Loss on Drying vs Karl Fischer
Loss-on-Drying (LOD, USP <731>) and Karl Fischer titration (KF, USP <921>) both measure 'moisture' — but not the same moisture, and they are not interchangeable. Substituting LOD for KF (or vice versa) where the monograph or spec requires the other is a recurring 21 CFR 211.84 finding.
How does Loss on Drying vs Karl Fischer apply to your shop floor?
Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.
01What each method actually measures
| Method | What it detects | Reference |
|---|---|---|
| Loss on Drying (LOD) | Total volatile mass loss on heating — water + residual organic solvents + volatile decomposition products | USP <731>, Ph. Eur. 2.2.32 |
| Karl Fischer (KF) | Water specifically, via chemical reaction (iodine + SO₂ + water) | USP <921>, Ph. Eur. 2.5.12 |
This is the single most important sentence on this page: LOD is a gravimetric measurement of everything that leaves the sample on heating, and KF is a chemical measurement of water. They agree only when water is the only volatile component present.
02When LOD and KF agree
- Non-hygroscopic, non-solvated crystalline materials with no residual solvents (many finished excipients).
- Materials with a certified low residual-solvent profile per ICH Q3C class 1/2 controls.
- Well-characterised legacy APIs whose monograph explicitly permits either method.
03When LOD and KF diverge (and which to trust)
- Hydrates (water of crystallisation) — KF measures the water; LOD may or may not, depending on drying conditions and the hydrate's dehydration temperature. Use KF.
- Hygroscopic salts — surface water plus bound water; KF gives the true water content, LOD gives 'water plus whatever else volatilises at that temperature.'
- APIs with residual class-2/3 solvents (methanol, IPA, DMSO, DMF) — LOD sees both; KF ignores non-water. Use KF when the spec is water; use both if you need water AND solvents.
- Lyophilised biologics — trace bound water; KF is the standard; LOD damages the protein.
- Herbal/botanical extracts with essential oils — LOD overestimates 'moisture' because oils volatilise; KF isolates the water for the moisture spec.
- Materials where decomposition begins below the drying temperature — LOD includes decomposition products, KF doesn't.
04Which does the monograph require?
USP General Notices 5.50 makes the monograph method authoritative. If the USP monograph specifies KF, LOD is not a valid substitute — regardless of correlation studies you might have done in-house. The same applies to Ph. Eur., JP, and USP-NF excipient monographs. When the spec says 'Water, USP <921>: NMT 0.5%' the required method is KF; running LOD and reporting the result as 'water' is a §211.84 finding and a Data Integrity finding under MHRA guidance.
05Operational differences
| Attribute | LOD | KF |
|---|---|---|
| Instrument | Oven + analytical balance | KF titrator (coulometric or volumetric) |
| Sample size | 1–5 g typical | 50–500 mg (coulometric) / 200 mg–2 g (volumetric) |
| Time per test | 3–4 h (105 °C typical) | 3–10 min |
| Sensitivity | ~0.1% moisture | ppm-level (coulometric) |
| Selectivity | Non-selective (any volatile) | Water-selective |
| Method transfer difficulty | Low | Higher (moisture control critical) |
| Typical use case | General moisture screen; non-hygroscopic solids | Regulated water spec; hydrates; hygroscopics; biologics |
06In the dispense engine
The moisture value drives the correction from anhydrous basis to as-is. If the recipe target is anhydrous and the assay is anhydrous, the moisture correction pulls the anhydrous target up to what the operator weighs: as-charged = anhydrous / (1 − moisture).
Which moisture number goes into that formula? The one that matches the material's actual behaviour: KF-measured water for hydrates and hygroscopic materials, LOD for simple non-solvated solids. Using the wrong number quietly under- or over-doses by the difference between the two — often 1–5% for hydrates.
07Common mistakes
- Substituting LOD for KF because 'the numbers were close last time' — a §211.84 finding waiting to happen.
- Free-text CoA field 'moisture' without the method — recipient site guesses wrong basis.
- KF titrant not standardised on the day of use → water value drifts silently.
- Sample exposure to atmospheric humidity before weigh-in on hygroscopic materials → gain, not loss; both methods over-report.
- LOD at a temperature that partially dehydrates a hydrate → LOD reports 'moisture', but you've measured half of the water of crystallisation.
- Averaging LOD and KF values because they disagree — never do this; pick the correct method and use it.
08How V5 Ultimate handles method selection
Frequently asked questions
Q.Can we do a bridging study to justify LOD in place of KF?+
Sometimes, if the monograph permits alternate methods with equivalency demonstration (some USP monographs allow this, most do not). Even then the bridging study must show equivalence across the full spec range, across seasonal humidity, and across suppliers. The default assumption is: use the monograph method.
Q.For anhydrous salts with residual crystallisation solvents, which method?+
KF for water; a residual solvents test (USP <467>, GC headspace) for solvents. LOD alone conflates the two and is unsuitable.
Q.Do supplier CoAs always declare the method?+
They should, but many don't. Establish a receiving rule: CoAs missing the method get flagged for supplier clarification before acceptance. A missing method is not a passing CoA.
Q.Is method choice a change-control matter?+
Yes. Switching a material from LOD to KF (or vice versa) changes the reported moisture value and cascades into every dispense correction. Route via change control with impact on retrospective batches assessed.
Primary sources
Further reading
V5 Ultimate ships with the Loss on Drying vs Karl Fischer controls already wired in — audit trail, e-signatures, validation evidence. Free trial, no credit card, onboard in days, not months.