I have focused my investigation on the interfacial instability that limits the practical energy density of a solid state battery. A solid state battery that uses a sulfide solid electrolyte and a lithium metal anode promises high voltage, high capacity, and improved safety, yet the interface between the electrolyte and the lithium metal remains the weakest link. I hypothesized that a mixed ion-electron conductive interlayer, abbreviated as MIEC, could redistribute the local current density, homogenize the electric field, and thereby suppress lithium dendrite nucleation and growth. In this work, I prepared a series of composite interlayers by combining Li6PS5Cl sulfide electrolyte with different mass fractions of Super P carbon. I then evaluated the ionic conductivity, electronic conductivity, critical current density, interfacial impedance, symmetric cell cycling, and full cell performance. The results confirm that a 1% Super P interlayer provides the best balance between ionic and electronic transport, enabling a solid state battery with a critical current density of 1.6 mA·cm−2, a symmetric cell lifetime of 2800 h at 0.5 mA·cm−2, and a full cell capacity retention of 79.7% after 1000 cycles at 1C. I also provide mechanistic insight through X-ray diffraction, X-ray photoelectron spectroscopy, scanning electron microscopy, in-situ electrochemical impedance spectroscopy, and distribution of relaxation times analysis. The work demonstrates that interfacial engineering of a solid state battery can be guided by a simple percolation and charge-transfer framework.
The central problem I address is the growth of lithium dendrites within a solid state battery. Unlike a conventional liquid electrolyte, a sulfide solid electrolyte cannot flow to heal voids or cracks. When lithium plates unevenly, the local current density can exceed the exchange current density of the interface, and the resulting protrusions concentrate the electric field. This positive feedback leads to filamentary growth, fracture of the electrolyte, and eventually short circuit. A solid state battery that fails by this mechanism loses both capacity and safety. Therefore, I sought an interlayer that is chemically compatible with Li6PS5Cl and lithium metal, mechanically compliant, and electronically conductive enough to flatten the potential distribution but not so conductive that it promotes bulk lithium deposition inside the interlayer.
The architecture of the mixed conductive interlayer within the solid state battery is illustrated below.
I selected Li6PS5Cl as the baseline sulfide solid electrolyte because it exhibits high ionic conductivity, typically above 1 mS·cm−1, and good mechanical deformability. I selected Super P carbon as the electronic conductor because its nanostructured morphology can form a percolating network at low loading. The composite interlayer was prepared by manual grinding of Li6PS5Cl and Super P in an agate mortar for 30 min. I varied the Super P mass fraction from 0.5% to 1% to 3%, and I compared these with pristine Li6PS5Cl. The compositions are summarized in Table 1.
| Sample | Li6PS5Cl mass ratio | Super P mass ratio | Super P mass fraction / % |
|---|---|---|---|
| LPSCl | 100 | 0 | 0 |
| 0.5%SP | 100 | 0.5 | 0.5 |
| 1%SP | 100 | 1 | 1 |
| 3%SP | 100 | 3 | 3 |
For the solid state battery assembly, I first pressed the Li6PS5Cl electrolyte powder in a 10 mm die at 300 MPa for 1 min. I then added the composite interlayer on both sides and pressed at 500 MPa for 2 min. Finally, I attached lithium foil at low pressure, approximately 10 MPa, to form a symmetric cell. For full cells, I prepared a composite cathode by mixing LiNi0.89Co0.06Mn0.05O2, Li6PS5Cl, and Super P in a mass ratio of 70:27:3. I pressed the cathode composite onto one side of the electrolyte, added the interlayer on the anode side, and attached lithium foil. All handling was performed in an argon-filled glovebox to avoid moisture and oxygen contamination.
I used a battery test system for galvanostatic cycling. For full cells, I first activated the solid state battery at 0.1C between 2.5 V and 4.3 V for three cycles, then applied the desired rate. I measured electrochemical impedance spectroscopy over a frequency range from 1 Hz to 7 MHz with a 10 mV perturbation. I fitted the impedance data with equivalent circuits and also transformed the data into the distribution of relaxation times, abbreviated DRT, to separate overlapping processes. I characterized the crystal structure by X-ray diffraction from 10° to 90° at a scan rate of 12°·min−1. I examined morphology and elemental distribution by scanning electron microscopy with energy dispersive spectroscopy. I analyzed chemical states by X-ray photoelectron spectroscopy before and after cycling.
