Facile Formation of Oxygen‐Vacancy Gradient Enables In Situ Uniform Prelithiation in Vanadium Oxide Thin‐Film Batteries

ABSTRACT

Vanadium oxide (VO_x) is a promising cathode material for thin‐film all‐solid‐state lithium‐ion batteries (TF‐ASSLIBs) owing to its high theoretical capacity and compatibility with microelectronic fabrication. However, its lithium‐free nature results in low initial Coulombic efficiency and poor cycling stability, necessitating an effective prelithiation strategy. Conventional methods are limited by sluggish lithium diffusion kinetics, causing inhomogeneous lithium distribution with surface accumulation and insufficient bulk penetration. Herein, we engineer an oxygen vacancy (O_V) gradient in VO_x, with O_V concentration increasing from surface to interior, to homogenize lithium distribution through two synergistic effects: (i) O_V creates additional Li⁺ diffusion pathways, accelerating prelithiation kinetics; (ii) O_V gradient progressively lowers diffusion barriers, enabling deep Li⁺ penetration and uniform distribution. Notably, this O_V gradient forms spontaneously during VO_x deposition on Pt current collectors via moderate oxygen adsorption of Pt, which generates a higher O_V concentration near the Pt interface, thereby facilitating uniform and efficient prelithiation during subsequent in situ electrolyte deposition. TF‐ASSLIBs comprising an O_V‐gradient VO_x cathode|LiPON electrolyte|NiO anode exhibit substantially enhanced electrochemical performance over control devices, delivering higher initial Coulombic efficiency (76.5% vs. 43.5%), superior areal capacity (37.5 vs. 6.5 µAh cm⁻²), and better cycling stability (91.5% vs. 73.4% capacity retention@2000 cycles). This work provides a facile and scalable strategy for developing high‐performance cathodes and TF‐ASSLIBs.

Oxygen vacancy gradients, enabled by current‐collector‐driven engineering, facilitates uniform, efficient cathode prelithiation, boosting key electrochemical metrics of thin‐film all‐solid‐state lithium‐ion batteries.

1. Introduction

Thin‐film all‐solid‐state lithium‐ion batteries (TF‐ASSLIBs) are promising power sources for miniaturized electronics [1, 2, 3, 4, 5]. Vanadium oxide (VO_x) is an attractive cathode material due to its high theoretical capacity and compatibility with microelectronic fabrication [6, 7, 8, 9, 10, 11]. However, its intrinsic lithium‐free nature inevitably leads to low initial Coulombic efficiency (ICE), low specific capacity, and poor cycling stability [12, 13, 14]. Consequently, prelithiation is critical to mitigate these issues and enhance electrochemical performance.

Electrochemical prelithiation is a conventional approach for VO_x film [12, 15, 16]. A half‐cell configuration with VO_x cathode, metallic Li anode, and liquid electrolyte is employed in this method [12, 15, 16]. Lithium transports from the Li anode into the VO_x cathode through discharge operation. After prelithiation, the half‐cell is disassembled, and the cathode is retrieved for subsequent full TF‐ASSLIB preparation. While this method can deliver sufficient lithium into VO_x, it suffers from low efficiency and poor scalability. Moreover, the high electrochemical activity of VO_x triggers excessive solid electrolyte interphase (SEI) formation and severe interfacial polarization, further degrading battery performance [12, 17].

As an alternative, in situ prelithiation via subsequent electrolyte deposition during TF‐ASSLIB fabrication offers high efficiency and seamless process integration since prelithiation is performed without interrupting device preparation [18, 19, 20, 21]. However, limited lithium diffusion kinetics result in inefficient and inhomogeneous distribution, with lithium predominantly accumulating at the cathode surface and failing to penetrate the bulk [14, 22, 23, 24, 25]. This surface confinement hinders full cathode utilization and degrades performance. Furthermore, the resultant inhomogeneity induces localized stress gradient during cycling, accelerating capacity fade [9, 22, 23, 26]. While strategies such as nanostructuring and conductive network design have been explored to overcome diffusion limitations [12, 13, 27, 28], their complexity and incompatibility with microfabrication processes hinder practical adoption in TF‐ASSLIBs.

