Elastomeric closures: extractables, leachables, and stopper chemistry in peptide vial storage

Discussions of peptide container chemistry usually stop at the glass. Borosilicate composition, delamination, and surface hydrolytic resistance are well-documented topics, and they matter. But by mass of chemically active formulation, the rubber stopper is by a wide margin the most complex component of a sealed peptide vial. A Type I borosilicate vial is essentially three or four oxides in a fused network. A pharmaceutical elastomeric closure is a cured polymer matrix carrying a dozen or more discrete additives — vulcanizing agents, accelerators, activators, fillers, plasticizers, antioxidants, and pigments — several of which are small, mobile, and at least partially soluble in aqueous media.

For a lyophilized peptide sitting under nitrogen, this is a slow and largely theoretical concern. For a reconstituted stock stored for weeks in contact with the closure, or for a vial subjected to repeated punctures, the closure becomes the dominant source of small-molecule contamination in the solution. Understanding what is in an elastomer, what migrates out of it, and what that means analytically is a prerequisite for interpreting purity data on stored material.

What a pharmaceutical elastomer actually contains

The base polymer in the overwhelming majority of modern injectable closures is halobutyl rubber — either bromobutyl or chlorobutyl. Butyl rubber is a copolymer of isobutylene with a small fraction of isoprene; halogenation of the residual isoprene unsaturation provides reactive sites for crosslinking. Halobutyl was adopted over natural rubber and over earlier polyisoprene formulations for two reasons: extremely low gas and moisture permeability, and the absence of the latex proteins that made natural rubber problematic.

Around that base polymer sits the compound formulation:

• Curing (vulcanizing) system. Sulfur-based cures, resin cures, or peroxide cures. Sulfur cures require accelerators — thiazoles, thiurams, dithiocarbamates, sulfenamides — and activators, typically zinc oxide plus a fatty acid such as stearic acid. Accelerator residues and their transformation products are among the most commonly detected leachables.
• Fillers. Carbon black, calcined clay, silica, or talc, at loadings that can approach 30–40% by weight. Fillers control modulus, tear strength, and coring behavior. They are largely immobile but carry their own trace metal profile.
• Plasticizers and processing aids. Paraffinic oils and low-molecular-weight polymers that aid mixing and molding. These are unbound and comparatively mobile.
• Antioxidants and antiozonants. Hindered phenols and amines that protect the polymer during cure and storage. Their oxidation products are frequently what shows up in an extractables study, rather than the parent compound.
• Pigments. Iron oxides, titanium dioxide, or carbon black, contributing to the trace metal background.

A modern low-extractable formulation may deliberately omit whole classes of these — peroxide-cured or resin-cured compounds eliminate the accelerator burden entirely, at some cost in physical properties. The tradeoff between mechanical performance and chemical cleanliness is the central design tension in closure selection.

Extractables versus leachables

The two terms are routinely used interchangeably in casual discussion, and they are not the same thing.

Extractables are the compounds that can be forced out of the closure under exaggerated conditions — aggressive solvents, elevated temperature, extended contact, sometimes pH extremes. An extractables study is a characterization exercise. Its purpose is to enumerate the universe of species the material is capable of releasing, so that analytical methods can be built to detect them. Extractables studies are typically run by the closure manufacturer and reported in a technical data package. They are performed with orthogonal detection — headspace GC-MS for volatiles, GC-MS for semivolatiles, LC-MS with both positive and negative electrospray for nonvolatiles, and ICP-MS for elemental species.

Leachables are the compounds that actually migrate into a specific formulation under real storage conditions over the real shelf life. Leachables are always a subset of extractables, usually a small one, and their identity depends heavily on the solution: pH, ionic strength, the presence of surfactant, and above all the presence of organic co-solvent. A leachables study is a stability exercise, run on the actual product in the actual container.

The practical implication for research peptide work is that an extractables data sheet describes a worst case that is deliberately unrepresentative. It is useful for knowing what to look for. It is not a prediction of what is in the vial.

What migrates, and what governs it

Migration out of a crosslinked elastomer is diffusion-limited, and the rate depends on the size and polarity of the migrant, the crosslink density of the matrix, temperature, and the partition coefficient between the rubber and the solution.

