HK1039344A - Improved epoxide polymerization process - Google Patents

Description

Improved epoxide polymerization process

Technical Field

The present invention relates to an improved process for polymerizing epoxides using a high activity substantially amorphous double metal cyanide complex catalyst, the polyether polyols obtained having an extended processing window in the preparation of molded and slabstock polyurethane foams. More particularly, the present invention relates to the use of an aprotic lewis acid (preferably in combination with water) during the epoxide polymerization process in order to reduce the high molecular weight tail content of the synthesized polyether polyol as desired.

Background

Polyurethane polymers are prepared by reacting di-or polyisocyanates with polyfunctional isocyanate-reactive compounds, in particular hydroxyl-functional polyether polyols. There are many recognized polyurethane polymers in the prior art such as cast elastomers, polyurethane RIMs, microcellular elastomers, and molded and slabstock polyurethane foams. Each of these types of polyurethanes presents special problems in formulation and processing.

The two highest volume types of polyurethane polymers are polyurethane molding and slabstock foam. In slabstock foams, the reactive components are delivered to a moving conveyor belt and allowed to foam freely. The resulting foam blocks, often 6-8 feet (2-2.6m) wide and high, can be sliced into sheets for seating, carpet backing and other applications. The molded foam may be used in foam parts having a corrugated profile, such as cushions for automobile seats.

In the past, polyoxypropylene polyether polyols for slabstock and molded foam applications have been prepared by the base-catalyzed oxypropylation of suitable hydroxyl initiators such as propylene glycol, glycerin, sorbitol, and the like to give the respective polyoxypropylene diols, triols, and hexaols. As already mentioned in the literature, in the base-catalyzed propoxylation reaction propylene oxide is rearranged to allyl alcohol. Monofunctional, unsaturated allyl alcohols carry a hydroxyl group capable of reacting with propylene oxide, and its continued derivatization and propoxylation will result in a greater amount of unsaturated polyoxypropylene monols having a broad molecular weight distribution. As a result, the actual functionality of the polyether polyols produced is significantly lower than the "nominal" or "theoretical" functionality. Furthermore, the monoalcohol derivatives will have lower practical limits on the molecular weight obtained. For example, a base-catalyzed 4000Da (daltons) molecular weight (2000Da equivalent weight) diol may have a measured unsaturation of 0.05meq/g, and thus contain 30 mole percent of unsaturated polyoxypropylene monol species. The actual functionality obtained will be only 1.7, rather than the "nominal" functionality of 2 expected for polyoxypropylene diols. As the molecular weight increases, the problem becomes more pronounced, and it is not possible to prepare polyoxypropylene polyols having equivalent weights greater than about 2200-2300Da using conventional base-catalyzed processes.

Double metal cyanide ("DMC") complex catalysts such as zinc hexacyanocobaltate complexes were found to be catalysts for propoxylation reactions about 30 years ago. However, their high cost, coupled with the very modest activity and difficulty of removing large amounts of catalyst residues from the polyether product, has prevented its commercialization. However, the unsaturation degree of polyoxypropylene polyols produced by these catalysts was found to decrease.

The comparatively general polymerization activity of these conventional double metal cyanide-complex catalysts has been regarded as a problem by the person skilled in the art.

More recently, researchers at the company ARCO chemistry have produced DMC coordination catalysts with outstanding activity, as described in U.S. Pat. Nos. 5,470,813,5,482,908,5,545,601 and 5,712,216, which have also been found to produce polyether polyols having degrees of unsaturation in the range of 0.002 to 0.007meq/g (levels previously obtainable only by using certain solvents such as tetrahydrofuran). The polyoxypropylene polyols so produced have been found to react in certain applications (primarily cast elastomers and microcellular foams) in a quantitative manner different from existing "low" unsaturation polyols. However, such polyols are often not as simple to replace in molded and slabstock foam formulations in their base catalyzed analogs. For molded foams, for example, the foam density is increased to such an extent that: the necessary crushing operation of the foam after moulding has proven difficult, if not impossible. For both molded and slabstock foams, collapse often occurs, making it impossible to produce such foams. Even if the high actual functionality of such polyols is intentionally reduced by the addition of low functionality polyols to achieve a similar actual functionality as the base-catalyzed polyol, these adverse results may likewise occur.