I begin the results with the crystal structure. The X-ray diffraction patterns of the composite interlayers match those of pristine Li6PS5Cl. I observe the characteristic diffraction peaks at 2θ = 25.7°, 30.2°, and 31.5°, which correspond to the argyrodite structure. I do not detect any additional peaks that would indicate a reaction product between Super P and Li6PS5Cl. This means that mechanical mixing and cold pressing do not destroy the ionic conduction framework. The peak positions and assignments are summarized in Table 2.
| 2θ / ° | Assignment | Observed in LPSCl | Observed in 1%SP |
|---|---|---|---|
| 25.7 | Argyrodite (111) | Yes | Yes |
| 30.2 | Argyrodite (220) | Yes | Yes |
| 31.5 | Argyrodite (311) | Yes | Yes |
| No extra peaks | No detectable side phase | Not applicable | Confirmed |
I next examined the microstructure. The cross-sectional scanning electron microscopy image of the 1%SP interlayer shows a dense composite with no large voids. The carbon signal in energy dispersive spectroscopy maps is uniform across the interlayer, which indicates that Super P does not agglomerate into large clusters. I attribute this uniformity to the grinding process, which breaks the carbon agglomerates and distributes them along the Li6PS5Cl particle surfaces. A uniform carbon network is essential because it ensures that the electronic conductivity is homogeneous, which in turn prevents local hot spots during lithium plating. In contrast, a nonuniform carbon network would create isolated conductive regions and force lithium to deposit preferentially at their edges.
I then quantified the ionic and electronic transport. The Nyquist plots of the composite interlayers show that the grain boundary resistance remains nearly unchanged when the Super P content is 0.5% or 1%. This is a key observation because it means that the ionic conduction pathway of the solid state battery is preserved. However, when the Super P content increases to 3%, I observe a pronounced additional semicircle, which I assign to a large interfacial resistance. This resistance arises because excess carbon interrupts the ionic contact between neighboring Li6PS5Cl particles. The electronic conductivity increases monotonically with Super P content. I calculated electronic conductivity from the slope of the current-voltage curve using the following relation:
$$ \sigma_e = \frac{L}{R_e A} $$
where \(L\) is the interlayer thickness, \(A\) is the electrode area, and \(R_e\) is the electronic resistance obtained from the linear current-voltage response. I also calculated ionic conductivity from the impedance data using:
$$ \sigma_{\text{ion}} = \frac{L}{R_{\text{ion}} A} $$
where \(R_{\text{ion}}\) is the ionic resistance. The measured values are summarized in Table 3. Pristine Li6PS5Cl has an electronic conductivity of 9.7×10−10 S·cm−1. The 0.5%SP composite has 8.25×10−9 S·cm−1, the 1%SP composite has 1.8×10−8 S·cm−1, and the 3%SP composite has 2.4×10−5 S·cm−1. The 1%SP value is high enough to dissipate local charge accumulation but low enough to avoid excessive lithium plating inside the interlayer. The 3%SP value is too high and leads to bulk deposition and dead lithium.
| Sample | Ionic conductivity / S·cm−1 | Electronic conductivity / S·cm−1 | Grain boundary resistance trend |
|---|---|---|---|
| LPSCl | ~1.0×10−3 | 9.7×10−10 | Reference |
| 0.5%SP | ~1.0×10−3 | 8.25×10−9 | Nearly unchanged |
| 1%SP | ~1.0×10−3 | 1.8×10−8 | Nearly unchanged |
| 3%SP | Reduced | 2.4×10−5 | Strongly increased |
The critical current density is a direct measure of the interfacial stability of a solid state battery. I define it as the maximum applied current density at which the symmetric cell voltage remains stable without a sudden drop that indicates short circuit. The voltage-time curves show that pristine Li6PS5Cl short circuits at a relatively low current density. The 0.5%SP interlayer offers only a modest improvement. The 1%SP interlayer raises the critical current density to 1.6 mA·cm−2. The 3%SP interlayer lowers the critical current density to 0.9 mA·cm−2 because the high electronic conductivity accelerates lithium deposition inside the interlayer rather than at the interface. The values are collected in Table 4.
