What Freeze-Drying Actually Stabilizes

ON THIS PAGE

1. Key findings
2. Methodology
3. From solution to solid
4. Which degradation pathways the drying closes
5. Transit, and the number this article will not state
6. Storage after receipt
7. Limitations
8. Sources

A research peptide arrives as a white cake because water is the reagent that would otherwise take it apart. Freeze-drying takes the water out. What that buys is narrower, and better measured, than the storage rules of thumb this article examines, and the most familiar of them, that staying below the glass transition temperature is what keeps the molecule intact, does not survive a look at the literature it is supposedly drawn from.

Methodology

Four named sources carry the load in this article, listed with their fetch dates at the end. Every sentence below that reports a measured quantity or a published conclusion says which of them it came from. Where no source supports a number, this article says so rather than supplying one.

From solution to solid

Lyophilization runs in three stages: freezing, primary drying by sublimation, and secondary drying by desorption. Freezing converts the solution to an ice matrix, and the freezing rate and endpoint temperature set the ice crystal size and much of the dried cake's structure. Primary drying then pulls the ice off directly under vacuum, at a product temperature held low enough that the cake does not collapse. It is the longest phase of the cycle. Secondary drying raises the product temperature under deep vacuum to strip the unfrozen water still adsorbed to the solid.

How dry is dry

Secondary drying commonly brings water content below 0.5% by weight, and can be completed in three to six hours at a shelf temperature of 40 or 50 °C, according to Tchessalov and colleagues in the 2023 update to their freeze-drying design guidance. The same paper is careful about the direction of the rule: moisture specifications should follow a product's own stability data, because a particular product may be most stable at some intermediate residual water content rather than at the driest point achievable.

A second figure, 1% to 3% water by weight, is often set against that one. Derek Duncan gives exactly that range as the typical pharmaceutical freeze-drying target in La Vague n°83, published by A3P in October 2024. The two numbers are not in conflict, because they are not the same quantity. Below 0.5% is what a designed secondary-drying step can achieve under the conditions Tchessalov and colleagues describe. The 1% to 3% range is where typical cycles are designed to land, and Duncan's surrounding discussion says why it is not lower: large molecules can be over-dried, the same intermediate-moisture caution Tchessalov and colleagues raise. Tchessalov and colleagues separately note that residual water content specifications are written in no-more-than form. Reading a cycle's achievable floor as a target, or reading a target as evidence that a cycle ran wet, is the error.

What the excipients are for, and which is which

Excipients in a lyophilized formulation do at least two unrelated jobs, and conflating them muddles what each excipient is for.

Sucrose and trehalose stay amorphous through drying and are used as lyoprotectants, forming the glassy matrix the peptide sits inside. Mannitol is not in that class. Tchessalov and colleagues group mannitol with glycine as a crystalline bulking agent, a solute whose crystallization during freezing is wanted rather than avoided, precisely because a crystalline phase gives the cake mechanical structure at product temperatures above the amorphous glass transition. Their guidance is explicit that crystallization is desirable for the bulking agents and should be prevented for the buffer and the lyoprotectant. Listing mannitol alongside sucrose and trehalose as a protective sugar mixes up the excipient holding the cake's shape with the one shielding the molecule.

The mechanism usually offered for the sugars is water replacement: sugar hydroxyls substituting for the water hydrogen bonds lost on drying. That account is a hypothesis with mixed support rather than a settled finding. Chang and colleagues, reporting in the Journal of Pharmaceutical Sciences in 2005 on an IgG1 antibody and recombinant human serum albumin dried with sucrose, trehalose or sorbitol, found that storage stability correlated reasonably well with how much native structure survived drying, concluded that glass dynamics alone does not account for sugar stabilization, and judged the evidence for a specific sugar-protein interaction to be not compelling.

Which degradation pathways the drying closes

Moving a peptide from a mobile solution to a rigid solid restricts the molecular motion that chemical and physical degradation require. The pathways do not all close equally.

Hydrolysis and oxidation

Hydrolysis, the cleavage of peptide bonds by water, falls the furthest, because the reactant itself is what has been removed. Oxidation of methionine, tryptophan and cysteine side chains is reduced rather than eliminated: oxygen still permeates a vial, and the usual counter is an inert headspace of argon or nitrogen plus opaque or amber glass against light.

Structure and aggregation

Drying is itself a stress. Secondary and tertiary structure can be perturbed during freezing and water removal, and partial unfolding is a route to aggregation in the solid state. This is what the Chang finding above bears on directly: in that work, how much native structure a formulation retained through drying was the better predictor of how it stored.

Why the glass transition rule is weaker than it sounds

Below its glass transition temperature, an amorphous solid is a rigid glass with very low molecular mobility, and the conventional instruction is to store well below Tg. Tg is a real and useful process boundary. As a predictor of shelf stability it performs badly. Cicerone, Pikal and Qian reported in 2015 that using the gap between Tg and storage temperature to predict relative degradation rates can produce errors of up to a thousandfold, and that correlations built on global glass dynamics fail to carry across different formulations. What does carry, in their account, is beta relaxation: the fast, local molecular motions that correlate with degradation rates consistently across more than a dozen proteins and formulations, with no large formulation-dependent shifts. The practical reading is that Tg tells you where the cake stops behaving like a solid, and tells you much less than is usually claimed about how fast the molecule inside it degrades.

