WO2000020106A2

WO2000020106A2 - Modified porous metal surfaces

Info

Publication number: WO2000020106A2
Application number: PCT/NL1999/000623
Authority: WO
Inventors: John Wilhelm Geus; Adrianus Gijsbert Kock; Adrianus Jacobus Van Dillen; Maria Johanna Henrica Virginie Kerkhoffs
Original assignee: Universiteit Utrecht; U-Cat B.V.
Priority date: 1998-10-07
Filing date: 1999-10-07
Publication date: 2000-04-13
Also published as: NL1010267C2; WO2000020106A3; AU6233199A; JP2002526239A; EP1128900A2

Abstract

The invention relates to coarsely porous metal structures in which a thin surface layer which exclusively contains pores having sizes of from about 2 to 0.01 νm is appplied to a metal structure, and to a method for manufacturing coarsely porous metal workpieces in which particles having sizes of from 2 to 0.05 νm of metals or alloys, of compounds to be reduced to metals or alloys, or of ceramic materials are applied in the pore mouths of the coarsely porous metal sublayer.

Description

Title: Modified Porous Metal Surfaces

The invention relates to porous metal structures, such as tubes or plates, in which the diameter of the pores in the surface layer is to be adjusted within relatively narrow limits, and to porous metal sublayers, the surface of which is modified in such a manner as to be suitable for applying either ceramic membranes or porous or nonporous metal membranes.

According to the known state of the art , porous metal structures, plates or tubes, are used in many cases in methods in which finely divided solids are separated from liquids. For the separation of solid particles from gas streams, too, porous metal structures are often used. In general, particles having sizes up to about 3 μm can be properly separated from gases or liquids. Although the separation of smaller solid particles is much needed, no suitable mechanically strong and relatively inexpensive porous materials are available for this purpose. Furthermore, for the separation of particles from liquids, the availability of filter materials having fine pores resistant to alkaline liquids is attractive. In that case, membrane layers completely built up from metals or alloys are preferably used. However, filter elements made from oxidic materials, such as titanium dioxide or zirconium dioxide, may also be used. Roughly speaking, porous metal structures are technically made in two different manners. In the first procedure, the starting metals are those which have a relatively low melting point and therefore sinter at relatively low temperatures. A mold of the desired shape and dimensions is filled with preferably globules of the metal of which the porous structure has to consist. In general, good results are obtained with molds of graphite, although, in principle, other materials can also be used. By heating the filled mold at a temperature at which the metal particles sinter, the desired porous structure is obtained. The weight of the metal particles is sufficient to bring the elementary metal particles into contact with each other in such a manner that the particles sinter together. The atmosphere in which sintering has to take place, strongly depends on the metal to be sintered. Metals and alloys such as copper, silver, nickel and bronze, which can be completely reduced at low temperatures, for instance 600°C, will be heated in a reducing gas stream. The oxide layer at the surface of the metal or alloy particles will be removed by reduction, so that a strong metallic bond is formed between the particles. With a difficulty reducible, but lower melting metal such as aluminum, the oxide layer present at the surface cannot be reduced with hydrogen. For sintering aluminum to porous structures, therefore, heating has to be effected in a very good vacuum. In addition to an extremely low residual pressure of oxygen, a very low residual pressure of water vapor is also essential for sintering very base metals such as aluminum. The weight of the aluminum globules pushes away the thin oxide layer present at the surface of the aluminum at elevated temperature, thereby deforming the metallic aluminum, and the metallic aluminum of particles touching each other can come into contact. The metallic aluminum readily sinters. The work can also be done in an inert gas with extremely low contents of oxygen, carbon dioxide and water vapor.

With highly melting metals or alloys which are hard to reduce, the work has to be done in a different manner. In that case, a much higher pressure must be applied to push away the oxide layer from the metal or the alloy on the contact surfaces between particles touching each other. Small metal or alloy particles therefore have to be pressed into a mold of the desired shape and dimensions. With larger molds, a rubber mold and hydrostatic pressing, for instance, can be advantageously used. After pressing, the "green" molded piece is heated for some time in a reducing gas stream at a highly elevated temperature. The neck-shaped contact sites between the elementary metal particles thereby grow out.

