Process for making polyether alcohols having oxyethylene units by polymerization of ethylene carbonate in the presence of double metal cyanide catalysts

Abstract

Ethylene carbonate is polymerized by itself or together with another cyclic monomer such as 1,2-propylene oxide in the presence of a double metal cyanide catalyst. Most of the ethylene carbonate adds to the chain to form a terminal carbonate group, which decarboxylates to produce a hydroxyethyl group at the end of the polymer chain. The polymerization of more ethylene carbonate onto the chain end results in the formation of poly(ethyleneoxy) units. Therefore, the process provides a method for making poly(ethyleneoxy) polymers without the need to polymerize ethylene oxide. The process is useful for making polyethers that are useful as water-absorbable polymers, surfactants and as raw materials for polyurethanes. The process is also useful for increasing the primary hydroxyl content of a polyether.

Claims

1. A process for preparing a polyether containing ether-linked ethyleneoxy units, comprising copolymerizing a mixture of 5-35 weight-% ethylene carbonate and 65-95 weight-% 1,2-propylene oxide, based on the combined weight of the ethylene carbonate and 1,2-propylene oxide, wherein the ethylene carbonate is dissolved in the 1,2-propylene oxide prior to being polymerized, in the presence of an initiator compound having one or more oxyalkylatable groups and a double metal cyanide catalyst and in the substantial absence of ethylene oxide, wherein 80 to 98% of the ethyleneoxy units are ether-linked.

2. The process of claim 1 wherein the oxyalkylatable groups are hydroxyl, primary amino or secondary amino groups, or a mixture of any two or more of such groups.

3. The process of claim 2 wherein the oxyalkylatable groups are hydroxyl groups.

4. The process of claim 3 wherein the initiator contains from 2 to 8 hydroxyl groups per molecule and has a hydroxyl equivalent weight of from 70 to 2500.

5. The process of claim 3 wherein the initiator is a polyether that has a hydroxyl equivalent weight of from 500 to 2000.

6. The process of claim 1 wherein carbon dioxide is removed at least once during the polymerization.

7. The process of claim 1, wherein the product contains carbonate groups.

8. The process of claim 1, wherein the product contains terminal poly(ethyleneoxy) groups.

9. The process of claim 8, which contains terminal primary hydroxyl groups.

10. The process of claim 1, further comprising continuing the polymerization in the presence of ethylene carbonate alone or together with up to 25% by weight of 1,2-propylene oxide to introduce terminal poly(ethyleneoxy) groups and increase the primary hydroxyl content of the polyether.

Description

EXAMPLES 1-6 AND COMPARATIVE SAMPLE A

(1) A polyether polyol (Comparative Sample A) having a target molecular weight of 3300 g/mol is made as follows: A 700 molecular weight poly(propylene oxide) triol is charged to a reactor. Enough of a zinc hexacyanocobaltate catalyst complex is added to provide 75 ppm of catalyst complex in the final product. 4% of propylene oxide (based on the weight of the starting triol) is added, and the reactor contents are heated to 150° C. The pressure inside the reactor is monitored until a decrease in pressure indicates that the catalyst has become activated. Activation is seen after 10-20 minutes. Once the catalyst is activated, propylene oxide is fed into the reactor over 3.5 hours. Enough propylene oxide is fed over this time to produce the target molecular weight. After the propylene oxide feed is completed, the reaction is digested at 100° C. until a constant reactor pressure is obtained, indicating that all of the propylene oxide has been consumed. The product is recovered and stripped to removal volatiles including residual monomer. Molecular weight is determined by GPC. The number average molecular weight (M.sub.n) of the product is 2770, the weight average molecular weight (M.sub.w) is 2900 and the polydispersity (M.sub.w/M.sub.n) is 1.05. Hydroxyl number is 51.7, consistent with a hydroxyl equivalent weight of 1085. 9% of the terminal hydroxyl groups are primary hydroxyls, which form due to some head-to-head polymerization of the propylene oxide. Unsaturation is measured to be 0.07 meq/g. Viscosity at 25° C. is 602 cSt.

(2) Polyol Examples 1, 2 and 3 are made in the same manner, except that, after the catalyst is activated with propylene oxide, the subsequent propylene oxide feed contains 1%, 5% or 10% by weight, respectively, of ethylene carbonate. In Examples 1-3, the ethylene carbonate is pre-dissolved into the propylene oxide to form a compatible mixture in which the propylene oxide and ethylene carbonate do not react with each other. A pressure rise in the reactor is seen during the feed step, followed by an initial drop in reactor pressure due to the polymerization of the monomers. A second pressure rise is then seen during the digestion step. This latter pressure rise is indicative of carbon dioxide production and is seen to become larger as the proportion of ethylene carbonate in the feed is increased from Example 1 to Example 2 to Example 3. This confirms that ethylene oxide has polymerized and decarboxylation has occurred. The resulting products are characterized as indicated in Table 1.

