Direct Modeling of the Interfacial Resistance in All‐Solid‐State Battery

ABSTRACT

Interfacial reconstruction and its associated high resistance govern the performance of all‐solid‐state batteries (ASSBs). However, indirectly inferring interfacial potentials from bulk band alignments masks the true solid–solid electrochemistry, causing orders‐of‐magnitude discrepancies in predicting space‐charge layer (SCL) resistances and impeding interface screening. Herein, by traversing 310 distinct interfaces from ∼29,000 literatures, we develop a non‐empirical numerical procedure that directly maps lithium‑ion redistribution to interfacial resistance by integrating ligand‑field theory with the SCL model. Considering electric potential differences and intrinsic carrier properties during interfacial reconstruction via a modified ligand‐field splitting strength (MLFSS) descriptor yields unprecedented bridging between modeling and measurement, reducing predicted resistance discrepancies from over ten orders of magnitude to within two. On this basis, we resolve the highly system‐dependent controversy over oxide interfacial resistances by identifying extreme MLFSS disparities (>3.5 eV) as the decisive factor, while emphasizing ion‑intercalation sulfides (<0.2 eV) as cathodes for their intrinsic SCL suppression. The predictive capability of this tunable criterion is validated in an all‐sulfide V_0.5Cr_1.5S₄/Li₁₀GeP₂S₁₂/75% Li₂S‐24% P₂S₅–1% P₂O₅/Li prototype. The resulting ultralow interfacial resistance of 8.8 Ω cm² ensures superior cycling stability at an active‐material energy density of 562 Wh kg⁻¹, establishing a practical paradigm for breaking the energy and kinetics trade‐off in ASSBs.

We introduce a non‐empirical numerical procedure bridging atomic‐scale electronics with macroscopic resistance in all‐solid‐state batteries, reducing prediction errors from 10 orders of magnitude to two. Validated by an all‐sulfide prototype where ultralow resistance ensures superior cycling stability at a record energy density, this framework transforms the space‐charge layer concept into a quantitative tool for precise interface tailoring.

1. Introduction

Employing inorganic solid electrolytes (SEs), all‐solid‐state batteries (ASSBs) promise exceptional safety and energy density, offering a definitive solution to the flammability risks of liquid electrolytes [1, 2]. The recent advent of SEs with liquid‐like ionic conductivities (∼10⁻² S·cm⁻¹, e.g., Li₁₀GeP₂S₁₂ [3, 4]) has shifted the performance bottleneck from bulk ionic transport to the interface. While prevailing theories attribute this impedance to space‐charge layer (SCL) formation [5, 6], interfacial chemical reactions [7, 8], and poor physical contact [9, 10], these mechanisms alone fail to reconcile the contradictory resistance values reported across identical material combinations [11, 12]. Recently, interfacial contact and conduction issues have been optimized via advanced preparation technologies, and chemical instability is mitigated through atomic‐scale structure engineering [13, 14, 15, 16, 17]. For instance, Lee et al. [13] used flash lamp annealing to induce near‐surface reconstruction, achieving a stable and efficient cathode/solid electrolyte interphase, which contributed to improved performance. However, since conventional interpretation methodologies relying solely on contact mechanics or chemical side reactions are insufficient to describe the true interfacial electrochemistry, the above discrepancies still persist [18, 19]. Instead, resolving these anomalies requires a paradigm shift toward understanding the intrinsic SCL effect [20, 21], which is governed by the coupled redistribution of ions and electrons beyond simple chemical or physical defects [22].

The interfacial potential difference ($\mathrm{\Delta }\varphi$) is the primary driver of lithium‐ion redistribution and SCL resistance (Figure 1a) [23]. Although the redistribution is also affected by the combination of cathode and SE, coupled with surface adsorptions, interface defects, and other factors, the intrinsically higher chemical potential (viz., lower $\varphi$) of the SE has led to the experimentally observed depletion of lithium ions on its side [24, 25]. Advanced techniques, including electron holography [26], in situ differential phase contrast scanning transmission electron microscopy [25], in situ ellipsometry [27], operando Raman spectroscopy [28] and Kelvin probe force microscopy [29], have been employed to investigate phenomena related to lithium ion/charge depletion, yet a priori prediction of electric potential distributions remains challenging due to complex chemical environments and dynamic, state‐of‐charge‐dependent evolution [30, 31]. So far, the SCL at cathode/SE has been extensively explored through computational/numerical methods [32, 33]. Conventional SCL models, approximating $\mathrm{\Delta }\varphi$ from ground‐state carrier chemical potentials, fail to accurately describe cathode/SE interfacial resistance under operational ion transfer/redistribution [34]. This leads to profound inaccuracies in predicted resistances, often erring by several orders of magnitude and hindering rational design (Table S5). For instance, in oxide‐based systems where severe SCL effects are not typically anticipated, the measured interfacial resistances of LiCoO₂/Li_1.2Al_0.2Ti_1.8(PO₄)₃ and LiCoO₂/Li₇La₃Zr₂O₁₂, both influenced by combined SCL and contact contributions, remain at the same level [35, 36, 37]. Yet predictions considering only SCL yield resistances differ by four orders of magnitude [34]. Recalculation using indirect $\mathrm{\Delta }\varphi$ confirms this gap is intrinsic, not technical (Figure 2a, Diagram IV).

FIGURE 1.

Calculating interfacial SCL resistance using the ligand‐field and space‐charge layer integrated numerical procedure. (a) Electric ($\varphi$), chemical (μ), and electrochemical ($\stackrel{\sim }{\mu }$) potential profiles for Li⁺, e–, and neutral Li in ASSB. ${\stackrel{\sim }{\mu }}_{e^{-}}$ is dominated by the Fermi level (E_F). Superscripts C and A represent the cathode and anode, respectively. (b) Schematic evolution of electric potential ($\varphi$) and lithium‐ion concentration ($c_{\mathrm{L}{\mathrm{i}}^{+}}$) with distance (x^k ) in the SCL. (c) Computational workflow for determining the normalized resistance (${\stackrel{\sim }{R}}_{\mathrm{SCL}}$) at the cathode interface. The splitting parameter (Dq) and coefficient (n) are derived via the ligand‐field model (see Code Availability). The procedure iteratively updates the lithium‐ion concentration ($c_{{Li}^{+}}^k$), electric field strength (E^k ), calculation step (Δx ^{k + 1}), and electric potential (${\varphi }^{k+1}$) at the interface. Upon satisfying the convergence criterion ($\mathrm{\Delta }{\varphi }^{k+1}\le {10}^{-8}$ V), the ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ is calculated.