DMC-catalyzed polyoxypropylene polyols have a particularly narrow molecular weight distribution, as can be seen from the gel permeation chromatogram of the polyol sample. The molecular weight distribution is often much narrower than, for example, similar base-catalyzed polyols (especially in the higher equivalent weight range). Polydispersities of less than 1.5 are generally obtained, with polydispersities in the range of 1.05 to 1.15 being common. In view of the low unsaturation and low polydispersity, it is surprising that DMC-catalyzed polyols have not proven to unexpectedly (drop-in) replace base-catalyzed polyols in polyurethane foam applications. Because the propoxylation reaction of the new DMC catalysts is highly efficient when used, it would be highly desirable to be able to produce DMC-catalyzed polyoxypropylene polyols that can be used in polyurethane molding and slabstock foam applications without causing excessive foam densification or foam collapse.

Summary of The Invention

It has now surprisingly been found that polyether polyols containing polymerized propylene oxide and mimicking the performance of base catalyzed analogs in polyurethane molding and slabstock foams can be obtained by using a highly active substantially amorphous double metal cyanide complex catalyst if during the polymerization of propylene oxide an effective amount of one or more non-protic lewis acids is present during the polymerization. The reduction in the amount of high molecular weight tail impurities in the polyether polyol may be further enhanced by the use of water in combination with a lewis acid.

The beneficial effects of the presence of a non-protic Lewis acid during the polymerization process are not yet foreseen according to the disclosure of Japanese laid-open Specification No. 2-265922. This published Japanese patent application teaches that adding a Lewis acid treating agent to a polyether prepared by a ring-opening reaction in the presence of a compound metal cyanide complex catalyst will deactivate the catalyst. Contrary to this teaching, applicants have found that the use of a low amount of an aprotic lewis acid does not significantly reduce the rate of polymerization, but rather is more effective in suppressing the production of high molecular weight tail impurities that can cause blistering.

Detailed Description

Intensive investigations into the chemical and physical properties of polyoxypropylene polyols have found that, despite the narrow molecular weight distribution and low polydispersity of polyols catalyzed by essentially amorphous, highly active double metal cyanide complex catalysts, small amounts of the high molecular weight fraction still result to a large extent in excessive foam densification (stabilization) and foam collapse.

Comparison of the gel permeation chromatograms of base-catalyzed and DMC-catalyzed polyols shows a significant difference. For example, base-catalyzed polyols show significant "lead" portions of low molecular weight oligomers and polyoxypropylene monols prior to the main molecular weight peak. After this peak, the weight percent of higher molecular weight species dropped dramatically. Similar chromatography of DMC-catalyzed polyols revealed exactly symmetrical peaks with very few low molecular weight "leading" portions, but with higher molecular weight portions (high molecular weight "trailing"), indicating the presence of very high molecular weight measurable species. Due to the low concentration of these high molecular weight substances, generally below 2-3 wt% of the total amount, the polydispersity remains low. However, extensive research has revealed that high molecular weight species, despite their low concentration, are the primary cause of the exceptional performance of DMC-catalyzed polyols in molded and slabstock polyurethane foam applications. It is speculated that these high molecular weight materials act like surfactants, which alter the solubility and phase-out of the growing polyurethane polymer in the isocyanate-polyol reaction.

By fractional distillation and other techniques, it has been determined that the high molecular weight tail can be separated into two molecular weight fractions based on the different effects of these fractions. The first fraction, referred to herein as the "medium molecular weight tail", consists of polymer molecules having a molecular weight in the range of about 20,000Da to 400,000Da and greatly varies the foam density of molded foams and High Resilience (HR) slabstock foams. The higher molecular weight fraction (hereinafter referred to as the "ultra-high molecular weight tail") significantly affects the collapse in molded foams and slabstock foams that are of ordinary and High Resilience (HR).

A completely effective method for avoiding the generation of a high molecular weight tail during propoxylation with DMC complex catalysts is not known in the prior art to date. The use of processes such as continuous addition of starter in both batch and continuous polyol preparation, as disclosed in WO97/29146 and U.S. Pat. No.5,689,012, has been shown to be partially effective in reducing the amount of high molecular weight tail in some cases. However, the remaining portion is still not optimal if the polyether polyol is used to make polyurethane foams. An industrially acceptable method of removing or destroying the high molecular weight tail has not been developed. Peroxide-induced cleavage resulting in destruction of high molecular weight species is somewhat effective, but also cleaves species of the desired molecular weight. Fractionation using supercritical carbon dioxide is effective for some polyols but not for others and is too costly to be commercially acceptable.

