Cr /silica catalyst

FIGURE 98 Melt indices of polymers made with titanated catalyst. Cr/silica was dried at the various temperatures, allowed to adsorb Ti(0-/Pr)4, and then rinsed several times with hexane. The amount of titanium adsorbed, and the resultant polymer melt index, were influenced by the drying temperature. 

FIGURE 124 Properties of polymers made with R R-activated catalysts. Cr/silica-titania was calcined 3 h at 871 °C in N2, CO, or CS2, then in air 2 h at the indicated temperature. The MW and the fluidity (similar to melt index) of the resultant polymers are plotted. Fluidity is the inverse of the melt viscosity, measured at 0.1 s 1. 

FIGURE 136 The melt index of polymer is sometimes increased by the presence of fluoride on the catalyst. (Cr/silica was impregnated with (NH4)2SiF6 and activated at the indicated temperature.)... 

It is well established that the presence of fluoride on a Cr/silica-titania catalyst increases its 1-hexene incorporation efficiency, and lowers the amount of very low-MW material [605]. Table 43 shows one example of this. Two catalysts, Cr/silica and Cr/silica-titania, were activated at 600 °C, with and without fluoride. These catalysts were tested for polymerization activity in the presence of 0.5 mol L 11-hexene. In the absence of titania, no difference was observed in the amount of 1-hexene incorporated into the polymer. However, Cr/silica-titania exhibited a major increase in 1-hexene incorporation when it contained fluoride. 

Several determinations of the number of propagation centers by the quenching technique have been carried out (98, 111). As a quenching agent methanol, labeled C14 in the alkoxyl group, proved to be suitable in this case. The number of active centers determined by this technique at relatively low polymerization rates (up to 5 X 102 g C2H4/mmole Cr hr at 75° and about 16 kg/cm2) (98, 111, 168) in catalysts on silica was about... 

The Phillips Cr/silica catalyst is prepared by impregnating a chromium compound (commonly chromic acid) onto a support material, most commonly a wide-pore silica, and then calcining in oxygen at 923 K. In the industrial process, the formation of the propagation centers takes place by reductive interaction of Cr(VI) with the monomer (ethylene) at about 423 K [4]. This feature makes the Phillips catalyst unique among all the olefin polymerization catalysts, but also the most controversial one [17]. 

Surely the Phillips Cr/silica polymerization catalyst, discovered by J. P. Hogan and R. L. Banks in the early 1950s (/), can claim one distinction—to have been one of the most studied and yet most controversial systems ever. Today we seem to be debating the same questions posed over thirty years ago, being no nearer to a common view. 

In contrast, the academic community has been astonishingly prolific on what might be considered an industrial topic. Researchers groping in this industrial darkness rarely refrained from drawing sweeping conclusions about the commercial catalyst from a few vacuum line experiments. And there has been a tendency to regard Cr/silica as monolithic, whereas in fact an industry has developed around its diversity. 

The Phillips Cr/silica polymerization catalyst is prepared by impregnating a chromium compound onto a wide pore silica and then calcining in oxygen to activate the catalyst. This leaves the chromium in the hexavalent state, monodispersed on the silica surface. Chromium trioxide (Cr03) has been impregnated mast commonly, but even a trivalent chromium salt can be used since oxidation to Cr(VI) occurs during calcining. 

Spectroscopy has not proven to be very conclusive in solving this problem. Similarities between the visible spectrum of the calcined catalyst and that of bulk dichromates have been noted (5,12-14). In the end, however, there is always doubt about the interpretation of spectra because no adequate reference data exist for these surface bound species (76). Krauss and coworkers have carefully studied the luminescence of Cr/silica and concluded that at least a portion of the chromium is present as chromate (75). 

One interesting result of stripping off Cr(VI), whether chromate or dichromate, is that it leaves pairs of hydroxyls, even if the catalyst has been calcined previously at 800°C to remove OH pairs. Despite being paired, these hydroxyls do not condense easily when heated a temperature of 800°C is required to remove all of them. However, they do react with chromyl chloride vapor at 200°C to yield a good deal of chromate. (Ordinarily silica calcined at 800°C forms no chromate when it reacts with chromyl chloride.) This suggests that even at 800°C Cr/silica may contain a considerable amount of chromate. 

