The Chemistry of Viscoelastic Foam: Polyol, Isocyanate, and the Reaction That Creates Memory

Memory foam does not “remember” anything. It is a polymer network whose time-dependent mechanical response arises from the chemistry of its synthesis — the specific combination of raw materials, reaction conditions, and molecular architecture that determines whether the resulting foam relaxes slowly under load or springs back immediately, whether it stiffens dramatically in cold conditions or barely changes, whether it lasts three years or twelve. Every property discussed on this site traces back to chemistry. This article covers that chemistry from the ground up: the raw materials, the reaction, and how the molecular structure determines macroscopic behaviour.

1. The Two Key Raw Materials

Polyurethane foam — the material category that includes both memory foam and HR foam — is formed by the reaction of two primary components: a polyol and an isocyanate. Understanding what each contributes to the final polymer network is the foundation of understanding why different foam formulations behave differently.

Polyol: the flexible backbone

A polyol is a molecule with multiple hydroxyl groups (–OH). In polyurethane synthesis, polyols form the long, flexible chain segments of the polymer network — the sections between cross-link points that are responsible for the foam’s elastic and viscoelastic behaviour.

Two classes of polyol are used in flexible polyurethane foam:

• Polyether polyols: formed by ring-opening polymerisation of propylene oxide (and sometimes ethylene oxide) onto a starter molecule. The resulting chains are flexible, hydrophobic (water-resistant), and have good low-temperature flexibility. Polyether polyols dominate consumer mattress foam production because of their hydrolysis resistance — ether linkages are stable in the presence of moisture, which is relevant for a product used in a humid sleep environment over years.
• Polyester polyols: formed by condensation polymerisation of dicarboxylic acids with diols. Polyester-based foams have excellent mechanical properties — high tensile strength, good abrasion resistance — but are susceptible to hydrolysis (the ester linkage is cleaved by water over time). Polyester polyols are used in specialty applications where their mechanical performance advantage justifies the hydrolysis risk, but are uncommon in consumer sleep foam.

The molecular weight of the polyol chain is one of the most important design variables in foam formulation. Higher molecular weight polyols produce longer, more mobile soft segments — lower glass transition temperature, more rubbery behaviour, greater tendency toward viscoelastic response. Lower molecular weight polyols produce stiffer, less mobile segments — higher Tg, more elastic behaviour.

For viscoelastic (memory) foam, polyols are selected with molecular weights and chain architectures that place the foam’s glass transition temperature in the range of 10–25°C — the bedroom temperature range where the body-warming effect will drive meaningful softening and conformance. This is a deliberate design choice, not an incidental property.

Isocyanate: the cross-linker

Isocyanates are highly reactive molecules containing one or more isocyanate groups (–N=C=O). In polyurethane synthesis, the isocyanate reacts with the polyol’s hydroxyl groups to form urethane linkages (–O–CO–NH–) — the cross-links that connect the flexible polyol chains into a three-dimensional network.

Two isocyanates dominate flexible foam production:

• TDI (toluene diisocyanate): a mixture of 2,4- and 2,6-isomers, TDI is the most widely used isocyanate in flexible foam. It produces foams with a good balance of mechanical properties and is well-suited to the continuous slabstock production process used for most commodity foam. TDI-based foams are the backbone of the global mattress foam industry.
• MDI (methylene diphenyl diisocyanate): MDI produces harder, more dimensionally stable foams than TDI. It is used in high-resilience (HR) foam production and in the moulded foam applications where precise dimensional control is required. MDI-based HR foams are increasingly used in premium mattress applications for their superior mechanical properties and lower volatile emissions compared to TDI foams.

The ratio of isocyanate to polyol — expressed as the isocyanate index — controls the degree of cross-linking in the final network. An isocyanate index above 100 (slight excess of isocyanate) produces a more densely cross-linked, stiffer network. An index below 100 produces a less densely cross-linked, softer, more compliant network. For viscoelastic foam, the isocyanate index is tuned to produce the desired combination of soft segment mobility and hard segment constraint that gives the foam its characteristic slow-recovery behaviour.