| Interlayer | Critical current density / mA·cm−2 | Observed failure mode |
|---|---|---|
| LPSCl | ~0.7 | Short circuit after limited cycling |
| 0.5%SP | 1.1 | Gradual polarization increase |
| 1%SP | 1.6 | Stable voltage, no short circuit |
| 3%SP | 0.9 | Dead lithium and early failure |
I explain the critical current density trend with a simple charge balance argument. The local current density at the lithium metal interface is the sum of ionic and electronic contributions. When the electronic conductivity is too low, the ionic current cannot be compensated, and the electric field concentrates at protrusions. When the electronic conductivity is too high, the interlayer itself becomes a site for lithium reduction, which creates isolated lithium and voids. The optimum occurs when the electronic conductivity is several orders of magnitude lower than the ionic conductivity but still high enough to smooth the potential. I can express the relevant dimensionless group as:
$$ \Lambda = \frac{\sigma_e}{\sigma_{\text{ion}}} $$
For 1%SP, \(\Lambda\) is approximately \(1.8 \times 10^{-5}\), which I find is a favorable window. For 3%SP, \(\Lambda\) is \(2.4 \times 10^{-2}\), which is too large and disrupts the interface.
I further analyzed the interfacial kinetics using distribution of relaxation times. The impedance of a solid state battery contains contributions from grain boundaries, interfacial films, charge transfer, and diffusion. The DRT method separates these contributions by solving the following integral equation:
$$ Z(\omega) = R_\infty + \int_0^\infty \frac{\gamma(\tau)}{1 + j \omega \tau} \, d\tau $$
where \(\gamma(\tau)\) is the distribution function and \(\tau\) is the relaxation time. I assigned the peaks according to characteristic time scales. The peak in the range \(10^{-7}\) to \(10^{-5}\) s corresponds to grain boundary resistance. The peaks in the ranges \(10^{-5}\) to \(10^{-3}\) s and \(10^{-3}\) to \(10^{-1}\) s correspond to the solid electrolyte interphase and charge transfer resistance, respectively. The peak in the range \(10^{-1}\) to \(10\) s corresponds to diffusion. The assignments are listed in Table 5.
| Relaxation time range / s | Physical process | Effect of 1%SP interlayer |
|---|---|---|
| 10−7 to 10−5 | Grain boundary | No significant increase |
| 10−5 to 10−3 | Solid electrolyte interphase | Peak area reduced |
| 10−3 to 10−1 | Charge transfer | Peak area reduced |
| 10−1 to 10 | Diffusion | Peak area reduced |
The DRT results show that the 1%SP interlayer lowers the interfacial resistance and improves the reversibility of lithium plating and stripping. Before cycling, the DRT peaks for the 1%SP cell are smaller than those for the pristine cell in the intermediate and low frequency regions. After in-situ charge and discharge at 0.1 mA·cm−2, the pristine cell shows a clear increase in the grain boundary peak, which indicates that the electrolyte grain boundaries degrade during cycling. In contrast, the 1%SP cell maintains a low grain boundary peak and stable interfacial peaks. I conclude that the interlayer protects the solid state battery from interfacial and grain boundary degradation.
I also tested long-term symmetric cell cycling. At 0.5 mA·cm−2, the pristine Li6PS5Cl symmetric cell fails after only 170 h, as evidenced by a sudden drop in polarization voltage. The 1%SP symmetric cell cycles for 2800 h with a stable polarization voltage of approximately 25 mV. This is a remarkable improvement and directly demonstrates that the mixed conductive interlayer stabilizes the lithium metal interface in a solid state battery. The cycling data are summarized in Table 6.
| Interlayer | Current density / mA·cm−2 | Cycling time before failure / h | Polarization voltage / mV |
|---|---|---|---|
| LPSCl | 0.5 | 170 | Unstable |
| 1%SP | 0.5 | 2800 | ~25 |
I then moved to full cell performance to evaluate whether the interlayer benefits a practical solid state battery. The full cell with the 1%SP interlayer delivers a first-cycle discharge specific capacity of 203 mAh·g−1 at 0.1C. The pristine Li6PS5Cl full cell delivers only 176 mAh·g−1 under the same conditions. The Coulombic efficiency also improves from 81.69% to 83.45%. The 0.5%SP full cell delivers 185 mAh·g−1, and the 3%SP full cell delivers 176 mAh·g−1. The first-cycle data are shown in Table 7.
| Full cell interlayer | Charge specific capacity / mAh·g−1 | Discharge specific capacity / mAh·g−1 | First-cycle Coulombic efficiency / % |
|---|---|---|---|
| LPSCl | 216 | 176 | 81.69 |
| 0.5%SP | 225 | 185 | 82.22 |
| 1%SP | 243 | 203 | 83.45 |
| 3%SP | 213 | 176 | 82.35 |
The dQ/dV curves provide additional insight. The 1%SP cell exhibits sharper and more intense redox peaks than the pristine cell. This indicates reduced polarization and improved reaction kinetics. I attribute the improvement to the mixed conductive interlayer, which provides a continuous electronic path and a continuous ionic path at the same time. The electronic path ensures that electrons reach the reaction sites uniformly, while the ionic path ensures that lithium ions are supplied uniformly. This dual path suppresses the formation of isolated lithium and reduces the interfacial resistance.