FIGURE 2.

Screening of promising cathode/solid electrolyte interface combinations. (a) Quantitative SCL modeling of interfacial resistances approaching realistic contact. (b) Band alignment model for studying resistances. Band alignment of 12 OCAs, 4 OCOs, 14 SCAs, and diverse SEs aligned relative to the Li‐1s reference state. SEs include thio‐LISICON, LGPS, argyrodite (3 S‐SEs); garnet, perovskite, anti‐perovskite, NASICON, LiPON (5 O‐SEs); and hexagonal/cubic close‐packed halides (2 H‐SEs). Gray areas indicate electron‐accepting (cathodes/coatings) or ‐donating (SEs) energy ranges. Solid and dotted lines represent average VBMs in discharge and charge states, respectively. The LGPS VBM is set to 0 eV (red box). Parenthetical S, M, and I denote semiconductor, metal, and insulator, respectively. (c) Interfacial Li⁺ and e⁻ migration driven by energy‐level differences between cathodes (or coatings) and SEs, where E _CBM and E _VBM denote average band edges.

The root of this failure lies in the disconnect between macroscopic thermodynamics and microscopic electronic structure. As Goodenough posited [38], lithium chemical potential differences ($\mathrm{\Delta }\varphi$) are fundamentally governed by the electrochemical potential of electrons ($\mathrm{\Delta }{\stackrel{\sim }{\mu }}_{{\mathrm{e}}^{-}}$, related to the Fermi level under ligand fields [39], Equation (S6)), which is dictated by the local ligand field. However, a quantitative bridge linking these local ligand‐field variations (at a reconstructed interface) to the macroscopic electric potential and resulting SCL resistance has been conspicuously absent. An open question persists: Is it possible to build a direct, non‐empirical numerical model that accurately maps atomic‐scale electronic reconstruction to interfacial resistance?

Herein, we develop a numerical procedure integrating the SCL model with a Modified Ligand‐Field Splitting Strength (MLFSS) descriptor that quantifies the limit of the interfacial reconstruction electric potential. The MLFSS establishes an electronic structure‐based criterion that directly maps the atomic‐scale interfacial reconstruction limit to macroscopic electric potentials. Unlike conventional models that treat interfaces as static contacts, our approach bridges the longstanding gap between theory and experiment, reducing the prediction error by over ten orders of magnitude. On this basis, we establish the MLFSS disparity as the universal physical determinant of interfacial SCL resistance, clarifying the distinct interfacial behaviors that were traditionally conflated with material classes. Our analysis reveals that extreme MLFSS disparities (>3.5 eV) are the fundamental driver of the divergent resistances in oxides, whereas the modest disparities (<0.2 eV) in ion‐intercalation sulfides intrinsically suppress the SCL effect. Guided by this electronic criterion, we design and synthesize a novel Li₂V_0.5Cr_1.5S₄ (LVCS) cathode. In a full‐cell all‐sulfide prototype (V_0.5Cr_1.5S₄/Li₁₀GeP₂S₁₂/75% Li₂S‐24% P₂S₅–1% P₂O₅/Li), the interface exhibits intrinsic ultralow resistance and enables an energy density of 562 Wh kg⁻¹, surpassing current chalcogenide systems by >20%. Practical interfacial issues indeed involve more factors accompanied by relevant strategies, for example, mechanical integrity, which can be improved through single‐crystalline and cobalt‐free design [40, 41], and optimized manufacturing methods using dry electrodes with ionic conductive binders [42, 43]. Herein, from the theoretical perspective, our procedure empowers rational cathode/electrolyte selection while establishing a predictive interface database for accelerated ASSB development.

2. Results

2.1. Methodological Principles: Direct Modeling of Interfacial Resistance

To resolve the profound inaccuracies inherent in conventional modeling, we develop a non‐empirical numerical procedure that integrates the SCL model with ligand‐field theory. The complete mathematical derivation and numerical algorithms are provided in Section S1. This procedure directly maps electronic structures to interfacial resistances through three logically integrated stages:

2.1.1. SCL Formation and Carrier Equilibrium

Upon contact, the chemical potential disparity between the cathode and solid electrolyte (SE) drives carrier migration [44], forming a non‐neutral SCL region [45]. The electrochemical potential (${\stackrel{\sim }{\mu }}_i$, i is the carrier, Figure 1a) is adopted to describe the carrier migration behavior in this region [46]. At equilibrium ($\mathrm{\Delta }{\stackrel{\sim }{\mu }}_{\mathrm{L}{\mathrm{i}}^{+}}=0$), the lithium‐ion concentration at k position ($c_{\mathrm{L}{\mathrm{i}}^{+}}^k$) in SE follows the Boltzmann distribution governed by the interfacial electric potential difference ($\mathrm{\Delta }{\varphi }^k={\varphi }^k-{\varphi }^{\text{bulk}}$, Figure 1b):

$$c_{{\text{Li}}^{+}}^k=c_{{\text{Li}}^{+}}^{\text{bulk}}\exp \left(-\frac{e\mathrm{\Delta }{\varphi }^k}{k_{\mathrm{B}}T}\right),$$ (1)