2. The Foaming Reaction: How Gas Becomes Structure

Polyurethane foam is not simply a solid polymer — it is a cellular structure: a polymer network containing a large volume fraction of gas-filled cells. The cellular structure is created during the polymerisation reaction itself, through a simultaneous gas-generation process.

The blowing reaction

When isocyanate reacts with water (present in the formulation as a co-reactant), it generates carbon dioxide gas as a by-product:

R–N=C=O + H₂O → R–NH₂ + CO₂↑
R–NH₂ + R’–N=C=O → R–NH–CO–NH–R’ (urea linkage)

The CO₂ produced by this reaction forms the bubbles that become the foam cells. The timing of this gas generation relative to the polymerisation reaction is critical: if the polymer network gels too quickly before sufficient gas is generated, the foam will be dense and poorly expanded. If the network gels too slowly, the gas bubbles will coalesce into large irregular cells or escape entirely before the network solidifies.

The balance between the polymerisation rate and the blowing reaction rate is controlled by catalyst selection and loading — a key formulation variable that determines cell size, cell size distribution, and the open/closed cell ratio of the final foam.

Open vs closed cells

Foam cells can be open (the cell walls rupture during foam rise, connecting adjacent cells) or closed (the cell walls remain intact, enclosing individual gas pockets). Most flexible polyurethane foam used in mattress applications is predominantly open-cell — the cells rupture during the controlled “crush” step that follows foam rise, creating an interconnected cell structure that allows air to flow through the foam under compression.

Open-cell structure is important for both mechanical and thermal performance. Mechanically, open cells allow air to escape when the foam is compressed and re-enter when it recovers — the air movement contributes to the foam’s energy dissipation and recovery behaviour. Thermally, open cells allow airflow through the foam matrix, reducing thermal resistance. Closed-cell foams trap air in individual pockets, behaving more like solid polymer at low compression and retaining heat more effectively — generally undesirable for sleep surface applications.

3. Soft and Hard Segments: The Molecular Origin of Viscoelasticity

The key to understanding why memory foam behaves differently from HR foam lies in the block copolymer structure of the polyurethane network — specifically, the relationship between soft segments and hard segments.

Soft segments

The polyol chains in the polyurethane network form the soft segments — long, flexible sections between hard segment cross-link points. Above their glass transition temperature, the soft segments have sufficient thermal energy for conformational rearrangement — the polymer chains can rotate around backbone bonds, adopting new configurations over time. This chain mobility is the molecular origin of stress relaxation and creep: under sustained load, the soft segments gradually rearrange to accommodate the applied deformation, reducing the internal stress while maintaining the applied strain.

The rate of this rearrangement depends on the energy barrier for conformational change — which is temperature-dependent. At lower temperatures, fewer chain segments have sufficient thermal energy to overcome the rotational barrier; rearrangement is slower and less complete. This is why memory foam is stiffer at low temperatures — not because the individual polymer bonds are different, but because the rate of stress-relaxing chain motion is thermally limited.

Hard segments

The urethane linkages (–O–CO–NH–) and urea linkages (–NH–CO–NH–) formed by the isocyanate reactions constitute the hard segments. These segments are rigid and polar — they form strong hydrogen bonds with adjacent hard segments, creating physical cross-links (hard-segment domains) that act as temporary junction points in the network.

The hard segments serve two functions:

• Structural cross-links: they constrain the soft segment chains and prevent the foam from flowing irreversibly under load (which would produce unrecoverable compression set rather than viscoelastic deformation).
• Energy dissipation sites: the formation and disruption of hydrogen bonds between hard segments under cyclical loading dissipates mechanical energy as heat — the molecular mechanism behind the foam’s high loss factor (tan δ) and low resilience.

The ratio of soft segment to hard segment content — controlled by polyol molecular weight, polyol functionality (number of –OH groups per molecule), and isocyanate index — is the primary determinant of the foam’s position on the viscoelastic spectrum from slow-recovery memory foam to fast-recovery HR foam.

4. Additional Formulation Components

A commercial polyurethane foam formulation contains more than just polyol and isocyanate. The additional components — catalysts, surfactants, additives — control the reaction kinetics and the final foam structure in ways that are as important as the primary reactants.

