WO2007028375A1

WO2007028375A1 - Inherent and reliable selective method for directly synthesising hydrogen peroxide from oxygen and hydrogen with the aid of a catalytically coated wettable porous membrane

Info

Publication number: WO2007028375A1
Application number: PCT/DE2006/001584
Authority: WO
Inventors: Roland Dittmeyer; Karel Svajda
Original assignee: Dechema Gesellschaft Für Chemische Technik Und Biotechnologie E.V.
Priority date: 2005-09-08
Filing date: 2006-09-07
Publication date: 2007-03-15
Also published as: DE102005042920A1; EP1926683A1

Abstract

The invention relates to a method for directly synthesising a hydrogen peroxide from hydrogen and oxygen with the aid of a catalytically coated wettable porous membrane. According to a prior method, the explosion hazard was very high during a hydrogen peroxide production due to a hydrogen and oxygen mixture in a gaseous phase. The aim of said invention is to use a known per se gas/liquid membrane contactor for synthesis, thereby avoiding a direct contact of reaction partners in a gaseous phase. Certainly, the reversed procedure, opposite to the prior usual contacting method, makes it possible to supply oxygen through a membrane and also reduce a hydrogen concentration in the reaction mixture by rendering inert at least a half of the dissolved oxygen concentration expected inside the membrane at a gas/liquid boundary phase by means of nitrogen.

Description

description

Inherently safe selective process for the direct synthesis of hydrogen peroxide from hydrogen and oxygen with a catalytic coated wettable porous membrane and apparatus for carrying out the process

The invention relates to a process for the direct synthesis of hydrogen peroxide from hydrogen and oxygen with a catalytically coated wettable porous membrane, and to an apparatus for carrying out the method.

Hydrogen peroxide is an important industrial chemical; the currently installed production capacity amounts to 1.2 million tonnes in Europe and 860,000 tonnes in the US (as of 2004), - consumption in 2003 was just over 1 million tonnes in Europe. Capacity utilization is in some cases over 90% and the trend is rising, as the growth rate per year for Europe is estimated at 4-5% for the next few years and at 2-3% for the USA. Hydrogen peroxide is widely used in the bleaching of wood, pulp, paper, textiles, fats and oils, bleaching agents for the detergent industry, and hair bleaching cosmetics, as well as cleaners and disinfectants. In the chemical and pharmaceutical industries hydrogen peroxide serves as a selective and environmentally friendly oxidant, especially for epoxidations and hydroxylations. It is used in a highly purified form in the semiconductor industry for the production of printed circuits and for the cleaning of semiconductor surfaces. Other areas of application include metalworking and, with further increasing importance in the future, the detoxification and deodorizing of waste water and exhaust air streams. Hydrogen peroxide is produced industrially today predominantly by an indirect reaction of hydrogen and oxygen, which uses alkylanthraquinone as a medium. In a first catalytic hydrogenation step, the alkylanthraquinone is reacted in an organic solvent mixture to the Alkylanthrahydrochinon. Separately, this reduced compound is then oxidized with oxygen to regenerate the alkylanthraquinone and obtain hydrogen peroxide as a product. This is followed by extraction of the hydrogen peroxide with water to give a 15-35% solution, which is purified and concentrated by distillation to up to 70% as needed.

In sum, the indirect manufacturing route, in which an organic transfer medium is first reduced, then oxidized, makes the process relatively complex and causes high capital and operating costs. A considerable disadvantage results from the non-negligible solubility of the alkylanthraquinone in the aqueous extraction phase, which firstly leads to a loss of this expensive working medium and secondly to organic impurities of the hydrogen peroxide, which can lead to undesired reactions during the work-up by distillation. Another problem is the solubility of the aqueous extraction phase in the organic alkylanthraquinone solution. If the latter is recycled to the oxidation state after separation of the aqueous phase, areas of aqueous phases (so-called "pockets") can form in the organic phase in which dangerously high hydrogen peroxide concentrations are present.

