For most R&D teams, in-house vs outsourced battery failure analysis is not an either-or choice. Keep fast electrical checks in-house and send teardown and materials characterization to a specialist lab. Build a full in-house capability only when failures come often enough to keep the instruments and analysts busy.
This guide is for engineers and managers with failed cells on the bench, from a cycling test, a swollen pouch or a field return, who need a root cause and have to decide who finds it. It compares both routes on cost, time, instruments, IP and independence, and shows the hybrid split many teams use.
A note before we start: Canrud offers Canrud’s failure and root cause analysis service, so we have a stake in this question. We have tried to say plainly where an in-house lab is the better choice.
What does battery failure analysis involve?
Battery failure analysis works out why a cell failed. It moves from non-destructive to destructive steps, so that evidence is recorded before the cell is opened. The battery failure analysis techniques fall into three stages.
1. Non-destructive checks
The cell is inspected and measured while it is still sealed: visual condition, swelling, mass and dimensions, open-circuit voltage, resistance and impedance. X-ray or CT imaging shows the inside of the cell, such as electrode alignment, deformation or foreign particles, before anything is cut.
2. Teardown and characterization
The cell is discharged under control and opened in an inert atmosphere, usually an argon glovebox, because lithium and many cell materials react with air and moisture. The analyst inspects the electrodes, separator and casing, then takes samples for characterization:
- SEM to image surfaces and particles at high magnification.
- EDS to map which elements are present, for example metal contaminants or deposits.
- XRD to check the crystal structure of the active materials.
- ICP-OES to measure elemental composition, for example metals that have moved from cathode to anode.
3. Root cause reasoning
The findings are matched to a mechanism, such as lithium plating, separator damage, a contaminant particle, electrolyte loss or a build defect like misaligned electrodes. This step needs experience as much as instruments: the same image can point to different causes depending on the cell history.
The in-house route: what it takes
Doing failure analysis in-house means owning every stage above. You need:
- A glovebox or dry room and safe tools for opening cells.
- Access to imaging and characterization instruments (CT, SEM/EDS, XRD, ICP), either your own or a shared university or company facility.
- Analysts trained in both cell chemistry and the instruments.
- Safe storage and handling for damaged cells, including test chambers with fire suppression and remote monitoring.
The costs are mostly fixed: space, electrical capacity, calibration, maintenance and salaries. They do not go away in months when nothing fails. In return, you control the schedule, can open a cell the day it fails and never ship a sample out.
In-house fits when: failures are frequent and steady, the chemistry is highly confidential, you already own most of the instruments, and fast answers matter more than the setup cost.
The outsourced route: what you get and what you give up
Outsourcing turns those fixed costs into a per-project cost. You get instruments you do not have to buy, analysts who see many failure modes across many cell types, and capacity when several failures arrive at once. An outside report is also independent, which helps when you need to discuss a failure with a cell supplier or customer.
What you give up: damaged cells must be packed and shipped under hazardous materials rules, you wait in the lab’s queue, and you share samples and some data. You also rely on the lab to explain its reasoning, not just its conclusion.
US rules for damaged, defective or recalled lithium cells are set out in 49 CFR 173.185. Agree packaging with the lab and your carrier before you ship.
Outsourcing fits when: failures come in bursts, you lack one or more key instruments, you need an independent report, or your team’s time is better spent on design than on analysis.
In-house vs outsourced battery failure analysis compared
| Factor | In-house | Outsourced |
|---|---|---|
| Upfront cost | High: lab space, glovebox, instruments, safety systems | None: no equipment to buy |
| Running cost | Fixed, even when nothing fails | Paid per project |
| Speed to start | Same day, if staff are free | After shipping and queue time |
| Instrument range | Limited to what you own or can book | Wide; whatever the lab runs |
| Expertise | Deep on your own cells, narrower overall | Broad across many cell types and failures |
| IP exposure | Lowest; nothing leaves the site | Managed with NDA, coded samples, data terms |
| Independence | Internal view only | Third-party report, useful with suppliers or customers |
| Scaling | Hard to absorb sudden volume | Capacity on demand |
How long does battery failure analysis take?