According to the present state of the art, it is not possible very well to make porous metals which only have pores less than 2 μm. However, such porous metal filter elements are technically also less attractive, because of the great transport resistance. For the thickness of the porous element cannot be made very small without detracting from the mechanical strength of the part . With a finely porous material having a thickness of for instance 3 mm, the transport resistance is high. Technically speaking, it is therefore much more favorable if a thin layer having much narrower pores is applied onto a coarsely porous sublayer. The object of the present invention is therefore to provide a coarsely porous metal having a thin surface layer which contains pores with sizes of from about 2 to 0.01 μm and absolutely no wider pores. As referred to herein, thin is a layer thickness of from 1 to less than 30 μm. The object of the invention also comprises a coarsely porous metal sublayer, a thin porous layer having pores of from 2 to

0.01 μm being applied in or onto the pore mouths.

Depending on the use of the porous metals having a finely porous surface layer, completely metallic layers or a ceramic layer will be chosen on or in the pore mouths of the coarsely porous metal sublayer. A special use of porous layers is the increase in the selectivity of chemical reactions. In the art, it is desired that the chemical reactions can be used between two compounds which, in case of reaction with themselves, give rise to undesired products. An example is the aldol condensation of two different aldehydes. Because the aldehydes can also react with each other, three different compounds are obtained here, of which only one is desired. When, however, a catalyst for the desired reaction is applied in the surface of a membrane having narrow pores, one reactant is supplied through the membrane and the other reactant is passed along the side of the membrane where the catalyst is present, there is an increased possibility of reaction between the different aldehydes.

According to the invention, particles having sizes of from 2 to 0.01 μm of metals or alloys, of compounds to be reduced to metals or alloys or of ceramic materials are applied in the pore mouths of a coarsely porous metal sublayer. In general, the starting material for providing the small particles in the pore mouths of the coarsely porous metal will be a suspension of the small particles in a suitable liquid. The object is then to cause the suspension to flow over the sublayer in a fairly uniform thickness, thereby preventing the suspension from locally penetrating deeply into the coarse pores of the sublayer, while other parts of the surface of the coarsely porous sublayer do not come into contact with the suspension. According to the present invention, there are two procedures for avoiding this drawback.

According to the first of the above two methods according to the invention, a uniform distribution of the suspension over the coarsely porous sublayer can be obtained by starting from a suspension of the particles to be applied, the viscosity of which is adjusted by adding agents otherwise known according to the state of the art, such as hydroxy- methyl cellulose or polyvinyl alcohol . The viscosity of the suspension is adjusted in such a manner that the suspension does not rapidly penetrate into the coarse pores of the sublayer. According to the invention, such a "dip coating" of the coarsely porous metal sublayer with a suspension of the fine particles to be applied leads to the filling of the pore mouths with the small particles.

According to the second of the above methods according to the invention, the pores of the coarsely porous metal can be filed nearly completely with a liquid or with a waxy material having a low melting point . A suspension of the small particles to be applied can then be caused to flow over the filled metal sublayer. According to a preferred form of the method according to the invention, the pores are filled with a liquid which is immiscible with the liquid of the suspension of the particles to be applied, and which has a density higher than that of the suspension. After drying, the liquid or the waxy material is removed. With tubes, a waxy material will preferably be used, and the horizontally disposed tube will rotate about its axis. After the outflow of the suspension, the rotation is continued until the liquid of the layer applied has been evaporated. The invention also comprises "spin coating" of a porous flat metal body filled with a waxy material . This involves rotating the porous metal body to be treated and bringing the suspension of the particles to be applied in the middle on the surface of the porous body. The centrifugal force causes the suspension to spread over the surface of the porous body as a thin layer. In order to obtain pores of from 2 to 0.01 μm, particles having a diameter of about the same sizes are required. The manufacture of metal or alloy particles having the above sizes is technically known for a number of metals or alloys. Thus, by decomposition of metal carbonyls in a gas stream and collecting the particles in oil, small iron, nickel or cobalt particles can be prepared. These particles can be magnetically separated from the oil. A large series of metal and alloy particles can also be prepared by finely divided application of a suitable precursor of the metal or the alloy to a support, such as finely divided silicon dioxide or aluminumoxide , and reduction of this precursor at elevated temperature in a gas stream suitable for the purpose. The sizes of the metal or alloy particles can be controlled by adjusting the load of the support. Sintering of the metal or the alloy is prevented by the support. After reduction, the support can be dissolved in lye, after which a suspension of extremely small metal or alloy particles is left, which in ferromagnetic metals or alloys can be readily separated from the liquid magnetically. For non- ferromagnetic metals, the same procedure can be used with an inert thermostable support . Because now the separation of metal particles from the liquid after dissolving the support cannot be effected magnetically, the small particles will now preferably be flocculated. Subsequently, the particles can be separated, and the impurities can be removed by washing. After resuspending the washed material, clusters of small particles are obtained, the sizes of which in the suspension according to the known state of the art to be applied to the coarsely porous metal, can be adjusted to the sizes of the pores of the coarsely porous metal. According to the state of the art, ultrasonic treatment can be used, a liquid flow with a high velocity gradient or a liquid flow in which cavitation is generated. The work can also be done ^"with a colloid mill or an ultraturrax. Finally, according to the present invention, small ferromagnetic particles can be covered with a layer of another metal, such as copper or silver.

