Residual moisture in lyophilized peptides: what a Karl Fischer number actually measures

A certificate of analysis for a lyophilized peptide typically carries three or four numbers that receive nearly all the attention: chromatographic purity, mass confirmation, peptide content, and sometimes counterion identity. Water content, when it appears at all, is usually a single line near the bottom — a percentage, occasionally a method name, rarely anything else. It is treated as a formality.

It is not a formality. Residual moisture is one of the few CoA parameters that is simultaneously a quality attribute of the material as manufactured and a predictor of how that material will behave over the following months. Water is a reactant in hydrolysis, a participant in deamidation, and a plasticizer that governs molecular mobility in the amorphous solid. A peptide with the same sequence, same purity, and same counterion can show substantially different stability behavior depending on whether the cake it was dried into holds one percent water or six. Understanding what the reported number represents — and what it does not — is the difference between reading that line and using it.

Why water in the solid state is not inert

Lyophilized peptide is an amorphous glass, not a crystal. The molecules are arranged without long-range order, frozen into a metastable state by the removal of solvent. In that state, the property that governs chemical degradation rates is molecular mobility: how much local motion the matrix permits at storage temperature.

Water is the dominant plasticizer of such a matrix. Adding water lowers the glass transition temperature, the point at which the solid changes from a rigid glass to a rubbery state with markedly greater mobility. The relationship is steep. A few percent of additional moisture can depress the glass transition temperature of an amorphous solid by tens of degrees. A formulation whose glass transition sits comfortably above ambient temperature when dry can, at elevated moisture, have a glass transition near or below the temperature at which the material is actually being stored. When that happens, degradation rates that were previously governed by glassy-state kinetics accelerate, and the solid behaves less like a dry powder and more like a concentrated solution.

Water also participates directly in the chemistry. Hydrolysis of labile bonds requires it. Deamidation of asparagine proceeds through a succinimide intermediate whose ring opening is a hydrolysis step. Aspartimide-related rearrangements have similar requirements. In each case the reaction is not merely permitted by residual water — its rate depends on water activity in the solid.

The relationship between moisture and degradation is not always monotonic, which is worth keeping in mind. In some systems, very low residual moisture has been characterized as destabilizing rather than stabilizing, because a minimal hydration shell contributes to conformational integrity, and stripping it drives structural rearrangement or aggregation on reconstitution. Studies of protein and peptide lyophilizates have repeatedly identified an optimal moisture window rather than a monotonic “drier is better” rule. The practical consequence is that a very low reported water content is not automatically a better result; it is a different result, and whether it is better depends on the molecule and the formulation.

What Karl Fischer titration measures

Karl Fischer titration is the reference method for water determination, and it is specific for water in a way most alternatives are not. The underlying reaction consumes water stoichiometrically: iodine and sulfur dioxide react in the presence of water and an alcohol, with a base present to maintain conditions, producing iodide and an alkyl sulfate. One mole of water consumes one mole of iodine. That stoichiometry is what gives the method its specificity — the titration is not measuring mass loss or a bulk property, it is measuring a chemical reaction that only water drives.

Two implementations are used. Volumetric Karl Fischer delivers iodine from a standardized reagent through a burette and suits samples with relatively high water content, typically above roughly one percent. Coulometric Karl Fischer generates iodine electrochemically in situ and measures the charge required, which allows much smaller quantities of water to be determined accurately. For lyophilized peptides, where sample amounts are small and water content typically sits in the low single-digit percent range or below, coulometric determination is generally the appropriate choice.

Sample introduction matters as much as the titration itself. Direct dissolution of the solid into the titration cell is common but requires that the sample dissolve fully and that nothing in it interferes. An oven accessory offers an alternative: the sample is heated in a sealed vessel and liberated water is carried by dry gas into the cell, leaving the solid behind. This avoids dissolution problems and sidesteps certain interferences, though it introduces its own variable in the chosen oven temperature, which determines how much tightly bound water is driven off.