Amorphous VO_x possesses a three‐dimensional disordered network where Li⁺ transport relies on percolation through interconnected free‐volume pockets rather than the ordered crystallographic channels occuring in the crystalline counterparts [29, 30, 31]. Within this disordered structure, O_Vs generate additional free volume through local structural relaxation, effectively expanding voids and creating isotropic diffusion pathways with reduced activation barriers. This mechanism significantly enhances Li⁺ diffusion kinetics and substantially improves prelithiation efficiency [30, 31, 32]. In this article, we engineered an O_V gradient in VO_x with O_V concentration increasing from surface to interior, facilitating lithium penetration and improving distribution homogeneity. Notably, this gradient spontaneously forms during VO_x deposition on Pt current collectors by exploiting Pt's moderate oxygen adsorption capability, which creates higher O_V concentration near the Pt interface and lower concentration far away from the current collector. This enables uniform and efficient prelithiation of VO_x films during subsequent solid‐state electrolyte deposition. It is demonstrated that TF‐ASSLIBs incorporating this O_V‐gradient cathode exhibit substantially enhanced performance over control devices. This work provides a facile and effective strategy to develop high‐performance cathodes and TF‐ASSLIBs.

2. Results and Discussion

2.1. O_V Gradient for Homogeneous Prelithiation

To address the inhomogeneity issues while preserving the microelectronics‐compatible merit of in situ prelithiation via subsequent electrolyte deposition, we propose O_V gradient engineering in the cathode. This strategy enables high‐density, homogeneous prelithiation through two synergistic effects: (i) increasing O_V concentration provides additional Li diffusion pathways, accelerating Li⁺ diffusion and prelithiation kinetics; (ii) O_V gradient progressively lowers diffusion barriers, enabling long‐range Li⁺ penetration and ensuring homogeneity.

2.2. Current‐Collector‐Driven O_V Gradient Engineering

A key insight is that oxygen adsorption varies significantly among current collector materials, offering a means to modulate O_V distribution in VO_x films. For instance, Pt current collectors exhibit high oxygen adsorption, whereas indium tin oxide (ITO) lacks this capability [33, 34] (Figure 1a). During VO_x deposition, Pt extracts oxygen from the growing film, creating enhanced O_V concentration near the interface. This effect attenuates progressively with increasing distance from the current collector, thus producing an O_V gradient distribution (Figure 1b) that contrasts sharply with the uniform distribution obtained on ITO. The resulting O_V gradient facilitates Li⁺ diffusion during lithium phosphorus oxynitride (LiPON) deposition, thereby enhancing prelithiation efficiency and homogeneity (Figure 1c).

FIGURE 1.

Schematic illustration of current‐collector‐driven O_V gradient engineering for enabling in situ cathode prelithiation. (a) ITO and Pt current collectors with different oxygen adsorption capability. (b) Strong oxygen adsorption capability of Pt current collector induces generation of O_V gradient distribution in VO_x film. (c) O_V gradient distribution enables high‐efficiency and homogenous prelithiation in VO_x.

Guided by this strategy, we prepared pristine and O_V‐gradient‐engineered VO_x cathodes using ITO and Pt current collectors, respectively. TF‐ASSLIBs configured as Pt or ITO | VO_x | LiPON | NiO | ITO were fabricated via physical vapor deposition. For clarity, devices with Pt and ITO bottom current collectors are designated LIB@Pt and LIB@ITO, respectively.