where ${\varphi }^{\mathrm{bulk}}$ and $c_{\mathrm{L}{\mathrm{i}}^{+}}^{\mathrm{bulk}}$ is the electric potential and Li⁺ concentration in bulk SE, respectively. ${\varphi }^k$ is the electric potential at k position, e is the elementary charge, k_B is the Boltzmann constant, and T is the Kelvin temperature. It is worth noting that while the standard Poisson–Boltzmann (PB) theory assumes a dilute solution by neglecting ion–ion correlations (a known limitation in highly concentrated liquid electrolytes), its application is fundamentally justified here. Because the SCL interfacial resistance is overwhelmingly bottlenecked by the carrier depletion region where the local mobile Li⁺ concentration drops exponentially ($c_{\mathrm{L}{\mathrm{i}}^{+}}^k\ll c_{\mathrm{L}{\mathrm{i}}^{+}}^{\mathrm{bulk}}$), the pertinent interfacial zone operates deep within an ultra‐dilute point‐defect regime, rendering steric crowding effects effectively negligible (a detailed quantitative justification is provided in Section S1.1). Crucially, the lithium chemical potential difference is dominated by the electrochemical potential difference of electrons ($\mathrm{\Delta }{\stackrel{\sim }{\mu }}_{{\mathrm{e}}^{-}}$) between electrodes [38], which corresponds to their Fermi level difference (ΔE_F) [39]. Thus, mapping Fermi levels under realistic contact conditions is a prerequisite for obtaining the upper limit of $\mathrm{\Delta }{\varphi }^k$ (Section i of Figure 1c).

2.1.2. Quantifying the Interfacial Potential Limit via a Ligand‐Field Descriptor

To quantify the electronic reconstruction that bulk models overlook (Figure 2b; Section S2), we introduce a Modified Ligand‐Field Splitting Strength (MLFSS, Section ii of Figure 1c) descriptor to determine the theoretical upper limit of the initial interfacial potential ($\mathrm{\Delta }{\varphi }_{\mathrm{MLFSS}}^0$) for the SCL model:

$$\mathrm{MLFSS}=nDq+\lambda E_{\mathrm{g}},$$ (2)

where n is the transition‐metal (TM)‐d splitting coefficient, Dq is the splitting strength derived from bond‐length variations [47], E _g is the band gap, and λ distinguishes the electron‐acceptance mechanism (valence vs. conduction band). This descriptor precisely bridges microscopic electronic states with the interfacial potential limit. Crucially, the quantitative fidelity of the MLFSS is inherently predicated on the precise description of the ground‐state electronic structure. Recognizing the high sensitivity of transition‐metal localized states to methodological parameters (e.g., functional choice and Hubbard U corrections), this procedure strictly necessitates rigorously calibrated density functional frameworks to prevent spurious Fermi level shifts and ensure reliable resistance predictions (Section S2 and Figure S1).

2.1.3. Resolving Resistance From Potential Profiles

With $\mathrm{\Delta }{\varphi }_{\mathrm{MLFSS}}^0$ established, the spatial evolution of the potential is resolved via the one‐dimensional Poisson equation, where ρ is the net charge density ($\mathrm{e}(c_{\mathrm{L}{\mathrm{i}}^{+}}^k-c_{\mathrm{L}{\mathrm{i}}^{+}}^{\mathrm{bulk}})$) and ε is the permittivity:

$$\frac{{\mathrm{d}}^2\varphi }{\mathrm{d}{x}^2}=-\frac{\rho }{\epsilon }.$$ (3)

To solve this numerically, we introduce the electric field intensity (E^k ) [48] to relate the potential profile to the locally uncompensated charges:

$$E^k=\sqrt{\frac{2{\mathrm{k}}_{\mathrm{B}}T}{\epsilon }\left[\left(c_{\mathrm{L}{\mathrm{i}}^{+}}^k-c_{\mathrm{L}{\mathrm{i}}^{+}}^{\mathrm{bulk}}\right)-\frac{\mathrm{e}{c}_{\mathrm{L}{\mathrm{i}}^{+}}^{\mathrm{bulk}}{\varphi }^k}{{\mathrm{k}}_{\mathrm{B}}T}\right]}$$ (4)

By iteratively updating $c_{\mathrm{L}{\mathrm{i}}^{+}}^k$, E^k , and ${\varphi }^k$ via a staged Taylor series expansion (see Section S1.3), we determine the SCL charge density (ρ_SCL =   − εE ⁰). Finally, a normalized SCL resistance (${\stackrel{\sim }{R}}_{\mathrm{SCL}}$) is defined relative to the standard LiCoO₂/Li₁₀GeP₂S₁₂ (LCO/LGPS) interface to quantify the impedance arising from carrier depletion:

$${\stackrel{\sim }{R}}_{\mathrm{SCL}}=\frac{R_{\mathrm{SCL}}^{\mathrm{Cathode}\left(\mathrm{or}\mathrm{coating}\right)/\mathrm{SE}}}{R_{\mathrm{SCL}}^{\mathrm{LCO}/\mathrm{LGPS}}}.$$ (5)

This rigorous procedure, implemented in our IR‐LFSCL code (https://github.com/IR‐LFSCL), enables resistance predictions that quantitatively mirror experimental trends across diverse material systems (Section iii of Figure 1c).

2.2. Limitations of the Conventional Band Alignment Model

Accurately predicting interfacial resistance in ASSBs requires a holistic understanding of the coupled ionic and electronic phenomena. While the electronic component parallels well‐understood semiconductor band bending, and the ionic component is driven by the resulting electric potential difference ($\mathrm{\Delta }\varphi$) [49], existing models often oversimplify this interaction. By indirectly deriving $\mathrm{\Delta }\varphi$ solely from ground‐state carrier chemical potentials, these models reduce the interface to simple Coulombic interactions, leading to resistance predictions that deviate from experimental values by orders of magnitude. Consequently, rational interface design has been hindered by a reliance on empirical trials rather than predictive theory.

