Oxidative Polymerization in Wood Finishes: The Radical Chain Mechanism, Drier Chemistry, and Why Tung Oil Outperforms Linseed Oil

This article is part of the wood finish curing guide — covering all four cure mechanisms: evaporative, oxidative polymerization, coalescence, and chemical cross-linking.

Oxidative polymerization is the chemical reaction that converts liquid drying oils — linseed, tung, danish oil, oil-based polyurethane — into solid, crosslinked polymer films. It is driven by atmospheric oxygen reacting with unsaturated carbon-carbon double bonds in the oil’s fatty acid chains, generating free radicals that link the oil molecules into a three-dimensional network. The same reaction that creates a durable finish on wood also generates heat in a soaked rag — the exothermic nature of oxidative polymerization is what drives spontaneous combustion in finishing rags and what makes thick oil coats stay tacky indefinitely.

Understanding the mechanism at the level of fatty acid chemistry explains every practical performance difference between oil finishes: why tung oil produces a harder, more water-resistant film than linseed oil; why cobalt-based metallic driers accelerate cure dramatically; why teak and rosewood resist oil finishes; and why thick application defeats the reaction entirely.

---

What Is Oxidative Polymerization in Wood Finishes?

Polymerization is the process by which small molecules (monomers) join together to form large molecules (polymers). Oxidative polymerization is specifically triggered by oxygen — the oil molecules do not react with each other spontaneously but do so when oxygen provides the initiating energy through a radical chain mechanism.

Drying oils contain fatty acid chains with unsaturated carbon-carbon double bonds (C=C). These double bonds are chemically reactive sites — molecular locations where oxygen can attack and initiate chain reactions. The number and configuration of double bonds in the oil’s dominant fatty acids determines how fast it polymerizes, how hard the final film becomes, and how resistant it is to water and abrasion. Not all oils contain sufficient reactive double bonds to polymerize usefully — this is the distinction between drying oils (useful as wood finishes) and non-drying oils (cooking oils, mineral oil).

The result of oxidative polymerization is a crosslinked thermoset polymer: the individual oil molecules are connected to each other through covalent bonds formed during the reaction, creating a three-dimensional network that cannot be re-dissolved by the original solvents. This distinguishes oxidatively cured films from evaporatively cured films (lacquer, shellac) where no chemical bonds form during cure and the film can be re-dissolved by solvent.

---

The Radical Chain Mechanism — Initiation, Propagation, Termination

Oxidative polymerization proceeds through three stages of a free radical chain reaction. Each stage has distinct chemistry and can be influenced independently by catalyst selection.

Stage 1 — Initiation

Initiation begins when an oxygen molecule (O₂) attacks a doubly allylic carbon-hydrogen bond adjacent to a C=C double bond in the fatty acid chain. This attack produces a carbon-centred free radical (R•) and a hydroperoxide (ROOH). Free radicals are highly reactive species with an unpaired electron that seek to complete their electron pair by reacting with nearby molecules.

The initiation step is the rate-limiting stage of oxidative polymerization — it is slow without catalysis because the initial oxygen attack requires activation energy. This is the basis of the induction period observable in pure, undried oils: the oil appears to do nothing for an extended time after application, then hardens relatively quickly once enough free radicals are present to propagate the chain reaction.

Cobalt driers (cobalt naphthenate or cobalt octoate) function specifically as initiation catalysts. Cobalt ions (Co²⁺/Co³⁺) undergo redox cycling that decomposes hydroperoxides into free radicals at dramatically lower activation energy than uncatalyzed initiation. Even trace concentrations of cobalt drier — typically 0.05–0.15% cobalt metal based on oil weight — shorten the induction period from hours to minutes and accelerate overall cure by an order of magnitude. BLO (boiled linseed oil) owes its faster cure versus raw linseed oil almost entirely to its cobalt drier content.