According to a preferred form of the method according to the invention, complex cyanides with two alloy components are applied onto a support soluble in lye, to prepare small alloy particles. Thus, for instance, for iron-nickel alloys, Ni₂Fe(CN)₆ can be applied to, for instance, aluminum oxide. In this manner, atomic mixing of the alloy components in the precursor is effected.

Another possibility of preparing small metal particles is to start from a finely divided precursor without support and to react the precursor into the metal . Preferably, the starting precursor is one yielding the metal at a relatively low temperature, because at a highly elevated temperature or during a prolonged thermal treatment the small metal particles sinter together. According to the invention, the starting material is preferably oxalates of copper, silver, or nickel, which are decomposed in an inert or a reducing gas stream. In this manner, other salts of organic acids can also yield small metal particles. In this manner, small metal particles are obtained which do not cohere strongly and, for instance, can be excellently dispersed by ultrasonic vibration.

It is highly important that the presence of larger clusters of small metal particles in the suspension is prevented. To this end, according to the present invention, the suspension, before being applied to the surface of the coarsely porous metal, is filtered over a fine filter.

An advantage of starting from metal particles is that no strong shrinkage occurs, which is the case when the starting particles are particles to be reduced to metals or alloys. Nevertheless, it can be advantageous to apply small reducible particles in the mouths of the pores of a coarsely porous metals structure. After heating in an inert or oxidizing gas, the particles applied and fixed in the pore mouths can then be reduced to the corresponding metals. This is excellently done with the oxides of silver, copper and nickel, which can be relatively readily reduced. For silver oxide, heating in the air is even sufficient to effect a reaction into the metal. According to the invention, therefore, particles are applied which have sizes of from 2 to 0.01 μm, and which are to be reduced to the desired metal in the pore mouths, after which, if necessary, the material is reduced after a thermal pretreatment in a non-reducing gas stream.

According to the invention, particles not to be reduced to metals can also be applied in the pore mouths of the coarsely porous metal. In general, the sizes of the pores obtained with particles not to be reduced to metals can be more properly controlled. If the sizes of the pores are to be closely adjusted, non-reducible particles will therefore be preferably used, or the particles will not be reduced to the corresponding metals. For working in alkaline liquids, oxides such as titanium dioxide or zirconium dioxide will be used, which are resistant to alkaline liquids. Non-reducible oxides, such as for instance silicon dioxide, aluminum oxide, titanium dioxide or zirconium dioxide, are very attractive if separations are to be carried out in reducing gas streams at higher temperatures. Small metal particles sinter under such conditions, so that either large pores are formed or the pore mouths are completely clogged.

In most cases, small metal particles can be properly adhered to the walls of the pores of the coarsely porous metal. Under reducing conditions, an intermetallic bond can be formed. When preheating at high temperatures under non- reducing conditions, oxidic materials can also form firm bonds with the oxidized surface of metals. However, there are a number technically important cases in which no proper bond can be effected.

It is therefore often difficult to strongly bind solid particles to the surface of metals. Of course, metal particles or particles of oxides reducible to metals can be applied in the pore mouths of the porous metal, and then a thermal treatment can be carried out in a reducing gas stream. If the temperature and the gas atmosphere can be adjusted in such a manner as to reduce the surface of the porous metal and of the small metal particles applied, a metallic bond can be realized between the surface of the porous metal and the metal particles. This also applies to the case that reducible oxides are applied in the pore mouths. It has been found, however, that the bond between the metal particles and the walls of the pores of the metal is so strong is that large cracks are often formed in the metal layer. Moreover, it is very difficult during reduction to prevent strong sintering of the metal particles. Strong sintering leads to a rough surface with relatively wide pores. For the reduction of the surface of many metals, high temperatures are required, in which connection strong sintering of the metal particles is unavoidable. The surface of stainless steel, a material very attractive for many uses, is very difficult to reduce. This requires very high temperatures. On the other hand, strong sintering cannot be properly avoided with metals which are readily reducible, such as copper.