Band alignment based on DFT calculations has emerged as a visualization approach to assess the electrochemical potential step at cathode/SE interfaces [50], which adopts the band energy‐level difference between the non‐contact cathode and SE bulk materials to estimate interfacial resistance (Figure 2a, Diagram IV). To evaluate its general validity, we systematically compute the band properties of 310 distinct cathode (or coating)/SE interfaces extracted from 29,689 theoretical and experimental reports on ASSBs over the last 5 years (Figure S3). These involve 13 oxide cathodes (OCAs), four oxide coatings (OCOs), 14 sulfide cathodes (SCAs), three sulfide SEs (S‐SEs), five oxide SEs (O‐SEs), and two halide SEs (H‐SEs) (Table S2). By aligning the Valence Band Maximum (E _VBM) and Conduction Band Minimum (E _CBM) relative to the Li‐1s Kohn–Sham state (Figures 2b and S4–S8), we assess the band energy‐level differences between cathodes, and SEs ($\mathrm{\Delta }{E}_{\mathrm{band}}^{\mathrm{cathode}}=E_{\mathrm{VBM}}^{\mathrm{cathode}}-E_{\mathrm{VBM}}^{\mathrm{SE}}$) that theoretically dictate resistance in the absence of external current. OCAs exhibit a small average $\mathrm{\Delta }{E}_{\mathrm{band}}^{\mathrm{OCAs}}$ with S‐SEs at both discharge (−0.715 eV, Figure 2b) and charge states (−1.144 eV, Figure S9), implying continuous electron depletion from the SE and inherently high resistance. Conversely, insulating OCOs exhibit large average $\mathrm{\Delta }{E}_{\mathrm{band}}^{\mathrm{OCOs}}($1.576 eV), consistent with their function as electron‐blocking protective layers (Figure 2c).

This phenomenon can be further quantitatively confirmed by the resistance calculated by the numerical approach combining band alignment with the SCL model (R _band, Table S4). Here, the relative resistances of OCAs (or OCOs) and SEs $\left({\stackrel{\sim }{R}}_{\mathrm{band}}^{\mathrm{OCAs}(\mathrm{or}\mathrm{OCOs})/\mathrm{SE}}=\frac{R_{\mathrm{band}}^{\mathrm{OCAs}(\mathrm{or}\mathrm{OCOs})/\mathrm{SE}}}{R_{\mathrm{band}}^{\mathrm{LCO}/\mathrm{LGPS}}}\right)$ are used to facilitate the comparison of their magnitudes. ${\stackrel{\sim }{R}}_{\mathrm{band}}^{\mathrm{OCAs}/\mathrm{S}-\mathrm{SEs}}$ are generally several orders of magnitude larger than ${\stackrel{\sim }{R}}_{\mathrm{band}}^{\mathrm{OCOs}/\mathrm{S}-\mathrm{SEs}}$ (Table S4), confirming that the introduction of OCO can reduce the interfacial resistance and protect the S‐SE. Since OCOs are engineered to suppress interfacial chemical reactions, their interfacial impedance is predominantly governed by the SCL effect. The accurate prediction of OCO resistance by our model serves as robust evidence that SCL physics dictates performance once chemical instability is mitigated. Consistent with the general consensus on their favorable chemical compatibility, SCAs generally maintain low interfacial resistances with S‐SEs (with the average $\mathrm{\Delta }{E}_{\mathrm{band}}^{\mathrm{SCA}}$ of 1.419/1.018 eV at the discharge/charge states, Figure 2b).

However, strictly relying on this “non‐contact” model reveals a critical flaw: it estimates interfacial potential from bulk energies without accounting for the realistic electrochemical environment, which leads to profound contradictions. For instance, the experimental resistance of the LiNi_0.8Co_0.1Mn_0.1O₂ (NCM811)/LGPS interface is over two orders of magnitude higher than that of LiCoO₂/Li₃InCl₆ [51, 52]. Yet, the conventional band alignment model erroneously predicts the opposite trend (Table S5). Resolving this requires moving beyond fixed energy levels to dynamically model the coupled evolution of lithium‐ion concentration, electric field strength, and potential distribution under realistic interfacial reconstruction (Figure 2a, Diagram V).

2.3. Ligand‐Field‐Integrated Model for Quantitative SCL Resistance Prediction

To bridge the gap between static theoretical approximations and dynamic experimental realities, we develop a numerical procedure that integrates the Modified Ligand‐Field Splitting Strength (MLFSS, Equation (2)) descriptor with the SCL model. Unlike conventional methods that rely on indirect potential estimations, the MLFSS descriptor directly quantifies the upper limit of the interfacial electric potential difference ($\mathrm{\Delta }{\varphi }^0$, Tables S5). Crucially, this descriptor acts as a position‐sensitive boundary condition that captures the localized structural distortions at the immediate interface, overriding invariant bulk properties. For example, while the bulk valence band alignments (ΔE _band) for ideal non‐contact LiCoO₂/LGPS and NCM811/LGPS interfaces are −1.777 and −0.384 eV, respectively, their MLFSS values shift dramatically to −0.487 and −0.814 eV upon realistic contact reconstruction (Figure S10). This inversion highlights the critical impact of localized structural deformations. Consequently, predicting the relative resistance of NCM811 against LCO using the bulk band alignment yields a severely distorted logarithmic value of $\lg ({\stackrel{\sim }{R}}_{\mathrm{band}}^{\mathrm{NCM}811/\mathrm{LCO}})=-11.077$. By contrast, our model, grounded in the contact‐reconstructed MLFSS, yields $\lg ({\stackrel{\sim }{R}}_{\mathrm{SCL}}^{\mathrm{NCM}811/\mathrm{LCO}})=2.572$. This corrects a massive discrepancy of over ten orders of magnitude and aligns highly consistently with the experimental measurement ($\lg ({\stackrel{\sim }{R}}_{\exp }^{\mathrm{NCM}811/\mathrm{LCO}})=0.887$).

Integrating MLFSS into the SCL numerical model allows us to dynamically simulate the coupled evolution of lithium‐ion concentration ($c_{\mathrm{L}{\mathrm{i}}^{+}}^k$, Equation (1)) and electric field strength (E^k , Equation 4) across the interface. This iterative process, governed by the Poisson equation (Equation (3), culminates in the calculation of a normalized SCL resistance (${\stackrel{\sim }{R}}_{\mathrm{SCL}}$, Equation (5)), a metric we define to strictly evaluate the impedance arising from carrier depletion. Applied to our dataset of 310 OCA/SCA/OCO and SE interfaces (Table S4) using key material parameters (Table S8), this approach demonstrates transformative predictive power. It successfully narrows the discrepancy between predicted and measured interfacial resistances to within two orders of magnitude. This represents a fundamental leap over conventional band‐alignment models, where errors frequently spanned beyond 10 orders of magnitude. Such quantitative fidelity across a vast chemical space validates the model's ability to capture the essential electrochemistry of solid–solid interfaces.

