Trimethylolpropane (TMP) is incorporated into carboxyl-functional polyester resins for powder coatings as a trifunctional branching species that alters both pre-cure melt rheology and post-cure network topology. The point in the polycondensation sequence at which TMP is charged—before esterification, after the first water distillation cut, or after reaching a defined acid value plateau—controls the average distance between branch points along the polyester backbone and the terminal functionality available for crosslinking with triglycidyl isocyanurate (TGIC), β-hydroxyalkylamide (HAA), or blocked isocyanates. A resin produced by charging the entire TMP content with the initial monomer mixture has an acid value distribution biased toward low molecular weight carboxyl-terminated oligomers because the trifunctional monomer introduces branching points at a stage when linear chain extension is still limited. In contrast, a split-feed addition in which TMP is injected after the reactor has achieved a steady water distillation rate permits linear oligoester segments to form before branching occurs, producing a narrower functionality distribution. The resulting gel time as measured by ISO 8130-6:2021 may shift from 90 s to 150 s at 200°C for the same total TMP content, depending on feed timing. Crosslink density of cured films is typically measured by dynamic mechanical thermal analysis following ASTM D4065-20, where storage modulus values above the glass transition temperature correlate with network molar mass between crosslinks; solvent swelling using tetrahydrofuran at 25°C is also used for qualitative comparison. Published data for this specific configuration is limited because resin producers rarely disclose the detailed feed sequence, but industrial quality-control records from carboxyl-functional polyester lines indicate that gel time distribution can widen by 10–15 s when the TMP feed point is not controlled to within ±5 min of the target acid value window.

What Does Lower Acid Value at the Time of TMP Addition Do to Final Crosslink Density?

When TMP is added after the acid value has fallen to 40–60 mg KOH/g in a standard 60:40 isophthalic acid/terephthalic acid polyester resin, the trifunctional monomer is no longer able to distribute branching points uniformly along all oligomer chains. Instead, TMP reacts preferentially with carboxyl-terminated chain ends that are already sterically accessible, creating terminal branch points rather than mid-chain tetrafunctional nodes. The consequence for powder coating cure is a network in which the elastically effective strand length is shorter than would be predicted from the average functionality alone. Published data for this specific configuration is limited; laboratory parallel studies on polyester resins with identical molar ratios show that terminal branching increases the storage modulus at 180°C by 15–25% compared with mid-chain branching at the same TMP level, while the glass transition temperature measured by differential scanning calorimetry per ASTM D3418-21 shifts upward by 3–6°C due to reduced free volume. The acid value at gelation is not the sole determinant; the distribution of TMP between oligomer populations matters. A reactor charge sequence that introduces TMP as a molten stream at 120°C after the polycondensation temperature has reached 230–240°C will not fully randomize unless the agitator tip speed exceeds 2.5 m/s and the water removal rate remains above 0.8 kg/h for a 2,000 L vessel. If the monomer is added too late, unreacted TMP may remain in the resin, creating free hydroxyl groups that compete with carboxyl groups for TGIC, increasing extractables as measured by ISO 8130-7:2019 beyond 0.5% after 15 min at 180°C.

During production-scale extrusion of powders formulated with TMP-modified polyester resins, the feed sequence effect is amplified by the short residence time distribution of a 32:1 L/D twin-screw extruder operating at 90–110°C. A resin with early-stage TMP loading typically exhibits a lower pre-cure melt viscosity, which can reduce screw torque by 8–12% relative to late-stage TMP loading at the same total branching content. However, the lower viscosity does not translate into better coalescence at the film surface; panel flow tests and wave-scan analysis of films cured at 200°C for 10 min indicate that early-stage TMP resins can produce a smoother appearance only when the curing schedule is extended to 15–20 min, because the lower initial modulus permits longer leveling. The practical processing window for TMP feed sequence is therefore bound by two failure modes: excessive melt viscosity during compounding when TMP is introduced late and insufficient sag control when TMP is introduced early. Incompatibility with certain amine-based flow modifiers should be considered when residual TMP is present, because free hydroxyl groups can adsorb amine additives and shift the gel time unpredictably. This is not a hypothetical boundary; batch-to-batch variance of gel time on a commercial powder coating line has been observed to increase from ±5 s to ±12 s when late-stage TMP addition is used without a subsequent vacuum stripping stage to remove residual monomer.