It has now been unexpectedly found that the problem of reducing the high molecular weight tail in polyether polyols obtained by the use of a substantially amorphous highly active double metal cyanide complex catalyst can be readily solved by the presence of an effective amount of a non-protic lewis acid in the alkoxylation reaction. A lewis acid is any molecule or ion capable of binding another molecule or ion (often referred to as an electrophile) by forming a covalent bond with two electrons from a second molecule or ion. Lewis acids are therefore electron acceptors. The term "aprotic" as used herein means other than capable of acting as a proton (H)⁺) A lewis acid other than the source.

In one embodiment of the present invention, the selected lewis acid is soluble in the polyether polyol produced. However, insoluble non-protic lewis acids, including lewis acids supported or otherwise immobilized on an insoluble substrate, may also be used if desired. Particularly preferred lewis acids for the purposes of the present invention include halides (i.e. fluorides, chlorides, bromides, iodides) of elements selected from the group consisting of aluminum, manganese, iron, cobalt and zinc. Other elements including, for example, halides of boron, iron, titanium, tin, chromium, magnesium, vanadium, hafnium, zirconium, and the like, may also be used. The lewis acid may contain substituents other than halogen groups. Specific examples of suitable lewis acids include, but are not limited to, zinc chloride, zinc bromide, zinc fluoride, aluminum trichloride, aluminum tribromide, aluminum trifluoride, stannous chloride, stannous bromide, ferric chloride, ferrous chloride, boron trifluoride, magnesium chloride, manganese dichloride, cobalt chloride, and the like, and mixtures thereof.

The amount of non-protic lewis acid should be sufficient to reduce the amount of high molecular weight tail in the polyether polyol to the desired extent. At the same time, however, the use of an excess of lewis acid should generally be avoided in order to maintain a higher catalytic activity. Typically, the amount of lewis acid present during the polymerization process needs to be adjusted so that the activity of the double metal cyanide complex catalyst, as measured by the weight of epoxide participating in the reaction per unit time at a given temperature, does not decrease by more than 20% compared to the catalytic activity under the same conditions without the lewis acid. In general, it is desirable to use an amount of the Lewis acid effective to result in a polyether polyol product having properties in the supercritical foam test or in the dense foam test that are more similar, preferably substantially similar, to the properties of a similar polyether polyol produced by conventional KOH catalyzed polymerization than would be the case in the absence of the Lewis acid. The most suitable amount will, of course, vary depending upon the lewis acid selected, the polymerization conditions, and the composition, amount, and activity of the double metal cyanide complex catalyst, but can readily be determined by routine experimentation. Typically, the Lewis acid is used in an amount of 0.1 to 200ppm by weight based on the weight of polyether polyol prepared (which is generally equal to the total weight of epoxide and initiator). Concentrations of Lewis acid in the range of from 0.5 to 50ppm are generally preferred, especially when the catalyst concentration is in the range of from 5 to 50 ppm. It is often advantageous to use a lewis acid in the range of about 0.1 to 1.0: the weight ratio of the catalyst. The Lewis acid is most preferably added to the polymerization reactor all at once, together with the initiator and double metal cyanide complex catalyst, prior to introducing the epoxide and initiating the polymerization. If desired, the Lewis acid and catalyst may be premixed or otherwise mixed prior to addition to the reactor.

In a particularly preferred embodiment of the present invention, water is also present in the double metal cyanide complex-catalyzed epoxide polymerization. The presence of water appears to greatly increase the efficiency of the lewis acid and improve the reproducibility of the results obtained (for reasons not yet fully understood). That is, the amount of high molecular weight tail impurity produced and the amount of this impurity varies from batch to batch generally much lower when water is present with the lewis acid than when water is strictly excluded. The amount of water is not considered critical but should be high enough to achieve the desired reduction in high molecular weight tail or improvement in reproducibility. Since water can be used as an initiator for the polymerization reaction to produce difunctional polyoxyalkylene glycol species, the use of excess water should generally be avoided when polyether polyols having a functionality other than 2 (e.g. 3) are the desired product. If a difunctional polyether polyol is the desired product, then the amount of water used should be taken into account when calculating the amount of difunctional initiator needed to achieve a given molecular weight during the polymerization. Higher amounts of water also deactivate the catalyst or interfere with facile catalyst activity. The optimum amount of water can be readily determined by routine experimentation, but is generally in the range of about 1-500ppm based on the weight of the polyether polyol. When zinc chloride is used as the Lewis acid in an amount of from 5 to 25ppm based on the weight of the polyether polyol, and the double metal cyanide complex catalyst is present in an amount of from 20 to 40ppm based on the weight of the polyether polyol, the amount of water present is preferably from about 5 to 100 ppm. One skilled in the art will recognize that the catalysts, initiators, solvents and lewis acids used in the present invention often contain water (unless strictly excluded) as an impurity in an amount sufficient to improve the effectiveness of the lewis acid described above. The water can of course also be added separately from these reaction components or else be introduced.