Considerably simpler and more economical than the alkylanthraquinone route is the direct oxidation of hydrogen with oxygen to hydrogen peroxide (direct synthesis). A simplified reaction scheme of direct oxidation of Hydrogen with oxygen to hydrogen peroxide is shown in FIG.

This method, which has been known since the beginning of the 20th century, is described inter alia in US Pat. No. 4,832,938 B1. However, the previous attempts at technical implementation have led to industrial accidents, which are ultimately attributable to the inherent risk of explosion of this method. This results from the very wide explosive range of hydrogen / oxygen gas mixtures, which extends between 4.7 and 93.9 vol .-% hydrogen. Furthermore, it is known that the dilution with inert gases, such as nitrogen, the lower explosion limit of the hydrogen / oxygen mixture - on an inert gas-free basis - only slightly shifts. The explosion limits are also within the usual pressure and temperature ranges (1 - 200 bar, 0 - 100 ⁰ C) only slightly variable. And even if the reactants are brought together in a ratio that is outside the explosive range in the state of homogeneous mixing, the blending process itself would inevitably have to undergo the explosion zone at least temporarily. For these reasons, the risk of explosion associated with a direct contact of hydrogen and oxygen in the gas phase is technically not easily controlled.

To avoid the risk of explosion by mixing hydrogen and oxygen in the gas phase, various strategies have been proposed (see below (A) to (D)).

(A) Choudhary et al. in US Pat. No. 6,448,199 Bl describe the use of a hydrophobic, hydrogen-only permeable composite membrane for separating gaseous hydrogen from an oxygen-saturated reaction solution. As starting material, porous ceramic tube membranes are used, as they are used for ultrafiltration. The membranes have on the inner side a cover layer of γ-alumina, α-alumina or zirconia. On this electroless plating is applied by electroless plating a gas-tight layer of a palladium alloy, preferably with silver, copper, gold or ruthenium as a second metal, which is permeable only to hydrogen. Subsequently, a thin palladium film of 0.1 to 1 .mu.m thickness is deposited in addition, which is oxidized by the action of an oxidizing agent (O ₂ , N ₂ O, H ₂ O ₂ , HClO ₄ , HOCl, KMnO ₄ , K ₂ Cr ₂ O ₇ ) on the surface becomes. Finally, a hydrophobic layer of polyfluorocarbons, silicone rubber or polysulfones is applied to the membrane with the aid of solvents and crosslinking agents. Thus produced membranes avoid the need for mixing of hydrogen and oxygen in the gas phase and thus allow a safe execution of the reaction. With complete conversion of the hydrogen transported through the membrane, a hydrogen peroxide selectivity of 83.2% could be achieved, but with low productivity, so that only very low concentrations of hydrogen peroxide by 0.02 wt .-% were achieved. The oxidation of the palladium film is required to ensure a high selectivity to hydrogen peroxide. Despite the safety advantages, this system has two serious disadvantages for practical application:

1) a low productivity, which is explained by the limitation of the reaction rate by the slow diffusion of hydrogen through the dense metal film at the comparatively low reaction temperature, and

2) a relatively large required amount of palladium for the film, which is associated with high costs. (B) Pujado, in US Pat. No. 6,768,013, describes another concept for the inherently safe performance of direct hydrogen peroxide synthesis: it relies on the use of specific, e.g. B. highly-fluorinated organic solvent, such as Tetrafluorooctan (z. B. about 0.7 gO ₂ / L at 25 ⁰ C and 1 atm) with very good solubility for the reactants complete oxygen and hydrogen, a low solubility for the product hydrogen peroxide and almost Immiscibility with a second aqueous acceptor phase in the reactor for the hydrogen peroxide formed. The reactants can be introduced into the organic phase in different ways without direct gas-phase contact with each other, - in order to avoid their accumulation in the reactor gas phase beyond, this is purged with inert gas. The extraction of the hydrogen peroxide formed in the organic phase of a suspended catalyst in the aqueous phase is preferably carried out in the direct synthesis reactor itself (reactive extraction), but can also be done separately in a downstream apparatus. The organic phase has a much higher density (by 1.5 g / ml) than the aqueous phase to facilitate phase separation. The method has opposite to the indirect oxidation, z. B. on the Antrachinonverfahren advantages based essentially on the better solvent properties. However, the required special organic solvents (eg 3M ™ - performance fluids) can be expected to incur significant costs, and the equipment required to carry out the reaction in a three-phase emulsion / slurry reactor is considerable.

