Multilayer Interfacial Composite Membrane

A set of experiments was performed to prepare interfacially polymerized permselective membranes from a selection of reagents. The composite support membrane was the same in each case. Interfacially polymerized layers were formed on the support using the sets of reagents shown in Table 1.

Composite support membranes were prepared by first casting a microporous membrane on a polyester web, then coating the microporous membrane with a thin sealing layer. The casting solution, consisting of 17.5 wt % UDEL.RTM. P3500 (Amoco Co, Marietta, OH) in 82.5 wt % dimethylformamide, was doctored onto a moving non-woven, polyester web. The casting speed was 3.5 m/min and the knife gap was 178 .mu.m. The belt passed into a water bath, which precipitated the polymer to form the microporous membrane. The belt was then collected on a take-up roll, the membrane was washed overnight to remove any remaining solvent and dried to form the microporous membrane. The dip-coating operation was then performed as follows. The polysulfone membrane was fed from a feed roll through a coating station containing 1.8 wt % polydimethylsiloxane (Wacker Silicones Co., Adrian, MI), in 98.2 wt % 2,2,4-trimethylpentene. The coated membrane passed through a drying oven, and was wound up on a product roll. This operation coated the traveling microporous membrane with a liquid layer, 50 to 100 .mu.m thick. After evaporation of the solvent, a polymer film, 0.5 to 2 .mu.m thick, was left on the membrane. Similar coating operations were carried out using polymethylpentene, TPX.RTM. MX 002 (Mitsui Co., New York, NY) and a polyamide copolymer, Pebax.RTM. 4011 (Atochem Inc, Glen Rock, NJ) as the sealing layer.

The finished thickness of the sealing layer of the composite support membranes was obtained by comparing its nitrogen flux with values obtained from isotropic films of known thickness. The sealing layer was checked for integrity by measuring the ratio of the oxygen and nitrogen permeabilities. Only those supports where the measured selectivity was close to the intrinsic selectivity of the coating polymer were used in the interfacial polymerization experiments.

The composite support membranes that passed the integrity test were tested for hydrogen and carbon dioxide permeability using a permeation test cell. The permeate gas flow rate was measured with bubble flowmeters, then converted into a normalized permeation rate at a standard pressure difference (1 cmHg) and a standard membrane area (1 cm.sup.2). The pressure on the feed side of the cell was 200 psig. The permeate side was at atmospheric pressure. The test cell had a membrane area of 36.8 cm.sup.2. The results are summarized in Table 2.

An interfacially polymerized permselective layer was formed on the composite support membrane of Example 1 that had a silicone rubber sealing layer. The preparation technique was as follows. The interfacially polymerized layer was made by submerging the composite support membrane in a hexane solution containing 0.5 wt % toluene-2,4-diisocyanate.

After thirty minutes, the composite support membrane was removed from the hexane bath and held vertically for one minute to drain excess solution. The composite support membrane was then submerged in an aqueous polyethyleneimine (PEI) solution, containing 1.0 wt % PEI (MW 70,000) in 99.0 wt % distilled water. After five minutes, the membrane was removed from the aqueous bath and air-dried at room temperature.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous PEI solution containing 1.0 wt % PEI (MW 10,000) and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous piperazine solution containing 1.0 wt % piperazine and 1.0 wt % potassium hydroxide in 98.0 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous 1,3-propane diamine solution containing 1.0 wt % dimamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous 1,6-hexane diamine solution containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous 1,9-nonane diamine solution containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous 1,3-phenylene diamine solution containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

The procedure of Example 2 was repeated using a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride and an aqueous 1,4-phenylene diamine solution containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 3.

Examples 2 and 3 were prepared from a polyamine and a diisocyanate, and a polyamine and a trifunctional agent, respectively. As can be seen, the membranes prepared using polyamines, preferred for the preparation of reverse osmosis membranes, did not form adequate gas separation membranes. The selectivities obtained are essentially those of the silicone rubber sealing layer.