2.3. Characterization of O_V Gradient in VO_x

To determine the chemical states and V/O stoichiometry, quantitative x‐ray photoelectron spectroscopy (XPS) depth profiling is performed on 140‐nm VO_x films deposited on ITO (pristine) and Pt (engineered) current collectors. As shown in Figure S1, the pristine film exhibits predominantly V⁵⁺ states throughout the thickness (Figure S1a) [35]. In contrast, the engineered film displays mixed V⁵⁺/V⁴⁺ valence states [35, 36], and the V⁵⁺ content decreases while the V⁴⁺ content increases progressively from the surface toward the Pt interface (Figure S1b).

Table S1 summaries the V/O ratio under various etching depths. The pristine VO_x film maintains a nearly constant V/O atomic ratio of ∼0.4 across the entire depth, consistent with stoichiometric V₂O₅. In contrast, the engineered film exhibits a continuous gradient with V/O increasing from 0.415 at the surface to 0.463 at 100 nm depth. Notably, the V/O ratio (0.463) at 100 nm approaches the theoretical value for V₂O₄ rather than V₂O₅. This confirms the progressive reduction from V₂O₅ toward V₂O₄ with depth.

To investigate the influence of the current collector on O_V distribution, O 1s XPS spectra are acquired at various etching depths for both pristine and engineered VO_x films (Figure 2a,b). The spectra are deconvoluted into two characteristic components: O_V (∼531.7 eV) and lattice oxygen (O_L, ∼530.3 eV) [37]. For the pristine film, the O_V/O_L ratio remains nearly constant (∼0.14) throughout the entire depth range, indicating homogeneous oxygen stoichiometry. In contrast, the engineered film exhibits a pronounced depth‐dependent gradient, with the O_V/O_L ratio increasing from 0.26 at the surface to 0.33 near the Pt interface. These results confirm that the Pt current collector promotes oxygen extraction from the VO_x film, generating a steep O_V gradient with the highest vacancy concentration at the Pt/VO_x interface.

FIGURE 2.

O 1s XPS spectra of VO_x films at different etching depths. (a) Pristine and (b) engineered VO_x.

Direct current (DC) polarization measurements (Figure S2) demonstrate that the engineered VO_x film exhibits an electronic conductivity (σ_e) of ∼2.0 × 10⁻³ S cm⁻¹, which is two orders of magnitude higher than the pristine VO_x (∼6.3 × 10⁻⁵ S cm⁻¹). Electrochemical impedance spectroscopy (EIS, Figure S3a) together with the fitting results (Figure S3b and Table S2) further confirms improved charge transfer kinetics, as evidenced by reduced charge‐transfer resistance (∼70 vs. 500 Ω). In addition, galvanostatic intermittent titration technique (GITT) measurements suggest that the engineered VO_x film exhibits higher Li⁺ diffusion coefficient than the pristine one (Figure S4). Thus, the O_V gradient significantly enhances both electrical and ionic conductivities, which are highly desirable for subsequent prelithiation. Benefiting from these improved kinetics, the engineered VO_x film exhibits enhanced electrochemical performance than the pristine one under the same operating conditions, including higher reversible capacity (∼210.4 vs. 141.9 mAh g⁻¹ at 0.3 C), improved rate capability (∼128.0 vs. 68.9 mAh g⁻¹ at 3.0 C), and superior cycling stability (∼95.6% vs. 89.9% capacity retention over 3000 cycles, Figure S5). It should be noted that the non‐stoichiometric nature of the engineered VO_x film results in a relatively lower potential compared with stoichiometric V₂O₅ [38].

Relative to the pristine counterpart, the engineered film exhibits much higher electronic conductivity (Figure S2) and Li⁺ diffusivity (Figure S4). This substantial improvement originates from two synergistic mechanisms: (i) Defect‐enhanced transport: O_Vs function as intrinsic n‐type dopants, substantially increasing free‐carrier concentrations and boosting electronic conductivity [39]. Additionally, these vacancies lower the activation energy for charge transfer across the electrode/electrolyte interface, thereby reducing charge‐transfer resistance and accelerating interfacial kinetics [40]. (ii) Phase‐enhanced conduction: Metallic VO₂ phases formed within the engineered VO_x matrix possess conductivity orders of magnitude higher than stoichiometric V₂O₅, drastically enhancing electronic transport while further reducing interfacial impedance [41, 42]. Collectively, O_Vs and metallic VO₂ phases cooperate to enhance both bulk and interfacial transport, with their combined effects yielding multiplicative gains.