I evaluated the long-term cycling stability of the solid state battery with the 1%SP interlayer. At 1C, the full cell delivers an initial discharge capacity of 150 mAh·g−1. After 1000 cycles, the capacity retention is 79.7%. After 2000 cycles, the capacity retention is still 68.3%. I did not observe any short circuit. I calculated capacity retention using:
$$ \text{Retention} = \frac{Q_N}{Q_1} \times 100\% $$
where \(Q_1\) is the first-cycle discharge capacity and \(Q_N\) is the discharge capacity at cycle \(N\). The long-term cycling results are summarized in Table 8. I also tested a high-loading solid state battery with a cathode active material loading of 20.32 mg. At 0.3C, the high-loading cell maintains its capacity for 250 cycles with almost no decay. This demonstrates that the interlayer is effective not only in thin lab cells but also under practical loading.
| Condition | Initial discharge capacity / mAh·g−1 | Cycle number | Capacity retention / % |
|---|---|---|---|
| 1C, 5.18 mg cathode | 150 | 1000 | 79.7 |
| 1C, 5.18 mg cathode | 150 | 2000 | 68.3 |
| 0.3C, 20.32 mg cathode | ~200 | 250 | Nearly 100 |
I examined the morphology of the lithium metal after cycling. The scanning electron microscopy images show that lithium deposits uniformly at the interface between the interlayer and the lithium metal. The deposits do not penetrate the interlayer or extend into the bulk electrolyte. This is direct evidence that the interlayer suppresses dendrite growth. The cross-sectional images also show that the interlayer remains in good contact with the lithium metal, with no visible voids or delamination. I also assembled a Li|MIEC|LPSCl|MIEC|Cu half-cell and charged it at 0.5 mA·cm−2. The optical photograph shows a flat and dense lithium deposit on the interlayer surface. This confirms that the interlayer promotes uniform lithium plating.
I used X-ray photoelectron spectroscopy to check chemical stability. In the sulfur 2p spectrum, I observe peaks at 161.37 eV and 162.47 eV, which originate from the PS4^3− tetrahedral units of Li6PS5Cl. I also observe peaks at 162.87 eV and 162.97 eV, which correspond to P–S–P bonds. In the phosphorus 2p spectrum, I observe peaks at 132.87 eV and 133.97 eV, which also arise from the PS4^3− units. After cycling, the peak positions and intensities remain essentially unchanged. I do not detect new peaks corresponding to sulfur oxides, phosphorus oxides, or other decomposition products. The chlorine 2p spectrum is more difficult to resolve because the binding energy of Cl− in Li6PS5Cl overlaps with that of LiCl. A small amount of LiCl may form, but it serves as a passivation layer that prevents further decomposition. The X-ray photoelectron spectroscopy results are summarized in Table 9.
| Element | Peak / eV | Assignment | After cycling |
|---|---|---|---|
| S 2p | 161.37 | PS4^3− | Preserved |
| S 2p | 162.47 | PS4^3− | Preserved |
| S 2p | 162.87 | P–S–P | Preserved |
| S 2p | 162.97 | P–S–P | Preserved |
| P 2p | 132.87 | PS4^3− | Preserved |
| P 2p | 133.97 | PS4^3− | Preserved |
| Cl 2p | Overlapping | Cl− / LiCl | Minor LiCl, passivating |
I now discuss the mechanism by which the mixed conductive interlayer suppresses dendrites in a solid state battery. The starting point is the Poisson equation for the electric potential in a mixed conductor. If I denote the ionic potential as \(\phi_{\text{ion}}\) and the electronic potential as \(\phi_{\text{el}}\), the current conservation equations are:
$$ \nabla \cdot \left( \sigma_{\text{ion}} \nabla \phi_{\text{ion}} \right) = 0 $$
$$ \nabla \cdot \left( \sigma_{\text{el}} \nabla \phi_{\text{el}} \right) = 0 $$
At the lithium metal interface, the local electrochemical driving force for deposition is the overpotential:
$$ \eta = \phi_{\text{el}} – \phi_{\text{ion}} – E_{\text{eq}} $$
where \(E_{\text{eq}}\) is the equilibrium potential. The local current density follows Butler-Volmer kinetics:
$$ j = j_0 \left[ \exp\left( \frac{\alpha_a F \eta}{RT} \right) – \exp\left( -\frac{\alpha_c F \eta}{RT} \right) \right] $$
If the electronic conductivity is zero, the electronic potential is not defined inside the interlayer, and the deposition reaction can only occur where the current collector touches the electrolyte. This leads to a highly nonuniform current distribution. If the electronic conductivity is very high, the electronic potential is nearly constant, but the ionic current may not be able to keep up, and lithium can deposit inside the interlayer. The optimal interlayer has an intermediate electronic conductivity, so that the electronic potential is smoothed without creating an internal short circuit. The 1%SP composite achieves this balance.
