Residual Moisture In A Lyophilized Vial, And How It Is Measured

Lyophilization ends when the process ends, not when the water runs out. Secondary drying strips water that is adsorbed onto the amorphous solid, and that removal slows asymptotically: the last fraction of a percent costs disproportionate time at elevated shelf temperature. Cycles are therefore stopped at a target, and the target is a residual moisture specification expressed in percent water by mass.

Karl Fischer titration is the reference technique for that measurement. It is a chemical assay rather than a drying experiment: water participates in a defined reaction with iodine and sulphur dioxide in an alcohol solvent, and the quantity of iodine consumed is stoichiometric with the quantity of water present. Because the reaction responds to water specifically, the result is not confounded by solvent or by material that simply evaporates when heated.

The measurement matters because water is the mobile phase for most degradation chemistry in a solid. Amide hydrolysis, deamidation and other bond-level changes need water available to the peptide backbone, and their rates in a dried solid track residual moisture far more closely than they track the nominal storage temperature.

The two Karl Fischer configurations

Volumetric and coulometric Karl Fischer share the same chemistry and differ in how iodine reaches the vessel. Volumetric titration adds iodine from a standardised burette and suits samples containing milligram quantities of water. Coulometric titration generates iodine electrochemically inside the cell, so the charge passed is the measurement, and its working range sits at microgram quantities. Lyophilized material sits at the coulometric end, because a whole small vial may hold less than a milligram of water in total.

A related method, loss on drying, weighs a sample before and after heating and attributes the whole mass difference to water. For a freeze-dried peptide that assumption is unsafe, because residual organic solvent from synthesis and cleavage leaves the same way. Loss on drying and Karl Fischer on the same cake can therefore disagree, and the gap between them is itself informative rather than an error in either.

What a number looks like when it is a record

Percent moisture alone is an adjective. The same cake can read differently depending on sample mass, on whether an oven accessory was used to drive water out of the solid into the cell, and on the oven temperature chosen. A usable record names the method and the conditions: for example, 1.8% water by mass by coulometric Karl Fischer, 22 mg sample, oven accessory at 110 °C, per USP <921> Method Ic.

• Method and mode: coulometric or volumetric, and the pharmacopoeial method number.
• Sample mass in milligrams, because the detection floor is absolute and not proportional.
• Direct transfer into the cell, or transfer via a heated oven accessory with its set temperature.
• Number of replicates, since single determinations on small samples carry visible scatter.
• Whether the figure is for one vial or a composite of several.

Typical published specifications for lyophilized peptide material fall in the range of roughly 1% to 5% water by mass, with tighter limits where the formulation is known to be moisture-sensitive. A specification stated without a method is not comparable to one stated with a method, even when the two numbers look alike.

The failure mode: the room is wetter than the sample

The characteristic Karl Fischer failure is contamination of the sample by atmospheric water during handling, which inflates the reported moisture. A 20 mg cake at 2% water by mass contains about 400 µg of water in total. Ambient air at 22 °C and 50% relative humidity holds roughly 10 mg of water per litre, so a few millilitres of room air entering the cell, or a hygroscopic cake sitting open on a balance pan for a minute, is enough to move the result by an amount that matters.

It shows up as a high result with poor agreement between replicates, and as a drifting blank between determinations. The controls against it are a measured cell blank before each run, transfer inside a dry glovebox or by oven accessory so the solid is never exposed, and replicate determinations whose spread is reported rather than averaged away. A single moisture figure with no blank and no replicate spread behind it is a weaker record than it appears.

Where residual moisture sits in the published record

Residual moisture is not among the specifications Aurum publishes. The material listed is independently assayed for purity by reverse-phase HPLC and for identity by mass spectrometry, and those are the assays the published figures come from. A water content figure is a separate determination on a separate sample, and no such figure appears in the published record for a lot.

Our published purity figures are reverse-phase HPLC purity at a stated wavelength. That is a statement about the relative area of chromatographic peaks, and it is silent on how much water sits in the cake alongside the peptide. Neither figure substitutes for the other: a cake can be high in HPLC purity and high in residual moisture at the same time, because the two measurements are looking at different things.

Sterility, endotoxin and pyrogen testing are not performed and not claimed, and residual moisture has no bearing on any of them. A dry cake is a poor environment for microbial growth, which is sometimes read as a substitute assurance. It is not one. Low water activity slows growth; it does not establish that a container was filled under conditions that would support a sterility claim.

Common questions

Why is any water left at the end of drying?

Because the last water is bound to the solid rather than frozen as ice. Removing it is desorption, not sublimation, and desorption slows as the remaining water becomes more tightly held. Driving it to zero requires temperature and time that a cycle is not usually willing to spend.

Does a lower moisture figure always indicate a better cake?

Not universally. Some amorphous solids are physically less stable when dried very aggressively, and over-drying has been reported to destabilise certain formulations. The specification exists as a range for that reason, not as a target of zero.

Can moisture be read from the appearance of the cake?

No. A visibly collapsed or shrunken cake suggests something went wrong thermally, but an intact white cake gives no reading of water content. The measurement requires an assay.

Is Karl Fischer affected by residual synthesis solvent?

Some solvents interfere with the reaction chemistry and some do not, which is why method validation includes a specificity check for the solvents plausibly present. Residual solvent is measured separately, by gas chromatography against the limits in ICH Q3C.

How much sample does a determination consume?

Tens of milligrams per replicate, and the sample is destroyed. For a vial containing a few milligrams of peptide, a moisture determination means the vial is spent.