Rigid polyisocyanurate foam formulated from trimellitic anhydride-modified aromatic polyester polyols derives its char retention from the overlapping contributions of aromatic ester density, branch-point concentration, trimerization crosslink density, and the thermal stabilisation of the decomposing polymer network. The polyol component is typically produced by the transesterification of dimethyl terephthalate process residues or the glycolysis of post-industrial polyethylene terephthalate glycolysate, followed by a partial condensation with benzene-1,2,4-tricarboxylic anhydride at temperatures between 180 °C and 220 °C under vacuum. The resulting aromatic polyester polyol carries a measurable acid number, a hydroxyl number, and a branched architecture that alters the rheological response of the formulated polyol blend. In continuous boardstock lamination, this modification affects mixhead pressure, cream time, fibre time, and the final isocyanurate conversion. Char retention, defined here as the residual mass after controlled thermal decomposition under an inert atmosphere or the protective carbonaceous layer formed during cone calorimeter exposure, is not an inherent property of the polyol alone but emerges from the reaction between the modified polyol and polymeric diphenylmethane diisocyanate at isocyanate indices typically exceeding 180. The residual mass is reported in accordance with ISO 11358-1:2022, and the reaction-to-fire contribution is assessed under ISO 5660-1:2015. Laboratory and line-scale observations show that the same polyol can produce markedly different char retention depending on mix quality, cure temperature, water content, and catalyst neutralisation, which is why char retention must be evaluated as a system property rather than a raw material constant.

What Role Does Trimellitic Anhydride Play in the Solid-State Char Morphology of Polyisocyanurates?

Trimellitic anhydride introduces a trifunctional aromatic node at concentrations typically between 5 wt% and 25 wt% of the neat polyol, raising the number-average functionality of the polyester diol or polyol from a linear value near 2.0 to a branched value above 2.5. The benzene-1,2,4-tricarboxylic anhydride unit contributes three potentially esterifiable carboxyl sites; after mono- or diesterification, the residual carboxylic acid group remains available for isocyanate reaction or acid-base interaction with potassium carboxylate trimerization catalysts. During high-temperature degradation, the trimellitic anhydride-derived aromatic rings increase the formation of graphitic precursors and reduce the evolution of volatile aliphatic fragments that would otherwise leave only low-density char. In differential scanning calorimetry and thermogravimetric analysis, increased aromatic ester concentration shifts the maximum mass-loss rate to higher temperatures and increases the residual mass at 800 °C under nitrogen, although the exact shift depends on free glycol, water, and catalyst residues. The solid-state char produced from a TMA-rich polyisocyanurate network is denser and exhibits fewer shrinkage cracks when the trimerization conversion is sufficient to lock the aromatic branch points into the polyisocyanurate lattice. In cone calorimeter tests according to ISO 5660-1:2015 at an irradiance of 50 kW/m², the carbonaceous layer acts as a transport barrier to volatiles and oxygen, which lowers the second heat release rate peak and extends time to flashover in room-corner scenarios, but the relationship is strongly dependent on foam density, facing, and blowing agent. The trifunctional node is therefore not merely a chain extender; it is a char-forming structural unit that modifies both the cure network and the solid-state degradation pathway.

Under inert thermal decomposition, polyurethane and polyisocyanurate foams undergo at least three distinguishable mass-loss events. The first event, between 200 °C and 300 °C, corresponds to residual urethane bond scission and the evolution of trapped blowing agent; the second, between 350 °C and 500 °C, is associated with isocyanurate ring degradation and ester carbonyl cleavage; the third, above 500 °C, involves aromatisation and char consolidation. In TMA-modified aromatic polyester polyols, the ester carbonyl concentration is higher than in aliphatic polyester diols, and the aromatic rings serve as intrinsic char precursors. The residual mass at 800 °C under nitrogen as measured by ISO 11358-1:2022 typically increases with increasing aromatic ring content and with increasing isocyanurate index up to a point where unreacted isocyanate volatilisation competes. Laboratory data reported in polyurethane fire performance literature indicate that the non-oxidative char residue of rigid polyisocyanurate foams can range from approximately 35% to 55% of initial mass, with TMA-containing systems falling at the upper portion only when the foam is adequately cured and the blowing agent is fully removed before testing. The oxidative char retention measured after cone calorimetry is generally lower because surface oxidation converts part of the carbonaceous residue to carbon monoxide and carbon dioxide, and the remaining char is no longer a simple pyrolysis residue. This distinction is operationally important because a formulation optimised for inert-atmosphere char yield may not be optimised under oxidative fire conditions, and both measurements are required for technical validation of char retention claims.