It depends on how deep the analysis goes, how many cells are involved and how fast samples move. Non-destructive screening is the quickest stage. A full teardown with SEM, XRD and ICP takes longer, because each technique needs its own sample preparation and its results are read together.
Outsourcing adds shipping time and the lab’s queue, while in-house work adds any wait for shared instruments. Whichever route you choose, ask for a written timeline per stage before work starts, so you can make decisions after screening instead of waiting for the full report.
What is included in a battery failure analysis report?
A useful report lets someone who was not in the lab follow the reasoning from evidence to cause. Expect:
- Sample log: cell IDs, condition on arrival and the history you supplied.
- Electrical data: voltage, resistance or impedance before opening.
- Imaging: photos and X-ray or CT images before teardown.
- Teardown observations: what each component looked like, with photos.
- Characterization results: SEM, EDS, XRD and ICP data with sample locations.
- Root cause statement: the most likely mechanism and how strongly the evidence supports it.
- Recommended actions: what to change in design, materials or process, and what to test next.
If a report states a root cause without the evidence behind it, ask for the data. The same rule applies to an in-house report.
How to protect your IP when you send cells out
Treat outsourced analysis like any other confidential supplier relationship:
- Sign an NDA before any cell or data leaves your site.
- Label samples with codes, not project or customer names.
- State in writing who owns the raw data and the report.
- Agree whether samples are returned or destroyed, and get confirmation.
- Share the cell history the analysis needs, not full design files.
For what to check before you pick a lab, see these red flags when choosing an analysis lab. For contract terms in more depth, read our guide on protecting your IP when outsourcing tests.
The hybrid route most R&D teams settle on
Most R&D teams split the work by stage. The early checks are cheap and fast to do yourself, while the later stages need instruments and experience that are expensive to keep idle.
| Keep in-house | Send to a specialist lab |
|---|---|
| Logging the failure and cell history | Opening cells in an inert atmosphere |
| Visual inspection and photos | SEM/EDS, XRD and ICP-OES |
| OCV, resistance and impedance | CT imaging, if you do not own a scanner |
| Safe quarantine of failed cells | Root cause report, when independence matters |
This keeps your engineers close to the data and makes the outside lab’s work faster, because the cells arrive with clean history and baseline measurements.
Which route fits your team?
- Failures are rare or come in bursts: outsource, and keep screening in-house.
- Failures are frequent, steady and highly confidential: build in-house capability, and outsource only the instruments you lack.
- You need a result a supplier or customer will accept: use an independent lab for that case, even if you have an in-house lab.
Know the limits of either route. Failure analysis gives the most likely root cause, not always a certain one. If a cell burned in thermal runaway, much of the evidence may be destroyed and the report may only narrow the possible causes. In those cases, testing more cells from the same lot is often the next step.
Frequently asked questions
Can we do part of the failure analysis in-house and outsource the rest?
Yes, and most R&D teams do. Keep electrical screening, logging and photos in-house, where they are fast and cheap. Send cell opening and materials characterization to a lab that already runs a glovebox, SEM, XRD and ICP. Agree the hand-over data format first so nothing is measured twice.
How do we ship damaged or swollen lithium cells to a lab?
Damaged and defective lithium cells have their own US shipping rules, separate from normal cells. They sit in paragraph (f) of 49 CFR 173.185 and include stricter packaging. Agree the packaging and carrier with the lab before you pack anything, and use staff trained to ship hazardous materials.
Should we send a good cell along with the failed ones?
Where you can, yes. A healthy cell from the same build or lot gives the analyst a baseline. Without one, it is harder to tell a failure feature from something every cell in that batch has. Send the cell history too: cycling data, conditions and when the failure appeared.
Does outsourcing failure analysis put our IP at risk?
Some risk exists, but you can manage it. Sign an NDA before shipping, label samples with codes instead of project names, and set data ownership and sample return or destruction in writing. Share the cell history the analysis needs, not your full design file. A good lab will expect all of this.
| Need a root cause for failed cells?
The Canrud team runs teardown and characterization for battery R&D teams. Tell us the cell format, chemistry and failure history, then send us your failed cells. Each analysis is quoted after a short scope call, all costs listed upfront. |