4. Conclusions
This catalyst system, i.e. Cr(AcAc)3 or other trivalent Cr salts deposited onto an acidic oxide carrier, has been found to polymerize ethylene at high activity, comparable to the commercial Phillips Cr oxide catalysts. This is an advantage in commercial operations because hexavalent chromium has been classified as carcinogenic. The catalyst is only active when the carrier contains strongly acidic sites, and also when in the presence of a trialkylaluminum cocatalyst. The purpose of the carrier is clearly to increase the electron deficiency of the chromium and the purpose of the cocatalyst must be to alkylate the chromium. However, the cocatalyst may also reduce the initially trivalent chromium, possibly to a divalent form. Thus the 60 year old debate about the valence of the active site on the commercial Phillips catalysts also spills over into this related system as well. At present, we have no information as to whether the alkylated active site is trivalent or divalent. Only the most acidic carriers were found to activate Cr(AcAc)3 to form a polymerization catalyst. Of the many supports tested, a fluoride-treated silica-alumina, a fluoride- and chloride-treated silica alumina, and a sulfate-treated alumina seemed to perform best in these experiments. However, as has been discovered in metallocene activationstudiesusing these same types of carriers [14–17], there are probably many other variations that are at least equal in their performance. It is clear from the polymer MW distribution that these Cr(AcAc)3 catalysts containanextremely wide variety of active sites compared to the usual Phillips chromium oxide/silica catalysts. No doubt this indicates a variety of ligand environments around the chromium sites after exposure to aluminum alkyl cocatalyst. These ligands could come from the initial organic ligand (AcAc), or through interaction with the acidic support, or they could even be new compounds made by reaction ofthe aluminum alkyl cocatalyst with the AcAc or the support. The similarity of the MW distributions shown in Fig. 11 and Table 6, where four very different chromium (III) organic salts were used, suggests that the initial organic ligand does not play a major role in determining the character of the active sites. Neither did the MW distribution change as the Cr(AcAc)3 catalyst was heated from 90 ◦C, where it probably contained three AcAc ligands per Cr, up to nearly 600 ◦C, where the initial organic ligands should have been removed or pyrolized (see Table 2 and supplemental Figs. S1 and S3). Furthermore, Fig. 12 shows almost identical MW distributions from Cr(AcAc)3 and also from chromium oxide when both were supported on the same sulfated alumina and treated triisobutylaluminum. All of these observations suggest a minor, if any, role for the AcAc ligand in determining the character of the active site population. It seems likely that the AcAc ligands are removed upon contact with the alkylaluminum cocatalyst