2.4. Characterization of Prelithiation in VO_x

Cross‐sectional transmission electron microscopy (TEM) images of completed cells (Figures 3a–e and S6) reveal well‐defined layered structures with sharp interfaces in both LIB@ITO and LIB@Pt devices. High‐resolution TEM (Figure 3b–d,f–h) and selected‐area electron diffraction (SAED) demonstrate distinct structural evolution between pristine and engineered VO_x cathodes. Pristine VO_x remains amorphous after LiPON deposition, which is consistent with the x‐ray diffraction (XRD, Figure S7) and SAED (Figure S8) observations. In contrast, the engineered VO_x film exhibits metallic vanadium diffraction rings [43], indicating a phase transformation driven by facilitated lithiation kinetics. During LiPON deposition, O_Vs gradient serve as preferential Li⁺ diffusion channels, leading to metallic vanadium formation via the following reaction:

$$2\mathrm{x}{\text{Li}}^{+}+2\mathrm{x}{\mathrm{e}}^{-}+{\text{VO}}_{\mathrm{x}}\to \mathrm{V}+\mathrm{x}{\text{Li}}_2\mathrm{O}$$

FIGURE 3.

TEM and SAED analysis of VO_x in LIB@ITO and LIB@Pt devices. (a) Cross‐sectional TEM image of the LIB@ITO device. (b) TEM image of the LiPON/VO_x/ITO layer in LIB@ITO device. HRTEM images as well as SAED patterns of the VO_x layer at the (c) LiPON/VO_x interface and (d) VO_x/ITO interface. (e) Cross‐sectional TEM image of the LIB@Pt device. (f) TEM image of LiPON/VO_x/Pt layer in LIB@Pt device. HRTEM images as well as SAED patterns of the VO_x layer at the (g) LiPON/VO_x interface and (h) VO_x/Pt interface.

XPS analysis in Figure S9 further confirms this transformation. Notably, metallic vanadium can be observed at the Pt/VO_x interface (Figure 3h), indicating lithium diffusion to the current collector. This provides clear evidence that O_V gradient engineering effectively enhances Li diffusion dynamics. The in situ formed vanadium networks enhance electrical transport within the cathode [44, 45], thereby improving the battery rate performance.

Time‐of‐flight secondary ion mass spectrometry (TOF‐SIMS) is used to probe elemental distribution profiles. Three‐dimensional elemental mapping reveals distinct spatial distributions of V and O between the LIB@ITO and LIB@Pt devices (Figures 4a,b,d,e and S10). In LIB@ITO, the VO_x film shows homogeneous oxygen distribution (Figures 4b and S10a), whereas LIB@Pt exhibits progressive oxygen depletion in the VO_x layer (Figures 4e and S10b). This agrees well with the XPS depth profiles as discussed above. Lithium mapping (Figure 4c–f) also reveals marked differences, with LIB@Pt showing significantly higher lithium content than LIB@ITO. This pronounced Li enrichment confirms superior prelithiation efficiency. The much higher Li content in the engineered VO_x, coupled with the ultralow electrochemical potential of lithium, results in a significantly lower open‐circuit voltage for LIB@Pt (∼ −0.59 V) than LIB@ITO (∼0.07 V, Figure S11) [46, 47].

FIGURE 4.

Spatial distribution of V, O, and Li in the LIB@ITO and LIB@Pt devices. 3D maps of (a) V, (b) O and (c) Li element for the LIB@ITO device and (d) V, (e) O, and (f) Li for the LIB@Pt device. Evolution of the V and Li contents along the LiPON/VO_x films on the (g) ITO and (h) Pt/ITO current collectors.

