How to Diagnose LiPo Pouch Cell Swelling in OEM Products
Technically Reviewed By
Dr. Maximilian Weber
Chief Scientist
Dr. Maximilian Weber is THOR Power’s Chief Scientist and a senior expert in lithium battery technology. His technical review focuses on battery safety, performance optimization, energy density and custom battery solution development.
Last technical review: September 2026
Written By
Victor Xiong
President of OEM Division & Custom Battery Specialist
Victor Xiong holds a Master’s degree from The Chinese University of Hong Kong, Shenzhen. He leads THOR Power’s OEM Division and focuses on custom battery solutions for global device brands, product developers and industrial customers.
Last technical review: August 2026
IMMEDIATE SAFETY GATE
Stop normal testing and follow the approved lithium-battery safety and EHS procedure if a pouch shows visible ballooning, leakage or odor, unusual heat, mechanical damage, enclosure pressure, abnormal open-circuit voltage, abnormal self-discharge, or a known abuse event. Isolate the sample under the approved process. Do not puncture, flatten, squeeze, rework or recharge it merely to continue a comparison.
QUICK ANSWER
LiPo pouch-cell swelling cannot be approved or rejected from one caliper reading. Compare the same cell or population at a controlled SOC, temperature, rest time, location and preload; separate reversible breathing from residual growth and external pack stack-up; then validate both free-space clearance and, where expansion is constrained, the qualified force or pressure envelope.
This guide is for OEM engineering, quality and sourcing teams investigating thickness growth after charge, storage, cycling or field use. For first-time mechanical selection, start with the 3.7V LiPo battery models by capacity and size. If the cell itself is still being selected, review custom LiPo battery options only after the device cavity, voltage class and use profile are defined.
Approval Questions
1. Is the thickness increase measured at the same SOC, rest time and temperature? 2. Has pack stack-up been separated from real pouch-cell growth? 3. Do electrical, thermal, charge and lot records point to the same root cause? 4. Can the approved cavity, compression, insulation and inspection method be repeated in production?
Case Transparency Note
The swelling examples in this article are representative composite scenarios based on recurring pouch-cell diagnosis patterns. They explain engineering decisions without identifying a specific customer, product, lot or confidential THOR Power project.
The Real Problem: A Thicker Pouch Is a Symptom, Not a Root Cause
A buyer receives a pouch cell specified at 4.60 mm. After charging, it measures 4.78 mm; after controlled discharge and rest, it returns to 4.63 mm. Later in life the same reference-condition measurement reaches 4.92 mm. Those three numbers do not describe the same mechanical event.
The high-SOC increase may contain reversible electrode breathing. The later reference-condition rise may indicate progressive residual growth. A finished pack can also become locally thicker because of PCM, insulation, label overlap, adhesive, foam, folded tabs, wire routing or enclosure features. A sound investigation separates those contributors before changing cell capacity, rejecting a supplier lot or redesigning the housing.
CUSTOMER PAIN POINT
If a cover becomes tight, do not immediately specify a thinner cell or blame ‘bad chemistry.’ First identify where thickness or force increased, under which controlled condition, how large the change is relative to measurement uncertainty, and whether the source is the pouch body, finished-pack stack, harness or enclosure.
1. Localize What Changed and Define the Decision Question
‘Is the battery swollen?’ is too vague for release. Localize the change and write a controlled question—for example: ‘Does free pouch-body thickness at 50% SOC, 23 ± 2 °C and 24 hours rest remain inside the approved drawing after the defined life test?’ Map the pouch body, protected-end stack, harness and enclosure contact separately before assigning a cause.
Object: bare pouch body, finished protected pack or installed assembly – do not mix datasets.
Reference condition: SOC, charge/discharge history, temperature, rest time and whether the sample was free or constrained.
Location: defined points or full profile; separately report seal edges, tab fold, PCM and local stack maxima.
Mechanical condition: free/lightly loaded thickness, constant-force displacement or constant-gap reaction force.
Comparison: control cells, same cell over life, another lot, supplier drawing or project-specific release limit.
Decision: safety containment, supplier corrective action, enclosure redesign, additional validation or production release.
Charging evidence must use the exact voltage class and approved charge profile. A 4.20 V-class and 4.35 V-class pouch cell are not interchangeable by assumption; use the 4.2 V vs 4.35 V Lithium Battery guide when high-SOC behavior or charger compatibility is part of the investigation.