(C) A third approach to the inherently safe performance of hydrogen peroxide direct oxidation was reported by UOP at the autumn 2004 annual meeting of the American Institute of Chemical Engineers (AIChE) in Austin, TX, USA. Here, a commercial, with hydrophobic porous polymer hollow fiber membranes Hydrogen is applied to the inside of the membrane contactor on the inside, while a conventional catalyst (for example palladium on activated carbon) is located in the jacket space through which an aqueous, oxygen-saturated reaction solution flows. Since the hollow fiber membranes are not wetted, the liquid can not penetrate into the pores and a high specific gas / liquid phase interface and a high volumetric gas / liquid mass transfer coefficient is achieved, which is a prerequisite for a technically interesting productivity. Of disadvantage, however, are:

1) The lack of resistance of the common polymer membranes to organic solvents which have proved advantageous for the reaction (eg methanol), and

2) a non-uniform, relatively long diffusion path for hydrogen from the gas / liquid phase boundary on the geometric surface of the hollow fibers to the (or in the pores of) in the shell space distributed catalyst (s).

3) In addition, the introduction of the catalyst into the mantle space of a polymer hollow fiber bundle requires special constructive solutions and may prove to be a weak spot, especially if regeneration of the catalyst is necessary, which will not be possible in situ due to the low temperature resistance of the polymer membranes.

A related approach based on the use of a porous hydrophobic catalyst layer in conjunction with a porous membrane for controlled gas metering is described in WO 99/02264. Here, the porous membrane serves for the controlled supply of water Substance in the reaction layer. A macroporous support may be disposed between both layers to impart the necessary rigidity to differential pressure operation throughout the structure. The oxygen needed for the reaction is supplied on the side of the hydrophobic catalyst layer in a suitable solvent, while the hydrogen is metered through the porous membrane. The carrier and the membrane for metering hydrogen are preferably hydrophobic; Furthermore, the membrane may be provided with a protective layer over the catalyst layer to prevent a gradual removal of catalyst.

The weak point of this method results from the use of a hydrophobic reaction layer. This requires that the contacting of the catalyst by the reactants takes place in the gas phase. For this purpose, the oxygen must pass into the gas phase at the transition to the reaction layer. The resulting oxygen partial pressure in the reaction layer and the oxygen to hydrogen molar ratio, which is important for the selectivity of the reaction, depend in complex ways on the transport properties of the gas metering membrane, the pressure difference across this membrane, and the mass transfer of oxygen to the gas phase at the point of contact with the reaction layer and can therefore only be adjusted poorly to the desired value range (sufficient oxygen excess).

(D) A fourth approach, which avoids these disadvantages, was developed by Centi et al. 2003 presented. Here, too, a gas / liquid Membrane contactor is proposed to avoid direct contact of the two reactants in the gas phase, but on the basis of hydrophilic porous, ceramic or carbon-coated ceramic tube membrane. In this case, the reaction solution saturated with oxygen and conducted in the mantle space wets the membrane and penetrates the membrane. Her in the pores, on the other hand, an overpressure of hydrogen acts in the tube interior. By suitably setting this overpressure in conjunction with a decreasing over several layers of the asymmetric membrane size of the gas side to the liquid side is achieved that only the thin, finely porous cover layer is filled with liquid and the gas / liquid phase boundary is within the porous membrane at the transition the fine-pored cover layer to coarse-pored carrier sets. The catalyst is in this cover layer in the form of nanoparticles deposited on the pore wall. FIG. 2 shows the functional principle of the catalytic gas / liquid membrane contactor. The contact of the reactants with the active catalyst takes place in contrast to the previously described concept in dissolved form. A porous membrane for the controlled metering of hydrogen is not needed.