The double metal cyanide catalysts used in combination with the lewis acid are substantially amorphous (meaning that no strong, sharp peaks are observed in the X-ray diffraction pattern of the catalyst) and consist of a double metal cyanide, an organic complexing agent and a metal salt. The catalyst has very high polymerization activity; that is, it is capable of polymerizing propylene oxide at 105 ℃ at a rate in excess of 3g (more preferably 5g) propylene oxide per minute per 250ppm catalyst (based on the total weight of initiator and propylene oxide). Double metal cyanide complex catalysts meeting these requirements and methods for their preparation are described in detail in U.S. Pat. Nos. 5,470,813,5,482,908,5,545,601 and 5,712,216, each of which is incorporated herein by reference in its entirety.

The double metal cyanide is most preferably zinc hexacyanocobaltate, while the metal salt (used in excess in the reaction to form the double metal cyanide) is preferably selected from the group consisting of zinc halides (zinc chloride is particularly preferred), zinc sulfate and zinc nitrate. It is believed that the metal salt in the catalyst is not present as a free metal salt, but is somehow intimately associated or complexed with the double metal cyanide matrix of the catalyst. The metal salt contained in the catalyst may in certain embodiments of the invention have the same chemical properties as the non-protic lewis acid introduced during the polymerization process. For example, the metal salt and the lewis acid may both be zinc chloride. However, for reasons not yet fully understood, modifying the catalyst synthesis to retain higher levels of residual zinc chloride in the isolated catalyst will not tend to inhibit the formation of high molecular weight tails as effectively as zinc chloride is added directly to the polymerization reaction mixture. Thus, in a preferred embodiment of the present invention, the non-protic Lewis acid is introduced separately from the double metal cyanide complex catalyst.

The organic complexing agent is desirably selected from the group consisting of alcohols, ethers and mixtures thereof, with water-soluble aliphatic alcohols such as t-butanol being particularly preferred. Double metal cyanide complex catalysts are desirably modified with polyethers as described in U.S. Pat. Nos. 5,482,908 and 5,545,601.

The catalyst concentration is generally selected so that sufficient catalyst is present to polymerize the epoxide at a desired rate or over a desired time. For economic reasons and to avoid catalyst removal from the polyether polyols produced, it is desirable to minimize the amount of catalyst used. The activity of the catalyst used in the present invention is extremely high; catalyst concentrations in the range of 5-5ppm (based on the total weight of active hydrogen-containing initiator and epoxide) are generally sufficient.

The process of the present invention is particularly useful for the polymerization of propylene oxide alone, since propylene oxide homopolymerization is particularly prone to the formation of undesirably large high molecular weight tails. However, the process can also be used to polymerize other epoxides such as ethylene oxide, 1-butene oxide, and the like, alone or with other epoxides. For example, copolymers of ethylene oxide and propylene oxide may be produced.

The active hydrogen-containing initiator may be any material known in the art capable of being alkoxylated with an epoxide using a double metal cyanide complex catalyst and is selected based on the desired functionality and molecular weight of the polyether polyol product. Typically, the initiator (also referred to as "starter") will be oligomeric and have a number average molecular weight in the range of 100 to 1000 and a functionality (number of active hydrogens per molecule) of 2-8. Alcohols (i.e. organic compounds containing one or more hydroxyl groups) are particularly preferred for use as initiators.

The polymerization reaction may be carried out by any alkoxylation method known in the art using double metal cyanide complex catalysts. For example, a conventional batch process may be used in which the catalyst, lewis acid and initiator are charged to a batch reactor. The reactor is then heated to the desired temperature (e.g., 70 ℃ to 150 ℃) and an initial portion of the epoxide is introduced. Once the catalyst has been activated, the remaining portion of the epoxide is gradually added with good mixing of the reactor contents and the reaction proceeds until the desired molecular weight of the polyether polyol product is reached, as indicated by the pressure drop and consumption of the initial epoxide charge. Then, if necessary, the Lewis acid is removed from the polyether polyol by adsorption, ion exchange, or the like. In addition, the lewis acid may remain in the polyether polyol, as it is not expected to affect the performance of the polyether polyol in many end-use applications, particularly if present in lower amounts. The initiators, monomers and polymerization conditions described in U.S. Pat. No.3,829,505, the entire contents of which are incorporated herein by reference, can readily be used in the process of the present invention with minor variations.