Figure 4g,h shows the ToF‐SIMS absolute depth profiles of Li and V across the VO_x layers for both the LIB@ITO and LIB@Pt devices. These results provide direct quantitative evidence of distinct lithium distribution behaviors. For the LIB@Pt device with engineered VO_x, the lithium signal intensity remains consistently high and nearly uniform throughout the entire VO_x layer, indicating unimpeded Li⁺ diffusion from the LiPON electrolyte and homogeneous distribution within the VO_x matrix. In stark contrast, the LIB@ITO device with pristine VO_x exhibits a pronounced lithium concentration gradient, with signal intensity decaying dramatically from the LiPON/VO_x interface toward the VO_x/ITO interface. Moreover, the overall lithium content is substantially lower than that in the engineered layer, directly confirming restricted Li⁺ transport and limited penetration depth within the pristine VO_x. XPS depth profiling further confirms these findings (Figure S12).

Quantitative ToF‐SIMS and XPS analyses collectively confirm that the oxygen vacancy gradient in the engineered VO_x effectively promotes uniform and deep lithium prelithiation through two synergistic mechanisms: (i) the elevated O_V concentration generates additional free volume and expands interconnected diffusion channels, accelerating Li⁺ transport kinetics [30, 31, 32]; (ii) the depth‐dependent gradient progressively lowers local diffusion barriers from the surface toward the Pt interface, enabling long‐range Li⁺ penetration and ensuring homogeneous distribution throughout the entire film thickness.

Based on quantitative analysis and characterization, approximately 1.3 × 10⁻⁷ mol cm⁻² of oxygen is transferred from the VO_x film to the Pt surface to establish the observed vacancy gradient, with an estimated oxygen penetration depth of ∼40 nm into the Pt layer (Figure S13). Further details are provided in the Supporting Information.

2.5. Electrochemical Performance of TF‐ASSLIBs

Figure 5a,b shows galvanostatic charge/discharge (GCD) voltage profiles of the LIB@ITO and LIB@Pt devices between 0.01–2.6 V at 1.0 µA cm⁻². The LIB@Pt device achieves an ICE of 76.5% that rapidly rises to 98.0% within three cycles, which substantially outperforms LIB@ITO (43.5% to 94.5%). Additionally, LIB@Pt maintains 76.0% of its initial discharge capacity (∼21.5 µAh cm⁻²) over the first three cycles, while LIB@ITO retains only 42.9% (∼3.6 µAh cm⁻²). The superior performance of LIB@Pt originates from sufficient, homogeneous prelithiation enabled by the O_V gradient. In contrast, LIB@ITO exhibits only trace and ununiform lithium incorporation (Figures 4g and S12) that fails to compensate for irreversible lithium loss, leading to low Coulombic efficiency and rapid capacity fade [48].

FIGURE 5.

Electrochemical performance of TF‐ASSLIBs. GCD curves of the (a) LIB@ITO and (b) LIB@Pt devices for the initial four cycles at a current density of 1.0 µA cm⁻². (c) GCD curves of the LIB@Pt device at different current densities. (d) Cycling performance of LIB@ITO and LIB@Pt devices at 6.0 µA cm⁻². (e) GCD curves of the thick‐electrode LIB@Pt device under different current densities. (f) Capacity ratio of thick‐electrode LIB@Pt/LIB@Pt and thick‐electrode LIB@ITO/LIB@ITO at the same current rate. (g) Cycling performance of thick‐electrode LIB@ITO and LIB@Pt devices at 6.0 µA cm⁻².