When TMA-Containing Polyol Blends Encounter High-Pressure Impingement Mixing Equipment

High-pressure impingement mixers used for discontinuous pour-in-place or continuous lamination applications require viscosity-matched polyol and isocyanate streams to achieve proper mixing. TMA-modified aromatic polyester polyols generally exhibit Brookfield viscosities at 25 °C between 800 mPa·s and 6,000 mPa·s, and the formulated polyol blend can exceed 1,500 mPa·s when flame retardants, smoke suppressants, and surfactants are added. Because the viscosity of these polyols rises steeply below 20 °C, heated storage tanks and temperature-controlled transfer lines are specified; common set points range from 40 °C to 60 °C. At the mixhead, component temperatures are normally held at 30 °C to 40 °C with an allowable variation of ±2 °C, because a colder TMA-modified polyol stream increases impingement pressure and reduces mixing quality, while a hotter stream accelerates premature urethane and trimerization reactions in the mixing chamber. Production-scale equipment operating at 120 bar to 180 bar polyol pressure and 120 bar to 180 bar isocyanate pressure with output ranges of 10 kg/min to 60 kg/min has shown that an increase in TMA content of 5 wt% can raise mixhead pressure by 5% to 15% at constant temperature, requiring recalibration of the metering pumps. Inadequate mixing produces coarse cells, internal voids, and localised index variations, all of which reduce the uniformity of the carbonaceous layer during fire exposure and lower the reproducibility of char retention measurements; this is not a laboratory-scale issue but a full-scale process limitation observed on high-output laminators. The practical control envelope for these mixers is therefore narrower than the bulk viscosity specification would suggest, and temperature drift of more than ±2 °C in the polyol line can produce measurable differences in the char residue at 800 °C.

The addition of TMA-modified polyester polyol is not a simple drop-in viscosity adjustment. Its residual acid number, which in commercial products may range from 0.5 mg KOH/g to 4.0 mg KOH/g, consumes part of the potassium carboxylate trimerization catalyst and can interact with amine co-catalysts, shifting the gel-to-cure balance. In formulations that use water as a chemical blowing agent, the acid groups can catalyse the isocyanate-water reaction, producing carbon dioxide and evolving heat earlier than expected; this shifts the cellular structure and changes the char morphology. Therefore the water addition is often limited to 0.1% to 0.5% by total polyol mass when high acid number TMA polyols are present, and the catalyst package is adjusted using potassium octoate or potassium acetate at 0.3% to 1.0% by total polyol mass. The concentration of low-molecular-weight glycols in the TMA-modified polyol also affects char retention: free diethylene glycol acts as a volatile fuel source, whereas aromatic terminal groups increase residual char. The acceptable upper limit for free glycol in char-critical applications is often cited in polyol datasheets as 3% by mass, though published data for this specific configuration is limited and must be verified against the final foam density and index. This interaction between acid number, water, and free glycol means that char retention cannot be controlled solely by increasing TMA content; the formulation must be balanced through titration of the blended polyol before production.