On Cr/silica catalysts, however, the effect of hydrogen is minor in comparison. And surprisingly, there is no evidence of hydrogenation on Cr/silica catalysts, Hydrogen must shorten the chains in some other still mysterious way. 

This diversity of sites explains why the molecular weight distribution (MWD) of polymers produced by Cr/silica is broad (71). Model calculations which assume a single type of active site usually predict Mw/Mn 2,4 but in reality Mw/Afn = 6-15 is common, and 20-30 can be achieved with catalyst modifications. The distribution is also broader than that generally obtained from Ziegler catalysts, for which Mw/Afn = 3-6 under similar conditions. Experience with organometallic compounds suggests that a broad MWD may be a general feature of catalysts which terminate by -elimination. 

Zakharov et al. have used a radio tagging technique to measure the active site density in which polymerization is killed with labeled methanol (72, 73). They found only about 1 % or less of the chromium to be active, or about one tenth of Hogan s number. But because they calcined Cr/silica at only 400-500°C, their catalyst was probably only one-tenth as active. So the two studies are not necessarily in conflict. As expected, the active site density found by tagging increases with time during a polymerization run. 

That the catalyst really does fracture during polymerization can be easily demonstrated. When a coarse Cr/silica, e.g., 60-80 mesh (150/zm), polymerizes ethylene to a yield of 1000 gg 1, each catalyst particle produces a polymer particle of approximately the same shape but about 1000 times larger. Cutting into the polymer granule will not expose the original catalyst particle. Even under a microscope the fragments are too small to be seen easily in the polymer background. 

These small fragments are almost impossible to see in the polymer, even under a microscope. However, evidence of the fracturing is seen by examining the ash via mercury intrusion. An example is shown in Table IV, where Davison Grade 952 Cr/silica was allowed to polymerize ethylene to various yields. The polymer was then burned away, and the porosity of the catalyst ash determined by mercury intrusion. 

Although titania is not itself a good carrier for Cr(VI), its presence in small amounts on Cr/silica catalysts does have a promotional effect on both activity and termination rate (74, 75). The beneficial effect probably results from a change in the electronic environment on the chromium, which possibly becomes linked to the titania during calcining. 

Two ways of incorporating titania onto Cr/silica catalysts are used, and each has certain advantages. In the simplest method the silica surface is coated with a layer of titania by allowing a titanium ester to react with the hydroxyl groups. 

Figure 13 demonstrates the promotional effect of titania on the activity of Cr/silica catalysts. These samples were made by coprecipitation. The chromium was then added and each sample was calcined at 760°C to form surface attached Cr(VI). For comparison, titania concentrations are expressed as Ti atoms per square nanometer of surface, even though a good part of the titania may actually be in the bulk. 

Fig. 16. The relative melt index potential (RMIP) of a series of cogelled Cr/silica titania catalysts rises and then falls with calcining temperature, indicating first dehydroxylation then sintering. However, the more titania in the catalyst, the more easily it sinters and therefore the lower the temperature at which peak RMIP develops.

Many experiments suggest that the promotional effect of titania is due to formation of Ti—O—Cr bonds, and that this is very dependent on subtle variations in the preparation of the catalyst. For example, most titanium salts are ineffective as promoters when impregnated onto CR/silica as an aqueous solution, but anhydrous titanium esters, which react with silanols, are often highly effective. A uniform monolayer of titania is probably not achieved in the first case. 

Fig. 18. A drop in surface area marks the onset of sintering in a series of Cr/silica-titania catalysts calcined in dry air or CO. Sintering is less severe in CO.

Fig. 19. The termination rate, plotted here as relative melt index potential (RMIP), reflects the extent of surface dehydroxylation in two series of Cr/silica-titania catalysts, calcined in (Y) air or ( ) CO and then air to reoxidize the chromium, both at the temperatures shown. The third series ( ) shows the additional benefit of low-temperature attachment. It was calcined in CO at the temperatures shown, then air at a lower temperature (760°C).