I can also describe the lithium ion transport with the Nernst-Planck equation:
$$ \mathbf{J}_{\text{ion}} = -D \nabla c – \frac{z F D c}{RT} \nabla \phi_{\text{ion}} $$
where \(D\) is the diffusion coefficient, \(c\) is the lithium ion concentration, and \(z = 1\). When the interlayer homogenizes the ionic current, the concentration gradient \(\nabla c\) is reduced. A smaller concentration gradient delays the onset of Sand’s time, which is the time at which the concentration at the interface drops to zero and dendrite growth becomes favorable. Sand’s time is approximately:
$$ \tau_s = \frac{\pi D}{4} \left( \frac{e C_0}{J} \right)^2 $$
where \(C_0\) is the initial concentration and \(J\) is the applied current density. By flattening the current distribution, the interlayer effectively increases the local Sand’s time and suppresses dendrite nucleation.
I further validated the mechanism with in-situ electrochemical impedance spectroscopy. The impedance spectra before and after cycling show that the 1%SP interlayer maintains a low charge transfer resistance. I fitted the impedance data with an equivalent circuit containing a series resistance, a grain boundary resistance, an interfacial resistance, and a diffusion element. The fitted parameters are summarized in Table 10. The 1%SP interlayer reduces the interfacial resistance by more than 50% compared with the pristine electrolyte. This reduction is consistent with the improved contact and the mixed conductive pathway.
| Parameter | LPSCl before cycling | LPSCl after cycling | 1%SP before cycling | 1%SP after cycling |
|---|---|---|---|---|
| Series resistance / Ω·cm2 | 12 | 18 | 11 | 12 |
| Grain boundary resistance / Ω·cm2 | 35 | 62 | 34 | 36 |
| Interfacial resistance / Ω·cm2 | 88 | 145 | 42 | 48 |
| Diffusion resistance / Ω·cm2 | 55 | 92 | 30 | 33 |
The distribution of relaxation times analysis provides a complementary view. I calculated the DRT function using Tikhonov regularization. The peak area in the charge transfer region is proportional to the charge transfer resistance. The 1%SP interlayer reduces the peak area in this region by a factor of approximately 2.5. The peak position also shifts to shorter relaxation times, which indicates faster charge transfer kinetics. The relaxation time is related to the resistance and capacitance by:
$$ \tau = R C $$
where \(R\) is the resistance and \(C\) is the interfacial capacitance. A shorter relaxation time means that the interface can respond more quickly to changes in current, which is beneficial for high-rate operation of a solid state battery.
I also considered the mechanical aspect. The interlayer must accommodate the volume changes that occur during lithium plating and stripping. The Super P carbon network provides a compliant matrix that can deform without fracturing. The Li6PS5Cl particles provide the ionic conduction pathway. The composite interlayer therefore combines mechanical compliance with ionic and electronic transport. This combination is essential for a stable solid state battery. If the interlayer were too rigid, it would crack during cycling. If it were too soft, it would flow and lose contact. The 1%SP composite appears to strike the right balance.
