The most frustrating problem in solid-state battery development: after 50 cycles your capacity dropped 30% and impedance climbed 5×, but you don't know whether the cathode or the anode failed. You open the cell for SEM and XRD and find suspicious features on both sides — a suspected CEI layer on the cathode surface and voids in the anode. Which one do you fix first?
A three-electrode cell can tell you the answer in real time, without taking the battery apart. This is the first article in a series demonstrating how three-electrode testing pinpoints failure causes through four real-world cases.
Case 1: Rapid Capacity Fade With the Root Cause in an Unexpected Place
The Problem
A research group ran Li | Li₆PS₅Cl | NMC811 cells at 60 °C. The first 20 cycles looked normal, then capacity started dropping fast:
- Cycle 1: 165 mAh/g
- Cycle 20: 158 mAh/g (−4%)
- Cycle 50: 115 mAh/g (−30%)
Two-electrode EIS showed total impedance climbing from 45 Ω·cm² to 310 Ω·cm². Initial diagnosis: interfacial problem, but which interface?
How Two-Electrode Data Misled the Team
Following the usual logic — sulfide electrolytes are vulnerable to oxidation at high voltage — the team first tried to fix the cathode side:
- Trial 1: Coat NMC with LiNbO₃ (2 months)
- Trial 2: Replace the electrolyte with Li₅.₅PS₄.₅Cl₁.₅, which has a wider oxidation window (1.5 months)
- Result: Capacity fade improved slightly but was still fast (still dropped to ~120 mAh/g by cycle 50)
3.5 months burned.
Three-Electrode Data Revealed the Truth
In month four they assembled a three-electrode cell. The result:
| Measurement | Initial impedance | After 50 cycles | Growth |
|---|---|---|---|
| Full cell | 45 Ω·cm² | 310 Ω·cm² | +265 |
| Cathode vs ref | 22 Ω·cm² | 51 Ω·cm² | +29 |
| Anode vs ref | 23 Ω·cm² | 259 Ω·cm² | +236 |
89% of the impedance growth was at the anode. The cathode coating work was essentially wasted effort.
The Root Cause
Further analysis found that at 60 °C, the lithium anode slowly forms Li₃P and Li₂S decomposition products at the interface with Li₆PS₅Cl (reductive decomposition). This SEI layer has poor conductivity and accumulates with cycling, eventually pulling impedance up.
Solution: Add a buffer layer with better reductive stability at the anode side (Li₃InCl₆ or Mg-doped Li), or lower the working temperature to 40 °C.
Had they started with a three-electrode cell, the 3.5-month detour would have been completely avoidable.
Key Takeaways
This case shows:
- Don't guess based on "common sense": Sulfide oxidation at high voltage is real, but it doesn't mean that's your battery's main problem
- Time is money: 3.5 months could have produced many experiments or a paper
- Value of three-electrode: One test clarifies direction and avoids blind trial-and-error
Case 2: Capacity Fade With Stable Impedance — The Problem Is Not at the Interfaces
The Problem
Li | Li₁₀GeP₂S₁₂ (LGPS) | LiCoO₂ cell, room-temperature cycling at 1C.
- Capacity dropped steadily from 135 mAh/g to 100 mAh/g over 100 cycles
- But EIS barely changed: stayed around 60 Ω·cm² the whole time
This is unusual: flat impedance means the interfaces did not degrade noticeably, so why did capacity drop?
The Challenge of Routine Troubleshooting
The team tried multiple directions without finding the cause:
- Checked cathode formulation: conductor ratio normal, binder showed no degradation
- Checked anode: lithium surface looked normal visually, no obvious discoloration
- Checked electrolyte: no cracking after pellet pressing
- Post-mortem SEM: both cathode and anode surfaces had some irregular features, but couldn't tell which was the main problem
Another 1.5 months gone.
Three-Electrode Data
| Measurement | Initial | After 100 cycles | Change |
|---|---|---|---|
| Cathode impedance | 28 Ω·cm² | 32 Ω·cm² | +4 |
| Anode impedance | 32 Ω·cm² | 28 Ω·cm² | −4 |
Neither interface degraded noticeably. This rules out the usual interfacial impedance accumulation.