Figures S14 and 4c display the rate performance of the devices, respectively. As the current rate increases from 1.0 µA cm⁻² to 6.0 µA cm⁻², the discharging capacity decreases from 4.1 to 1.4 µAh·cm⁻² for the LIB@ITO device and from 20.5 to 11.8 µAh·cm⁻² for the LIB@Pt device, corresponding to 65.9% and 42.4% degradation. The LIB@Pt device displays better rate performance than the LIB@ITO one mainly ascribed to its faster reaction kinetics. This can be confirmed by the EIS spectra shown in Figure S15, where the LIB@Pt device shows smaller charge‐transfer resistance compared with the LIB@ITO device.

Figure 4d illustrates the cycling characteristics of the devices at 6.0 µA cm⁻². After 2000 cycles, the reversible capacity decreases from 1.3 to 1.0 µAh cm⁻² for the LIB@ITO device and from 12.5 to 10.8 µAh cm⁻² for the LIB@Pt device, corresponding to 73.4% and 91.5% capacity retention, respectively. The superior stability of LIB@Pt originates from its homogeneous lithium distributions, which stabilizes the cathode structure by accommodating volume strain during repeated lithiation/delithiation. Conversely, LIB@ITO suffers from inhomogeneous prelithiation, leading to localized stress accumulation and accelerated capacity fade [9, 22, 23, 26].

Although increasing electrode thickness in TF‐ASSLIBs improves areal capacity, it often degrades rate capability and cycling stability due to sluggish lithium transport kinetics [10, 12, 49, 50]. Our O_V gradient engineering strategy effectively mitigates this trade‐off. To demonstrate this advantage, we fabricated thick‐electrode TF‐ASSLIBs with cathode and anode thicknesses twice those of the standard devices. As shown in Figures S16 and 5e, the thick‐electrode LIB@Pt device delivers 37.5 µAh cm⁻² at 1.0 µA cm⁻² and retains 17.6 µAh cm⁻² at 6.0 µA cm⁻², corresponding to a 47.0% retention rate. In stark contrast, the LIB@ITO counterpart exhibits severe capacity loss, decreasing from 6.50 to 1.70 µAh cm⁻² (∼26.0% retention) under identical conditions. Moreover, the LIB@Pt device achieves nearly linear areal capacity scaling with electrode thickness (Figure 5f), indicating fast kinetics enabled by sufficient and homogeneous prelithiation.

The thick‐electrode LIB@Pt device also maintains excellent cycling stability (Figure 5g), retaining 91.1% capacity after 2000 cycles, which is comparable to the standard LIB@Pt (91.5% retention). Conversely, the thick‐electrode LIB@ITO device retains only 59.8% (1.84 to 1.10 µAh cm⁻²), significantly worse than its standard counterpart (73.4% retention). These results underscore that our O_V gradient engineering simultaneously achieves high areal capacity, rate capability, and cycling stability in thick‐electrode TF‐ASSLIBs, with performance metrics comparable to state‐of‐the‐art VO_x‐based devices [10, 11, 12, 20] (Figure S17 and Table S3).

2.6. Current Collector‐Dependent Prelithiation in VO_x

To investigate effects of the current‐collector materials on cathode prelithiation, Ti, Cr, and Cu are also evaluated alongside Pt and ITO discussed above. According to the literature, oxygen adsorption strength typically follows the order: Ti > Cr > Pt > Cu [51, 52, 53, 54]. Consequently, the relative O_V content at the VO_x/Ti interface is higher (∼82.1%) than that at the VO_x/Pt interface (∼35.3%), as characterized in Figure S18. However, despite the stronger oxygen adsorption capability of Ti (or Cr), their interaction with VO_x leads to formation of a compact and insulating interfacial oxide layer, as evidenced by the XPS data of Figure S19. This dense passivation layer obstructs subsequent oxygen diffusion from VO_x into the metal layer, thereby inhibiting sustained oxygen adsorption and limiting further O_V generation within the VO_x film. In contrast, Pt undergoes only partial oxidation owing to its chemical inertness, as demonstrated in Figure S20. The resulting Pt oxide is loosely structured and embedded within the Pt matrix, possessing limited capacity to impede oxygen diffusion. Thus, oxygen adsorption and attendant vacancy formation can proceed continuously. These different interfacial behaviors cause that the O_V content at the VO_x surface for the VO_x/Ti system is lower (∼17.8%) than that for the VO_x/Pt (∼26.0%), as quantified in Figure S21.