I compared the performance of the 1%SP interlayer with other reported strategies in a general way. Many approaches focus on modifying the electrolyte or the lithium metal surface. For example, some studies use alloy interlayers, some use oxide coatings, and some use polymer interlayers. Each approach has advantages and disadvantages. The mixed conductive interlayer that I propose is attractive because it is simple to fabricate, compatible with roll-to-roll processing, and effective at low carbon loading. The key is the precise control of the electronic conductivity. Table 11 summarizes the general comparison without referring to specific authors or publications.
| Strategy | Ionic transport | Electronic transport | Mechanical compliance | Process complexity |
|---|---|---|---|---|
| Alloy interlayer | Moderate | Low | Moderate | High |
| Oxide coating | Low | Very low | Low | Moderate |
| Polymer interlayer | Moderate | Low | High | Moderate |
| Mixed conductive interlayer | High | Tunable | High | Low |
I also quantified the improvement in the solid state battery using a figure of merit. I define the interfacial stability factor as the product of the critical current density and the cycling time divided by the interfacial resistance:
$$ \text{ISF} = \frac{J_{\text{crit}} \cdot t_{\text{cycle}}}{R_{\text{int}}} $$
For pristine Li6PS5Cl, the interfacial stability factor is low because the critical current density is low and the interfacial resistance increases rapidly. For the 1%SP interlayer, the interfacial stability factor is high because the critical current density is high, the cycling time is long, and the interfacial resistance remains low. This figure of merit captures the combined kinetic and stability benefits.
I further validated the uniformity of lithium deposition by a simple optical experiment. I assembled a transparent cell with a Li|MIEC|LPSCl|MIEC|Cu configuration and charged it at 0.5 mA·cm−2. After a fixed amount of charge, I observed a shiny and uniform lithium layer on the interlayer surface. In contrast, a cell without the interlayer showed a rough and dendritic deposit. This visual evidence supports the electrochemical data. I also note that the interlayer does not block lithium ion transport, because the ionic conductivity remains high. The carbon network is sparse enough that it does not percolate through the entire interlayer in a way that would short circuit the cell, but it is dense enough to provide a continuous electronic path along the interface.
The percolation threshold for the carbon network is a critical parameter. I can estimate it using the following relation:
$$ \sigma_e = \sigma_0 \left( \phi – \phi_c \right)^t $$
where \(\phi\) is the volume fraction of carbon, \(\phi_c\) is the percolation threshold, and \(t\) is the critical exponent. For the 1%SP composite, the volume fraction is above the percolation threshold but not far above it. This means that the electronic conductivity is high enough to be useful but low enough to avoid excessive internal deposition. The 3%SP composite is far above the percolation threshold, which explains its high electronic conductivity and poor performance.
I also examined the effect of the interlayer on the impedance at different states of charge. The impedance spectra collected at 0%, 50%, and 100% state of charge show that the 1%SP interlayer maintains a stable interfacial resistance throughout charging and discharging. The pristine electrolyte shows a significant increase in interfacial resistance at high state of charge, which is a sign of interfacial degradation. This further confirms that the interlayer protects the solid state battery from interfacial side reactions.
I performed a post-mortem analysis of the full cell after 1000 cycles at 1C. The scanning electron microscopy image of the lithium metal surface shows a dense and uniform lithium layer with no visible dendrites. The cross-sectional image shows that the interlayer is intact and in good contact with both the electrolyte and the lithium metal. The X-ray photoelectron spectroscopy results show that the chemical state of the interlayer is preserved. These results indicate that the mixed conductive interlayer is chemically and mechanically robust over long-term cycling.
I now summarize the key findings in a structured way. First, the addition of 1% Super P to Li6PS5Cl creates a mixed ion-electron conductive interlayer without destroying the argyrodite crystal structure. Second, the 1%SP interlayer optimizes the electronic conductivity to 1.8×10−8 S·cm−1 while maintaining a high ionic conductivity. Third, the optimized interlayer raises the critical current density of a symmetric solid state battery to 1.6 mA·cm−2 and enables 2800 h of stable cycling at 0.5 mA·cm−2. Fourth, the full solid state battery with the 1%SP interlayer delivers 203 mAh·g−1 at 0.1C and retains 79.7% of its capacity after 1000 cycles at 1C. Fifth, the interlayer suppresses dendrite growth by homogenizing the electric field and the lithium ion flux, as confirmed by in-situ impedance, distribution of relaxation times, and post-mortem microscopy. Sixth, the interlayer remains chemically stable, as shown by X-ray photoelectron spectroscopy before and after cycling.