Catalysts

Two catalyst types are used in flexible foam production:

• Amine catalysts: tertiary amines (such as DABCO or DMEA) catalyse both the polyol-isocyanate reaction and the water-isocyanate blowing reaction. Their relative loading controls the balance between gelation and foaming — a critical kinetic parameter.
• Tin catalysts: organotin compounds (such as dibutyltin dilaurate, DBTL) specifically catalyse the polyol-isocyanate urethane reaction without significantly affecting the blowing reaction. Tin catalyst loading primarily controls gelation rate and the final degree of cross-linking.

For viscoelastic foam, catalyst selection and loading are tuned to produce a slow, controlled gelation that allows the foam to fully expand before the network sets — and to produce the specific hard segment domain morphology that gives the foam its high loss factor.

Surfactants

Silicone surfactants stabilise the foam during rise by reducing surface tension at the gas-liquid interface, preventing cell coalescence and controlling cell size distribution. The surfactant type and loading affect cell uniformity — fine, uniform cells are generally associated with better mechanical property consistency than coarse or bimodal cell distributions.

Chain extenders and cross-linkers

Low-molecular-weight diols and triols are added to the formulation to control the hard segment content and cross-link density independently of the isocyanate index. Chain extenders (difunctional) increase hard segment content without increasing cross-link density; cross-linkers (trifunctional or higher) increase both. For memory foam, cross-linker selection is used to tune the Tg and the temperature sensitivity of the foam’s mechanical properties.

Antioxidants and stabilisers

As covered in the Foam Degradation article, polyurethane undergoes oxidative chain scission over time. Antioxidant additives — typically hindered phenols and phosphites — interrupt the radical chain mechanism of oxidative degradation. Their selection and loading determine the foam’s oxidative stability over its service life. Premium foam formulations use antioxidant systems designed for extended service life; budget formulations use minimal antioxidant loading to reduce raw material cost.

5. From Reaction to Foam: The Production Process

Commercial flexible foam is produced primarily by two processes: continuous slabstock and moulded foam. Understanding the differences explains why foam cut from slabstock behaves differently from moulded comfort layers in some applications.

Continuous slabstock

In slabstock production, the liquid foam formulation is dispensed continuously onto a moving conveyor, where it rises freely in an open mould to form a large rectangular bale (typically 1–2 m wide, 1 m tall, and many metres long). The bale is then cut into sheets of the required thickness using horizontal band saws or wire cutters.

Slabstock foam has a natural density gradient — the core of the bale rises faster and reaches higher temperatures during the exothermic reaction than the skin zones, producing slight density variation from core to skin. This variation is typically small in well-controlled production but can affect consistency in low-quality operations.

Moulded foam

In moulded foam production, the liquid formulation is injected into a closed mould of the desired final shape and allowed to cure under controlled temperature and pressure. Moulded foam allows more precise control of density, cell structure, and final dimensions than slabstock. It is used for contoured comfort layers, anatomically shaped pillows, and applications where dimensional precision is critical.

Moulded foam typically has a denser skin layer (formed at the mould surface during cure) than slabstock — this skin layer contributes to the perception of a firmer top surface on some moulded foam products.

Summary

Memory foam’s behaviour — its slow recovery, temperature sensitivity, pressure redistribution over time — is not mysterious. It is the direct consequence of the polymer chemistry established during synthesis: the molecular weight and type of the polyol chains, the isocyanate’s cross-linking chemistry, the ratio of soft to hard segments, and the thermal activation of chain mobility that makes these properties temperature-dependent.

The formulation variables that determine whether the resulting foam is a slow-recovery viscoelastic material or a fast-recovering HR foam are well-understood at the molecular level. The next article in this series examines those variables directly — what specifically differs between a memory foam formulation and an HR foam formulation, and how those differences produce the mechanical behaviours that distinguish them at the macroscopic level.

The Sleep Mechanic is a materials engineer with hands-on R&D experience in cushioning materials and viscoelastic polymers. Sleep Science Lab applies materials engineering analysis to sleep surfaces — because “it feels comfortable” is not an explanation.