Surprisingly, it has now been found that by covering the surface of the walls of the pore mouths of the porous metal structure with a porous or nonporous layer of an oxide which firmly adheres to the pore wall and to the metal or oxide particles to be applied, it is possible to anchor the small particles in the pore mouths. Figure 1 is a schematic view of the metal structure. According to the invention, therefore, a coarsely porous metals structure is used, in which the walls of the pore mouths are covered with a thin porous or nonporous layer of an oxide which firmly adheres both to the pore wall and to the metal to be applied, the alloy or the oxide. As referred to herein, thin is a thickness of less than 3 μm. In general, enamel -forming oxidic compounds are very suitable as bonding layer. It has been found that oxides such as silicon dioxide or titanium - dioxide very firmly anchor both metal particles and oxidic particles to the surface of the pore walls. According to a preferred form of the invention, therefore, porous metal structures are used in which a thin layer of silicon dioxide or titanium dioxide or another enamel -forming compound is applied to the walls of the pore mouths.

According to a preferred form of the method according to the invention, a thin layer of silicon dioxide is applied to the walls of the pore mouths by applying a layer of silicon rubber to the walls of the pore mouths. Pyrolysis of the dried layer of silicon rubber leads to a silicon dioxide layer, which is nonporous after heating at a high temperature and is porous at lower temperature (for instance 600°C) . In an analogous manner, according to a preferred form of the method according to the invention, titanium dioxide can be firmly adhered to the walls of the pore mouths by pyrolyzing a layer of the titanium equivalent of silicon rubber. Also a layer of water glass applied to the walls of the pore mouths leads to a strong interaction with metal, alloy and oxide particles .

According to the invention, therefore, the surface of porous metal structures is modified by applying a thin bonding oxide layer to the walls of the pores, after which via "dip coating" the pores in the surface are filled with a thin porous layer of coherent solid particles, the diameter of which is smaller than the diameter of the pores in the surface of the metal . The solid oxide particles having sizes of from about

2 to 0.01 μm, which are thus applied in the mouths of the pores of the metal by "dip coating", can be obtained according to the known state of the art. Very suitable is a preparation of small particles by the decomposition in flowing air or in an oxidizing gas stream of metal oxalates, such as nickel oxalate or zinc oxalate, or mixing oxalates of two or more metal ions, such as nickel -magnesium oxalate. By decomposition of oxalates in the air, the corresponding (mixing) oxides are obtained. Also by flame hydrolysis of for instance chlorides, oxide particles can be prepared which are eminently suitable for the manufacture of metal surfaces modified according to the present method. For the preparation of silicon dioxide by flame hydrolysis, the reaction is SiCl₄ + 2H₂0 = Si0₂ + 4HC1 In an analogous manner, many other metal oxides, such as aluminum oxide and titanium dioxide, are produced on an industrial scale.

According to the invention, "dip coating" is carried out by using conventional techniques. A suspension of the oxide particles to be applied is prepared in for instance a water or a water/ethanol mixture to which for instance methyl cellulose is added to adjust the viscosity. The presence of larger conglomerates in the suspension with which "dip coating" is carried out must be avoided. Larger conglomerates lead to a rougher surface. According to the invention, the suspension is pre-filtered over a filter having pores of 25 μm or smaller pores.

When "dip coating" is carried out with particles of titanium dioxide or zirconium dioxide having an average size of 0.05 μm, a membrane is obtained which is eminently suitable for microfiltration, using pores of from 0.05 to 10 μm. Both the metal and the oxide particles are resistant to an alkaline medium. According to a preferred form of the method according to the invention, therefore, "dip coating" is carried out with a suspension of titanium dioxide or zirconium dioxide .

In the art, membranes can be used very advantageously which have much narrower pores than those occurring in porous metal or alloys and much narrower than the above pores of from 2 to 0.05 μm. In this case, pores of from 10 to 0.5 nm are involved. By means of such membranes, dissolved compounds can be separated from liquids, and gas molecules having a different molecular weight or having different sizes can be separated. The advantage of such porous materials has long since been recognized, and different materials capable of carrying out such separations have been technically proposed.

