Karl Fischer Titration (KF)

A 1 g sample of an active pharmaceutical ingredient (API) that holds 0.1% water contains 1 mg of water. The United States Pharmacopeia (USP) chapter 921 recommends sizing the sample so at least 2 mg of water is titrated, so that sample is too small for a typical volumetric setup. That arithmetic is the main reason low-water results need a different setup.

This entry defines the method, shows how to choose between its two forms, and gives the working to check a supplier's figure yourself. The calculations are built from the published USP chapter, so you can rerun them.

Iodine and sulfur dioxide react with water in an alcohol medium, and the amount of iodine used gives the amount of water. In volumetric KF the iodine arrives in a reagent solution. In coulometric KF it is generated by electric current in the cell.

Volumetric KF depends on the reagent's strength, its titer, also called the water equivalence factor F (mg of water per mL of reagent). USP allows F to be set against Purified Water, sodium tartrate dihydrate, or a traceable commercial standard. Sodium tartrate dihydrate is 15.66% water (36.04 ÷ 230.08).

Substances that react with iodine, such as ascorbic acid, mercaptans or peroxides, give falsely high results, so the method must be shown to suit the material. Check that the COA names a method suited to the material.

Water in an API is a small number that moves large ones. It changes the assay you calculate, the quantity you weigh for a batch, and whether a moisture-sensitive material is safe to store.

Yet a line on a certificate of analysis (COA) can read only "Water (Karl Fischer): 0.4%", with no method, no basis and no sample size behind it.

ICH Q6A says a water test matters when a substance is hygroscopic (moisture-absorbing), degraded by moisture, or a stoichiometric hydrate.

Use coulometric KF below about 0.1% water, volumetric KF above about 1%, and either between (a working rule, not a USP limit). The deciding factor is whether you can weigh enough sample to titrate at least 2 mg of water.

Which KF form fits which water level

USP estimates the minimum sample as F × C × V ÷ KF. C is the percent of the buret (the graduated reagent tube) used, V is the buret volume in mL, and KF here is the expected water in percent. USP recommends F × C × V of at least 200, and a reagent with F of 2.0 or less for water under 1%.

Below about 0.1% water, volumetric KF needs a sample of 2 g or more, so coulometric KF is the better choice. USP calls coulometry a micro-method and requires the sample to be compatible with the cell chemistry. Above about 1% water, volumetric KF is straightforward.

Worked through with a 10 mL buret and 10% of it used:

Divide the as-is assay by the dry fraction: anhydrous assay = as-is assay × 100 ÷ (100 − water %). The same idea, subtracting KF water, underlies the mass-balance purity calculation for reference standards.

Worked examples:

• 98.0% as-is with 2.0% water: 98.0 × 100 ÷ 98.0 = 100.0%.
• 90.0% as-is with 8.0% water: 90.0 × 100 ÷ 92.0 = 97.8%.

The same arithmetic tells you how much to weigh. See Assay on Anhydrous Basis. For peptide APIs, read the water figure next to Peptide HPLC Purity vs. Content.

Run these five checks on every COA, and ask the supplier for the sample weight and reagent F where the COA omits them. Each has a pass and a fail. The first three are below.

1. Method named and suited to the material. Pass: the COA cites a pharmacopeial water method suited to the material (for KF, USP chapter 921 Method Ia or Ic, or the pharmacopeia the specification names). Fail: "moisture" with no method, or a method not shown to suit the material.
2. Lot-specific. Pass: a measured percent tied to your lot number. Fail: only the specification limit.
3. Basis stated. Pass: assay is marked as-is or anhydrous. Fail: unmarked.

4. Arithmetic repeats. Pass: converting as-is to anhydrous with the formula above matches the COA within rounding. Fail: it does not.
5. Method fits the level. Pass: water below 0.1% was run coulometrically, or the supplier states reagent F, buret volume and sample weight and the weight is at least F × C × V ÷ water % (from USP chapter 921). Fail: a volumetric run at or below 0.1% water with a sample under about 2 g, or no sample weight available on request.

USP says precision depends largely on excluding atmospheric moisture, so seal the cell and weigh hygroscopic material quickly. For what a COA can and cannot prove, see 21 CFR Part 111: What a Supplier's COA Can and Can't Prove and ICH Q7 Explained.

KF and loss on drying disagree because LOD is a weight loss on heating, while KF reacts with water. LOD is the weight lost when a sample is heated. USP reserves the LOD heading for cases where "the loss sustained on heating may be not entirely water", and directs articles containing water of hydration to the water methods the monograph names (titrimetric, azeotropic or gravimetric).

So a solvent-wet batch reads higher by LOD than by KF. ICH Q6A says "a detection procedure that is specific for water (e.g., Karl Fischer titration) is preferred."

• Myth: Coulometric Karl Fischer needs a titer, just as volumetric does. Fact: USP states that calibration of a coulometric instrument is not necessary, because the current consumed is measured absolutely.
• Myth: The water percent on a COA can be dropped straight into an assay. Fact: The assay must be marked as-is or anhydrous, and the conversion between them has to repeat within rounding.
• Myth: Volumetric KF suits every water level. Fact: Below about 0.1% water it needs a sample of 2 g or more, which is why coulometric KF is the better choice there.

Karl Fischer titration is among the quality tests for Aarise Pharmaceuticals' APIs, and its test methods reference the USP and the European, British and Indian Pharmacopoeia (EP, BP, IP). The tests are described in its Pharmaceutical API Quality Testing Guide.

Available documentation may include a COA, specifications and safety data sheet (SDS/MSDS) where applicable, and the company confirms what is available for each product at the quotation stage. It also supplies evaluation quantities to qualified B2B buyers.