Interferences exist and are worth knowing. Strongly basic samples can shift the pH of the titration medium outside the range where the reaction proceeds cleanly. Thiols and some sulfur-containing species react with iodine directly, producing an apparent water content that is too high. Aldehydes and ketones can react with the methanol in classic reagent formulations, generating water as a side product and giving a drifting, inflated endpoint; ketone-specific reagents exist for this reason. None of these are exotic in peptide work, and a method that has not been assessed against the specific material is producing a number of uncertain provenance.

Loss on drying is a different measurement

Loss on drying is often reported in place of Karl Fischer, and the two are frequently treated as interchangeable. They are not.

Loss on drying measures mass lost when a sample is held at a defined temperature for a defined time. Everything volatile under those conditions contributes: water, residual solvents from synthesis or purification, volatile counterion species, degradation products that leave the sample as gas. In a peptide that has been through reversed-phase purification, residual acetonitrile and trifluoroacetic acid are plausible contributors, and volatile acid counterions in particular can be driven off under drying conditions. A loss-on-drying result is therefore an upper bound on water content, not a measurement of it.

The bias can also run the other way. Tightly bound water may not be released at the drying temperature used, in which case loss on drying underreports even the water fraction. Thermogravimetric analysis, which tracks mass continuously as temperature rises, provides more information about the shape of the loss curve but shares the fundamental limitation: it measures mass, not water. Pairing it with an evolved-gas technique restores specificity, but that combination is uncommon outside development laboratories.

The practical reading is straightforward. When a CoA reports water content, the method matters, and Karl Fischer with a stated implementation is a substantively stronger claim than a loss-on-drying percentage. When the method is not stated at all, the number should be treated as approximate.

Why the number moves after it is measured

A residual moisture value describes the material at the moment of testing, in the container it was tested from. Lyophilized peptides are hygroscopic — the amorphous solid has considerable capacity to take up water from the atmosphere — and the value can change with handling in ways that have nothing to do with the manufacturing process.

Several routes contribute. Ambient exposure during weighing or aliquoting allows direct uptake, and the rate depends on relative humidity, which varies substantially by season and location. Stoppers used in vial closure are not perfect barriers; elastomeric closures have finite moisture permeability, and stoppers that have been steam-sterilized retain water that can subsequently desorb into the headspace and be taken up by the cake. Over long storage, moisture transfer from closure to product is a documented mechanism, which is why stopper drying conditions are a controlled parameter in lyophilization practice rather than an afterthought.

The most common handling error is a temperature one. A vial removed from cold storage and opened before it has equilibrated to room temperature will condense atmospheric moisture on cold interior surfaces, and the cake will take that water up directly. The amount involved is not trivial relative to the low percentages under discussion. Allowing sealed vials to reach ambient temperature before opening is a small procedural step with a measurable effect on the moisture content of what remains in the vial.

For analytical sampling, the same physics dictates that a moisture determination should be performed on material with minimal ambient exposure, transferred under dry conditions where practical. A Karl Fischer result generated from a sample that sat open on a balance for several minutes in a humid room is measuring the room as much as the material.

Reading the number in context

A water content figure becomes interpretable when it is paired with three other things: the method used to obtain it, the formulation it applies to, and the storage condition the material is intended for. Karl Fischer coulometry on a stated sample mass supports a different level of confidence than an unlabeled percentage. A value from a peptide lyophilized without excipients means something different from the same value in a bulking-agent formulation, because the glass transition behavior of the two matrices differs. And a moisture level that is unremarkable for material held at low temperature may sit closer to a mobility threshold for material stored warm.

None of this makes water content a parameter to optimize in isolation. It is one term in a stability picture that also includes sequence liabilities, counterion identity, oxygen and light exposure, and the container closure system. But it is the term most often reported without being read, and among CoA parameters it is unusual in describing not only what the material is but how it is likely to change.