According to the state of the art, membranes consisting of polymers are generally used for the separation of different gas molecules. A drawback of polymeric membranes is the fact that such membranes cannot readily be cleaned by heating the membranes at high temperatures. Nor can, in general, a great pressure difference readily be applied over polymeric membranes. Finally, polymeric membranes cannot properly be used at elevated temperatures. In particular when the thermodynamic equilibrium is to be moved by selectively removing a specific component, the work must in general be done at elevated temperatures. If membranes are to used at higher temperatures and larger pressure differences, the membranes must be made from metals or ceramic materials. For technical uses, the mechanical properties of metals are, in general, much more attractive than the properties of ceramic materials. It is extremely difficult, however, to produce coarsely porous metal plates or tubes having on the surface a thin layer with pores having sizes of from 10 nm to less than 1 nm. In the art, therefore, membrane modules are mostly used which are completely made from ceramic material. According to the state of the art, a thin ceramic layer having very fine pores is applied to a ceramic layer having coarser pores. One or more ceramic intermediate layers are further often applied, the size of the pores being decreased stepwise, starting from the top layer with the narrowest pores. A drawback of such completely ceramic membranes is that clamping in technical apparatus is difficult, certainly when clamping takes place at low temperature, and the membrane must be used later at elevated temperature.

In the art, as observed above, coarsely porous metal tubes or plates are often used. The size of the pores in such materials can be varied to a minimum size of about 5 μm. For this reason, it had been proposed before, in International Patent Applications WO 91/12879 and WO 92/13637, that a thin finely porous ceramic layer is applied to the surface of coarsely porous metals. The thin porous ceramic layer closes the coarse pores of the metal. Important is the difference in coefficient of expansion of the metal and the ceramic layer. Of ceramic materials, such as aluminum oxide, the coefficient of expansion is about 7xl0^"6 per °C , while that of metals, such as iron and copper, is about 15xl0^"6 per °C. Because of this difference in thermal coefficient of expansion, only thin ceramic layers can be used. In connection with the permeability of the membrane layer, attempts are made to realize a leak-tight membrane with a thinnest possible layer ceramic material . Accordingly, the use of a thin ceramic layer is not objectionable.

The above patent application WO 91/12879 proposes to fill the pores of the porous metal with a liquid having a higher density than water. It is also possible to fill these pores with a material having a relatively low melting point, such as a wax or tallow. Subsequently, a sol of a suitable material, such as silicon dioxide, aluminum oxide or titanium dioxide, is caused to flow over the surface of the porous metal, after which the sol is dried after it has optionally been caused to gel . In order to prevent cracks from being formed in the dried layer, drying must be effected very carefully. After the sol or the gel has been dried, the liquid or the waxy material is removed, after which the finely porous ceramic layer is stabilized by heating at a temperature of, for instance, from 450 to 650°C .

Although this procedure yields interesting membrane composites, drying of the gelled or non-gelled sol is a very critical operation. For this reason, WO 92/13637 proposes to flow a solution of a material like silicon rubber or the titanium equivalent of silicon rubber over a porous metal, the pores of which are filled with a heavy liquid or a waxy material, and to remove the solvent by evaporation. This results in a very thin layer of silicon rubber, which very strongly adheres to the metal surface. Subsequently, the heavy liquid or the waxy material is removed, after which the organic constituents of the silicon rubber are removed by heating in the air or in an oxygen-containing gas stream. The result is a very thin layer, highly porous silicon dioxide, which very strongly adheres to the metal. The silicon dioxide contains pores of from 0.8 and 1.2 nm. It is important that the finely porous ceramic membrane layer is formed at high temperatures. When cooling, the layer comes under compressive stress; ceramic materials can take up a compressive stress much better than a tensile stress.

Although the last discussed method yields very good membranes, the quality of the porous metal sublayers is an immense problem. In general, the surface of porous metal tubes and plates commercially produced in the above discussed manners is relatively rough, while pores of from 15 to 20 μm also occur in the surface (to a small extent) . A study with a scanning electron microscope has shown that the surface of commercial porous stainless steel structures contains a number of large pores having diameters up to 20 μm. In qualitatively good structures, few relatively large pores occur in the surface. In order that pores having such sizes can be bridged, relatively thick ceramic membrane layers are required. As a result, the presence of a relatively small fraction of large pores leads to a substantially lower permeability of the membrane layer. It is therefore technically highly important to provide porous metal sublayers containing no pores with sizes greater than about 10 μm.

According to the known state of the art, attempts have been made to make the surface of commercial porous metals smoother by applying a network of very thin wires (for instance with a diameter of from 5 to 10 μm) and to eliminate pores with greater sizes. In practice, however, it has been found that this is an expensive procedure, while the result often leaves much to be desired. A study of a large number of commercial structures in the scanning electron microscope has shown that in this case, too, the large pores cannot be removed sufficiently. Attempts have also been made to improve the quality of the surface by spraying with small metal particles. Although it has been found that the coarse pores can be eliminated by this method, this is at the expense of a greater roughness of the surface. This renders it difficult to cover the surface with a closing thin layer of a finely porous ceramic material . Nor do cloths of thin metal wires yield a surface suitable for applying thin porous ceramic layers.