Mid-Stage TMP Feed as a Process Lever for Blocked Isocyanate Cure

In systems using caprolactam-blocked isocyanate crosslinkers, the cure reaction consumes polyester hydroxyl groups rather than carboxyl groups, and the placement of TMP in the polyester backbone becomes a direct determinant of the number of available primary hydroxyl groups. A polyester resin with a hydroxyl value of 25–35 mg KOH/g and an acid value below 8 mg KOH/g can be produced by adding TMP after the first water removal stage, then holding the reactor at 240°C until the acid value falls below 10 mg KOH/g. This mid-stage TMP feed preserves more terminal hydroxyl functionality than an initial-charge TMP feed, because the trifunctional monomer is not permitted to form extensive cyclic or internal ester structures during the earliest stages. The effect on crosslink density with uretdione or caprolactam-blocked isophorone diisocyanate is measurable as an increase in solvent resistance after cure. In methyl ethyl ketone double rubs following ASTM D4752-20, films prepared with mid-stage TMP feed can withstand 150–200 double rubs before substrate breakthrough, whereas early-stage TMP feed often reaches only 80–120 double rubs at the same 200°C/20 min cure. The reverse effect occurs when the TMP feed is delayed until the acid value has fallen below 15 mg KOH/g; here, the hydroxyl value may not reach the target because TMP reacts with residual acid groups and forms a network of low molecular weight polyesters that are not effective crosslink sites. The narrow processing window for mid-stage TMP feed—typically 15–25 min after the onset of water distillation in a 2,000 L batch—requires in-process acid value titration at 15 min intervals. Published data for this specific configuration is limited, but industrial practice recognizes that mid-stage TMP addition reduces the risk of gelation in the reactor while still improving crosslink density relative to early-stage addition.

When Late-Stage TMP Addition Is Used with Hydroxyalkylamide Cure

Late-stage TMP addition creates a polyester resin with higher hydroxyl functionality and higher melt viscosity, which influences the specification limits for β-hydroxyalkylamide (HAA) cure such as Primid XL-552. In HAA-cured systems, the stoicheiometric ratio is typically 95:5 to 105:5 resin to HAA; an excess of hydroxyl groups from free TMP can shift the actual stoichiometry outside this range unless the hydroxyl value is corrected. Because HAA cure proceeds through an esterification reaction that liberates water, residual TMP-related hydroxyl groups can increase cure shrinkage and cause pinholes in films thicker than 80 µm. The effect is most pronounced when the powder coating is applied to preheated steel panels at 150–170°C metal temperature, where the gel time shortens below 60 s and the water evolution rate exceeds the diffusion rate through the film. Crosslink density in HAA-cured films is usually assessed by the storage modulus in the rubbery plateau, with values of 15–30 MPa at 200°C reported for standard systems; late-stage TMP feed can raise the rubbery plateau modulus by 10–20% but simultaneously lowers impact resistance because the network becomes less capable of dissipating stress. Mandrel bend performance per ISO 1519:2019 may degrade from 3 mm to 5 mm at equivalent film thickness, indicating that crosslink density alone does not define mechanical suitability. The operational boundary for late-stage TMP feed in HAA-cured powders is therefore limited to films below 60 µm unless a flexibilizing agent such as a low-viscosity saturated polyester is blended into the formulation.

Thermal analysis of cured films from TGIC-cured polyester resins reveals that the choice between early-stage and late-stage TMP feed changes the breadth of the glass transition region as well as the absolute crosslink density. A film prepared from an early-stage TMP resin often exhibits a glass transition onset at 58–63°C and a half-width of 12–16°C when scanned at 10 K/min under ASTM D3418-21. The same total TMP content shifted to a late-stage feed can produce a glass transition onset at 64–69°C with a narrower half-width of 8–11°C, reflecting a more homogeneous network. The rubbery plateau storage modulus at 200°C, determined by ASTM D4065-20, is frequently observed in the range of 10–25 MPa for TGIC-cured films, with the upper end corresponding to resins in which TMP branching was introduced during the middle to late stages of polycondensation. The crosslink density calculated from rubber elasticity theory is then estimated at 1.0–2.5 mmol/cm³, depending on the assumed front factor and the exact network functionality. Solvent swelling ratios in xylene at 25°C corroborate the DMTA data, but the correlation is not linear because the polyester backbone composition itself influences the polymer-solvent interaction parameter. The practical consequence is that two resins with the same acid value and the same nominal TMP content can differ in crosslink density by 15–30% purely as a function of feed sequence, a difference that is readily detected by MEK double rubs, reverse impact, and humidity resistance testing under DIN EN ISO 6270-2:2018.