The Problem With Two-Electrode Cells
You run a solid-state cell for 80 cycles. Capacity drops from 180 mAh/g to 120 mAh/g. Impedance climbs from 40 Ω·cm² to 310 Ω·cm². Now answer this question: which electrode failed?
In a two-electrode configuration you cannot answer it. The cell voltage you measure is the difference between cathode potential and anode potential:
V_cell = E_cathode − E_anode
One number, two unknowns. A 250 Ω·cm² impedance rise could be:
- Lithium dendrites penetrating the separator layer at the anode
- A resistive interphase growing between the sulfide electrolyte and the NMC cathode
- Contact loss at the cathode composite from volume change during delithiation
- Any combination of the above
Each of these demands a completely different fix. Dendrite suppression means raising stack pressure or adding an interlayer. Cathode interphase growth means coating the active material (LiNbO₃, Li₂ZrO₃). Contact loss means reformulating the composite or changing the pressing protocol. Guess wrong and you burn two months.
The three-electrode cell resolves this by adding a reference electrode — a third terminal that holds a stable, known potential and lets you measure each electrode independently against it.
How a Reference Electrode Works in a Solid-State Cell
The basic principle
A reference electrode is any electrode whose potential stays fixed regardless of the small currents flowing through the measurement circuit. In liquid systems this is straightforward — Ag/AgCl, saturated calomel, or a lithium wire immersed in electrolyte. In a solid-state cell, the reference must make ionic contact with the solid electrolyte without disturbing the main current path.
Once installed, your potentiostat measures three things simultaneously:
V_cell = E_cathode − E_anode (what you saw before)
V_working = E_cathode − E_reference (cathode alone)
V_counter = E_anode − E_reference (anode alone)
Now the 250 Ω·cm² rise resolves into two numbers. If the cathode contributes 30 Ω·cm² and the anode contributes 220 Ω·cm², you have a lithium-metal problem, not a cathode problem.
Three practical implementations
1. Embedded lithium wire (most common)
A 50–100 µm lithium wire is placed at the mid-plane of the solid electrolyte layer during assembly, with one end exiting through a side port in the mold. Li/Li⁺ is an excellent reference in solid-state systems — its potential is defined, stable, and needs no additional interface.
- Pros: Simple, cheap, thermodynamically well-defined (0 V vs Li/Li⁺)
- Cons: The wire creates a small mechanical discontinuity in the electrolyte; under 300+ MPa it can extrude or short to an electrode if misplaced. Requires careful placement during assembly.
2. Side-contact ring
A thin metal ring (Li-coated Cu, or Li foil) contacts the outer circumference of the electrolyte pellet, away from the active stack.
- Pros: No discontinuity in the ion path; mechanically robust under high stack pressure
- Cons: Sits outside the primary current-flow region, so at high current density it can report a potential offset from the true mid-plane value. Best for low-rate studies (<C/2).
3. Micro-reference on the separator edge
A lithographically-defined Li dot or a thin Li/SE micro-cell bonded at the pellet edge.
- Pros: Minimal geometric perturbation, good for high-rate work
- Cons: Fabrication complexity; typically only used in specialized labs with cleanroom access.
For most research groups, the embedded lithium wire is the right starting point — it's what the majority of published three-electrode ASSB studies use.
Placement: The Detail That Decides Whether Your Data Is Real
This is the part most groups get wrong the first time.
A reference electrode reports the potential at the point where it sits. If it's off-center — say 80% of the way toward the cathode instead of at the mid-plane — it picks up part of the cathode's ohmic drop and attributes it to the anode. The result looks like clean data but is systematically wrong.
Rules for placement
| Rule | Reason |
|---|---|
| Mid-plane, ±10% of electrolyte thickness | Symmetric ohmic split between the two half-cells |
| Radially outside the active stack footprint | Prevents the wire from carrying main-cell current |
| Electrically isolated from both current collectors | A short to either electrode makes all three measurements meaningless |
| Same Ar-glovebox handling as the rest of the cell | Li reacts with trace H₂O/O₂; a passivated reference drifts |
Sanity check before you trust the data
After assembly, measure the open-circuit potential of each half:
- E_anode vs reference should read ~0 V if the anode is lithium metal. If it reads 0.3 V, your reference is contaminated or the wire is partly passivated.
- E_cathode vs reference should match the expected OCV of the cathode material (e.g. ~3.7 V for NMC811 at 50% SOC). A reading 200 mV off means the reference is misplaced or drifting.
- Sum check: E_cathode − E_anode must equal the directly-measured V_cell to within 5 mV. If it doesn't, something is shorted.
Run this check on every cell. It takes 30 seconds and saves entire experimental campaigns.
Reading the Data: Four Diagnostic Patterns
Once you have clean three-electrode data, most failure modes reveal themselves as characteristic signatures.
Pattern 1 — Anode-dominated impedance growth
Signature: Anode half-cell impedance climbs steadily cycle-over-cycle; cathode stays flat.
Interpretation: Lithium-metal interface degradation. Either (a) SEI accumulation from electrolyte decomposition, or (b) contact loss from void formation during stripping.
Distinguishing between (a) and (b): Raise stack pressure by 50 MPa and re-measure. If impedance drops sharply, it was void formation (mechanical). If unchanged, it's chemical SEI growth.
Typical fixes: Higher stack pressure, Li-alloy interlayer (Mg, Ag), or a more reductively-stable electrolyte at the anode side.