Armed with this predictive procedure, we resolve the longstanding controversy regarding oxide interfaces. While the conventional view attributes the high resistance of OCA/S‐SE interfaces to O–S electronegativity differences [4], this fails to explain the anomalously low resistance of oxide coatings (OCOs) [6]. Our calculations identify the MLFSS disparity as the decisive factor: Featuring transition metals in lower oxidation states (+2 to +4), OCAs possess low Fermi levels and correspondingly low MLFSS values (−0.4 to −0.814 eV, Figure 3a). This induces a large $\mathrm{\Delta }{\varphi }^0$, driving severe SCL formation and resulting in high ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ (1–10⁴, Table S4). Conversely, OCOs with high‐valence transition metals (+4, +5) and energy levels (4d/5d) exhibit significantly higher average MLFSS (2.710 eV). This electronic configuration imparts insulator‐like characteristics that suppress electric potential drops and SCL formation, yielding ultralow ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ (<0.1). Therefore, the extreme |ΔMLFSS| of up to 3.673 eV between OCOs and OCAs is identified as the universal physical determinant of their divergent interfacial resistances (Figure 3b, Areas I and II).

FIGURE 3.

Quantification of normalized interfacial SCL resistance (${\stackrel{\sim }{R}}_{\mathrm{SCL}}$). (a) MLFSS distributions for 13 OCAs, 4 OCOs, and 14 SCAs; the inset shows average values. Gray areas indicate MLFSS ranges for oxides (OCAs and OCOs) and sulfides (SCAs). (b) Three‐dimensional scatter plot mapping ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ across 310 interfaces against MLFSS and SE ionic conductivity (σ). Sphere projected radii represent σ magnitudes. Colored regions denote specific material classes: yellow for OCAs (Li₂MnO₃, Li₄V₃O₈, LiMn₂O₄, LiCoO₂, LiNiO₂, LiMnO₂, LiFeO₂, NCM333, NCM523, NCM622, NCM811, LiNi_0.5Mn_1.5O₄, and LiFePO₄); blue for OCOs (Li₂ZrO₃, LiNbO₃, Li₄Ti₅O₁₂, and LiTaO₃); and green for SCAs (lithiated LiCuS₂, LiWS₂, LiMoS₂, LiCrS₂, LiMnS₂, LiNiS₂, LiVS₂, LiHfS₂, LiNbS₂, LiTiS₂, LiTaS₂, LiZrS₂, LiCoS₂, and LiFeS₂).

Notably, our model also accurately predicts the behavior of outliers such as the LiFePO₄ cathode. Despite being an OCA, its large band gap (>3.5 eV) results in a high MLFSS (2.249 eV), explaining its experimentally observed low interfacial resistance (${\stackrel{\sim }{R}}_{\mathrm{SCL}}\approx 0.01$) with S‐SEs compared to other OCAs (1–10⁴, Table S4). Thus, based on the ligand‐field theory combined with the SCL model, we quantitatively explain the nature of LiFePO₄ as a cathode with the lowest ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ among OCAs, realizing the design of high voltage OCAs with low interfacial resistances. While applying oxide coatings (OCOs) successfully mitigates resistance, this band‐aid approach introduces intrinsic penalties. The incorporation of inactive coating materials inevitably adds parasitic mass and volume, diluting the overall energy density. Furthermore, the poor ionic conductivity of OCOs necessitates precise manufacturing control to limit thickness (<10 nm [28, 53, 54]), increasing processing complexity and cost.

To circumvent these trade‐offs, we turn our attention to sulfide cathodes (SCAs). While the superior compatibility between sulfide cathodes and sulfide electrolytes is acknowledged in the field, previous understandings were largely empirical or attributed to similar anion frameworks. Our work advances this qualitative consensus into a predictive design rule. Guided by this tunable criterion, we hypothesize that SCAs can offer an intrinsic solution governed by their unique electronic structure. Our numerical procedure reveals that SCAs naturally overcome the limitations of oxides through two fundamental mechanisms: Firstly, the TM‐S bond lengths in SCAs (L_{TM ‐S} = 2.266–2.636 Å) are significantly larger than TM‐O bonds (L_{TM ‐O} < 2.112 Å), owing to the higher energy level of S‐3p orbitals compared to O‐2p. This structural difference results in a higher average MLFSS for SCAs (−0.222 eV, Figure 3a) compared to OCAs (−0.585 eV). Mathematically, this higher MLFSS minimizes the interfacial potential difference ($\mathrm{\Delta }{\varphi }^0$), thereby intrinsically suppressing the driving force for SCL formation. Secondly, SCAs exhibit a remarkably modest MLFSS disparity (|ΔMLFSS_SCAs|≈0.176 eV), which is orders of magnitude smaller than that of oxides. This implies that SCAs do not suffer from the trade‐off between resistance and conductivity (Area I vs. Area II in Figure 3b) seen in oxide systems. Consequently, the calculated ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ of SCAs against typical electrolytes (O‐SEs, S‐SEs, and H‐SEs) consistently falls within a low‐impedance regime (0.0001–0.5, designated as Area III). This finding fundamentally shifts the challenge of ASSB design; that is, the goal is no longer to desperately mitigate high interfacial resistance, but rather to strategically select SCA/SE combinations that optimally balance this inherent low resistance with high energy density (indicated by the red star in Figure 3b).

