Лако-красочные материалы — производство

1. a — and p-Set Acceleration

The so-called a — and p-set acceleration of curing for very alkaline PF resins for foundry core binders was pioneered in the early 1970s [22], although it had been discovered in the early 1950s [22]. In this application the addition of considerable amounts of esters or other chemicals in liquid form (a-set) or as a gas (p-set), such as propylene carbonate, methyl formate, glycerol triacetate, and others, was found to accelerate resin curing to extremely short times. This technique is now used extensively around the world for foundry core PF binders [22] and is being considered for wood adhesives [9] and rigid alkaline PF foams. The technique is applicable in the approximate pH range 7 to 14. The mechanism that makes PF curing acceleration possible has only been explained recently [9] and different explanations exist (see below); it is based on the carbanion behavior of the aromatic nuclei

Figure 4 Cure retarding at high pH and ester acceleration effect of NaOH and KOH-catalysed PF resins (ester = propylene carbonate). Note curve 5, the effect of 4 months aging of the PF resin of curve 1 on the extent and starting pH of the retardation effect. Compare the start of acceleration for curves 4 and 6, showing the differences between propylene carbonate and triacetin esters, and compare the starting point of acceleration at pH 5.5 and pH 7.1. The ‘‘bumps’’ on the curves at pH 8-11 are caused by methylene ether formation, decomposition, and rearrangement [9].

of phenate ions, leading to a more complex variant of the Kolbe-Schmitt reaction. The ester, or residue of its decomposition, attacks the negatively charged phenolic nuclei, and its reaction is not limited to the ortho and para sites, transforming the phenolic nuclei in a temporary condensation reagent of functionality higher than 3, leading to much earlier gelling. Furthermore, temporary condensation occurs not only according to the PF mechanism but also according to a second reaction superimposed on it [9,23] (Fig. 4).

Other explanations and mechanisms for this occurrence have also been advanced: determination by TMA of the average number of freedom of polymer segments between cross-linking nodes of PF resin hardened networks indicate that additive accelerated PF resin polycondensations and hardening present several different acceleration mechanisms. [23]. Some additives such as sodium carbonate appear to present a purely catalytic effect on the polycondensation reaction [23]. Other additives such as propylene carbonate pre­sent both a catalytic effect as well as including an increase in the average functionality of the system, due to further cross-linking, or alternative reactions in which the accelerator itself participates, leading to a tighter final network [23]. These alternative cross-linking reactions could be of a different nature, such as the propylene carbonate case in which the reaction appears to be related to a Kolbe-Schmitt reaction, or they could be similar to the accelerating effect due to the hydrolysis of formamide to formic acid and ammonia with the subsequent rapid reaction of the latter with two or more hydroxybenzyl alcohol groups of PF resols [23]. The rapid reaction of the — NH2 group of formamide with two hydro­xybenzyl alcohol groups of PF resols, a reaction which is also characteristic of urea and methylamine, also appears likely to occur. In some cases such as in formamide none of the two acceleration mechanisms detected appear to be due to catalytic action only, but both appear to be related to additional cross-linking reactions. Both liquid and solid phase 13C nuclear magnetic resonance (NMR) supporting evidence of the mechanisms proposed has been presented [23].

Further proof of complex reactions between propylene carbonate and phenolic nuclei leading to compounds in which the carbonic acid has attacked the phenolic ring has been presented [23] based on the 13C NMR spectrum of the product of the reaction of resorcinol with propylene carbonate, in the absence of formaldehyde. Resorcinol was chosen as its aromatic ring is a stronger nucleophile than that of phenol and thus if any reaction had occurred this would be less elusive and much more easily observed [23]. The reaction pro­ducts which appeared to be formed were carboxylic and dicarboxylic species. That they might be present was also derived by NMR [23]. It must be pointed out that such structures need to be only transitory and not permanent to obtain the same effects noted experimen­tally. Such a subsequent lability could be the reason why it is difficult to observe such linkages in the hardened resin except for faster reacting phenols where they can be observed due to early immobilization of the network which surely occurs.

It must also be remembered that in hot temperature curing of phenolic resins, their polycondensation is accelerated particularly on a wood substrate surface, first by heterogeneous catalysis effect by the cellulose [24], and secondly by the substrates subtracting water from the system and thus increasing its effective concentration, always a very important effect in polycondensation reactions [3,25]. Under these con­ditions the existence of the additional cross-linking mechanism will then be even more marked. It is also clear that if the anhydride exists it might decompose at higher temperature curing, with what type of further reactions it is not possible to say with the data available.