(3) TABLE-US-00001 TABLE 1 Comp. Property Sample A* Example 1 Example 2 Example 3 Wt. % ethylene car- 0 1 5 10 bonate in monomer feed M.sub.n, g/mol 2770 2780 2740 2710 M.sub.w, g/mol 2900 2930 2900 2870 Polydispersity 1.05 1.05 1.06 1.06 Hydroxyl number 51.7 51.6 51.8 53.6 Viscosity, cSt 602 615 611 630 Unsaturation, meq/g 0.007 0.007 0.004 0.004

(4) As can be seen from the data in Table 1, the inclusion of ethylene carbonate in the propylene oxide feed has little effect on molecular weight, polydispersity, hydroxyl number and product viscosity. Unsaturation values tend to be reduced, which is beneficial.

(5) These products are analyzed by MALDI-MS and NMR to determine whether and how the ethylene carbonate becomes incorporated into the polymer chain. The Example 1 product contains about 0.6 weight-% by weight ethyleneoxy units, which are distributed almost entirely as single ethyleneoxy units (i.e., bonded to two adjacent propyleneoxy units). This is indicative of a random or pseudorandom polymerization. Some carbonate units are present (about 0.14% of the total weight of the product), but only about 1 carbonate linkage is found for each 10 ethyleneoxy units. These results indicate that about 87% of the fed ethylene carbonate has reacted into the polymer backbone, and that about 90% of the reacted ethylene carbonate subsequently decarboxylated to form ether linked ethyleneoxy units. These results demonstrate that ethylene carbonate can substitute for ethylene oxide in this polymerization.

(6) The results of the MALDI-MS and NMR analyses of Examples 2 and 3 are as indicated in Table 2, together with the results for Comparative Sample A and Example 1.

(7) TABLE-US-00002 TABLE 2 Comp. Property Sample A* Example 1 Example 2 Example 3 Wt. % ethylene car- 0 1 5 10 bonate in monomer feed Ethyleneoxy groups/ 0.0 0.5 1.7 4.8 molecule Carbonate groups/ 0.0 0.05 0.2 0.6 molecule % Ether-linked 0.0 90 88 88 ethyleneoxy groups Propyleneoxy groups/ 66 58 55 56 molecule Weight-% ethyleneoxy 0.0 0.6 2.2 5.9 groups Weight-% ethyleneoxy 0.0 0.14 0.5 1.3 carbonyls Weight-% propylene- 97.6 96.6 94.5 90.2 oxy units Primary OH content, 9.0% 9.0% 13.5% 15.9% %

(8) The MALDI-MS data indicate that in Examples 2 and 3, the ethyleneoxy units are almost all distributed as single, double or triple units (i.e., 1, 2 or 3 ethyleneoxy groups in succession), which reflects that ethylene carbonate copolymerizes with propylene oxide in a random or pseudo-random manner. Some of the ethylene carbonate is polymerized at the end of the polymer chains, as indicated by the increasing primary hydroxyl content of Examples 2 and 3. The level of ethyleneoxy units increases with the ethylene carbonate concentration in the monomer feed.

(9) Examples 4-6 are made in the same manner as Examples 1-3, respectively, except that in each case the reactor is vented to remove accumulated carbon dioxide three hours after the start of the monomer feed and vented again at the end of the digestion. The products are analyzed as indicated with respect to Examples 1-3. Results are as indicated in Tables 3 and 4.

(10) TABLE-US-00003 TABLE 3 Comp. Property Sample A* Example 4 Example 5 Example 6 Wt. % ethylene car- 0 1 5 10 bonate in monomer feed M.sub.n, g/mol 2770 2801 2770 2730 M.sub.w, g/mol 2900 2920 2940 2880 Polydispersity 1.05 1.04 1.06 1.05 Hydroxyl number 51.7 52.0 52.4 53.1 Viscosity, cSt 602 621 621 618 Unsaturation, meq/g 0.007 0.007 0.006 0.006

(11) TABLE-US-00004 TABLE 4 Property Example 4 Example 5 Example 6 Wt. % ethylene carbonate in 1 5 10 monomer feed Ethyleneoxy groups/molecule 0.6 2.4 3.6 Carbonate groups/molecule 0.02 0.2 0.4 % Ether-linked ethyleneoxy units 97 92 89 Propyleneoxy groups/molecule 61 62 62 Weight-% ethyleneoxy groups 0.7 2.8 4.1 Weight-% ethyleneoxy carbonyls 0.1 0.5 0.9 Weight-% propyleneoxy units 96.7 94.3 92.6 Primary OH content, % 8.2% 13.7% 15.0%

(12) As can be seen from the data in Tables 3 and 4, the products of Examples 4-6 are similar to those of Examples 1-3, respectively. Similar to the foregoing, the MALDI-MS data indicates that in Examples 5 and 6, the ethyleneoxy units are almost all distributed as single, double or triple units. The ethyleneoxy units in Example 4 are distributed almost exclusively as single units, due to the lower concentration of ethylene oxide in the monomer feed. These results reflect that ethylene carbonate copolymerizes with propylene oxide in a random or pseudo-random manner. Decarboxylation occurs at least 90% of the time in all cases. Some of the ethylene carbonate is polymerized at the end of the polymer chains, as indicated by the increasing primary hydroxyl content of Examples 2 and 3.