2.4. Design and Validation of an Optimized SCA/S‐SE Interface

To demonstrate the predictive accuracy of our ligand‐field‐integrated numerical procedure, we executed a criterion‐driven selection to design a validation prototype. Among the 310 distinct interfacial systems derived from ∼29,000 literatures, our MLFSS model uniquely singles out LVCS as the premier candidate exhibiting a near‐ideal interfacial electric potential distribution. Consequently, although this material was originally constructed in our previous work targeting high voltage attributes (Figure S11) [55], it is reidentified here solely by the MLFSS criterion as the optimal platform to rigorously validate our theoretical model. Unlike the previous study, which focused on cathode material design, here we construct a full‐cell ASSB to specifically verify the predicted ultralow interfacial resistance.

The LVCS is predicted to maintain exceptionally low interfacial resistances with high‐conductivity S‐SEs (${\stackrel{\sim }{R}}_{\mathrm{SCL}}$< 0.1), specifically achieving a normalized resistance (${\stackrel{\sim }{R}}_{\mathrm{SCL}}$) of 0.0015 with LGPS (Table S4). This value represents a near‐ideal interface, starkly contrasting with the high resistance of conventional oxide systems. Subsequent DFT interface modeling elucidates the atomistic origin of this prediction (Figures S12–S14 and Table S11). Both electrostatic potential reduction (from −2.617 to −6.112 eV, Figure S13a,b) and Li traveling distance (< 0.5 Å, Figure S14a,b) in the outermost atomic layer of LGPS in LVCS/LGPS are much smaller than those in the widely studied LCO/LGPS system (Figures S13c,d and S14c,d), indicating a uniform Li distribution at the optimized LVCS/LGPS interface. Furthermore, the Li vacancy formation energies, E _v[Li] (Section S3), in LGPS exhibit a smooth gradient from the interface to the bulk. The values range from 2.052 to 2.605 eV at the interface, which are close to those in the LVCS bulk (2.031–2.386 eV, Figure 4d). This seamless energy landscape indicates the absence of thermodynamic barriers between LVCS and LGPS, facilitating rapid ion migration across the boundary.

FIGURE 4.

Designing the LVCS/LGPS combination for optimized interfacial resistance. (a) XRD pattern and (b) morphology and EDS elemental mapping images of the VCS particle sample. (c) Galvanostatic charge‐discharge profiles of ASSBs using LVCS and LGPS cycled at 25 mA g⁻¹ (1.0 V cutoff to prevent conversion). (d) Distribution of Li vacancy formation energies (E _V[Li]) for LVCS/LGPS and LCO/LGPS interfaces. Red regions indicate Li accumulation at the cathode side. (e) Nyquist plots of LVCS and LCO at 25 mA g⁻¹ after the first and 50th cycles; the inset shows the equivalent circuit model. (f) Cycling stability of the ASSB at 100 mA g⁻¹. (g) Literature comparison of average discharge voltage, specific capacity (based on active material), and energy density for ion‐intercalated chalcogenide cathodes in lithium‐ion batteries (circulars) and ASSBs (diamonds).

Experimental validation confirms the accuracy of our theoretical design. The LVCS cathode is synthesized by a solid‐phase method followed by a heating process (Experimental Section). The characteristic peaks in the X‐ray diffraction (XRD) pattern (Figure 4a) and the energy‐dispersive X‐ray spectroscopy (EDS) mapping results (Figure 4b) confirm the crystal structure and chemical composition. The all‐sulfide ASSB prototype employing the LVCS cathode and LGPS electrolyte delivers a stable reversible capacity of more than 220 mAh g⁻¹ over 50 cycles at 25 mA g⁻¹, as well as 190 mAh g⁻¹ over 150 cycles at 100 mA g⁻¹, confirming that the highly matched MLFSS parameters at the LVCS/LGPS interface impart intrinsic stability during cycling (Figures S15 and 4f). Moreover, it possesses considerable rate capability, such that when the current density increased to 500 mA g⁻¹, the discharge capacity maintained around 160 mAh g⁻¹, approximately two‐thirds of its capacity at 25 mA g⁻¹. One key reason is the intrinsic matched MLFSS between LVCS and LGPS, which minimizes the SCL‐driven Li⁺ depletion at the interface and facilitates rapid ion migration even under high‐current conditions (Figure S16).

Electrochemical impedance spectroscopy (EIS) at various cycling states further highlights the superior interfacial stability. As depicted in the Nyquist plots (Figure 4e), the ASSB employing the LVCS cathode exhibits lower initially bulk and interfacial resistances as compared to the conventional LCO‐based system. Corresponding DRT (distribution of relaxation times) curves also demonstrate relatively weak relaxation peaks across the entire frequency range (Figure S17). Its interfacial resistance of only 8.8 Ω cm² (355.6 Ω cm² for the LCO/LGPS system) should be a hallmark of intrinsically suppressed SCL effect as a consequence of modest electronic structure disparity (|ΔMLFSS| < 0.2 eV). After 50 cycles, the LVCS/LGPS interface maintains significantly lower resistance increments, indicating favorable and stable interfacial transfer processes, whereas the LCO/LGPS system suffers from severe resistance escalation. Relevant differences in ionic and electronic transport capabilities are also reflected by the above rate capability and cycling tests. This stark contrast underscores the exceptional interfacial compatibility and structural integrity of the LVCS/LGPS combination, which effectively mitigates space‐charge layer effects and interfacial degradation during cycling.

Compared to the pristine electrode particles, SEM imaging of the cycled sample still maintains the original microstructure, indicating its good structural stability during the charge/discharge process (Figure S18). Most significantly, the p‐type alloying strategy boosts the operating voltage to >2.6 V (Figures 4c and S19). Combined with the high capacity of the ion‐intercalation reaction (251 mAh g⁻¹, Section S4), this enables the VCS/LGPS/75% Li₂S‐24% P₂S₅–1% P₂O₅/Li ASSB to achieve an active‐material energy density of 562 Wh kg⁻¹. Although a conservative 45 wt% active material loading was employed here to ensure sufficient ionic percolation for strictly validating the interfacial resistance model, this active‐material‐level energy density is already competitive with commercial oxide cathodes (≈550 Wh kg⁻¹) and surpasses state‐of‐the‐art ion‐intercalated chalcogenide systems by over 20% (Figure 4g). While realizing high loadings (>80%) requires further electrode engineering (e.g., binder optimization) beyond the scope of this mechanistic study, the high intrinsic energy density of LVCS provides a solid material foundation for future high‐energy ASSBs.