Although the presence of a lewis acid is highly desirable in the initiation process (catalyst activation), in another embodiment of the invention, the lewis acid is added after initiation. Additional portions of the lewis acid may also be introduced as the polymerization reaction proceeds (e.g., during the addition of the epoxide).

Alternatively, a conventional continuous process may be used wherein the previously activated initiator/catalyst/lewis acid mixture is continuously added to a continuous reactor such as a Continuous Stirred Tank Reactor (CSTR) or a tubular reactor. A feed of epoxide is introduced into the reactor and the product is continuously withdrawn. The process of the present invention can also be readily adapted for continuous addition of starter (initiator) processes (batch or continuous operation), such as those described in detail in U.S. patent application serial No. 08/597,781 (filed 2/7/1996), now U.S. patent No. ___________, and U.S. patent No.5,689,012, both of which are incorporated herein by reference in their entirety.

The polyether polyols produced by the operation of the process of the present invention preferably have functionality, molecular weight and hydroxyl number suitable for use in molded and slabstock foams. The nominal functionality typically ranges from 2 to 8. Generally, the polyether polyol blend has an average functionality in the range of about 2.5 to 4.0. The equivalent weight of the polyether polyol is generally in the range of from slightly below 1000Da to about 5000 Da. The degree of unsaturation is preferably 0.025meq/g or less. The hydroxyl number is preferably from 10 to about 80. The blend may of course contain polyols having lower and higher functionalities, equivalent weights and hydroxyl numbers.

The properties of polyether polyols can be analyzed by testing these polyether polyols in the "density foam test" (TFT) and the "supercritical foam test" (SCFT). Polyether polyols passing these tests were found to perform well in industrial molding and slabstock foam applications without excessive densification and without foam collapse. The SCFT consists of preparing a polyurethane foam using a formulation that is specifically designed to amplify the differences in polyether polyol properties.

In SCFT, a foam prepared from a given polyether polyol is recorded as "settled" if the foam surface appears convex after jetting (blow-off), and as collapsed if the foam surface is concave after jetting. The amount of collapse can be recorded in a relatively quantitative manner by calculating the percent change in cross-sectional area across the foam. The foam formulation was as follows: polyether polyol, 100 parts; 6.5 parts of water; 15 parts of dichloromethane; 0.10 portion of Niax  A-1 amine catalyst; 0.34 portion of T-9 tin catalyst; l-550 siloxane surfactant, 0.5 part. The foam was reacted with 80/20 of a mixture of 2, 4-and 2, 6-toluene diisocyanate at an index of 110. The foam can be conveniently poured into a standard 1 cubic foot cake box, or a standard 1 gallon ice cream container. In these formulations, base-catalyzed polyether polyols prepared in the conventional manner, i.e., having a high secondary hydroxyl content, cause foam collapse to be about 10-20%, typically 15% + -3%, while polyether polyols prepared from DMC catalysts which contain too high a level of high molecular weight tail cause foam collapse to be about 35-70%.

When SCFT is used to assess the difference in foam stability, the density foam test (TFT) amplifies the difference in reactivity, which is reflected in the porosity of the foam. In the density foam test, the resin component consisted of 100 parts polyether polyol, 3.2 parts water (reactive blowing agent), 0.165 parts C-183 amine catalyst, 0.275 parts T-9 tin catalyst and 0.7 parts L-620 siloxane surfactant. The resin component was reacted with 80/20 toluene diisocyanate at an index of 105. The foam density is analysed by measuring the aeration in a conventional manner. Dense foams reduce air flow.

The analytical method used to measure the amount of high molecular weight tail in a given DMC-catalyzed polyether polyol is a conventional HPLC technique, which can be readily accomplished by one skilled in the art. The molecular weight of the high molecular weight fraction can be measured by comparing its elution time in a GPC column with that of a polystyrene standard of the appropriate molecular weight. It is well known that the high molecular weight fraction elutes from the GPC column more rapidly than the low molecular weight fraction and assists in maintaining a stable baseline, and it is convenient to direct the remainder of the HPLC eluent to waste after elution of the high molecular weight fraction, rather than passing it through a detector, burdening the latter. While many suitable detectors may be used, a convenient detector is an Evaporative Light Scattering Detector (ELSD), such as those commercially available.

In a preferred analytical method, a Jordi Gel DVB 103 angstrom column, 10X 250mm, 5 micron particle size, was used with a mobile phase consisting of tetrahydrofuran. The detector used was a Varex Model IIA evaporative light scattering detector. Polystyrene stock solutions were prepared by appropriate dilution of polystyrene of different molecular weights with tetrahydrofuran, forming standards containing 2, 5 and 10mg/l polystyrene.