Furthermore, the insulating nature of the TiO_x interlayer and its blocking effect on oxygen extraction collectively contribute to inferior electronic conductivity in VO_x/Ti relative to VO_x/Pt (Figure S22). Notably, the VO_x/Cu system displays the lowest electronic conductivity among all the samples mainly ascribed to Cu's minimal oxygen adsorption affinity and consequently negligible vacancy induction. This poor conductivity is detrimental to the electrochemical performance of the corresponding battery cells as described latter.

Figure 6a shows the open‐circuit voltages (OCV) of TF‐ASSLIBs with various current collectors. As previously discussed in Figure S11, OCV closely correlates with prelithiation ability (and thus cathode's lithium content). The Pt‐based device exhibits the lowest OCV. As discussed above, Pt's moderate oxygen affinity prevents the formation of a compact insulating barriers between the VO_x/Pt interface, enabling the establishment of a controllable O_V gradient throughout the VO_x film, ensuring efficient ionic transport, and thereby exhibiting superior performance in the prelithiation process. Conversely, the Cu‐based device shows the highest OCV due to its inertness toward oxygen adsorption. This analysis opens up a new quantitative method to evaluate oxygen adsorption capability by using OCV values.

FIGURE 6.

Current collector‐dependent prelithiation in VO_x. (a) Open circuit voltage of all TF‐ASSLIBs. (b) Initial four‐cycle Coulombic efficiency of all TF‐ASSLIBs. (c) Specific capacity of all TF‐ASSLIBs under different current densities. (d) Rate performance of all TF‐ASSLIBs. (e) Cycling performance of all TF‐ASSLIBs at 6.0 µA cm⁻².

Figure 6b–e presents the electrochemical characteristics of TF‐ASSLIBs with various current collectors. The Pt‐ and Cu‐based devices almost exhibit the best and worst performance, respectively, in terms of specific capacity, ICE (∼76.5% for the Pt‐based one vs. 37.8% for the Cu‐based one), rate characteristics and cycling stability under the same operation conditions. This stark contrast, arising from their different prelithiation efficiencies, clearly demonstrates the dominant role of prelithiation in TF‐ASSLIB performance.

3. Conclusion

In summary, this work presents a facile, microelectronics‐compatible strategy to achieve efficient and homogeneous prelithiation in lithium‐free cathodes for high‐performance TF‐ASSLIBs. By leveraging oxygen adsorption capability of the current collector, a gradient distribution of high‐concentration oxygen vacancies spontaneously forms during cathode deposition. This tailored O_V architecture significantly enhances prelithiation efficiency and uniformity during subsequent in situ electrolyte deposition. Furthermore, prelithiation can be precisely tuned through current collector selection. This strategy resolves the intrinsic trade‐off between high specific capacity and long‐term cycling stability/rate capability in TF‐ASSLIBs, resulting in the improvement of the overall performance. By elucidating the critical role of current collectors in regulating prelithiation, this work provides a versatile and scalable approach for developing high‐performance TF‐ASSLIBs.

