction and densification among LLZO grains owing to the flow characteristic of liquid LiBr during high-temperature sintering. Due to good wetting properties among LLZO grains, large contact area, and effective bonding effect, the incorporation of LiBr into LLZO matrix could thereby give rise to fast Li-ion migration in the LLZO-based electrolytes. Moreover, the interfacial resistance between the LLZO-based electrolyte and electrode would be reduced by improving wetting properties and effective contact at the microscopic level.¹⁵¹ The LiCoO₂∣LLZO-based composite electrolyte∣Li full cells delivered a specific discharge capacity of 121.5 mAh g⁻¹ in the first cycle, and achieved capacity retention of 91.8% at the 70th cycles. X. Yan et al. prepared Li₇La₃Zr₂O₁₂ (LLZO) composite solid electrolyte by a novel nanoparticle slurry coating method. The thin film solid electrolyte, composed of LLZO, LiN(CF₃SO₂)₂ (LiTFSI) Li-salt and some additives, was obtained by a wet coating method. The Li∣LLZO composite solid electrolyte∣LCO cell showed a discharge capacity of 119.3 mAh g⁻¹ at a current density of 0.21 mA cm⁻² after 45 cycles at room temperature.¹⁵²

When garnet Li₇La₃Zr₂O₁₂ (LLZO) is combined with a common positive electrode LiCoO₂, mutual diffusion of Co, La, and Zr between LLZO and LiCoO₂ occurs at the interface during a thermal treatment after depositing thin LiCoO₂ films on LLZO pellets. It is demonstrated that reaction phases due to mutual diffusion are easily formed between LLZO and LiCoO₂ at the interface.¹⁵³ In this case, electrolyte/electrode interface modification by intermediate coating has been suggested as an effective way to reduce interfacial resistivity.

T. Kato et al. deposited a thin Nb metal layer on Li₇La₃Zr₂O₁₂ (LLZO) pellet by radio frequency magnetron sputtering in an Ar atmosphere at room temperature. A thermal treatment following deposition of a Nb metal layer onto LLZO produced a low-resistivity amorphous layer. The interfacial resistivity between LLZO and LiCoO₂ decreased from 2600 Ω·cm² to 150 Ω·cm² depending on the thickness of the Nb metal layer. The interface modification approach using a Nb interlayer dramatically improved the discharge capacity and rate capability of a ASSLB.¹⁵⁴ S. Ohta and coworkers constructed a ASSB by screen-printing process using Nb doped LLZO (LLZONb) as a solid electrolyte and Li₃BO₃ as a intermediate layer between the LiCoO₂ (LCO) active cathode material and LLZONb. The Li₃BO₃, a lithium ion conductor, has a low melting point (ca. 700 °C) and is expected to act as a bonding material at the interface by sintering. With Li₃BO₃ layer sufficient interface contact between the LCO cathode layer and the LLZONb solid electrolyte can be easily achieved by an annealing process.¹⁵⁵

C. Wang et al. demonstrated an effective cathode/garnet interface with mixed ionic-electronic conductor through a unique slow pre-lithiation process, which dramatically improved the conductivity of a cathode for both electrons and ions as well as the cathode/garnet interface. A composite cathode of titanium sulfide (TiS₂) mixed with carbon nanotubes was slowly pre-lithiated, resulting in more than 20 times lower interfacial resistance. The all solid state batteries of Li metal anode∣garnet (Li_6.75La_2.75Ca_0.25Zr_1.75Nb_0.25O₁₂ ) SSE∣TiS₂ composite cathode can work at high temperatures from 100 °C to 150 °C for 400 cycles at current densities up to 1 mA cm⁻².¹⁵⁶

To address the high interfacial resistance between the solid electrolyte and cathode, B. Liu et al. developed a rapid thermal annealing method to melt the cathode and form a continuous contact. The ASSLB using thermally stable Li₇La_2.75Ca_0.25Zr_1.75Nb_0.25O₁₂ garnet SSE, lithium metal anode, and V₂O₅ cathode was demonstrated to operate at 100 °C high-temperature with 97% Coulombic efficiency and reliable safety as well as stable cycling performance. The resulting interfacial resistance of the solid electrolyte and V₂O₅ cathode was significantly decreased from 2.5 × 10⁴ to 71 Ω·cm² at room temperature and from 170 to 31 Ω·cm² at 100 °C.¹⁵⁷ E. A. Ilina et al. deposited 30Li₂O·47.5V₂O₅·22.5B₂O₃ glassy cathode on Li₇La₃Zr₂O₁₂ substrate applied to lithium middle-temperature ASSLBs. The interface features between cathode and SSE at various crystallizing temperatures (650 °C–850 °C) and holding times (0.5–5 min) of cathode on LLZO substrate was investigated. The cathode deposited on LLZO annealed at a proper temperature and holding time can create the optimal interface and good contact between electrode and electrolyte.¹⁵⁹

The composite cathode, which contains solid electrolyte to the cathode active material to improve ionic and electronic conduction, is another way to obtain a low electrode/electrolyte interfacial resistance. M. He et al. fabricated LiFePO₄ based flexible composite cathodes by introducing conductive frameworks consisting of succinonitrile and lithium salt LiTFSI, significantly improving the contact performance and interface stability between garnet solid electrolyte and LiFePO₄ cathode. The introduction of such flexible frameworks not only enables close contact between the cathode and the stiff garnet-structured Li_6.375La₃Zr_1.375Nb_0.625O₁₂ SSE, but also bridges every electrode and electrolyte particles together forming interconnected three dimensional ionic conductive paths. The ASSB of Li∣SSE∣LiFePO₄ with the flexible composite cathodes demonstrated an initial discharge capacity of 149.8 mAh g⁻¹ and the Coulombic efficiency of 99% after 100 cycles at 0.05 C at room temperature.¹⁵⁰ W. Zha et al. applied wet coating and hot pressing technique to form a continuous and homogeneous electrolyte layer on a cathode layer. The LLZTO-based electrolyte contained 90 wt% Li_6.4La₃Zr_1.4Ta_0.6O₁₂ (LLZTO) powder and 10 wt% PEO-LiTFSI. The composite cathode contained 55 wt% LiFePO₄, 20 wt% LLZTO powder, 10 wt% Super-P, and 15 wt% PEO-LiTFSI. The Li∣90LLZTO-10PEO∣LiFePO₄ batteries delivered a discharge specific capacity of 148.3 mAh·g⁻¹ after 50 cycles at 60 °C and a discharge specific capacity of 96.0 mAh·g⁻¹ after 50 cycles at 25 °C.¹⁵⁸