Stage 2 — Propagation

Once free radicals exist in the system, propagation proceeds rapidly. Each radical abstracts a hydrogen atom from an adjacent double bond in a neighbouring fatty acid chain, forming a new radical on that chain and a stable product on the original chain. The new radical immediately attacks another fatty acid chain, continuing the sequence. Each propagation step creates one new crosslink between two fatty acid chains, progressively building the polymer network.

Manganese driers (manganese naphthenate or manganese octoate) function primarily as propagation accelerators, maintaining radical concentration during the chain-extension phase. Cobalt-only drier systems can produce a “skin” that cures at the surface while propagation stalls beneath — manganese drier prevents this by sustaining radical density through the film thickness.

The propagation stage is exothermic: each new C-C or C-O bond formed releases approximately 80–120 kJ/mol of energy as heat. On a flat wood surface with large exposed area, this heat dissipates into the surrounding air. In a folded rag, the insulating geometry traps heat, allowing the exothermic reaction to raise temperature progressively — the same positive feedback loop that causes spontaneous combustion, explained mechanistically in the oil rag combustion guide covering the thermal runaway mechanism and disposal protocol.

Stage 3 — Termination

Termination occurs when two free radicals meet and combine, forming a stable covalent bond between two polymer chains. Each termination event creates one permanent crosslink while consuming two radicals. As the polymer network densifies and chain mobility decreases, the probability of two radicals meeting decreases — the reaction slows and eventually stops when remaining radical concentrations are too low to sustain propagation.

Zirconium driers function at the termination stage, promoting through-cure — ensuring that the base of the film (furthest from the oxygen supply at the surface) achieves adequate crosslink density. Without through-driers, cobalt-catalysed surface cure can produce a hard skin over a softer, incompletely cured underlayer. Zirconium through-drier is why modern oil-based polyurethane and alkyd varnish achieve relatively uniform hardness through the film thickness despite the oxygen-diffusion gradient from surface to substrate.

---

Which Finishes Cure by Oxidative Polymerization — and Which Don’t

Oxidative polymerization is only possible in oils with sufficient reactive unsaturation — measured by iodine value (IV), a standardised test that quantifies the number of reactive double bonds per unit mass of oil by measuring iodine absorption.

Iodine Value as a Drying Classifier

The iodine value threshold for “drying” classification is approximately 130 g I₂/100g — oils above this value have sufficient reactive double bonds to cure by oxidative polymerization into a film with practical durability. Semi-drying oils (100–130) cure too slowly and incompletely for use as standalone wood finishes without significant drier additions. Non-drying oils (below 100) do not cure by oxidative polymerization regardless of drier additions — the double bond density is insufficient to build a crosslinked network.

---

Conjugated vs Non-Conjugated Fatty Acids — Why Tung Oil Cures Differently

The position and arrangement of double bonds in a fatty acid chain matters as much as their number. Tung oil’s dominant fatty acid — alpha-eleostearic acid — has three double bonds arranged in conjugated configuration: three C=C double bonds are separated only by single bonds (C=C-C=C-C=C). Linseed oil’s dominant fatty acid — linolenic acid — also has three double bonds, but they are non-conjugated: separated by CH₂ groups (C=C-CH₂-C=C-CH₂-C=C).

Why Conjugation Matters

Conjugated double bonds are dramatically more reactive in radical chain reactions than non-conjugated ones. In a conjugated system, the pi electrons from adjacent double bonds are delocalized across the whole conjugated segment — this electron cloud is available from multiple attack points and stabilises the radical intermediate once formed, lowering the activation energy for propagation. Non-conjugated double bonds react as isolated units, each requiring independent radical attack.

The practical consequences are significant:

Cure rate: Tung oil cures measurably faster than linseed oil at identical temperatures and drier loadings, despite linseed oil having a higher iodine value. The conjugated structure of alpha-eleostearic acid outperforms the higher double bond count of non-conjugated linolenic acid in terms of radical reactivity.