2. Separate Reversible Breathing, Residual Growth and Pack Stack-Up
Reversible SOC breathing
Electrode dimensions change as lithium is inserted and removed. A pouch can therefore be thicker at one state of charge and move back toward the controlled reference thickness after discharge and rest. This reversible component is not automatically a defect, but it still consumes mechanical space and must remain inside the exact cell maker’s qualified behavior.
Progressive residual growth
If the reference-condition baseline rises over storage or cycling, the residual component is increasing. Possible contributors include gas-producing side reactions, SEI/interphase growth, electrode restructuring, lithium plating risk, seal or process abnormalities, electrical abuse and excessive temperature. A high-SOC snapshot alone cannot establish residual growth.
Finished-pack and enclosure stack-up
A local maximum can come from folded tabs, PCM, insulation, label overlap, adhesive, foam, wire exit or a housing rib. A center-point pouch reading does not prove the complete assembly fits, and a protected-end maximum does not automatically prove pouch-body swelling. For a custom LiPo battery, the controlled finished-pack drawing—not the nominal cell body alone—must define the release boundary.
CONCEPTUAL DECOMPOSITION
Observed thickness = reference-condition thickness + reversible SOC/temperature/load-history change + residual life/storage/abuse change + external pack stack-up + measurement error. Estimate each term only from controlled, traceable evidence.
3. Distinguish Free Thickness from Constrained Force
Free thickness and constrained force answer different mechanical questions. A lightly loaded thickness measurement shows how much space the pouch wants to occupy under a defined method. A constant-gap or enclosure-like test shows the reaction force or pressure generated when the product restricts that displacement. A rigid cover can therefore hide visible swelling while load rises against the display, PCB, adhesive bond or cell surface.
| Test boundary | What is controlled | What is measured | OEM question answered |
|---|---|---|---|
| Controlled light-preload / free-thickness | SOC, temperature, rest, platen area and small repeatable measurement force | Thickness or surface profile | Reversible breathing, residual growth and spatial distribution without the product becoming the clamp. |
| Constant-force / compliant fixture | Applied force or pressure | Thickness change vs SOC / life | How the pouch moves under a known mechanical condition. |
| Constant-gap / enclosure-like fixture | Gap or structural displacement | Reaction force / pressure vs SOC, temperature and life | How much load the product may experience when expansion is constrained. |
If total force F is measured over an effective contact area A, average pressure can be screened as P = F / A. That arithmetic does not prove local pressure is uniform; pressure mapping or a validated structural model may be required when nearby components have local load limits.
Published pouch-cell studies show that pressure can change transport and degradation, but each result remains specific to its chemistry, format, fixture and use profile. Never import a literature pressure as a universal LiPo limit [1,2].
4. Build a Controlled Thickness Measurement Method
A digital caliper can be useful for screening, but the instrument name is not the method. Soft pouches deform under contact load, and narrow jaws can create local indentation. Release-quality data requires a repeatable mechanical boundary.
| Control variable | Minimum record |
|---|---|
| SOC / charge history | Charge profile, termination condition, discharge endpoint and the defined reference-SOC method. |
| Temperature | Cell and room temperature, plus stabilization time before measurement. |
| Rest time | Time after charge/discharge and whether the cell was free or constrained during rest. |
| Mechanical method | Platen geometry/contact area, preload or pressure, dwell time and fixture ID. |
| Location | Drawing or coordinate map for pouch body, seal, tab/PCM region and local maxima. |
| Measurement system | Gauge ID, calibration, resolution, repeatability, operator and uncertainty/Gauge R&R. |
| Sample identity | Cell/pack serial, lot/date code, build revision, cycle/storage history, charger/firmware and visual condition. |
5. Treat Measurement Uncertainty as Part of the Result
A thickness change must be evaluated against both measurement-system capability and sample-to-sample variation. Gauge R&R or an equivalent MSA should quantify repeatability and reproducibility for the defined fixture, operator, location and preload. For a paired life comparison, separately define the uncertainty of ΔT = T_checkpoint − T_BOL; do not use measurement uncertainty as a substitute for lot variation, sample size or confidence intervals.
Evaluate each cell’s paired change ΔTᵢ = Tᵢ,checkpoint − Tᵢ,BOL. Define UΔ as the expanded measurement uncertainty of that paired change—for the synthetic example below, UΔ(k = 2) = ±0.04 mm. Report n, mean ΔT, standard deviation and a 95% confidence interval. If the shift is not clearly separated from both UΔ and the observed population variation, classify it as unresolved—not as proven growth and not as zero growth.