(13) The proportions of carbonate linkages in the products of Examples 4-6 are reduced, relative to the corresponding Examples 1-3, respectively. This effect is due to the reactor venting that is performed in Examples 4-6; this venting reduces reactor pressure and removed carbon dioxide from the reactor head space, each of which favors lower incorporation of carbonate linkages into the product. Still lower proportions of carbonate linkages can be produced by further reducing reactor pressure and/or more frequent venting. Conversely, higher proportions of carbonate groups can be produced by increasing the reactor pressure (relative to Examples 1-3).

(14) Polyurethane foams are made using the polyols produced in Comparative Sample A and Examples 3, 4 and 5. The foams are made by blending the ingredients listed in Table 5 below at room temperature and pouring the resulting mixture into a box lined with a thermoplastic film.

(15) TABLE-US-00005 TABLE 5 Ingredient Parts by Weight Polyol.sup.1 100 70% bis(dimethylaminoethyl)ether solution 0.05 33% triethylenediamine solution.sup.2 0.075-0.15  Silicone surfactant 0.7 Tin octoate.sup.2 0.18-0.20 Water 3.0 80/20 mixture of 2,4- and 2,6-toluene diisocyanate To 100 index .sup.1The polyols are the products of Comparative Sample A, Example 3, Example 4 and Example 5. .sup.2Catalyst amounts are varied to obtain comparable free-rise foam heights.

(16) Density, sag factor (ratio of 65% to 25% indentation force deflection), hysteresis loss and compression force depression at 50% compression are measured. Results are as indicated in Table 6.

(17) TABLE-US-00006 TABLE 6 Comp. Sample A Example 3 Example 4 Example 5 Property polyol polyol polyol polyol % ethylene carbonate 0 10 1 5 in propylene oxide feed Density, kg/m.sup.3 36.1 36.2 35.1 34.1 Sag factor 2.82 2.58 2.49 2.57 Hysteresis loss, % 71.4 70.4 70.8 70.8 CFD (50%), kPa 4.88 5.14 4.7 4.04

(18) The results in Table 6 show that the polyols produced in accordance with the invention process similarly to the control. Foam properties also are comparable to the control.

EXAMPLES 7 AND 8

(19) For Example 7, a polyol is prepared as follows: A 4000 molecular weight poly(propylene oxide) triol is prepared by adding propylene oxide onto a 700 molecular weight poly(propylene oxide) triol initiator in the presence of 150 ppm (based on product weight) of a zinc hexacyanocobaltate catalyst complex. The resulting polyol, including catalyst residues, is stored overnight under nitrogen. Then 5% of ethylene carbonate (based on the weight of the 4000 molecular weight triol) is added and the mixture is heated to 160° C. for about 3 days. During this time, a pressure increase is seen in the reactor, indicating that carbon dioxide is being generated.

(20) For Example 8, a polyol is made in the same manner, except that the ethylene carbonate is mixed with propylene oxide in a 5:1 weight ratio, and the reaction is stopped after only 9 hours. Again, a pressure increase indicative of carbon dioxide is seen during the polymerization.

(21) The products of Examples 7 and 8 are analyzed as described in the previous examples. Results are as indicated in Table 7 below.

(22) TABLE-US-00007 TABLE 7 Property Example 7 Example 8 M.sub.w 4261 4371 Hydroxyl number 39.5 38.5 Viscosity, cSt 562 794 Unsaturation, meq/g 0.013 0.018 % ethylene carbonate reacted 14.9 22.7 % decarboxylation 95.4 92.7 Primary hydroxyl groups, % 9.4 10.3

(23) These results demonstrate that poly(ethylene oxy) end-caps can be introduced onto a polyol by polymerizing ethylene carbonate in accordance with the invention. As before, a large percentage of the ethylene carbonate groups that polymerize onto the end of the polymer chain decarboxylate, so that only a small proportion of carbonate linkages are introduced into the molecule. The results from Example 8 suggest that ethylene carbonate polymerization proceeds more efficiently if the ethylene carbonate is copolymerized with a small amount of propylene oxide.