Once defined the nature of the accelerating mechanism induced by increased cross­linking and its existence through the determination of the increased tightness of the PF networks formed [9,23], it is necessary to address the nature of the other accelerating mechanism that appears to be common to both sodium carbonate and propylene carbo­nate. The apparent failure by different analyses [26] such 13C NMR to find any trace of C=O after purification of sodium carbonate accelerated PF resins indicates quite clearly that the sodium carbonate effect may well be purely catalytic and that the C=O is transformed during the reaction to another group, or even more likely that the C=O disappears from the system as CO2 or precipitates completely away as sodium hydrogen carbonate. The presence of the C=O has been clearly noticed in non purified samples of accelerated PF resins [9,23,27] but strictly speaking this in only proof of the additional cross-linking mechanism just discussed above or it could just be due to any carbonic acid salts still present in the system. That this mechanism exists is proven by the acceleration of hardening being marked for high molar ratio (formaldehyde/phenol > 2.5) PF resins in which all available ortho and para sites on the phenol are blocked by methylene or methylol groups. In this case a soft gel, and no subsequent rapid hardening is obtained. The mechanism involved could then be one of the two proposed up to now, namely the hydrogen carbonate ion intermediate activated complex and derived mechanisms [26,28] which present inherent disadvantages that have been outlined [23], which have been pro­posed without any evidence, and for which direct evidence would be rather difficult to gather, and the mechanism [23] based on rapid transesterification reactions of the hydro — xybenzyl alcohol group of a PF resol. This latter mechanism is based on the very facile transesterification of propylene carbonate with methanol through which dimethyl carbo­nate is rapidly obtained [23].

It is interesting to remark that other reactive materials which will readily undergo transesterification analogous to that of propylene carbonate with methanol are trialkyl borates, tetraalkyl titanates, and trialkyl phosphates in an alkaline environment. Also gas injection of methyl borate (and carbon dioxide) has been found to enhance the results of wood composites bonded with formaldehyde-based resins [29], just as the addition of propylene carbonate and glycerol triacetate have been shown to do in wood composites bonded with phenolic resins.

In the case of wood adhesives, first glycerol triacetate (triacetin) and secondly gua­nidine carbonate are the accelerating esters of choice yielding long pot-lives at ambient temperature and fast cure times at higher temperature, and are used in proportions vari­able between 3% and 10% of adhesive resin solids [30-32]. Propylene carbonate is unsui­table for wood adhesives application as it yields far too short pot-lives at ambient temperature. Methyl formate and other esters, including propylene carbonate, are used instead in foundry core binders where sometimes the proportion of ester accelerator used is up to an equal amount of the resin solids; hence the accelerator application technology is rather different from one field to another. Most other esters are much less effective accel­erators at higher temperature, or they shorten the ambient temperature life of the resin to such an extent that in practice the resin cannot be used [30-32]. Triacetin gives long pot — lives and short cure times instead due, among other reasons to its lower rate of hydrolysis at ambient temperature. Another series of compounds, some of which where finally found to yield sufficiently rapid acceleration at higher temperatures still coupled with increased strength of the cured resin as well as sufficiently long shelf-life at ambient temperature, were the salts of guanidine. Guanidine carbonate, guanidine hydrochloride, and guanidine sulphate were tried with positive results [32]. Guanidine carbonate appeared to be the best PF accelerator; both its accelerating capability remained acceptable, while the shelf life at ambient temperature of the PF and phenol-urea-formaldehyde (PUF) resins to which it had been added in different proportions was much longer and the performance in particle­board preparation was the same as triacetin [32]. Even in the case of some industrial higher condensation resins, their pot-life was as long as three weeks with the guanidine carbonate already incorporated in the resin [32].

It has repeatedly been established that the energy of activation of the reaction of polycondensation of PF resins, and also of urea-formaldehyde (UF), melamine — formaldehyde (MF), and other resins, is markedly influenced by the presence of wood [24,33-39]. In the presence of wood as a substrate, the energy of activation of the poly­condensation reaction, and hence of the hardening of PF and other resins is considerably lowered. This implies that resin polymerization and cross-linking proceeds at a much faster rate when the resin is in molecular contact with one or more of the wood constitu­ents [24,33]. It was indeed shown that catalytic activation of the hardening and advance­ment of a PF and other polycondensation resins induced by the wood substrate did exist and was a rather marked effect. The reason for the effect has been found to be due to the mass of secondary attraction forces binding the resin to the substrate [24,33]. These cause variations in the strength of bonds and intensity of reactive sites within the PF oligomer considered, an effect well known in heterogeneous catalysis for a variety of other chemical systems [40], bond cleavage and formation within a molecule being greatly facilitated by chemisorption onto a catalyst surface. This work indicated also which bonds in the PF resin were involves and what was the extent of the acceleration of the hardening reaction caused by such an effect [33].