3. Conclusions and Discussion

Our numerical procedure, integrating the SCL model [44, 45, 46, 48] with the ligand‐field‐derived MLFSS descriptor, successfully bridges the longstanding gap in ASSB interface modeling. By mapping the interfacial electric potential difference ($\mathrm{\Delta }{\varphi }^0$) directly from the electronic structure, we achieve resistance predictions that agree with experimental values within two orders of magnitude. This consistency across a diverse dataset of 310 interfaces validates our fundamental premise that local electronic reconstruction rather than bulk band alignment dictates solid‐solid interfacial electrochemistry. The technique is also generalizable to the SE/anode interface, but two complications place this task beyond the scope of the present work. (i) The physical contact morphology of the anode/SE interface and the existence of diverse side reactions have led to controversy in determining the SCL effect and its influence on interfacial resistance [56]. (ii) It is a great challenge to determine the MLFSS for the non‐TM anodes, such as Li or Li alloys, since they are composed of metallic bonds rather than TM‐centered ligand‐field structures.

It should be emphasized that our numerical procedure predicts the intrinsic SCL resistance, representing the theoretical lower limit of interfacial impedance governed by fundamental thermodynamics. In practical ASSBs, variations in Li concentration driven by interfacial chemical reactions inevitably induce local lattice distortions and structural changes, which introduce additional resistance. To account for this superimposed effect, we systematically evaluated the interfacial reaction energies across different combinations (Figure S20 and Table S6). For instance, the higher reaction driving force of the LiCoO₂/LGPS interface (−0.350 meV/atom) leads to a larger resistance increment (28.6%) compared to the highly compatible LVCS/LGPS interface (−0.078 meV/atom, 14.5% increment, Table S7). Crucially, while these reaction‐induced structural changes superimpose additional resistance to varying degrees, they do not alter the overarching resistance trends across different material classes. The MLFSS‐driven SCL effect remains the persistent and governing baseline that dictates the fundamental lower bound of interfacial resistance for interface combinations.

Besides, to guide the practical application of this model as a rational design tool, it is crucial to delineate its predictive fidelity on a per‐material‐class basis, as the residual prediction error is inherently coupled with interfacial chemical reactivity (Figure S20 and Table S6). For chemically compatible interface families, such as OCAs with halide SEs (e.g., LiCoO₂/Li₃YCl₆, ΔE[c_a , c_b ] = −0.023 meV/atom) or SCAs with sulfide SEs (e.g., Li₂V_0.5Cr_1.5S₄/LGPS, ΔE[c_a , c_b ] = −0.078 meV/atom), interfacial structural distortions are minimal. In these systems, the experimental resistance is overwhelmingly dominated by SCL depletion, rendering our MLFSS framework strictly quantitative (predicted values typically align within one order of magnitude of experiments, Table S5). Conversely, for highly dissimilar families like OCAs with sulfide SEs (e.g., NCM811/LGPS and LiCoO₂/Li₃PS₄, exhibiting massive reaction energies of −0.284 and −0.416 meV/atom, respectively), severe chemical reactions superimpose substantial additional resistance. For these incompatible interfaces, the MLFSS framework is trend‐correct; it establishes a rigorous comparative baseline and theoretical lower bound, while the superimposed reaction‐induced impedance accounts for the residual deviation of up to two orders of magnitude. This per‐material‐class classification empowers researchers to discern where the model offers direct quantitative metrics versus foundational trend screening.

While the current model captures the decisive trends and resolves the oxide paradox, absolute quantitative precision can be further refined by addressing complexities beyond the current idealized parameterization. Future iterations of this procedure should focus on three frontiers: (i) Electronic Theory Expansion. The current MLFSS is derived from a point charge model focusing on TM‐ligand electrostatic interactions. Integrating molecular orbital theory would enhance accuracy for systems with significant covalent characteristics (e.g., Li‐metal anodes with metallic bonding), which remain challenging for standard ligand‐field descriptions. (ii) Geometric Fidelity. The current SCL model does not account for the complex, inhomogeneous geometry of real interfaces, which significantly affects resistance [57]. (iii) Electrochemical‐Mechanical Coupling. Coupling the model with mechanical effects, such as stress induced by electrochemical cycling [58], would enable more accurate prediction of carrier redistribution. Notably, although the calculated space charge layer resistance currently differs in absolute magnitude from experimental values, primarily due to idealized parameterization, it consistently predicts correct trends across different material systems. This validates the model's fundamental mechanistic insight. Building a more precise materials parameter database and incorporating the aforementioned factors will be pivotal for achieving quantitatively accurate predictions of interfacial resistance.