Transit, and the number this article will not state

Lyophilization extends shelf life. It does not confer indefinite stability, and shipping is where that gets tested.

What a transit window actually depends on

The variables are the peptide itself, the excipient system, the residual water, the seal and headspace, and the temperature history inside the box. Exposure that pushes product temperature toward the glass transition raises mobility, and mobility is the thing the drying was meant to remove. Those are the mechanisms, and they are also why a single general number is not available.

None of the sources named below reports an ambient holding window for a specific lyophilized peptide, so this article does not state one. A vendor stating that a compound tolerates a given number of days at room temperature should be asked which stability study produced that number, on which peptide, in which formulation, and at what measured endpoint.

The pathways that stay open in the dry state

• Hydrolysis: reduced sharply, but trace residual water can still drive slow hydrolytic chemistry at susceptible bonds.
• Deamidation: asparagine and glutamine residues can deamidate in the solid state, altering charge and creating a route to aggregation.
• Oxidation: residual oxygen and light exposure still act on susceptible side chains.
• Aggregation: conformational change or partial unfolding can drive association into insoluble species, and mobility above the glass transition makes it faster.

Storage after receipt

Every pathway listed above is temperature dependent, which is the whole argument for cold storage: lower temperature slows the chemistry and reduces the mobility that the drying was designed to suppress. Ordinary laboratory cold storage means a refrigerator at 2 to 8 °C or a freezer at −20 °C or −80 °C. Which of those a given material calls for is a property of that material's own stability data, and this article names no source that establishes it compound by compound.

Moisture ingress is the failure mode that matters

A cycle designed to reach a fraction of a percent residual water is undone by a seal that lets ambient humidity back in. Vials are sealed under vacuum or inert gas with an elastomeric stopper and a crimped cap for that reason, and desiccant in a secondary container is a second line against humidity. Water taken back up does not only restart hydrolysis. It also acts as a plasticiser, lowering the glass transition temperature of the amorphous phase, which moves the material closer to the mobile regime at whatever temperature it is being held.

Once water goes back in

Adding solvent reverses the condition the entire process created. Molecular mobility returns to solution values, and the degradation pathways above resume at solution rates rather than solid-state ones. That is the material-science point, and it is the limit of what this article will say about it. None of the sources named below measured a holding time for a reconstituted research peptide, and this publication does not issue preparation guidance.

Limitations

Lyophilization is a stress process as well as a stabilising one, and freezing or drying can damage a molecule that a poorly chosen excipient system fails to protect. No universal formulation exists. Complete water removal is not achievable, and the residual fraction still matters in ways that depend on where it sits in the amorphous matrix. Two limits are specific to this article. The sources cited below are formulation and process science, not peptide-by-peptide stability data, so they establish mechanisms and design targets rather than shelf lives for any particular compound. And the FDA inspection guide on lyophilization of parenterals, which would have been the regulatory anchor here, did not load when it was tried.

Sources

• Tchessalov S, Maglio V, Kazarin P, Alexeenko A, Bhatnagar B, Sahni E, Shalaev E. "Practical Advice on Scientific Design of Freeze-Drying Process: 2023 Update." Pharmaceutical Research 40(10):2433–2455, 2023. Read on PubMed Central, accession PMC10661802, fetched 2026-09-01. Basis for the secondary-drying moisture range and for mannitol as a crystalline bulking agent.
• Cicerone MT, Pikal MJ, Qian KK. "Stabilization of Proteins in Solid Form." Advanced Drug Delivery Reviews 93:14–24, 2015. Read on PubMed Central, accession PMC4623959, fetched 2026-09-01. Basis for the thousandfold error in Tg-based prediction and for beta relaxation as the better correlate.
• Chang L, Shepherd D, Sun J, Ouellette D, Grant KL, Tang X, Pikal MJ. "Mechanism of protein stabilization by sugars during freeze-drying and storage: native structure preservation, specific interaction, and/or immobilization in a glassy matrix?" Journal of Pharmaceutical Sciences 94(7):1427–1444, 2005. Citation record and abstract confirmed 2026-09-01; the publisher full text returned HTTP 403 and was not read, so this article relies on the abstract only. Basis for the standing of the water-replacement account.
• Duncan D. "USP <922> Water Activity: A Better Approach for Lyo Moisture Determination." La Vague n°83, published by A3P, October 2024. Read at a3p.org, fetched 2026-09-01. Basis for 1% to 3% water by weight as a typical stated freeze-drying moisture target.

Named and not relied upon: the FDA inspection guide on lyophilization of parenterals, requested once at fda.gov on 2026-09-01, returned HTTP 404. Nothing in this article rests on it. The pharmacopoeial and ICH standards an earlier version of this page invoked are likewise named here without being cited for any specific figure, because no specific figure above was taken from them.