The Key Information in the Potential Curves
A three-electrode cell lets you plot the cathode and anode potential curves separately. The team found:
- Cathode potential curve: cycle 1 and cycle 100 nearly overlapped, meaning LiCoO₂ de/lithiation did not degrade noticeably
- Anode potential curve: at cycle 100, the end-of-charge potential was about 50 mV lower than at cycle 1
This means: the available lithium inventory was shrinking — some lithium was consumed by irreversible side reactions, leaving the anode "lithium-deficient."
The Root Cause
Opening the cell for ICP analysis revealed that about 15% of the LiCoO₂ particles in the cathode composite had cracked during cycling, and lithium in the cracks reacted with the electrolyte to form Li₂CO₃ (LGPS decomposes slightly at high voltage). That lithium was permanently lost.
Impedance stayed flat because the conductive network was still intact; but with less lithium inventory, capacity dropped.
Solution: Lower the charge cutoff voltage (4.2 V → 4.1 V), or optimize the cathode pressing process to reduce particle cracking.
Key Takeaways
- Not all failures show up in impedance: Lithium inventory loss and active-material structural degradation don't appear in EIS
- Potential curves contain more information: Small changes in curve shape and position are meaningful
- Three-electrode told you "don't waste time on the interfaces," which by itself saved months
Case 3: Sudden Short Circuit — Three-Electrode Early Warning 48 Hours Ahead
The Problem
Li | Li₇La₃Zr₂O₁₂ (LLZO) | NMC cell shorted abruptly mid-charge in a certain cycle, with voltage dropping nearly to 0 V.
Post-mortem: a lithium dendrite had pierced the ceramic electrolyte.
A standard two-electrode cell can only do post-mortem analysis: "Dendrites formed, cell is dead, be careful next time." But how to be careful next time? When will dendrites form? It's a complete black box.
How Three-Electrode Gave Early Warning
About 48 hours (roughly 10 cycles) before the short, the team observed:
The anode potential briefly crossed below 0 V vs Li/Li⁺ during charge — it dropped to −0.15 V for a few seconds, then recovered.
This is the signal that lithium deposition is exceeding the interface's transport capability: local current density is too high, lithium starts depositing non-uniformly and forms small metallic protrusions.
A two-electrode cell cannot see this — because the cathode potential is rising at the same time, so the difference between the two (cell voltage) is still in the normal range (e.g. 3.8 V). The user only sees "everything normal" until a sudden short.
What to Do
Once you see the anode undershoot signal, you can take immediate action:
- Lower the charge rate: From 1C down to 0.5C or 0.2C to reduce local current density
- Raise stack pressure: Add 50–100 MPa (if the mold allows) to improve interfacial contact
- Reduce depth of charge: Don't charge to 100% SOC; leave a safety margin
- Pause cycling: Rest for a few hours to let the interface re-equilibrate
In that case, the team adjusted the protocol right after seeing the undershoot signal:
- Charge rate from 1C down to 0.5C
- Charge cutoff from 100% SOC down to 90%
- Stack pressure from 300 MPa up to 400 MPa
After the adjustment, the next 200 cycles ran without another short and the cell continued stable cycling.
Key Takeaways
- From "post-mortem" to "real-time warning": This is three-electrode's most intuitive value
- 48-hour window is enough time to adjust parameters: You don't have to wait for the cell to die before you know there's a problem
- Lithium dendrites don't "appear suddenly": They have precursors, but two-electrode cells can't see them
Case 4: Both Electrodes Degrading — Sorting Out Priorities
The Problem
Li | β-Li₃PS₄ | LiNi₀.₈Co₀.₁Mn₀.₁O₂ (NCM811), 80 °C cycling (accelerated aging test).
After 100 cycles, capacity dropped from 180 mAh/g to 95 mAh/g and impedance rose from 50 Ω·cm² to 420 Ω·cm².
Post-mortem showed problems on both sides:
- Cathode surface blackened (suspected oxidative decomposition products from the electrolyte)
- Anode surface rough with voids (suspected reductive reaction between lithium and electrolyte)
Both interfaces are degrading, but you only have time and budget to optimize one. Which first?
The Challenge With Two-Electrode Data
Two-electrode testing can only tell you:
- Total impedance rose by 370 Ω·cm²
- Both cathode and anode "have problems"
But improvement resources are limited:
- Cathode coating takes 2 months (synthesis, characterization, assembly, testing)
- Anode interlayer also takes 2 months
- Doing both directions at once takes 4 months
If you pick the wrong one first, that's more months wasted.