Crosslink density: Conjugated systems form more crosslinks per double bond reacted because the radical intermediates are more stable and more likely to react with adjacent chains rather than terminating on the same chain. Higher crosslink density produces a harder, more water-resistant film. This is the molecular explanation for tung oil’s superior water resistance compared to linseed oil — not merely a difference in double bond count, but a difference in crosslink architecture. The full comparison of tung and linseed oil performance properties, including water resistance testing and fatty acid profiles, is in the tung oil vs linseed oil guide covering conjugated chemistry and product authentication.

Film character: Tung oil produces a film that is initially harder and more brittle than linseed oil, with higher initial water contact angle (better water beading). Linseed oil produces a more flexible film — useful in outdoor paints where flexibility prevents cracking under thermal cycling, but lower in hardness for woodworking applications.

---

What Disrupts Oxidative Polymerization — Failure Modes

Oxidative polymerization requires two things simultaneously: atmospheric oxygen diffusing into the film and reactive double bonds to react with. Disrupting either prevents cure. The failure mode manifests as permanently tacky finish — softness that never resolves because the crosslinked network cannot form.

Oxygen Starvation — Thick Coat Application

Oxygen diffuses into the oil film from the surface exposed to air. In a thin coat (0.05–0.1mm), oxygen reaches the entire film depth within hours, allowing cure throughout. In a thick coat (0.3mm+), the surface layer cures and forms a skin — a partially polymerized layer that dramatically reduces oxygen permeability. Oxygen can no longer diffuse to the film below the skin in sufficient concentration to sustain propagation. The subsurface oil remains in a low-crosslink-density state permanently, producing the characteristic “hard on top, soft underneath” failure.

The rescue protocol — mineral spirits dissolution of the skin followed by complete removal and restart — and the full diagnosis of each failure mode, is covered in the oil finish not drying guide covering the oxygen starvation mechanism. The same article addresses danish oil and BLO failure, as the BLO component in danish oil follows identical oxidative polymerization chemistry — detailed in the danish oil guide covering BLO base composition and cure behaviour.

Terpene Inhibition — Oily Species

Many tropical hardwoods — teak, rosewood, cocobolo, ipe, padauk — contain high concentrations of phenolic extractives and terpene compounds in their cell walls. These compounds function as natural antioxidants: they react preferentially with the radical intermediates generated during initiation, scavenging them before they can attack fatty acid double bonds and initiate propagation. The radical chain reaction is terminated before it begins.

The result is that oil finishes applied directly to these species fail to cure — the terpene antioxidants in the wood surface are more reactive with oxygen radicals than the oil’s fatty acids. The standard solution is a pre-treatment with acetone or naphtha to remove surface terpenes, followed by immediate finish application before the terpene concentration rebuilds. The protocol specifics vary by species.

Surface Contamination — Silicone and Wax

Silicone compounds and wax residues on wood surfaces create a physical barrier that prevents oil penetration into the wood and can also interfere with radical propagation at the finish-substrate interface. Oil finish applied over wax-contaminated surfaces cures correctly in the free film but fails at the adhesion interface — the cured film peels or lifts because the crosslinked network formed above the wax layer rather than within the wood surface. Surface preparation before any oil finish application must include complete wax and silicone removal.

Temperature Below Threshold

Radical chain reactions have significant temperature dependence — reaction rate roughly halves for every 10°C decrease (the Arrhenius relationship). At temperatures below approximately 10°C, oxidative polymerization proceeds so slowly that practical cure times become impractical, and metallic driers lose much of their effectiveness. Oil finishes applied in cold workshops may appear permanently tacky simply because cure is proceeding at a fraction of normal rate. Warming the workspace to above 18°C typically restores normal cure progression in recently applied coats — the reaction resumes rather than being permanently disrupted.

---

Oxidative polymerization is one of four cure mechanisms — the complete framework, including evaporative cure (lacquer/shellac), coalescence (water-based), and cross-linking (catalyzed finishes), is in the wood finish curing guide covering all mechanisms and their practical implications for repairability and compatibility. The exothermic nature of oxidative polymerization — the same reaction mechanism that generates heat in a finishing rag — is the basis of spontaneous combustion risk covered in the oil rag combustion guide covering the thermal runaway mechanism and iodine value as a risk predictor.

---