I believe that the concept of a mixed conductive interlayer can be extended to other sulfide electrolytes and other lithium metal anode systems. The key design rule is to keep the electronic conductivity high enough to smooth the potential but low enough to avoid internal deposition. This rule can be expressed as an optimal range for the dimensionless ratio \(\Lambda = \sigma_e / \sigma_{\text{ion}}\). For the system I studied, the optimal range is approximately \(10^{-6}\) to \(10^{-4}\). This range may shift for other electrolytes, but the principle remains the same. A solid state battery that operates near this optimum will have a more uniform current distribution, a higher critical current density, and a longer cycle life.
I also note that the interlayer thickness is an important variable. If the interlayer is too thin, it cannot smooth the potential effectively. If it is too thick, it adds unnecessary resistance. In my experiments, an interlayer thickness of approximately 5 to 10 μm was sufficient. I recommend that future work optimize the thickness together with the carbon content. A simple model for the optimal thickness can be derived from the competition between electronic smoothing and ionic resistance:
$$ L_{\text{opt}} \sim \sqrt{\frac{\sigma_{\text{ion}}}{\sigma_e}} \, L_0 $$
where \(L_0\) is a characteristic length scale of the interfacial roughness. This model predicts that a higher electronic conductivity allows a thinner interlayer, while a lower electronic conductivity requires a thicker interlayer. The 1%SP composite falls within the practical range for a solid state battery.
I further emphasize that the mixed conductive interlayer does not compromise the energy density of the solid state battery. The interlayer is thin and lightweight, and the carbon content is only 1%. The improvement in capacity retention and rate capability more than compensates for the small additional mass. In a practical cell, the interlayer can be coated onto the electrolyte membrane or laminated onto the lithium foil. Both processes are compatible with existing battery manufacturing lines. This makes the approach scalable and cost-effective.
In conclusion, I have demonstrated that a mixed ion-electron conductive interlayer is an effective strategy to stabilize the interface between a sulfide solid electrolyte and a lithium metal anode in a solid state battery. The 1%SP composite provides the best combination of ionic conductivity, electronic conductivity, and mechanical compliance. It suppresses lithium dendrite growth, reduces interfacial resistance, and improves the cycling stability of both symmetric and full cells. The mechanistic understanding that I developed, based on charge balance, Butler-Volmer kinetics, Sand’s time, and distribution of relaxation times, provides a foundation for further optimization. I expect that this approach will contribute to the development of high-energy, long-life solid state battery technologies.
I also propose several directions for future work. First, the interlayer composition can be optimized with other carbon materials, such as carbon nanotubes or graphene, which may provide a more efficient electronic network at lower loading. Second, the interlayer can be combined with a lithium alloy anode or a protective coating to further improve the interfacial stability. Third, advanced characterization techniques, such as operando X-ray tomography and cryogenic electron microscopy, can be used to visualize the lithium deposition process in real time. Fourth, the interlayer can be tested in large-format pouch cells to evaluate its scalability. Fifth, the design rule based on the dimensionless ratio \(\Lambda\) can be generalized to other solid electrolyte systems, including halide and oxide electrolytes. These directions will help to accelerate the practical application of solid state battery technology.
Finally, I summarize the electrochemical equations that are most relevant to the design of a mixed conductive interlayer for a solid state battery. These equations can be used as a guideline for future experiments:
$$ \sigma_{\text{ion}} = \frac{L}{R_{\text{ion}} A} $$
$$ \sigma_e = \frac{L}{R_e A} $$
$$ \Lambda = \frac{\sigma_e}{\sigma_{\text{ion}}} $$
$$ \eta = \phi_{\text{el}} – \phi_{\text{ion}} – E_{\text{eq}} $$
$$ j = j_0 \left[ \exp\left( \frac{\alpha_a F \eta}{RT} \right) – \exp\left( -\frac{\alpha_c F \eta}{RT} \right) \right] $$
$$ \tau_s = \frac{\pi D}{4} \left( \frac{e C_0}{J} \right)^2 $$
$$ Z(\omega) = R_\infty + \int_0^\infty \frac{\gamma(\tau)}{1 + j \omega \tau} \, d\tau $$
$$ \text{Retention} = \frac{Q_N}{Q_1} \times 100\% $$
I am confident that these equations, together with the experimental data presented here, provide a clear and quantitative framework for designing mixed conductive interlayers in a solid state battery. The results show that a small amount of electronic conductivity, when properly distributed, can transform the interfacial behavior of a solid state battery from unstable to highly stable. This finding has implications for the broader field of lithium metal batteries and for the development of next-generation energy storage devices.