The test specimens were prepared by weighing 0.1 grams of the polyether polyol into a 1 ounce bottle and then adding tetrahydrofuran to the sample to bring the total weight of the sample and tetrahydrofuran to 10.0 grams. 2. Samples of 5 and 10mg/l polystyrene calibration solutions were injected sequentially into the GPC column. The injection of each polyether polyol sample solution was then repeated, followed by the injection of each polystyrene standard. The peak areas of the polystyrene standards were electronically integrated, and the two electronically integrated peaks for each polyol candidate were electronically integrated and the average was calculated. The calculation of the high molecular weight tail (in ppm) was done by standard data manipulation techniques.

Having generally described this invention, an understanding can be enhanced by reference to certain specific embodiments which are provided by way of illustration only and not limitation unless otherwise specified.

Examples

A series of epoxide polymerizations were conducted using a semi-batch reactor with a nominal 1 liter volume. In each operation 167g of a basic charge of trifunctional polyether polyol having a hydroxyl value of 240mg KOH/g were introduced into the reactor as initiator (starter). A double metal cyanide complex consisting of zinc hexacyanocobaltate, t-butanol, zinc chloride and a polyether prepared according to US patent No.5,482,908 is then added to the reactor. The amount of catalyst used is normally 30ppm based on the final weight of the polyether polyol product, although other catalyst amounts may be evaluated. After catalyst addition, the initiator/catalyst mixture is normally extracted at 130 ℃ for 30-90 minutes using a full vacuum and nitrogen purge in an attempt to completely remove any water (no measurement of the actual amount of water present is made). This was used to establish a "waterless" baseline for each batch.

Once the extraction is complete, zinc chloride and water are added to the reactor. This is typically accomplished by using a premixed aqueous solution of zinc chloride. After the addition of the Lewis acid and water, the addition reaction of propylene oxide begins. The first portion of propylene oxide added is the initiator charge. Propylene oxide was added until the reactor pressure reached 35 psia. The feed of propylene oxide was then stopped and the reactor pressure was monitored. When sufficient epoxide reacted to drive the pressure down to half its peak volume (i.e., about 17.5 psig), the catalyst was considered to have been activated. Once activated, the propylene oxide feed was restarted at a rate of 6.5 g/min. A continuous epoxide feed was maintained until propylene oxide was added in an amount sufficient to obtain a final polyether polyol having a hydroxyl value of 56mg KOH/g.

The polyether polyol product was tested by two main methods. Various properties including hydroxyl number, molecular weight distribution (including quantitative measurement of high molecular weight species), unsaturation, and viscosity were determined analytically by common methods. The foam properties were evaluated by using the supercritical foam test method already described in the detailed description of the invention section of the specification. The degree of foam settling (or collapse) was compared to a control polyether polyol produced using a conventional KOH catalyzed alkoxylation process.

Key results are summarized in the following series of tables, each showing the effect of varying different key process parameters. All concentrations are given in ppm (parts per million by weight) based on the weight of the final polyether polyol product.

Table I illustrates the effect of using a combination of zinc chloride and water in reducing the amount of higher molecular weight species (> 200,000) in a polyether polyol product. When no water or zinc chloride was added (example 1), the polyether polyol failed the supercritical foam test and settled (collapsed) to a greater extent than the control polyether polyol produced using the KOH catalyst. However, when 5-10ppm ZnCl is present₂And 10ppm of added water (examples 2 and 3), the polyether polyol passed the supercritical foam test and performed substantially equally to the KOH catalyzed product. A significant reduction in the amount of impurities with molecular weights above 200,000, believed to be the primary cause of the collapse, was observed. Although the amount of medium molecular weight (40,000-60,000) materials increases when zinc chloride and water are introduced, these materials do not affect the foam properties. When the zinc chloride concentration is too high (example 4), the catalyst does not initiate. Therefore, under this particular polymerization condition and at a catalyst concentration of 30ppm, an effective range of zinc chloride concentration is about 5-25 ppm.

