The research begun roughly ten years ago into the reaction behaviour of hexahydrotoluylene diisocyanate produced an unexpected result: the discovery of the polycyclic iminooxadiazinediones. The knowledge gleaned resulted in a new class of hardeners for the coatings industry.
Polyurethane coatings formulated with aliphatic isocyanates can withstand the severest of tortures virtually unscathed.
|Drive your car from the spray booth directly to the petrol station (spilling as much petrol as possible over the fresh paint), through a sandstorm in Death Valley (at a surface temperature of 80 °C), then through the car wash (with the hardest of brushes) before heading off to Alaska …|
Coatings are intended as decorative protection for the surface of an object against external influences. A typical example are automotive coatings light stability, weatherability and resistance to just about every possible antagonist are required here.
Coatings shine in a number of colours
– including on cars
If someone interested in chemistry had asked 20 years ago what car he should choose if he wanted to be sure it would have a two-component polyurethane coating, the answer would have been: “You have to buy a Rolls Royce - Polyurethane guaranteed, regardless of the model.” Now even the less affluent can call a car with a polyurethane coating their own: Roughly 25% of all cars produced today have a polyurethane clearcoat.
Polyisocyanates – the hardeners
Toluylene diisocyanate (TDI) no longer harbors any major surprises, but the same cannot be said about its core hydrogenated derivative hexahydrotoluylene diisocyanate (H6TDI). The reason lies in the stereochemistry of the cycloaliphatic skeleton: Figure 2 shows all theoretical and actually existing H6TDI isomers.
Fig. 2.: Toluylene diisocyanate (TDI, left) and
its core hydrogenated, cycloaliphatic derivative (H6TDI): textbook examples of isomeism
While hexamethylene diisocyanate (HDI) is essentially without competition among the linear aliphates, a comparative look at the manufacturing processes for isophorone diisocyanate (IPDI) andH6TDI suggests taking a closer look at the latter (Figure 3).
Fig. 3.: Technical synthesis of IPDI (top) and H6TDI.
To go from monomer to physiologically safe polyisocyanate hardener for coatings, the chemist requires at least one additional step: modification (Figure 4).1)
Fig. 4.: From monomeric diisocyanate to polyisocyanate:
the most important structure types on the basis of idealised model reactions.
X = (Cyclo)alkyls (*only relevant as a coatings polyisocyanate if a di- or polyol is used (n > 2).
H6TDI is unusual here: Instead of making one of the NCO groups available for the modification reaction and the other for the later crosslinking to form the PU coating, both NCO groups frequently react during the modification reaction, resulting in low-NCO products that are unusable for coatings applications. Figure 5 shows a reaction originally planned as a trimerisation (isocyanurate formation, cf. Figure 4) as an example.
Fig. 5.: Mechanism for the formation of the main (Path A) and
by-product (Path B) with the catalyzed reaction of cis-1,3-cyclohexane diisocyanate.
Only those isomers whose NCO groups have a 1,3-cis configuration relative to one another are amenable to cyclopolymerization. Unfortunately this is the case of the major portion of the isomer mixture.
A closer look at the gel permeation chromatogram of the reaction mixture thus produced suggested, however, that in addition to unreacted monomer (the trans fraction) and undesired, high molecular weight cyclopolymer, another low molecular weight species was also present. Investigations with model compounds showed that this could not be one of the structures known from polyisocyanate chemistry shown in Figure 4.
Nitrogen or oxygen – which is the reaction centre?
A crystal structure analysis – not exactly a standard analysis for coatings – revealed the secret: The previously unknown compound proved to be a polycyclic iminooxadiazinedione – the first of its kind, incidentally. Its formation can be explained mechanistically as the discontinuation of the chain growth mechanism that provides the main product of the reaction, 1-nylon (Figure 5). The activation of one NCO group by the catalyst and the intramolecular attack on the adjacent NCO group is followed by the intermolecular reaction step, which involves another NCO group. The later can take place via either the oxygen or the nitrogen atom. The points of attack are highlighted in Figure 5 with negative partial charges.
In the standard case, the formation of 1-nylon, the nitrogen atom is the reaction center and cyclopolymerisation can proceed (Figure 5, Path A). If the oxygen atom reacts, however, chain formation does not continue; instead the 6-membered ring closes via the third NCO group, which is the last to react. Sterical reasons make this impossible along the isocyanurate formation path, so that a C=N fragment remains intact and the C-O double bond is opened (Figure 5, Path B). This opening is a rare exception in isocyanate chemistry.