In conclusion, our analysis uncovers a fundamental interfacial resistance contradiction in conventional designs: Oxides (OCAs) suffer from high resistance due to low MLFSS, while coatings (OCOs), despite low resistance, compromise energy density due to poor ionic conductivity. Based on our integrated numerical procedure, we propose two distinct strategies to resolve this trade‐off, utilizing the tunable criterion to design cathodes with operatively optimized resistances (Figure 5): (i) Strategy I: High‐Capacity Li‐Free Sulfide Cathodes (LF‐SCAs). This is an intrinsic solution. LF‐SCAs naturally possess a higher average MLFSS (−0.222 eV, Figure 3a) compared to oxides (−0.585 eV) due to the unique d‐p hybridization of TM‐S bonds [55]. This intrinsic property ensures low interfacial resistance without sacrificial coatings. Furthermore, their ability to undergo full (de)intercalation and conversion reactions unlocks high capacities, as demonstrated by our model‐validated LVCS prototype (Figure 4c). (ii) Strategy II: High‐Voltage Multi‐Functional Oxide Cathodes (MF‐OCAs). This is an electronic structure engineering solution driven by the tunable criterion. The goal is to design oxides that mimic the high‐MLFSS characteristic of insulators while maintaining electrochemical activity. A prime example identified by our model is LiFePO₄. Its large band gap (>3.5 eV) results in an exceptionally high MLFSS (2.249 eV, Figure 3a), yielding the lowest ${\stackrel{\sim }{R}}_{\mathrm{SCL}}$ (<0.140) among oxides (Figure 3b and Table S4). By targeting MF‐OCAs with similar high‐MLFSS/high‐voltage traits and pairing them with conductive carbon networks to offset electronic limitations [59], we can achieve high‐voltage operation with minimal SCL losses. Importantly, the physical logic linking wide band gaps, high MLFSS, and suppressed SCL resistance demonstrates excellent transferability far beyond conventional layered and spinel oxides, extending robustly to diverse structural families. Representative calculations on polyanionic and rocksalt wide‐bandgap cathodes, such as Li₂MnP₂O₇, Li₂FePO₄F, Li₂CoPO₄F, rocksalt‐Li₃NbO₄, reveal consistently high MLFSS values (e.g., 2.681 to 3.505 eV, Table S8) dictated by their insulating nature (λ = 1). Consequently, their predicted SCL resistances against sulfide electrolytes (e.g., LGPS) are remarkably low (Table S9), mirroring the LiFePO₄ paradigm. Conversely, for candidates like LiVPO₄F, where more delocalized V‐3d orbitals result in a noticeably narrower bandgap despite fluorination, the rigorous assignment of λ = 0 correctly predicts a negative MLFSS (−1.016 eV) and a highly resistive interface. This contrast not only confirms the broad structural transferability of the MLFSS framework but also reinforces the necessity of accurately evaluating the intrinsic electronic boundaries of emerging high‐voltage materials to unlock rational interface design. Because elevated voltage plateaus are fundamentally associated with profound orbital splitting, which often inherently widens the bandgap toward the insulating regime, a rigorous experimental or theoretical assessment of the intrinsic electronic conductivity is required to judiciously assign λ and prevent spurious SCL resistance estimations. These two strategies provide routes for designing cathode/SE combinations with low interfacial resistances and high energy densities, which are usually mutually exclusive in existing systems (Figure S21).

FIGURE 5.

Two strategies for designing cathode/SE combinations with low interfacial resistance and high energy density (capacity or voltage). The schematic illustrates replacing conventional OCAs with high‐capacity Li‐free sulfide cathodes (Strategy I) or high voltage multi‐functional oxide cathodes (Strategy II). The diagram highlights the intrinsic contradiction in current oxide systems: OCAs suffer from high interfacial resistance due to Li accumulation (low interfacial ionic conductivity), whereas OCOs benefit from low interfacial resistance but are limited by poor intrinsic ionic conductivity.

4. Experimental Section

4.1. Synthesis of LVCS Samples

LVCS material was synthesized by a two‐stage solid‐phase grinding method followed by heat treatment under vacuum. Specifically, V (0.01 mol, 99.99%, Acros), Cr (0.03 mol, 99.9%, Acros), and S (0.08 mol, 99%, Aladdin) were mixed and ground for 6 h. Then the mixture was sealed in a quartz tube under vacuum, heated slowly to 650°C, held for 24 h, and finally cooled to room temperature naturally to obtain V_0.5Cr_1.5S₄ (VCS). Subsequently, the above VCS powders were mixed with 0.027 mol Li₂S (99%, Innochem) and appropriate amounts of V and Cr; the previous synthetic steps were repeated to obtain LVCS.

4.2. Materials Characterizations

The phase structures of the as‐synthesized samples were characterized using an X‐ray diffractometer (XRD, Panalytical Empyrean) with Cu Kα radiation at a voltage of 60 kV. The morphologies were studied using a scanning electron microscope (SEM, Hitachi/SU8230) at a voltage of 5 kV with an Oxford XMaxN energy‐dispersive X‐ray spectroscopy (EDX) detector.

4.3. Assembly of ASSB

To increase the contact area and ensure an intimate contact interface between electrolyte and active materials, Li₇P₃S₁₁ electrolyte particles were anchored on lithiated LVCS by an in situ liquid‐phase approach, which can be found in previous work [60]. LVCS‐Li₇P₃S₁₁ as the active material, Li₁₀GeP₂S₁₂ (LGPS) as the solid electrolyte, and Super P as a conductive agent were mixed homogeneously by manual grinding in an optimal weight ratio of 45:50:5. This ratio was specifically selected to ensure sufficient ionic percolation within the cathode composite, thereby eliminating kinetic limitations and allowing for a rigorous validation of the interfacial resistance model. Bilayer solid electrolytes consisting of 100 mg of LGPS and 50 mg of 75% Li₂S‐24% P₂S₅–1% P₂O₅ were pressed together in a poly(tetrafluoroethylene) mold under 240 MPa successively. Then, the above‐mentioned composite cathodes were distributed on the LGPS side uniformly under 240 MPa. Finally, a metal lithium foil with a diameter of 10 mm was pressed on the opposite side under 360 MPa. The all‐solid‐state battery was secured by nuts and screws and tested without applying additional pressure.

4.4. Electrochemical Performance Measurements

The testing of charge/discharge and cycling performance was carried out on a multichannel battery test system (LAND CT‐2001A; Wuhan Rambo Testing Equipment Co., Ltd.) in the voltage range of 1.0–3.0 V (vs. Li⁺/Li) at various scan rates. The mechanism of capacity change was predicted via electrochemical impedance spectroscopy (EIS) measurements, which were conducted between 10⁶ and 10 Hz with an amplitude of 15 mV. The obtained Nyquist plots mainly consisted of a compressed semicircle in the middle‐frequency region, which corresponded to the charge transfer resistance (R _ct), and a sloped line in the low frequency region denoted Warburg resistance, related to Li‐ion diffusion into the bulk electrode. R _ct mainly stemmed from the interfacial resistance between LVCS (or LCO) and LGPS in the cathode layer [61, 62]. The Z′‐intercept in the high‐frequency range was ascribed to the Ohmic resistance (R _e), which stemmed from the resistance of the electrode and solid electrolyte layers. All test processes were carried out in a dry Ar‐filled glovebox at ambient temperature.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma74207‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.