A second main object of the present invention is therefore to provide a coarsely porous metal having such a surface that a thin porous layer of a ceramic material is sufficient to close all the coarse pores. Such a modified surface of a porous metal preferably has pores of from 2 to 0.1 μm. A further object of the present invention is to provide porous metals having a surface with pores which are completely or nearly completely closed with thin layers of metals or alloys which selectively pass specific gas molecules .

Starting from silicon rubber, a porous layer of silicon dioxide having pores of from 0.8 and 1.2 nm can be relatively readily applied to a thus modified porous metal surface. Figure 2 is a schematic view of the metal structure. In this manner, a membrane is obtained the pores of which are so small that a number of gas molecules can be separated by using a difference in transport with Knudsen diffusion. In order to obtain a completely dense layer, more than one layer of silicon dioxide will generally be applied via silicon rubber. It is also very attractive to apply one or more layers of silicon dioxide via sol -gel methods according to the known state of the art. The invention also comprises a first layer applied by pyrolysis of a layer of silicon rubber, followed by one or more layers according to a sol -gel method. It is known that by using sol-gel techniques, layers can be obtained the pores of which only pass hydrogen.

In a number of cases, the fact that specific metals selectively dissolve specific gas molecules can be advantageously used to separate gases. An example is silver in which oxygen selectively dissolves, while hydrogen selectively dissolves in palladium or palladium alloys. In order to avoid recrystallization, a palladium-silver alloy is preferably used. According to the present invention, a dense layer of a metal (the membrane metal) is applied in the pores of a porous metal modified with solid particles by carrying out "dip coating" with a colloidal suspension of the desired metal. It is also possible to fill the pores of the modified porous metal with a heavy liquid and to allow a sol of the desired metal to spread over the modified porous metal. After a thermal treatment in a gas stream, if necessary a reducing gas stream, the small metal particles sinter together into a dense little ball which completely closes the pore mouth. Because of the strong interaction with the thin bonding layer applied to the wall of the pore mouth, the membrane metal does not tend to detach from the surface of the pore wall. According to a preferred form of the method according to the invention, a silver sol is used prepared according to a method published by Carey Lea (Carey Lea, Am.J.Sci. (3) 37 (1889) 476) . This preparation method leads to a very concentrated silver sol . Surprisingly, it has been found that the silver particles preferably penetrate into the pores of the modified coarsely porous metal.

The modified porous metal surfaces according to the present invention are also eminently suitable for applying oxygen ion conductors. If only transport of oxygen is intended, an electronic conductor is eminently suitable. An opposed electron current then neutralizes the flow of oxygen ions through the layer. If, however, the use in a "solid oxide fuel cell" is intended, then an electronic insulator must be used. Now the oxidic layer will be preferably applied to the pore mouths, while a metal conductor which does not completely cover the surface is applied to the oxidic layer.

Example

Manufacture of a silicon dioxide membrane having very narrow pores

In this example, stainless steel filters made by the firm of Krebsόge, Radevormwald, Germany, are used. The filters consisted of sintered AISI 316L steel (Cr 16.5%, Ni 13.5%, Mo 2.5%, Si 0.7%, C <0.02%, Mn <0.2% and iron). The membrane layers were applied to disks having a diameter of exactly 24.0 mm and a thickness of 2 mm. The porosity of the filter was about 40%. In order that the disks can be clamped leak-tight for measurement of the transport, the disks were arranged in a nonporous stainless steel disk having a diameter of 48 mm and a thickness of 5 mm, in which a cylindrical hole of exactly 24.0 mm was provided. The porous disks were arranged in the cylindrical hole of the heated nonporous disk. After cooling, the porous disk appeared to be shrunk leak-tight in the nonporous ring.

The thus obtained stainless steel structure was thoroughly cleaned by ultrasonic vibration in acetone and subsequently by immersion in a boiling solution of 1.5 molar ammonia with 0.1 wt . % Triton-X in water for at least one hour. The disks were rinsed twice with boiling demineralized water and stored under ethanol . In order to obtain a good adhesion between the nickel oxide particles to be applied afterwards and the wall of the pores of the stainless steel, a silicon dioxide layer was applied to the wall of the pores. This was done by applying from 30 to 60 μl of a solution of RTV silicon rubber (Bison, Perfecta Int. Nederland) , 4.8 wt . % in ethyl acetate (Merck) or diethyl ether (Merck) to the porous metal disks. The silicon rubber layer resulting after drying was pyrolyzed to silicon dioxide in air at 723 K.