Worked example: when 0.03 mm is not yet evidence
The paired dataset below is synthetic. UΔ is an example expanded uncertainty for the paired change, while SD and the 95% CI describe sample variation and uncertainty of the group mean. The values demonstrate decision logic; they are not a universal LiPo allowance or THOR Power laboratory data.
| Checkpoint | Sample basis | Control paired result | Suspect paired result | Interpretation |
|---|---|---|---|---|
| Beginning-of-life reference | n = 10 per group | T_BOL = 4.60 mm; SD = 0.03 mm | T_BOL = 4.60 mm; SD = 0.03 mm | Common reference distribution; UΔ applies to later paired changes. |
| Early-cycle return to reference state | 10 paired cells per group | Mean ΔT = +0.02 mm; SD = 0.03 mm; 95% CI ≈ −0.001 to +0.041 mm | Mean ΔT = +0.03 mm; SD = 0.03 mm; 95% CI ≈ +0.009 to +0.051 mm | Intervals overlap and both shifts are small relative to example UΔ; unresolved. |
| After high-SOC storage, returned to reference state | 10 paired cells per group | Mean ΔT = +0.04 mm; SD = 0.03 mm; 95% CI ≈ +0.019 to +0.061 mm | Mean ΔT = +0.30 mm; SD = 0.05 mm; 95% CI ≈ +0.264 to +0.336 mm | Separated persistent shift; contain and investigate with electrical, thermal and lot evidence. |
A later residual shift becomes more credible when it correlates with capacity or usable Wh, DC resistance/pulse behavior, self-discharge, temperature and charge history. Thickness alone cannot identify chemistry or root cause. If runtime loss is part of the complaint, use the LiPo capacity and runtime calculation method rather than treating thickness as a proxy for usable energy.
6. Build a Root-Cause Evidence Matrix
Thickness growth is not synonymous with gas. Use competing hypotheses and request evidence that can separate them.
| Possible mechanism | Typical pattern | Evidence / next check |
|---|---|---|
| Reversible lithiation/delithiation breathing | Thickness moves systematically with SOC and largely returns at the same controlled reference state. | Repeat SOC-thickness loop with temperature, rest, preload and control cells fixed. |
| Gas generation / pouch inflation | Persistent volume or thickness increase; may be local or broad and may correlate with storage, voltage or temperature history. | Review charge limits, storage SOC/temperature, self-discharge, residual trend and supplier analytical evidence where appropriate. |
| SEI/interphase growth, electrode reorganization or cracking | Residual growth can accompany resistance rise and capacity/energy loss. | Trend DCIR/pulse data, delivered Wh/capacity and thickness; specialist diagnostics only where justified. |
| Lithium plating risk | Risk increases with aggressive charging, especially at low temperature; appearance alone is not diagnostic. | Audit cell temperature, charge current, voltage/taper behavior, maker limits and specialist evidence. |
| Formation / degassing / moisture / seal or process issue | Abnormality clusters by lot, production date or process/change point and can appear early in life. | Quarantine lot; review moisture control, formation/degassing/aging, sealing/leak checks, retained samples and change history. |
| Electrical / thermal application abuse | Overvoltage, deep discharge, hot spots, overcurrent or abnormal self-discharge. | Retrieve charger/BMS logs and repeat only within an approved instrumented final-device plan. |
| Finished-pack / enclosure stack-up | Local maximum at PCM, label, foam, harness or housing feature while cell-body measurement remains stable. | Profile the complete pack and run a local CAD/tolerance interference review. |
Temperature and high-SOC exposure often confound swelling complaints. Use the Battery Pack Thermal Management Guide to locate product hot spots, and use Reference 3 only as commercial NMC/graphite evidence that thickness, gas, resistance, capacity and lithium plating should be investigated together [3].
7. Calculate Enclosure Margin Without Double Counting
There is no universal swelling clearance. Build a local worst-case stack from controlled upper-bound inputs, then remove any allowance already included in the supplier’s maximum finished-pack dimension.
Mechanical margin = T_cavity,min − [T_pack,BOL,max + ΔT_SOC,max,qualified + ΔT_EOL,upper + T_stack,max + U_mech]
Require: Mechanical margin ≥ G_project.
Every added term must be a controlled upper-bound value under its stated tolerance, confidence or qualification basis—not a typical value. Omit any term already included in another maximum.
T_cavity,min — minimum local usable cavity after enclosure tolerance, keep-outs and relevant environmental dimensional change.
T_pack,BOL,max — maximum beginning-of-life finished-pack thickness, including the controlled PCM, insulation, label/tape and local assembly features.
ΔT_SOC,max,qualified — qualified maximum reversible SOC/temperature excursion not already included in T_pack,BOL,max.