This principle, also referred to as a catalytic diffuser, is described in the patent DE 100 64 622 A1 for the selective hydrogenation of nitrate in aqueous solution to nitrogen. Compared with the hydrophobic membrane contactor, the advantage of short diffusion paths for both reaction partners is the hydrogen peroxide direct synthesis, which leads to a higher achievable productivity compared to the concepts mentioned under (A) and (C).

However, the fact that the reactants pass from opposite sides into the reaction zone inevitably results in opposing concentration profiles of hydrogen and oxygen (compare FIG. 3). In a mode of operation according to the above-described 4th concept (D) with supply of hydrogen through the membrane, this results in a selectivity to hydrogen peroxide, which is too low for industrial application, which is essentially attributable to two effects. 1) Through the use of pure hydrogen and pure oxygen arise in the case of many solvents, such. As water, methanol or common organic solvents, similar solubility of the two gases in the reaction solution and because of the similar pressure on both sides of the membrane comparable concentrations at the opposite edges of the fine-pored outer layer. Because of this opposite nature of the concentration profiles, a rapid change in the ratio of dissolved oxygen to dissolved hydrogen leads to a change in the degree of oxidation of the surface of the catalyst particles within the reaction zone, which has an unfavorable effect on the selectivity. In this case, as indicated in FIG. 3, two zones can be distinguished: Near the outer surface of the membrane there is a large excess of oxygen, while near the gas / liquid phase boundary the conditions are reversed and hydrogen is present in excess. Here, the conditions are now too strong reducing, whereby the parallel formation of water, at higher concentrations of hydrogen peroxide and its hydrogenolysis is favored. The dividing line between these two regions was drawn in Fig. 3 at an oxygen / hydrogen molar ratio of about 2. According to Li et al. (L. Li, S. Gandhi, K. Van den Busche, Membrane Reactor for Direct Synthesis of Hydrogen Peroxide, AIChE National Meeting, Austin, TX, USA, Item No. 554a, November 10, 2004), selectivity decreases with decreasing ratio Strongly back oxygen / hydrogen; a value of 2 should therefore not be undercut if possible in order to achieve acceptable selectivities. 2) If, to a certain extent, catalytically active metal particles are also present in the coarse-pored carrier, which can often not be completely ruled out on the preparation side, any residual oxygen (cf., c * ₀₂ in Fig. 3) which is present at the gas / liquid phase boundary will be present the porous membrane from the liquid phase enters the gas phase, where it reacts with the excess of hydrogen present to water. As a result, the selectivity is also lowered.

In summary, it should be noted that all previous approaches to implementing an inherently safe hydrogen peroxide direct oxidation with reasonable effort and at the same time for the technical application of sufficient productivity and selectivity are not yet satisfactory. There is therefore still the task, the need for a simple, safe, efficient and therefore cost-effective method for the direct oxidation of hydrogen peroxide, not least for smaller applications in which an inexpensive on-site generation as needed economically and also for safety reasons could be interesting to satisfy.

This object has been achieved according to the invention so that while maintaining the principle of the catalytic diffuser explained under (D) and the associated contact of the catalyst from the liquid phase, the contact is reversed, d. H. Oxygen is now supplied through the membrane, and also the hydrogen concentration in the reaction solution is lowered by inerting with nitrogen to about half of the expected dissolved oxygen concentration at the gas / liquid phase boundary within the membrane.

In this way, an optimal material flow of the reactants in the direct oxidation of hydrogen achieved with oxygen to hydrogen peroxide in a catalytically coated, porous gas / liquid Membrane contactor, thereby achieving an inherent process safety and high productivity and selectivity to hydrogen peroxide required for industrial application.