A series of membranes was prepared using the same reagents and general procedure as in Example 6. The same time of contact between the composite support membrane and hexane phase, and between the hexane loaded support and the aqueous phase was used in each case.

An interfacially polymerized membrane was prepared as follows. A composite support membrane was a silicone rubber sealing layer was prepared and tested for integrity as in Example 1. The interfacially polymerized layer was made by submerging the composite support membrane in a hexane solution containing 1 wt % 1,3,5-benzene tricarbonyl trichloride.

After one minute, the composite support membrane was removed from the hexane bath and held vertically for one minute to drain excess solution. The composite support membrane was then submerged in an aqueous 1,6-hexane diamine solution containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water. After one minute, the membrane was removed from the aqueous bath and air-dried at room temperature.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 4.

The procedure as in Example 10 was repeated, using immersion and contact times of five minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 4.

The procedure as in Example 10 was repeated, using immersion and contact times of 10 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 4.

The procedure as in Example 10 was repeated, using immersion and contact times of 15 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 4.

The procedure as in Example 10 was repeated, using immersion and contact times of 30 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 4.

The procedure as in Example 10 was repeated, using immersion and contact times of 60 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 4.

A series of membranes was prepared using the same reagents and general procedure as in Example 10-15. In this case, the same immersion time was used for all the experiments, but the contact time between the reagents was varied.

An interfacially polymerized membrane was prepared as follows. A composite support membrane with a silicone rubber sealing layer was prepared and tested for integrity as in Example 1. The interfacially polymerized layer was made by submerging the composite support membrane in a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride.

After 30 minutes, the composite support membrane was removed from the hexane bath and held vertically for one minute to drain excess solution. The composite support membrane was then submerged in an aqueous 1,6-hexane diamine solution containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water. After one minute, the membrane was removed from the aqueous bath and air-dried at room temperature.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 5.

The procedure as in Example 16 was repeated, using an immersion time of 30 minutes and a contact time between the reagents of 5 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 5.

The procedure as in Example 16 was repeated, using an immersion time of 30 minutes and a contact time between the reagents of 10 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 5.

The procedure as in Example 16 was repeated, using an immersion time of 30 minutes and a contact time between the reagents of 30 minutes.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 5.

The data in Table 5 show that a contact time of only one minute was sufficient in this case for the interfacial polymerization reaction to occur.

Comparing the sets of examples 10-15 and 16-19, it appears that the increase in hydrogen/carbon dioxide selectivity with increasing contact times shown in Table 4 is the result of increased penetration of the BTC into the silicone rubber sealing layer.

The data in Table 5 show that the permeation rates of the membrane do not decrease with increasing contact time between the hexane-soaked composite support membrane and the aqueous 1,6-hexane diamine solution. This suggests that after the initial interfacial polymerized layer is formed, the reaction essentially ceases due to slow amine diffusion through the interfacial layer.

Interfacially polymerized composite membranes were prepared as in Example 17. The membranes were tested using the procedure described in Example 1 with the following pure gases: helium, hydrogen, oxygen, nitrogen, methane, carbon dioxide, ethane, propane, and butane. The results are summarized in Table 6.

The performance of the interfacial composite membrane is typical of the behavior of glassy polymers, in that small molecules permeate preferentially compared with larger molecules. The hydrogen/carbon dioxide selectivity of the interfacial composite membrane is exceptionally high.

A series of experiments was carried out using the same general preparation technique and immersion and contact times as in Example 17, but with various types of polymers for the composite support membrane sealing layer.

An asymmetric microporous support membrane was cast using the same recipe and procedure as in Example 1. The membrane was not coated with a sealing layer. An interfacially polymerized layer was formed on the uncoated support as follows. The support membrane was submerged in a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride.