Pattern 2 — Cathode-dominated impedance growth
Signature: Cathode half-cell impedance rises; anode is stable.
Interpretation: Cathode–electrolyte interphase (CEI) growth, typically from oxidative decomposition of the solid electrolyte at high cathode potential. Sulfides are particularly vulnerable above 2.5 V vs Li/Li⁺.
Confirming it: Cycle the same cell in a narrower voltage window (e.g. cut off at 4.0 V instead of 4.3 V). If the growth rate drops proportionally, it's oxidative.
Typical fixes: Coat the cathode active material (LiNbO₃, Li₂ZrO₃, Al₂O₃), or use a dual-layer electrolyte with an oxidatively-stable layer (e.g. halide) at the cathode.
Pattern 3 — Sudden anode potential excursion
Signature: Anode potential drops below 0 V vs Li/Li⁺ during charge, often abruptly.
Interpretation: Lithium plating exceeding the transport capability of the interface — the precursor to dendrite formation. This is the single most valuable measurement a three-electrode cell provides, because in a two-electrode cell it's completely invisible until the cell shorts.
Action: Reduce charge rate, raise temperature, or increase stack pressure. Map the onset current density as a function of temperature to find your safe operating window.
Pattern 4 — Both electrodes stable, capacity still fading
Signature: Neither half-cell impedance changes much, but capacity declines.
Interpretation: The loss is not interfacial. Likely candidates: active-material cracking in the cathode composite, loss of percolation in the carbon network, or lithium inventory loss to irreversible side reactions.
Next step: Post-mortem — cross-section SEM of the cathode composite, or ICP for lithium inventory. Three-electrode data has told you where not to look, which is still valuable.
Impedance Deconvolution: A Worked Example
Consider a Li | Li₆PS₅Cl | NMC811 cell after 100 cycles at 60 °C.
Two-electrode EIS gives a single semicircle plus tail, total 310 Ω·cm². You fit it with R_bulk + R_interface + Warburg and get R_interface = 265 Ω·cm². That number is the sum of both interfaces, and there's no honest way to split it.
Three-electrode EIS run at the same state:
| Measurement | R_bulk | R_interface |
|---|---|---|
| Full cell | 45 Ω·cm² | 265 Ω·cm² |
| Cathode vs ref | 22 Ω·cm² | 51 Ω·cm² |
| Anode vs ref | 23 Ω·cm² | 214 Ω·cm² |
The bulk resistance splits cleanly in half — expected, since the reference sits at the mid-plane. This is itself a validation that placement was correct.
The interface resistance splits 51 / 214. 81% of the interfacial loss is at the lithium anode. Cathode coating work would have been almost entirely wasted effort; the correct intervention is on the anode side.
This single measurement redirects months of work.
Practical Considerations
Which cells justify a three-electrode mold?
Three-electrode assembly takes roughly 2× as long as two-electrode and has a higher failure rate (a misplaced reference wire means scrapping the cell). It is not the default choice for every experiment.
Use three-electrode when:
- You're diagnosing an unexplained capacity fade
- You're developing a new electrolyte and need to know its stability window against both electrodes
- You're optimizing charge protocols and need to see the anode plating threshold
- You're writing a paper where reviewers will ask "which electrode?"
Stick with two-electrode when:
- Running long-term cycling on an already-characterized system
- High-throughput composition screening (you want speed, not mechanism)
- Routine QC against a known baseline
A common workflow: screen in two-electrode, then build three or four three-electrode cells of the top candidate to understand why it works.
Temperature and reference stability
At elevated temperature, lithium creeps. A reference wire that sat cleanly at the mid-plane at 25 °C can migrate under 300 MPa at 80 °C. If you're running above 60 °C:
- Re-verify the OCV sanity check after thermal equilibration, not just at assembly
- Consider the side-contact ring geometry instead of an embedded wire
- Re-check the sum rule (E_cathode − E_anode = V_cell) at intervals during long runs
Common failure modes during assembly
| What happens | Cause | Prevention |
|---|---|---|
| Reference reads open-circuit / noisy | Wire lost ionic contact | Press the wire gently into the electrolyte surface before the top layer goes on |
| All three potentials read the same | Wire shorted to an electrode | Keep the wire radially outside the stack footprint; check with a multimeter before pressing |
| Sum rule violated by >50 mV | Reference off-center or passivated | Re-scrape the Li wire in the glovebox immediately before placement |
| Reference drifts over hours | Trace moisture reacting with Li | Verify glovebox H₂O < 0.1 ppm; dry all components 24 h at 80 °C under vacuum |
Summary
A two-electrode solid-state cell tells you that something failed. A three-electrode cell tells you which electrode and by how much. For mechanism work, that distinction is the difference between a targeted fix and months of guessing.
The key requirements:
- Reference at the electrolyte mid-plane, radially outside the active stack
- Lithium metal as the reference material (0 V vs Li/Li⁺, no additional interface)
- Sanity-check every cell with the OCV sum rule before trusting any data
- Re-verify after thermal ramps if operating above 60 °C
Our three-electrode mold (MOL1D100A) provides a dedicated side port for the reference lead, a PEEK core rated to 200 °C, and die-steel plungers stable to 5 T stack force — designed so that the reference wire can be positioned and held at the mid-plane without extruding under pressure.
For working-cavity sizes outside the standard φ10 mm, or for 1000 MPa configurations, contact business@labcorematerials.com — quotes returned within 48 hours.
Related reading: Solid-State Battery Test Mold Selection and Usage Guide