From the process according to the invention, the characteristic of the concentration profiles in the catalytically active covering layer shown in FIG. 4 results qualitatively, which guarantees an excess of oxygen over hydrogen within the entire reaction zone. Since, according to the stoichiometry of the reaction shown in FIG. 1, the amount of hydrogen consumed per unit time by reaction must always be greater than or equal to the amount of oxygen consumed locally per unit time, the reaction can be carried out by suitably adjusting the parameters of the fine-pored cover layer (porosity, thickness , catalytic activity) and the experimental conditions (pressures, temperature, solvents) are carried out so that the entire, by diffusion over the outer surface of the membrane entering the outer layer of hydrogen reacts within this. As a result, undesired total oxidation of the hydrogen, which would lead to a reduction in the selectivity to hydrogen peroxide, can not occur in the dry region of the support, even if the preparation technique is not completely avoidable, even in the case of optimized preparation technology.

A second equally important measure to guarantee inherent safety is defined by the device claim.

In the following, the invention will be explained in more detail with reference to an embodiment. To show: Figure 1. Simplified Reaction Scheme of Direct Oxidation of Hydrogen with Oxygen to Hydrogen Peroxide.

Fig. 2. Catalytic gas / liquid membrane contactor - operating principle

FIG. 3. Concentration profiles in the catalytic diffuser with hydrogen supply through the membrane and saturation of the reaction solution with oxygen. CO ₂ , L stands for the concentration of oxygen in the phase core of the liquid phase, cH ₂ , L for the corresponding hydrogen concentration. In contrast, c * H ₂ and c * O _{2 represent} the concentrations in the liquid phase immediately at the gas / liquid phase boundary.

FIG. 4. Concentration profiles in the catalytic diffuser with oxygen supply through the membrane and saturation of the reaction solution with dilute hydrogen.

FIG. 5. Basic circuit of the proposed method for the safe direct synthesis of hydrogen peroxide in connection with the modified contacting principle of the catalytic diffuser according to FIG. 4.

Fig. 6. Laboratory autoclave for the investigation of the kinetic properties of the catalytic membrane for the direct synthesis of hydrogen peroxide. Operation: a) H ₂ in the membrane and O ₂ in the liquid; b) O ₂ in the membrane and H ₂ / N ₂ in the liquid.

Fig. 7. Increase in the concentration of hydrogen peroxide and water as a function of the reaction time in Stainless steel autoclave. Operation a) Dosing of pure hydrogen in the membrane and of pure oxygen in the gas space above the liquid phase.

Fig. 8. Increase in the concentration of hydrogen peroxide and water as a function of the reaction time in stainless steel autoclave. Operation a) Dosing of pure oxygen in the membrane and 9% hydrogen in nitrogen in the gas space above the liquid phase.

By the following detailed example, the invention will be described in detail and the advantages of the invention will be made clear. With regard to the figures 1 to 4, reference is made to the above.

In Fig. 5 a basic circuit of the method is outlined as well as way of adding the reactants oxygen and hydrogen explained, which ensures that nowhere in the reactor explosive mixtures can arise.

Accordingly, in a saturator 1 upstream of the catalytic membrane module, the reaction solution is saturated with a hydrogen-containing gas (for example hydrogen diluted with nitrogen) under a pressure of typically 20-200 bar. This solution is conveyed via a metering pump 2 into the membrane module 3 and flows there on the side of the fine-pored cover layer along the membrane surface. A compressed oxygen-containing gas is supplied to the membrane module on the coarse-pored carrier side of the membrane, wherein the pressure is typically chosen such that firstly the partial pressure of oxygen is about 2-3 times the partial pressure of hydrogen in the saturator 1, and secondly the total pressure difference between the Gas side and the liquid side is sufficiently high to displace the liquid from the pores of the coarsely porous carrier. Depending on the design principle of the module 3 and the design of the catalytic membrane, the guidance of the gas and liquid phases may differ in individual cases, but it always follows these general considerations.