Again (as mentioned in Section V,C) sulfur compounds perform better than CO, as can be seen in Fig. 20, because they are better dehydrating agents. When Cr/silica is reduced by COS or CS2 a black chromium sulfide forms. Reoxidation then converts it back to the hexavalent oxide. The catalyst retains no sulfur, but it often takes on a new reddish hue and the activity is greatly improved. This is probably an extension of the trend already observed in Fig. 10, which shows both activity and termination to increase as the catalyst is dehydrated. Perhaps the color change from yellow to orange, and finally to red for sulfided catalysts, indicates a transition from chromate to dichromate, or maybe just less coordination to hydroxyls. Adding water vapor to a sulfided catalyst completely reverses the benefit. 

Fig. 20. After being reduced at 870°C, three series of Cr/silica-titania catalysts yield highest termination rates (RMIP) after reoxidation at 600°C. Catalysts reduced in CS2 display best results because CS2 is the most effective dehydroxylating agent. Carbon monoxide is second best. Trivalent samples calcined in N2 also show the benefit of low-temperature reoxidation, but without the effect of increased dehydroxylation.

The polymerization behavior of Cr/alumina seems to reflect the higher hydroxyl population. More surface hydroxyls also means more sites available to support chromium, and alumina does stabilize about twice as much Cr(Vl) as silica. However, the higher chromium levels do not yield a more active catalyst. Cr/alumina is typically only one tenth as active as Cr/silica. Termination rates are also extremely depressed on Cr/alumina. Both effects could be attributed to the extra hydroxyls, which are thought to interfere with polymerization. 

Silica and aluminum phosphate have much in common. They are isoelec-tronic and isostructural, the phase diagrams being nearly identical even down to the transition temperatures. Therefore, aluminum phosphate can replace silica as a support to form an active polymerization catalyst (79,80). However, their catalytic properties are quite different, because on the surface the two supports exhibit quite different chemistries. Hydroxyl groups on A1P04 are more varied (P—OH and A1—OH) and more acidic, and of course the P=0 species has no equivalent on silica. The presence of this third species seems to reduce the hydroxyl population, as can be seen in Fig. 21, so that Cr/AP04 is somewhat more active than Cr/silica at the low calcining temperatures, and it is considerably more active than Cr/alumina. 

Like Cr/silica catalysts, the activity of Cr/AlP04 is improved with increasing calcining temperatures up to about 800°C. Again this probably reflects the condensation of surface hydroxyls, which are believed to interfere with polymerization. 

Alkylmetals (MR) have long been added to the reactor to increase the polymer yield of Cr/silica catalysts. They can be strong reducing agents or scavengers to remove the usual redox by-products (scheme 3). The induction time of Cr(VI)/silica is usually eliminated but the maximum activity is not greatly affected. 

The activity and termination rate of the Phillips Cr/silica polymerization catalyst have been examined in relation to the surface hydroxyl population. 

This paper examines some factors which affect not only the overall activity, but also the rate of termination of polyethylene chains growing on the Phillips Cr/silica polymerization catalyst. Although the theme of this symposium is not the termination but the initiation of polymer chains, the two aims are not inconsistent because on the Phillips catalyst the initiation and termination reactions probably occur together. They are both part of a continuous mechanism of polymerization. One possibility, proposed by Hogan, is shown below. The shift of a beta hydride simultaneously terminates one live chain while initiating another ... 

Catalysts. Two types of silica support were used in these experiments. Davison grade 952 silica had a pore volume of 1.6 cc/g and a surface area of about 280 nr/g. The other support was a coprecipitated silica-titania (3.3 wt% Ti02) having a pore volume of 2.5 cc/g and a surface area of about 450 m /g. Ordinarily both supports were first treated with chromium (III) acetate to yield 1 wt Cr. Activation was accomplished in a shallow bed fluidized by air or another gas predried through alumina columns. Gases other than air were also deoxygenated through columns of specially reduced Cr/silica-alumina catalyst. 

R/R Activation. Figure 5 shows that the enhanced dehydroxyl-ation by carbon monoxide also had a pronounced effect on the termination rate during polymerization. In these experiments, two series of Cr/silica catalyst samples were activated and allowed to polymerize ethylene to a yield of about 5000g PE/g. In one series the catalyst samples were simply calcined five hours in air as usual at the temperatures shown. The relative melt index potential (RMIP) has been plotted against activation temperature and the expected increase up to the point of sintering was observed. 

---