TABLE 1

Examples	Catalyst, ppm	ZnCl2,Ppm	Added water, ppm	40-60K Mw fraction, ppm	200-400K Mw fraction, ppm	Mw fraction > 400K, ppm	Foam test results,% Settlement	KOH control foam test,% Settlement
									11	30	0	0	238	42	7	44.2	13.9
2	30	5	10	1294	16	0	13.4	13
									3	30	10	10	1051	2	0	16.3	13.9
42	30	30	＜30	-	-	-	-	-

¹Control (comparative example)²Fail to cause

Table II illustrates the effect of using water in combination with zinc chloride. When the amount of water is low (examples 5 and 6), the foam properties of the polyether polyol are similar to those of a polyether polyol prepared without the addition of zinc chloride (example 1). When the water concentration was higher (example 9), the activation of the catalyst was inhibited and the polymerization reaction could not be successfully completed. These results indicate that a suitable range of water concentration is about 10-45ppm under the particular polymerization conditions being evaluated.

TABLE 2

Examples	Catalyst, ppm	ZnCl2,ppm	Added water, ppm	40-60KMw fraction, ppm	200-400KMw fraction, ppm	Mw fraction > 400K, ppm	Foam test results,% Settlement	KOH control foam test,% Settlement
									11	30	0	0	238	42	7	44.2	13.9
52	30	10	0	353	462	24	37.6	15.7
									6	30	10	6.7	680	21	0	46.8	13.8
3	30	10	10	1051	2	0	16.3	13.9
									7	30	10	20	1008	7	0	17.2	13.9
8	30	10	30	968	17	6	12.3	13.8
									93	30	10	50	-	-	-	-	-

¹Control (comparative example)²In the presence of ZnCl₂The initiator/catalyst mixture was then stripped under vacuum for 150 minutes to remove the water³Fail to cause

Table III shows the results relative to the amounts addedThe presence of zinc chloride and water, and the effect of varying the catalyst concentration. The results show that optimum ZnCl is present at each catalyst concentration₂The catalyst ratio, the amount of high molecular weight material is minimized. When the catalyst concentration was increased from 30 to 45ppm, the corresponding increase in zinc chloride and water levels was sufficient to effectively suppress the formation of high molecular weight species (examples 3,10 and 11). Under these polymerization conditions, ZnCl is preferred₂The weight ratio/catalyst is in the range of 0.2 to 0.5.

TABLE 3

Examples	Catalyst, ppm	ZnCl2,ppm	Added water, ppm	40-60K Mw fraction, ppm	200-400K Mw fraction, ppm	Mw fraction > 400K, ppm	Foam test results,% Settlement	KOH control foam test,% Settlement
									11	30	0	0	238	42	7	44.2	13.9
3	30	10	10	1051	2	0	16.3	13.9
									10	30	15	15	942	12	14	11.5	13.8
11	40	15	15	617	0	0	11.8	13.4

¹Control (comparative example)

A further series of epoxide polymerizations was carried out using a double metal cyanide complex catalyst of the same type as described in the preceding examples, in which the ratio of trifunctional polyether polyol initiator with respect to propylene oxide was varied at a hydroxyl number of 240mg KOH/g in order to obtain a final product with a molecular weight of about 3000. Different amounts of aluminium chloride or zinc chloride were introduced (except in example 16, which is a comparative experiment).

In examples 14-17, the reaction mixture was stripped under vacuum (at 105 ℃ C., except for example 15, which was stripped at 130 ℃ C.) before the initial charge of propylene oxide was added. The time required for the addition of propylene oxide is generally about 120 minutes, although in some cases longer times are required because of the reduced catalyst activity.

The polyether polyol product thus obtained was characterized in the same manner as described previously, the key results being summarized in table 4. Both zinc chloride and aluminum chloride were found to be effective in inhibiting the formation of high molecular weight impurities and improving the foaming properties of the polyether polyol in the supercritical foam test. Other lewis acids (zinc bromide, zinc acetate, zinc sulfate, ferric chloride, ferrous chloride) are less effective under these conditions.

TABLE 4

Examples	12	13	14	15	16*	17
							Lewis acid ppm	AlCl310	AlCl35	AlCl320	AlCl320	--	ZnCl210
Hydroxyl number, mg KOH/g	55.7	55.7	57.5	56.2	56.2	56.5
							Viscosity, cps (25 ℃ C.)	575	558	897	616	566	652
Polydispersity (Mw/Mn)	1.042	1.029	1.508	1.119	1.028	1.152
							Mw fraction > 100K, ppm	147	115	85	54	121	20
Mw fraction > 400K, ppm	6	5	nd	nd	10	nd
							% sedimentation rate of supercritical foam test	23.75	13.08	14.26	12.47	30.9	9.88
KOH control foam test,% Settlement	12.98	10.75	10.75	10.25	11	10.25

Control

nd = no detection

Claims (24)

1. An epoxide polymerization process comprising reacting an epoxide and an active hydrogen-containing initiator in the presence of (a) a substantially amorphous, highly active double metal cyanide complex catalyst comprising a double metal cyanide, an organic complexing agent, and a metal salt and (b) an effective amount of a non-protic lewis acid for a time and at a temperature effective to form a polyether polyol, wherein the polyether polyol contains a reduced amount of high molecular weight tail as compared to a similar polyether polyol prepared in the absence of the non-protic lewis acid.