Cone Calorimetry and Thermogravimetric Benchmarks for Char Retention

Char retention should be measured using a combination of thermogravimetric analysis, microscale combustion calorimetry, and medium-scale reaction-to-fire tests. Thermogravimetric analysis under ISO 11358-1:2022, with a specimen mass of approximately 10 mg, a heating rate of 10 K/min, and a nitrogen flow of 50 mL/min, provides the non-oxidative residue at 800 °C. The first derivative of the mass-loss curve identifies the temperature of maximum decomposition rate, which for rigid polyisocyanurate foams typically lies between 320 °C and 420 °C for the isocyanurate ring and between 500 °C and 650 °C for the aromatic ester char consolidation step. Microscale combustion calorimetry under ASTM D7309-21 uses a controlled pyrolysis method to evaluate heat release rate per unit mass, and the char residue after pyrolysis can be compared directly with TGA data. At the medium scale, ISO 5660-1:2015 specifies a horizontally mounted 100 mm × 100 mm specimen exposed to an external heat flux of 50 kW/m² with an exhaust oxygen analyser; the measured peak heat release rate, total heat release, and mass loss during 300 s to 600 s provide the primary char retention benchmarks. The protective char layer may reduce the peak heat release rate by 20% to 40% relative to an equivalent non-charing foam, but the reported value depends on foam density, skin formation, and the presence of glass facers. Surface flame spread and smoke development are frequently evaluated using ASTM E84-23; a Class A rating is achieved when the flame spread index is less than 25 and the smoke developed index is less than 450. The use of multiple instruments is essential because char retention is not directly measured by any single standard and must be interpreted from residual mass, heat release suppression, and visual char integrity.

The numerical thresholds in this table are standard-defined or widely used acceptance criteria; however, char retention is not a pass/fail parameter in ISO 5660-1:2015 or ASTM E84-23, and it must be interpreted as a comparative material property rather than a classification metric. Products evaluated under EN 13501-1 may use the results of ISO 11925-2:2020 and the single burning item test, EN 13823:2020, to establish Euroclass B, C, D, or E; the char contribution is indirect but critical in the smoke and heat release profile. Scanning electron microscopy of cone calorimeter residues from TMA-containing polyisocyanurate foams reveals a two-layer char structure: a porous outer crust and a dense inner layer. The dense layer is composed mainly of aromatic carbon, and its continuity correlates with the mass loss rate after the first 300 s of exposure. The presence of TMA-derived branch points increases the number of covalent connections per unit volume, reducing the extent of surface cracking during the drying and carbonisation stages. Image analysis of the char cross-section can quantify crack width, void fraction, and residual thickness. In formulations with insufficient isocyanurate conversion, the char layer exhibits through-thickness fissures that allow volatiles to escape unoxidised; this reduces the apparent char retention and allows flame spread. Published data for this specific configuration is limited, but morphological evidence supports the use of TGA residue as a screening tool only when combined with cone calorimeter mass-loss curves.

Tracking Kinetic Competition Between Urethane Formation and Isocyanurate Trimerisation in TMA-Rich Systems

The formation of a stable char depends on the conversion of a sufficient fraction of the isocyanate excess into isocyanurate rings before the urethane gel point freezes the network. In rigid polyisocyanurate processing, the urethane reaction between isocyanate and hydroxyl groups is catalysed by tertiary amines and tin carboxylates, while trimerization is catalysed by potassium carboxylates or quaternary ammonium salts. The two reactions compete for the same isocyanate. When the TMA content of the polyol is high and the residual carboxylic acid content is elevated, the acid groups can neutralise part of the potassium carboxylate catalyst, slowing trimerization at the beginning of the rise profile and allowing more urethane network to form. This creates a process conflict: if the cream time becomes too short because of urethane acceleration, the foam may close cells before trimerization completes; if the trimerization is too slow, the green foam may collapse or the final isocyanurate content may be insufficient for the expected char retention. Temperature control at the mixhead and on the lamination line is therefore critical; the trimerization rate is strongly temperature dependent, with an apparent activation energy in the range of 40 kJ/mol to 60 kJ/mol reported for potassium-octoate-catalysed isocyanurate formation. A rise in the lamination line platen temperature from 50 °C to 70 °C can increase the isocyanurate conversion and improve the char residue at 800 °C, but excessive heat can cause blistering, skin delamination, and cracking of the char layer, particularly when the blowing agent has not condensed fully. The kinetic balance must therefore be controlled through catalyst selection, acid neutralisation, and thermal history, not through TMA concentration alone.