4. Materials and Methods

4.1. Device Fabrication

To investigate how bottom current collectors' oxygen adsorption capability influences the cathode prelithiation, VO_x films were deposited on Pt and ITO current collectors, designated as engineered and pristine VO_x cathodes, respectively. Electrochemical properties of both films were characterized based on 2016‐coin type half cells with VO_x‐coated Cu foil (pristine sample) or Pt/Cu foil (engineered sample) as cathode, metal Li as anode, 1.0 M LiPF₆ in ethylene carbonate (EC) and ethyl methyl carbonate (EMC), (EC/EMC = 50/50 (v/v), from Sigma‐Aldrich) as electrolyte. Full TF‐ASSLIBs with a configuration of Pt or ITO bottom current collector | VO_x cathode | LiPON electrolyte | NiO anode | ITO top current collector were prepared on the ITO‐coated substrate by using physical vapor deposition (Hex Deposition System) at room temperature. The low deposition temperature ensures that the as‐deposited VO_x film retains an amorphous structure. Furthermore, the degree of prelithiation depends predominantly on the intrinsic Li⁺ diffusivity of the VO_x film rather than on thermally activated diffusion during film growth. All sputtering targets used in this work have a uniform diameter of 2 inches. The detailed deposition conditions were as below: First, a 100 nm thick Pt film was deposited on the substrate by using Pt target under Ar = 30 sccm at a power of 75 W. Then, a 140 nm thick VO_x film was deposited as the cathode by using a V₂O₅ target under Ar/O₂ = 16 sccm/2 sccm at 85 W. After that, a 500 nm thick LiPON was deposited as the electrolyte by using a Li₃PO4 target under N₂ = 13.4 sccm at 130 W. After that, a 90 nm thick NiO film was deposited as the anode by using a NiO target under Ar = 41.5 sccm at 120 W. Finally, a 240 nm thick ITO film was prepared as the top current collector by using an ITO target under Ar = 41.5 sccm at 120 W. To investigate the capability of the O_V gradient in enhancing the kinetics, thick‐electrode TF‐ASSLIBs comprising a 280‐nm VO_x cathode and a 180‐nm NiO anode were also fabricated, while all other layers maintained the same thicknesses as described above. In order to study the effects of bottom current collectors on prelithiation, devices with different bottom current collectors (including Ti, Cr, and Cu) were also prepared by using identical processes as described above.

4.2. Material Characterization

Chemical states of the films were analyzed by XPS (Thermo Fisher Nexsa). Ar⁺ etching with an energy of 2000 eV and a current of 10.0 mA was performed to characterize the depth profiles. Crystallinity of the films was examined by XRD (Rigaku SmartLab3) and SAED (Titan). Film thickness was measured by spectroscopic ellipsometry (Horiba UVISEL) and confirmed by TEM (ARM‐200F) equipped with energy dispersive x‐ray spectroscopy (EDX). The detailed preparation of the TEM specimens was described in Supporting Information. TOF‐SIMS (IONTOF GmbH) depth profiles were collected using a 30 keV Bi₃ ⁺ primary ion source with 2 keV 100 nA Cs⁺ cluster sputtering.

4.3. Electrical and Electrochemical Characterization

The DC conductivity of the engineered VO_x and pristine V₂O₅ films were characterized based on DC polarization by using electrochemical workstation (CHI 660E). The EIS of the batteries was measured by using an electrochemical workstation (CHI 660E) over a frequency range of 10⁻²–10⁶ Hz with an AC amplitude of 5 mV. The electrochemical behaviors of the batteries were investigated by using GCD (Arbin LBT20084) tests. GITT test was performed a low current density of 0.2 C. Each galvanostatic pulse lasted for 15 min, followed by a 45 min open‐circuit relaxation to ensure complete voltage stabilization.

Author Contributions

Jian Wang: conceptualization, methodology, software, data curation, investigation, validation, formal analysis, writing – original draft. Xincheng Lei: data curation, investigation, validation, formal analysis, writing – original draft, software. Yonghe Zhuang: resources, data validation and formal analysis. Xinlong Xia: data curation, formal analysis. Dong Su: writing – review and editing, conceptualization, methodology, software, investigation, validation. Xiaodong Huang: writing – review and editing, funding acquisition, validation, writing – original draft, project administration, resources, conceptualization.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: anie72777‐sup‐0001‐SuppMat.docx.

Data Availability Statement

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