Nickel oxalate dihydrate was prepared by mixing solutions of oxalic acid and nickel nitrate. A solution of nickel nitrate hexahydrate was rapidly added to a solution of oxalic acid dihydrate. The volumina of both solutions were equal; the final concentration of the reactants was 1.0 molar. The precipitation vessel was maintained at a fixed temperature of 293 K. After 30 min, the precipitate was filtered off and washed with demineralized water and ethanol. The filter cake was dried for 24 hours at 353 K in stationary air. The resulting material was milled in a ball mill with ethanol for 4 hours, after which the alcohol was evaporated at 393 K for at least 16 hours.

Nickel oxide particles were obtained by heating the thus prepared nickel oxalate in a rotating oven at a rate of 1 K/min to 723 K in a 100 ml/min 0₂/A (20/80) stream. After the sample was maintained for 5 hours at 723 K, it was cooled to room temperature. The particle size of the nickel oxide particles was determined by scanning electron microscopy and with an optical particle counter, Accusizer 770 Optical Particle Sizer with a detection limit of 1 μm. The last instrument indicated an average cluster size of 1.5 μm with a small fraction larger than 2 μm. The BET surface varied from 10 to 17 m² per gram. This corresponds with elementary nickel oxide particles of from 0.08 to 0.03 μm. Scanning electron microscopy indicated that the size of the individual nickel oxide particles was substantially smaller than 1 μm.

The nickel oxide powder was milled in a ball mill, suspended in alcohol for at least 4 hours. The resulting paste was diluted with ethanol and filtered over a 17 μm filter (Haver & Boecker) . The content of solid matter of the filtered suspension was brought to from 10 to 12 wt . % . Subsequently, the suspension was placed in a stirred vessel with deflectors and methyl cellulose (Genfarma) was added until an amount of 2 wt . % . The suspension was heated to 328 K with stirring, precisely above the gelling point. Subsequently, an equal volume of water was added with vigorous stirring, which led to a dramatic increase in the viscosity. The suspension was stirred for 10 min and then filtered through a 25 μm gauze under 2 to 4 bar pressure. The suspension was finally cooled to room temperature.

The carefully cleaned stainless steel disks were vertically mounted in a dip coating apparatus. A rate of 1 and of 3 mm/min was used. After dip coating, the disk was heated for 5 hours in an air stream of 100 ml/min at 723 K, after heating at a rate of 1 K/min in a special quartz oven. After cooling, the nickel oxide appeared to be very strongly bound in the pore mouths of the stainless steel . The average diameter of the pores was 2 μm. In order to obtain a total covering of all the coarse pores of the stainless steel, a layer of nickel oxide had to be applied, in general three times, by dip coating. The final thickness of the nickel oxide layer was then from 15 to 20 μm. The pores of the stainless steel were temporarily filled with 1 , 1 , 1-trichloroethane (Aldrich) , after which a solution of silicon rubber (Bison, Perfecta Int. Nederland) of from 8 to 10 wt . % in diethyl ether (Jansen) was flowed over the surface of the stainless steel disk. After the solvent had been evaporated, the silicon rubber film was air dried for at least one hour. After removal of the 1,1,1- trichloroethane, the covered stainless steel disk was maintained in an air stream for 5 hours at 723 K and then cooled. The heating and cooling was carried out at 1 K/min. After applying 1 to 5 silicon dioxide layers, the permeance of the resulting membrane systems was from 7 to 12 x 10 ^"5 mol/ (msecPa) . From the variation of the permeance with the pressure difference over the membrane system, some contribution of laminar flow appeared to be present. When the number of silicon dioxide layers was larger, a contribution of laminar flow was practically absent. In order to examine the reproducibility of the manufacture of membrane systems according to the invention, four different membrane systems were prepared in the same manner. When there was no substantial contribution of laminar flow, the permeance varied from 1.2 to 5.4 x 10^"6 mol/ (m²secPa) . This shows that membrane systems with such a high permeance can be reproducibly manufactured.

Claims

1. Coarsely porous metal structures, characterized by a thin surface layer which only contains pores having sizes of from about 2 to 0.01 μm, applied to the metal structure.