2. The process of claim 1 wherein the non-protic lewis acid is soluble in the polyether polyol.

3. The process of claim 1 wherein the non-protic lewis acid is a halide.

4. The process of claim 3 wherein the non-protic Lewis acid is a halide of an element selected from Al, Mn, Fe, Co and Zn.

5. The process of claim 1 wherein the Lewis acid is present in an amount of from 0.1 to 200ppm based on the weight of the polyether polyol.

6. The process of claim 1 wherein the double metal cyanide compound is zinc hexacyanocobaltate.

7. The method of claim 1 wherein the metal salt is selected from the group consisting of zinc halides, zinc nitrate, zinc sulfate, and mixtures thereof.

8. The method of claim 1 wherein the organic complexing agent is a water-soluble aliphatic alcohol.

9. The process of claim 1 wherein the polyether polyol has a functionality of from 2 to 8 and an equivalent weight of from 1000 to 5000.

10. The process of claim 1 wherein the double metal cyanide complex catalyst additionally comprises a polyether.

11. The process of claim 1 wherein water is additionally present.

12. A process for producing a polyether polyol comprising

In (a) a substantially amorphous zinc hexacyanocobaltate complex catalyst comprising zinc hexacyanocobaltate, a water soluble aliphatic alcohol, and a metal salt, the zinc hexacyanocobaltate complex catalyst is capable of polymerizing propylene oxide at 105 ℃ at a rate in excess of 5g propylene oxide per minute per 250ppm (based on the total weight of active hydrogen-containing initiator and propylene oxide) of zinc hexacyanocobaltate complex catalyst, (b) an effective amount of a non-protic lewis acid selected from the group consisting of halides of zinc, manganese, iron, cobalt and aluminum and mixtures thereof, and (c) reacting an epoxide and an active hydrogen-containing initiator in the presence of an effective amount of water for a time and at a temperature effective to form a polyether polyol, wherein the polyether polyol contains a reduced amount of high molecular weight tail compared to a similar polyether polyol prepared in the absence of a non-protic lewis acid and water.

13. The process of claim 12 wherein the non-protic lewis acid is selected from the group consisting of zinc chloride, aluminum chloride, and mixtures thereof.

14. The process of claim 12 wherein the water-soluble aliphatic alcohol is t-butanol.

15. The process of claim 12 wherein the non-protic lewis acid is present in an amount of from 0.1 to 200ppm based on the weight of the polyether polyol.

16. The method of claim 12, wherein the metal salt is zinc chloride.

17. The method of claim 12, wherein 5 to 100ppm of water is present.

18. The process of claim 12 wherein the epoxide is selected from the group consisting of propylene oxide and mixtures of propylene oxide and ethylene oxide.

19. The process of claim 12 wherein the polyether polyol has a functionality of from 2 to 8 and an equivalent weight of from 1000 to 5000.

20. The method of claim 12, wherein the temperature is in the range of 70 ℃ to 150 ℃.

21. The process of claim 12 wherein the concentration of the zinc hexacyanocobaltate complex catalyst is from 5 to 50ppm based on the weight of the polyether polyol.

22. A process for polymerizing an epoxide comprising:

reacting an epoxide and an active hydrogen-containing initiator in the presence of (a) a substantially amorphous double metal cyanide complex catalyst preparable from a double metal cyanide compound, an organic complexing agent, and a metal salt, wherein the polymerization reaction mixture further comprises (b) an added non-protic lewis acid; and

the polyether polyol is recovered and contains a reduced proportion of high molecular weight tail components compared to other polyether polyols prepared in a similar manner but in the absence of the added non-protic lewis acid (b).

23. Use of an added non-protic lewis acid to reduce the proportion of high molecular weight tail component in a substantially amorphous double metal cyanide complex-catalyzed epoxide polymerization reaction mixture compared to a polyether polyol prepared in a similar manner but in the absence of the added non-protic lewis acid.

24. A process according to claim 22 or a use according to claim 23, wherein the catalyst, non-protic lewis acid, reactants or reaction conditions are as defined in any one of claims 1 to 21.