Three-Electrode Data
| Measurement | Initial | After 100 cycles | Growth |
|---|---|---|---|
| Cathode impedance | 24 Ω·cm² | 145 Ω·cm² | +121 |
| Anode impedance | 26 Ω·cm² | 275 Ω·cm² | +249 |
Both interfaces are degrading, but the anode is worse (67% of total growth).
Optimization Strategy
Based on this data, the priorities are clear:
Short term (within 1 month):
- Fix the anode first: Add a Li-In alloy interlayer at the anode to suppress reductive decomposition of β-Li₃PS₄
- Expected effect: Bring anode impedance growth down from +249 to ~+100
Medium term (if the short-term solution works):
- Replace with an electrolyte that has a wider reductive window (e.g. Li₆PS₅Cl)
- This will improve both anode and cathode stability at once
Long term (if the first two steps are not enough):
- Then consider cathode coating (because after the anode is optimized, the cathode's 33% becomes the new primary issue)
Actual Result
The team followed this roadmap:
- 1 month later, the Li-In interlayer solution tested successfully; after 100 cycles total impedance was only 180 Ω·cm² (down 57% from the previous 420)
- Capacity retention rose from 53% to 78%
- No need to touch the cathode; the problem was solved to an acceptable level
Three-electrode data helps you prioritize and avoid the trap of "both are important."
Key Takeaways
- Quantified priorities: A 67/33 split is far clearer than "both are important"
- Staged optimization: Solve the primary issue first; don't try to fix everything at once
- Cost of validating assumptions: Guessing wrong about which to fix first wastes real money and time
Common Pattern Across All Four Cases
Limitations of Two-Electrode Data
| Case | What two-electrode showed | Actual situation | Time wasted |
|---|---|---|---|
| Case 1 | Large impedance rise | 89% at anode | 3.5 months (wrong direction) |
| Case 2 | Capacity drop but stable impedance | Lithium inventory loss | 1.5 months (blind troubleshooting) |
| Case 3 | Sudden short | Precursor 48 hours earlier | Cell scrapped |
| Case 4 | Both sides have problems | Anode accounts for 67% | Potentially 4 months (if doing both) |
Total avoidable detours: 9–10 months
Value of Three-Electrode
- Clarify direction: Tell you where to apply effort (Cases 1, 4)
- Rule out distractions: Tell you what not to worry about (Case 2)
- Early warning: Give you reaction time before the problem worsens (Case 3)
When Three-Electrode Is Most Worth It
Based on these four cases, strongly recommend three-electrode in the following situations:
✅ Diagnosing unexplained capacity fade (Cases 1, 2) ✅ Accelerated aging tests (Case 4, need to quickly pinpoint the dominant failure mechanism) ✅ High-risk chemistries (Case 3, electrolytes prone to shorting) ✅ Limited time and budget (Case 4, can't do multiple directions at once) ✅ Writing high-quality papers (reviewers will ask "which electrode?")
Next Steps
This article demonstrated three-electrode testing in real failure-analysis scenarios. The next articles in the series will explore:
- How to use three-electrode to measure electrochemical windows during new electrolyte development
- Key data for charging protocol optimization
- Temperature and pressure effect studies
- How failure modes evolve with cycling
Solid-State Battery Failure Analysis (Part 3): Data Quality and Paper Writing
- Quality control checklists for assembly and testing
- How to judge whether three-electrode data is trustworthy
- Complete workflow from data to publication
- Strategies for addressing common reviewer questions
Related Reading
- Three-Electrode Testing for Solid-State Batteries: Separating Anode and Cathode Behavior (detailed technical principles)
- How to Choose a Solid-State Battery Test Mold (selecting the right test tools)
Recommended Test Tool
The cases in this article used the Solid-State Battery Three-Electrode Test Mold, designed specifically for diagnostic testing:
- Side-exit port for the reference lead: Reference wire exits from the side, not interfering with the main compression path
- PEEK core: Electrically insulating, stable to 200 °C, chemically inert
- Split-body design: Convenient for precise reference wire placement
- Rated 500 MPa, customizable to 1000 MPa
Product inquiry: business@labcorematerials.com | Quote replied within 48 hours (USD, DDP to major US ports)