Potassium octoate is often preferred over potassium acetate in TMA-modified aromatic polyester polyol systems because it provides a more controlled trimerization exotherm at the high viscosities encountered in continuous lamination. The addition level is typically 0.5% to 2.0% by total polyol mass depending on the acid number of the polyester and the required index. Because TMA-modified polyols retain residual acidity, the use of strongly basic amine catalysts, such as 2,2′-dimorpholinodiethyl ether, may lead to partial carboxylate salt formation and catalyst deactivation if the amine is added before the acid groups are fully neutralised. In addition, hydrolysis-sensitive polyester backbones require the polyol blend to be protected from atmospheric moisture; storage vessels are kept under dry nitrogen and the water content of the blended polyol is maintained below 0.1% by mass to avoid premature ester hydrolysis and viscosity drift. On production lines where the polyol blend is recycled through the mixhead during interruptions, the accumulation of moisture and catalyst decomposition products changes the char retention of the resulting foam, and batch-to-batch variance is observed in thermogravimetric residue unless the recycle loop is sampled and adjusted. The interaction between residual acid, moisture, and catalyst is therefore one of the main sources of full-scale char retention variability in TMA-containing systems, and it is not captured by neat polyol characterisation alone.

For Acid Number, Free Glycol Content, and the Effective Isocyanate Index

The nominal isocyanate index is calculated as the ratio of actual isocyanate equivalents to hydroxyl equivalents multiplied by 100. However, in the presence of free carboxylic acid groups from trimellitic anhydride modification, the effective index is shifted because carboxylic acid groups consume isocyanate and release carbon dioxide, forming amide and urea intermediates. The acid number, expressed in mg KOH/g, quantifies the residual acidity. A polyol with a hydroxyl number of 220 mg KOH/g and an acid number of 2.0 mg KOH/g has an apparent equivalent weight that is lower than the hydroxyl-number-only calculation would indicate, meaning that the actual amount of isocyanate consumed by the polyol is higher. Failure to correct for acid consumption produces a foam with a lower effective excess of isocyanate for trimerization, reducing the isocyanurate content and lowering the char residue at 800 °C. Low-molecular-weight free glycols in the polyol, including diethylene glycol and ethylene glycol, participate in urethane reactions but contribute substantial aliphatic segments that volatilise at temperatures below 350 °C. Their presence increases the mass loss in the first decomposition step and reduces the efficiency of the aromatic char-forming network. In practice, char-optimised TMA-modified polyester polyols are specified with acid numbers between 1.0 mg KOH/g and 3.0 mg KOH/g and free glycol contents below 2.5% by mass; published data for this specific configuration is limited, and the values are often validated against a single lamination line with a fixed isocyanate index near 200 to 250. The effective index correction is not a minor adjustment; omitting it can reduce the isocyanurate yield sufficiently to compromise char retention in otherwise well-formulated systems.

On a continuous sandwich panel laminator, the TMA-modified polyol blend is applied to a lower metal facing at a pour rate calculated to produce a board core density of 38 kg/m³ to 60 kg/m³, while the isocyanate index may be set between 200 and 300 depending on the desired flammability classification and compressive strength. The lamination oven is usually divided into zones with platen temperatures increasing from 50 °C to 75 °C; residence time varies from 2 minutes to 6 minutes depending on line speed and panel thickness. Because the char retention of the final board is sensitive to the degree of isocyanurate conversion, insufficient oven residence time leads to a green panel that may show a higher high-temperature mass loss and lower char residue than the same formulation cured under longer conditions. Conversely, excessive oven residence time or temperatures above 80 °C can cause the formation of a brittle char precursor that cracks during cooling and creates pathways for smoke release during a subsequent fire. This processing window is typically narrower than ±5 °C for high-TMA, high-index systems, and the lamination line controls are therefore set with zone-to-zone tolerances of ±2 °C to maintain consistent char retention. The metal facer also affects char retention: a steel facer retains the foam and char mechanically, while an aluminium facer conducts heat away from the reaction zone but may melt during fire, changing the apparent char layer stability in full-scale tests. Core density is normally verified against ASTM D1622-20, and the measured density is used to normalise the cone calorimeter mass-loss data because low-density foams of identical formulation can show disproportionately lower char retention due to higher internal surface area and greater oxidative attack.