For an aqueous peptide solution at near-neutral pH, the migrants that matter are mostly small and moderately polar:

• Zinc. Zinc oxide activator makes zinc the single most abundant elemental leachable from sulfur-cured halobutyl. Zinc is relevant well beyond its own toxicological profile, because it is a coordinating metal. For peptides with histidine, cysteine, or clustered acidic residues, trace zinc can shift aggregation behavior and, in some systems, catalyze oxidative pathways.
• Accelerator residues and degradants. Mercaptobenzothiazole and its disulfide, thiuram fragments, and dithiocarbamate breakdown products. Several of these are thiol-reactive, which is the concern for cysteine-containing peptides — a reactive small-molecule thiol in solution provides a route to mixed disulfides that appear as new late-eluting peaks in reversed-phase analysis.
• Antioxidant oxidation products. Hindered phenol degradants are commonly detected by LC-MS and are usually chemically inert toward peptide backbones, but they contribute to the impurity peak count.
• Oligomeric silicone. From the silicone lubricant applied to the closure, discussed below.
• Volatile hydrocarbons. From paraffinic processing aids. Detected by headspace GC, generally chemically inert.

Organic co-solvent dramatically accelerates all of this. A peptide stock reconstituted in water or bacteriostatic water extracts very little from a modern closure. The same peptide in 20% acetonitrile, or in a DMSO-containing diluent, presents a far more favorable partition environment for hydrophobic migrants, and extraction can increase by an order of magnitude or more. Storage of organic-containing stocks in stoppered vials is a materially different chemical situation from storage of aqueous stocks, and it is a common unexamined variable in research handling.

Coatings, lubricants, and silicone

Nearly all closures carry a surface treatment, and it changes the extractables picture substantially.

Silicone oil is the traditional lubricant, applied to allow machine handling and to reduce insertion friction. It is effective and inexpensive. It also migrates. Silicone oligomers partition into solution readily, and free silicone at an interface is a well-documented nucleating agent for protein and peptide aggregation. In analytical work, silicone shows up as a characteristic 74-dalton-spaced ion series in mass spectra, and as a persistent background in low-level LC-MS. A peptide analyst who sees that series should suspect the closure before suspecting the sample.

Fluoropolymer laminates — a thin ETFE or similar film bonded to the product-contact face — are the alternative. The film is a physical barrier that reduces migration of essentially every organic extractable class by one to two orders of magnitude and eliminates the need for silicone on the contact surface. The cost is higher, the film adds stiffness that affects sealing behavior, and the barrier applies only to the laminated face; the plug sides and the punctured region are still bare elastomer.

Crosslinked surface treatments applied by plasma or by a proprietary cure represent a middle path, reducing both particulate shedding and silicone requirement without full lamination.

Puncture, coring, and the repeated-access problem

Every puncture does two things: it exposes fresh, untreated elastomer surface in the needle track, and it carries some risk of coring — the detachment of an elastomer fragment into the solution.

Coring risk scales with needle gauge, bevel geometry, insertion angle, and the number of prior punctures. Non-coring (pencil-point or side-port) needles substantially reduce it. Blunt or reused needles substantially increase it. A cored fragment is a visible or subvisible particulate, and it is also a fresh high-surface-area source of leachables sitting in the solution for the remainder of the storage period.

The fresh-surface effect is the subtler one. A laminated closure’s barrier is defeated locally at every puncture site. A vial punctured twenty times has a meaningful cumulative area of unlaminated elastomer in direct contact with solution. For a working stock accessed repeatedly over several weeks, this is one of the reasons extended-use stocks accumulate impurities faster than a simple first-order degradation model predicts.

What this means for interpreting stored-material data

None of this changes what a certificate of analysis on incoming lyophilized material says. A COA characterizes the peptide as manufactured, in a sealed vial, before any of these processes have had time to operate. Closure chemistry is a storage and handling question, not a manufacturing-quality question, and it should not be read as a reflection on the source material.

Where it does matter is in reconciling analytical results on stored solutions with expectations. Several patterns are worth recognizing. New small early-eluting peaks in reversed-phase chromatography of a stored aqueous stock, absent from the freshly reconstituted sample, are more often closure leachables than peptide degradants — they typically lack the UV absorbance ratio expected of a peptide bond. A silicone ion series in an MS trace points to lubricant, not to the sample. An unexplained increase in soluble aggregate in a stock that has been repeatedly accessed is consistent with either silicone-nucleated aggregation or trace zinc, and distinguishing between them requires a silicone-free control container.

The straightforward experimental control is a parallel storage arm in an alternative container — a glass ampoule, a fluoropolymer-lined cryovial, or a laminated closure where the baseline is siliconized. Species present in the stoppered arm and absent in the alternative are closure-derived by construction. That comparison is cheap, requires no specialized instrumentation beyond what the purity method already uses, and resolves a category of ambiguity that otherwise gets attributed to the peptide.

The closure is a formulation component, not inert packaging. Treating it as one — selecting for it, controlling for it, and reading analytical artifacts against it — is the difference between a stability observation about a peptide and a stability observation about a rubber stopper.