ΔT_EOL,upper — approved upper-bound residual life growth at the defined reference state, including its stated statistical or qualification basis.
T_stack,max — maximum foam, adhesive, mounting or other external mechanical stack not already included.
U_mech — applicable measurement/fixture uncertainty not already embedded in another upper bound; G_project — separate structural and assembly guard band based on project risk.
MECHANICAL RED LINE
Positive free-thickness margin does not approve a constrained design by itself. If the pouch contacts a compression pad or enclosure surface, also validate the exact cell maker’s allowed force/pressure range, distribution, interface material, worst SOC/temperature condition and life behavior.
8. Stop, Pause or Continue: Make the Decision Explicit
| Decision | Trigger examples | Required action before release |
|---|---|---|
| STOP normal testing | Visible ballooning, leakage, unusual heat/odor, mechanical damage, abnormal voltage/self-discharge, known abuse, or force that threatens the product structure. | Stop normal cycling; identify and isolate affected samples/lot under the approved safety process; escalate to qualified battery personnel. |
| PAUSE release / contain | Progressive reference-state growth, force rising beyond the project envelope, lot clustering, unexplained electrical degradation, result near a decision boundary or missing traceability. | Hold release; preserve sample/lot identity; review uncertainty, charger/temperature history and supplier evidence; compare retained/control samples; define corrective action and retest. |
| CONTINUE controlled evaluation | Repeatable reversible movement or small unresolved change inside the validated measurement band, with no correlated abnormal electrical/thermal behavior or product interference. | Continue only under the approved controlled test plan; trend multiple samples and maintain the same reference condition. |
| RELEASE | Safety, electrical, dimensional, force/pressure, enclosure, process and final-device requirements are closed for the production-intent configuration. | Approve controlled drawing, test plan, limits, change control and production monitoring. |
A sample can remain suitable for controlled investigation without being approved for production or shipment. Separate evidence collection from release authority.
9. Contain the Lot and Audit Supplier Evidence
When persistent growth is suspected, preserve traceability before samples are mixed, cycled further or destructively analyzed. The supplier response should go beyond saying that the cell is inside a nominal thickness.
| Evidence / containment item | Why it matters |
|---|---|
| Affected lot/date/trace ID, immediate containment and retained/control samples | Preserves identity, prevents mixing and creates a valid comparison group. |
| BOL distribution plus paired reference-state thickness/force trends | Shows production tails and separates reversible movement from progressive residual or force growth. |
| Formation, degassing, aging/quarantine, moisture, seal/leak and relevant process-change records | Tests early-life gas, process stability, seal integrity and change-point hypotheses without requiring unsupported conclusions. |
| Capacity/usable Wh, DCIR or pulse response, self-discharge and charge/temperature logs | Correlates mechanical change with electrical and thermal state. |
| Lot disposition, corrective action, validation result and revalidation trigger | Closes recurrence risk and prevents unapproved substitution or process drift. |
Use the PCM/BMS Selection Guide for protection-related evidence and the Battery Pack Prototype Testing Guide for production-intent validation and change control.
10. Release the Production-Intent LiPo System
Release the population and process—not one visually acceptable sample. Close five linked gates for the production-intent configuration: safety; qualified free-thickness and constrained-force envelopes; correlated electrical/thermal behavior; worst-case enclosure margin; and lot/process/change control. Record the evidence and revalidation triggers in the controlled prototype-to-production plan.
FINAL APPROVAL RULE
Do not approve a LiPo pouch cell because its thickness looks acceptable at one SOC, and do not reject it because one unconstrained reading increased by a fixed number of millimeters. Approve the exact production configuration only when free-thickness trend, constrained force/pressure, measurement uncertainty, electrical/thermal evidence, enclosure margin and supplier production controls all remain inside the project-specific validated envelope.
What the Buyer Should Send – and What the Supplier Should Return
Buyer input should identify the cavity and local load limits, sample/lot history, reference measurement method, charge and temperature exposure, plus capacity/Wh, DCIR and self-discharge evidence. The supplier response should return the controlled finished-pack drawing and BOL distribution, comparable paired thickness/force data, applicable charge and aging limits, retained/control comparisons, lot disposition, corrective action and change-control plan.
Key Takeaways
Safety triage comes before dimensional analysis; a suspect pouch is not a normal test sample.
Separate reversible breathing, residual reference-condition growth and external finished-pack stack-up.
Use multiple samples and a capable measurement system; small changes near Gauge R&R or uncertainty remain unresolved.
