Induction of viscosity increase during polyglycol addition to isocyanates

Two-component polyurethane adhesives often require accurate viscosity control for curing. When mixing the
components, polyols engage into addition reaction with isocyanates, which increases mol. wt. resulting in higher
viscosity and solidification. Multifunctional additives can be employed to assure cross-linking. However, viscosity
does not necessarily increase in parallel with polymerization and such discrepancy cannot be easily explained. In this
study, an oligomeric ether-ester macro diol (EEMD of 2700 g/mol) was used as a polyol, diluting it with polyethylene
glycols (PEG of 200 g/mol or 400 g/mol) for viscosity control. The polyols were mixed with a trifunctional adduct of
hexamethylene diisocyanate (HDI3) to initiate the addition reaction and viscosity was monitored.

Counter-intuitively, all the blends showed some thinning initially. The drop in viscosity was the most evident in
the PEG-free blend. On the other hand, this blend was more viscous than the other two. Nevertheless, its polymerization
proceeded slower as a result of lower HDI3 concentration and during the later stages viscosity of PEG 200 blend
became more similar. The initial viscosity drop was also evident in the blends with PEG. After the drop, viscosity
stagnated or was slowly recovering for quite some time. Only well after 1 hr of polymerization viscosities finally
exceeded the initial values, which were measured in the beginning of the test. Further viscosity increase approximately
followed linear kinetics, with trendline correlations of the average values better than R² = 0.938.
This suggests that some induction period is present before viscosity begins increasing despite ongoing
polymerization. It remains unclear, which factors might cause such induction effect. Inclusion of PEG seems to enhance
the effect, because the PEG-free blend simply thinned down initially and began thickening quite soon. Bubble formation
or compositional non-homogeneities in the blend are not likely to have a strong effect, since the blends were agitated
quite vigorously between the cone and plate of the viscometer when measuring each point.
It might be possible that supramolecular arrangements could yield the observed
induction of viscosity increase. However, more diverse formulations, extensive testing
and additional analytical techniques might be needed to further elaborate on this hypothesis.