Thermal Degradation Pathways in Potassium Octoate-Catalysed Trimellitic Anhydride-Modified Polyols

The potassium-catalysed isocyanurate ring is thermally more stable than the urethane linkage, but it is not inert. At temperatures between 350 °C and 450 °C, isocyanurate rings can decompose by ring opening to yield isocyanate, carbodiimide, or amine species that either volatilise or recombine into char. In TMA-modified aromatic polyester polyols, the aromatic ester linkages are also cleaved in this temperature range, producing aromatic acids and anhydrides that can undergo condensation oligomerisation and contribute to the carbonaceous residue. The residual carboxylic acid groups derived from incomplete trimellitic anhydride esterification may form anhydride bridges during heating, releasing water and creating additional crosslinks in the condensed phase. This dehydration step, visible in the TGA derivative curve as a shoulder between 300 °C and 400 °C, is considered a char-promoting event because it stabilises the network before the main isocyanurate decomposition. At higher temperatures, above 600 °C, the remaining aromatic carbon skeleton undergoes graphitisation, and the char yield at 800 °C correlates with the aromatic carbon density of the original formulation. The presence of potassium from the catalyst can affect the oxidation behaviour of the char: potassium may catalyse the oxidation of carbon at high temperatures, which slightly decreases the oxidative char retention in cone calorimeter tests but improves flame inhibition in the gas phase. This dual effect means that a formulation optimised solely for TGA residue under nitrogen may not be optimal under cone calorimeter oxidative exposure, and both measurements must be considered. The degradation pathway is further altered when the blowing agent is a hydrocarbon such as n-pentane, which leaves the foam before the main thermal events, or a hydrofluoroolefin, which can generate hydrogen fluoride and interact with potassium and char oxidation chemistry.

Flame retardants and smoke suppressants influence char retention in TMA-modified systems. Expandable graphite is added at 5% to 25% by total formulation mass in some boardstock and pipe insulation systems; its exfoliation creates a mechanically expanded char layer that compensates for otherwise weak char. However, graphite can interfere with the acid-base balance of TMA polyols, adsorbing potassium catalyst and changing the cream time. Melamine and ammonium polyphosphate are used in halogen-free formulations; their char-promoting mechanisms are different, and their presence can enhance the residue at 800 °C while changing viscosity. Liquid phosphate esters such as tris(2-chloro-1-methylethyl) phosphate reduce pH and can accelerate ester hydrolysis in moist polyol blends, requiring reformulation of the catalyst package. The resulting char retention cannot be predicted from additive content alone; cone calorimeter testing under ISO 5660-1:2015 and TGA under ISO 11358-1:2022 are necessary. The operational boundary for TMA-modified aromatic polyester polyols extends to storage, mixing, and testing. At polyol storage temperatures below 20 °C, the viscosity may exceed the capability of the transfer pumps, and partial crystallisation of aromatic ester segments can occur, requiring reheating to 60 °C for 24 hours before use. Exposure to relative humidity above 60% during storage or bulk transfer should be avoided because ester hydrolysis increases acid number over time and shifts the catalyst demand. Avoid combining TMA-modified polyols with unneutralised strongly basic amine catalysts at high loading, as the resulting acid-base reaction can form amide salts that precipitate in cool transfer lines and reduce the effective catalyst concentration. When post-consumer PET-derived glycolysate is used as the aromatic feedstock, variability in the content of terephthalic acid, ethylene glycol, and oligomeric species can alter the TMA incorporation efficiency and the final char retention; published data for this specific configuration is limited, and incoming lots should be assessed by acid number titration, hydroxyl number determination according to ISO 14900:2017, and residual mass at 800 °C under ISO 11358-1:2022 before full-scale production.