After 30 minutes, the composite support membrane was removed from the hexane bath and held vertically for one minute to drain excess solution. The composite support membrane was then submerged in an aqueous 1,6-hexane diamine solution containing 1.0 wt % diamine and 1.0 wt % potassum hydroxide in 98 wt % distilled water. After five minutes, the membrane was removed from the aqueous bath and air-dried at room temperature.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 7.

An asymmetric microporous support membrane was cast using the same recipe and procedure as in Example 1, with a silicone rubber sealing layer. An interfacially polymerized layer was formed on the composite support membrane using the same procedure as in Example 28.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 7.

An asymmetric microporous support membrane was cast using the same recipe and procedure as in Example 1, with a polymethylpentene sealing layer prepared from a 2.0 wt % solution of polymethylpentene (TPX.RTM. MX 002, Mitsui Co, New York, N.Y.) in 98 wt % cyclohexane. An interfacially polymerized layer was formed on the composite support membrane using the same procedure as in Example 28.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 7.

An asymmetric microporous support membrane was cast using the same recipe and procedure as in Example 1, with a sealing layer prepared from a 1.0 wt % polyamide copolymer (Pebax.RTM. 4011, Atochem, Inc., Glen Rock, N.J.) in 99.0 wt % butanol solution. An interfacially polymerized layer was formed on the composite support membrane using the same procedure as in Example 28.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 7.

Comparison of the results from examples 28-31 shows that an interfacially polymerized membrane prepared on a support without a sealing layer did not yield a gas separation membrane with good properties. The polyamide copolymer sealing layer also resulted in a membrane with poor hydrogen/carbon dioxide selectivity. However, in this case, the polyamide copolymer alone has a selectivity for carbon dioxide over hydrogen of about 15. That the finished interfacial composite is selective for hydrogen over carbon dioxide at all shows the powerful influence of the interfacially polymerized layer.

Two experiments was carried out following the same general preparation techniques and procedures as Examples 28-31. In this case, however, the support was immersed in the aqueous solution, then contacted with the organic solution.

An asymmetric microporous support membrane was cast using the same recipe and procedure as in Example 1. The membrane was not coated with a sealing layer.

An interfacially polymerized layer was formed on the support as follows. The support membrane was submerged in an aqueous solution of 1,6-hexane diamine containing 1.0 wt % diamine and 1.0 wt % potassium hydroxide in 98 wt % distilled water.

After 30 minutes, the composite support membrane was removed from the water bath and held vertically for one minute to drain excess solution. The composite support membrane was then submerged in a hexane solution containing 1.0 wt % 1,3,5-benzene tricarbonyl trichloride. After five minutes, the membrane was removed from the hexane bath and air-dried at room temperature.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 8.

An asymmetric microporous support membrane was cast using the same recipe and procedure as in Example 1, with a polyamide copolymer sealing layer as in Example 31. An interfacially polymerized layer was formed on the composite support membrane using the same procedure as in Example 32.

The resulting membrane was tested in a permeation cell using the same procedure as was used to evaluate the composite support membranes described in Example 1. The results are summarized in Table 8.

A test was performed to determine whether the interfacially polymerized layer extends into the matrix of the sealing layer. An interfacially polymerized membrane was prepared as in Example 17, with a silicone rubber sealing layer and an interfacial layer formed by the reaction of 1,3,5-benzenetricarbonyl trichloride with 1,6-hexane diamine. The resulting membrane was soaked in water for two hours, then dried and tested. The hydrogen permeation rate increased by a factor of three, whereas the other gas permeation rates increased by more than a factor of three. The membrane was then again soaked in water for two hours. This time, before the membrane was allowed to dry, the membrane surface was rubbed in an attempt to remove the interfacial polymerized layer. Upon retesting, the membrane properties were close to those of the silicone rubber/polysulfone composite support membrane in both permeation rates and selectivities. The interfacial layer had been removed, indicating that the polymerization reaction does not proceed substantially into the silicone rubber matrix.