Free thickness can appear stable while constrained reaction force rises.
Calculate free-space margin without double counting, then validate force/pressure separately when contact is intended.
FAQ: LiPo Pouch Cell Swelling
Is every increase in LiPo pouch-cell thickness abnormal?
No. Reversible SOC-dependent movement can occur. Confirm return to the controlled reference state, trend residual growth, and verify thickness and force against the project envelope.
How should an OEM measure LiPo pouch-cell swelling?
Proceed only after the safety gate. Define SOC, temperature, rest, location, flat contact geometry, preload, dwell time, instrument and uncertainty; map local maxima separately.
Can a hand-held caliper approve pouch-cell thickness?
A caliper can screen samples, but jaw area and operator force deform soft pouches. Production decisions need a validated method with suitable repeatability and reproducibility.
What is the difference between breathing and residual swelling?
Breathing is largely reversible SOC-dependent change. Residual swelling is the persistent increase remaining after SOC, temperature and rest are normalized.
How much clearance should I leave around a LiPo pouch battery?
There is no universal value. Compare the minimum cavity with the maximum finished stack, adding only qualified SOC, life and mounting terms not already included.
Can a rigid enclosure prevent swelling?
It can limit displacement while reaction force rises. Constrained designs therefore require qualified force/pressure evidence as well as free-thickness margin.
Does LiPo swelling always mean gas generation?
No. Gas is one contributor; reversible breathing, electrode/interphase change, lithium plating risk, process abnormalities and pack stack-up can also affect thickness.
Can a BMS or fuel gauge detect a swelling LiPo cell?
Some devices flag correlated conditions. TI BQ27Z846 can use impedance change as a potential-swelling indication, but it is not a thickness sensor and cannot replace mechanical validation [4].
Is one swollen sample enough to reject a production lot?
One abnormal sample can justify containment, but lot disposition needs traceability, controls/retains, capable measurement, distribution data and root-cause evidence.
Conclusion
LiPo pouch-cell swelling is not closed by one millimeter value. The release decision must separate reversible and residual change, distinguish signal from measurement and population variation, and determine whether an enclosure converts expansion into force. Correlating that mechanical evidence with electrical, thermal and manufacturing records prevents both unnecessary loss of energy density and release of a product whose rigid housing hides a growing load.
Need an Engineering Review for a Tight LiPo Cavity?
Send the controlled battery drawing, minimum cavity and structural limits, sample and lot IDs, charge and temperature history, reference-state thickness or force data, and relevant electrical evidence. THOR Power can help separate cell swelling, pack stack-up and enclosure pressure before production release.
Talk to a Battery EngineerTechnical References
- Wang H. et al. — The interplay between stack pressure, mechanical expansion and degradation pathways in lithium-ion batteries. Nature Energy 11, 1032–1042 (2026). Pressure-controlled Gr/NMC811 study; cell- and fixture-specific.
- Aufschlaeger A. et al. — High precision measurement of reversible swelling and electrochemical performance of flexibly compressed 5 Ah NMC622/graphite lithium-ion pouch cells. Journal of Energy Storage 59 (2023) 106483.
- Gasper P. et al. — Lithium loss, resistance growth, electrode expansion, gas evolution, and Li plating in commercial NMC-Gr pouch cells. Journal of Power Sources 604 (2024) 234494.
- Texas Instruments — BQ27Z846 Dynamic Z-Track Gauge datasheet, SLVSIU4 (April 2026). Optional impedance-correlated indication; not a direct thickness sensor.
- IEC 61960-3:2017 — Portable-cell performance, designation, marking and dimensions; Annex A covers laminate-film dimensions, not a universal swelling allowance.
- Jeevarajan J. — Safety and Long-Term Performance of Lithium-ion Pouch Cells. NASA NTRS (2012). Historical off-nominal swelling context; not a universal modern small-format model.
Evidence standard: Published research findings remain tied to the cited chemistry, cell format, fixture and test condition. Numerical examples and diagrams in this article are illustrative engineering screens, not universal LiPo swelling limits or THOR Power first-party performance claims. Final approval must use the exact production cell/pack documentation, a validated measurement system and controlled application data.
Dr. Maximilian Weber is THOR Power's Chief Scientist and a senior expert in lithium battery technology. His technical review focuses on battery safety, performance optimization, energy density and custom battery solution development.
Last technical review: September 2026
Victor Xiong holds a Master's degree from The Chinese University of Hong Kong, Shenzhen. He leads THOR Power's OEM Division and focuses on custom battery solutions for global device brands, product developers and industrial customers.