The iminooxadiazinedione structure formed is a dimer with respect to the number of incorporated H6TDI monomer units, but with respect to the number of incorporated NCO groups it is a trimer: an asymmetric trimer, AST.
Once aware of the unusual structure of the iminooxadiazinedione, the next question was: Is its formation limited to systems in which two NCO groups stem from one molecule, or can the formation of iminooxadiazinedione also be applied as a modification reaction to conventional isocyanates, for example the standard monomers HDI or IPDI? This question led away from the original approach, which had the objective of obtaining something useful from H6TDI. A search of the literature for other iminooxadiazinediones, without the prefix “polycyclic”, returned just a few hits. Only two of these concerned isocyanate modification reactions in a stricter sense. 2,3) The reaction of methyl isocyanate in the presence of catalytic quantities of trialkylphospine2) was previously known from industrial practice, but with HDI instead of methyl isocyanate as the reactant. The results can hardly be more different: Whereas HDI produces primarily uretdione, the iminooxadiazinedione is the main product when methyl isocyanate is reacted. The dependence shown in Table 1 of the selectivity of the reaction on the chain length of the alkyl group in isocyanate is evidence of the sensitivity of isocyanate reactions to even marginal changes to the substrate structure.
The reaction of HDI in the presence of CO2 and fluoride catalysts3) resulted in the first ever monocyclic iminooxadiazinediones produced from a diisocyanate. These are produced together with the long-known symmetric trimer structure, the isocyanurate, as well as the known oxadiazinetrione (comprising two moles NCO and one mole CO2, Figure 6).
Fig. 6. : HDI modification with fluoride-bearing catalysts.
As internal experiments demonstrated, the HDI reaction with fluoride catalysis - even in the absence of CO2 - produces a small amount of iminooxadiazinedione in addition to a lot of isocyanurate. A new catalyst comprising polyfluorides (Figure 6)4) significantly increased the fraction of AST.
A very simple approach chemically, but extremely effective for the modification of HDI: Whereas fluorides yield product mixtures in which the AST fraction is less than 20%, polyisocyanate hardeners with equimolar fractions of both trimer structures are achievable with polyfluoride catalysis. Study of the property profile of the new generation of HDI hardeners could now begin.
The very low viscosity of the HDI-AST is a major advantage. This applies to both the ideal structure, which can be isolated with some effort from oligomer mixtures4), and to the actual products, which – as a function of the degree of reaction - have varying fractions of higher molecular weight species (pentamers, heptamers, etc.) with identical or differently structured junction points (Table 1).
|R in R-NCO|
Table 1.: Substrate dependence of trialkylphosphine-catalyzed isocyanate modification. (Structure fractions in mole %).
HDI-ASTs are the lowest viscosity at least NCO-trifunctional isocyanate hardeners
In the past, coatings producers regulated the viscosity of their coating systems by adding solvent. Today and even more so in the future, statutory environmental protection requirements restrict this practice. The coatings industry’s answer is “high solids,” often with prefixes such as “very” or “ultra,” up to “100 %” instead of “high.”
Waterborne systems are another alternative. HDI-ASTs offer advantages here as well: Low viscosity is always beneficial for the dispersion of an inherently hydrophobic hardener in water. But viscosity isn’t everything; the number of NCO groups in the hardener molecule, the (average) NCO functionality, determines the quality of the final coating film. A trimer, asymmetric or not, is always superior to a dimer – more precisely, a uretdione – even if the latter has a significantly lower viscosity (Table 2).
150 +/- 80
350 +/- 100
700 +/- 100
1200 +/- 300
3000 +/- 750
10000 +/- 2000
Table 2.: Viscosity of typical HDI-based polyisocyanates in mPas at 23 °C.
“Ideal structure” stands for the respective species with the lowest molecular weight, simplified as in Fig. 3.
* Y = n-Butyl, ** extrapolated
Frank Richter, Bayer MaterialScience AG, Leverkusen
1)H. J. Laas, R. Halpaap, J. Pedain, J. Prakt. Chem./Chem. Ztg. 1994, 336, 185–200
2)K. H. Slotta und R. Tschesche, Chem. Ber. 1927, 60, 295–301
3)H. J. Scholl (Bayer AG) DE-A 3902078, 1989
4)F. Richter, J. Pedain, H. Mertes, C.-G.Dieris (Bayer AG) DE-A 19611849, 1996