2. A method for manufacturing coarsely porous metal structures according to claim 1, characterized in that particles having sizes of from 2 to 0.05 μm of metals or alloys, of compounds to be reduced to metals or alloys, or of ceramic materials are applied in the pore mouths of a coarsely porous metal sublayer.

3. A method according to claim 2, characterized in that the particles are applied by "dip coating" with a suspension of particles of from 2 to 0.05 μm.

4. A method according to claim 2, characterized in that the particles are applied by flowing a suspension of particles of from 2 to 0.05 μm over the coarsely porous metal sublayer, the pores of the metal being largely filled with a liquid.

5. A method according to claims 3 or 4 , characterized in that the suspension for applying to the coarsely porous metal sublayer is filtered over a filter having openings of maximally 17 μm.

6. A method according to claim 4, characterized in that the pores of the coarsely porous metal sublayer are largely filled with a liquid which is immiscible with the liquid of the suspension of small particles and has a greater density than the liquid of the suspension.

7. A method according to claim 4, characterized in that the pores of the coarsely porous metal sublayer are largely filled with a waxy material having a melting point below about 120°C .

8. A method according to claim 6, characterized in that "spin coating" is used to apply the suspension of small particles .

9. A method for preparing small metal or alloy particles by finely divided application of a suitable precursor of the metal or the alloy to a thermostable support soluble in lye, then reducing this precursor at elevated temperature in a gas stream suitable therefor and subsequently, if desired after cooling to room temperature, dissolving the support in lye, after which the small metal particles, if necessary after flocculation, are separated from the liquid and the contaminants are removed by washing.

10. A method according to claim 9, in which insoluble complex cyanides are applied to a suitable support material as precursor for alloy particles.

11. A coarsely porous metal structure, characterized in that the walls of the pore mouths are covered with a porous or nonporous layer of an oxide, which properly adheres both to the metal of the coarsely porous metal structure and to that of the small metal particles to be applied, after which, according to one or more of claims 2 through 8, metallic or non-metallic particles having sizes of from 2 to 0.05 μm are applied in the pore mouths.

12. A coarsely porous metal structure according to claim 11, characterized in that a thin layer of an enamel-forming oxide is applied to the wall of the pore mouths.

13. A coarsely porous metal structure according to claim 11, characterized in that a thin layer of silicon dioxide or titanium dioxide is applied to the walls of the pore mouths.

14. A method for applying a thin oxide layer according to claim 13, characterized in that a thin layer of silicon rubber or the titanium equivalent thereof is applied to the walls of the pore mouths of the metal structure, and that after drying of the layer, the organic constituents are removed by oxidation at a temperature of from about 400 to 700°C .

15. A coarsely porous metal structure according to claim 11, characterized in that improperly reducible oxides, such as silicon dioxide, titanium dioxide, or zirconium dioxide, are applied in the pore mouths.

16. A coarsely porous metal structure according to claims 1, 9 or 10, characterized in that a thin layer with pores having diameters of from 5 to 0.5 nm is applied in or on the modified pore mouths.

17. A coarsely porous metal structure according to claims 11, characterized in that the thin porous layer having pores of from 5 to 0.5 nm consists of silicon dioxide or titanium dioxide .

18. A method for applying a finely porous layer having pores of from 5 to 0.5 nm to a porous metal surface modified according to claims 1, 9 and 10, by applying a thin layer of silicon rubber to the surface and then pyrolyzing this layer.

19. A method for applying a finely porous layer having pores of from 5 to 0.5 nm to a porous metal surface modified according to claims 1, 9 and 10, by applying a thin layer of the titanium equivalent of silicon rubber to the surface and then pyrolyzing this layer.

20. A coarsely porous metal structure according to claims 1, 9 or 10, characterized in that a thin nonporous layer of a metal which selectively passes specific gas molecules is applied in or on the modified pore mouths.

21. A coarsely porous metal structure according to claim 20, characterized in that a thin nonporous layer of silver or an oxygen-permeable silver alloy is applied in or on the modified pore mouths.

22. A coarsely porous metal structure according to claim

20, characterized in that a thin nonporous layer of palladium or a hydrogen-permeable palladium alloy is applied in or on the modified pore mouths.

23. A coarsely porous metal structure according to claim 20, characterized in that a thin nonporous layer of an electroconductive oxide or mixing oxide which selectively passes oxygen is applied in or on the modified pore mouths.

24. A coarsely porous metal structure according to claim 20, characterized in that a thin nonporous layer of a non- electroconductive oxide or mixing oxide which selectively passes oxygen is applied on the modified pore mouths.