The liquid throughput through the membrane module 3 is controlled so that the total, located in the solution of hydrogen at the output of the module 3 is consumed. For this purpose, the dissolved hydrogen concentration of the liquid at the module outlet is continuously recorded. The emerging from the module reaction solution is expanded via a pressure relief valve 4 in a container 5 to ambient pressure, the escaping gas consists of unreacted oxygen and inert gases (hydrogen is at best in traces) and can optionally be delivered directly to the atmosphere. The saturated at ambient pressure with oxygen solution is cooled if necessary in the cooler 6 to remove the heat of reaction. About a splitter 7, a partial flow is removed as a product, the residual flow is recompressed by means of a circulation pump 8 to saturation pressure and returned to the saturator 1. By the relation to the ambient pressure high pressure in the saturator 1 it is ensured that the amount of oxygen introduced per unit of time with the recycled solution is small compared to the amount of hydrogen metered into the saturator 1 per unit of time, so that even in the saturator 1 explosive gas mixtures can not form , By suitable choice of the recycling ratio in the fragment 7, the concentration of the hydrogen peroxide solution produced can be adapted to the product-side requirements. Even with a possible failure of the membrane, no critical state is reached in this process, since membrane failure is immediately recognized on the basis of the pressure difference across the membrane and the supply of dissolved hydrogen via the metering pump 2 can be stopped immediately. Further improvement in yield is achieved by periodically modulating the gas pressure in the membrane at a sufficiently high frequency, thereby inducing a permanent renewal of the gas / liquid phase interface within the porous membrane, thus increasing the mass transfer rate.

Thus, a device-simple, safe, efficient and flexible method is available, which is particularly recommended for smaller system sizes. A significantly lower price of the hydrogen peroxide produced by this process compared to the established production methods is expected, which should consequently also improve the economy of many hydrogen peroxide-based processes and thus lead to a wider use of hydrogen peroxide as an environmentally friendly oxidizing agent.

Example: Direct synthesis of hydrogen peroxide in methanol using tubular membranes of α-Al ₂ O ₃ coated with monometallic palladium as a catalyst

Tube-shaped microfiltration membranes made of α-alumina from the Hermsdorfer Institute of Technical Ceramics (HITK) eV with 10 mm outer and 7 mm inner diameter and 100 mm length were coated with palladium according to the procedure described in FIG. The membranes consist of a coarse-pored carrier with an average pore size of 3 microns and about 1600 microns wall thickness. On the outside of the carrier there are two intermediate layers of about 30 microns thick and graded average pore diameters of 0.8 microns and 0.2 microns. As the last layer, the membranes have a fine-pored cover layer with a thickness of approximately 20 μm and an average pore diameter of 100 nm. The porosity of the individual layers decreases from the carrier layer. is about 25-30% for the support, 35-40% for the intermediate layers and 45-50% for the cover layer. After the deposition of palladium, finely divided, well-adhering, approximately hemispherical, palladium nanoparticles are predominantly present in the cover layer of the membrane, in lesser concentrations also in the two intermediate layers and to a lesser extent in the support. The size of the particles moves according to experience between 3 and 15 nm, the total amount of deposited palladium is under the usual conditions about 12-15 mg. Based on the outer pipe surface of approx. 30 cm2, this corresponds to a palladium loading of 4-5 g / m2. Extensive investigations using atomic absorption spectroscopy and an electron microprobe have shown that the membranes prepared by the above-mentioned method have a largely uniform distribution of the palladium over the circumference and in the axial direction, and that about 2/3 of the deposited palladium is present in the cover layer and find underlying intermediate layer.

The membranes were sequentially tested in various operations in a conventional stirred stainless steel autoclave, as shown schematically in Figure 6, for productivity and selectivity for the direct oxidation of hydrogen by oxygen to hydrogen peroxide. The conditions were as follows:

Reactor volume: 1.5 1

Fluid volume: 1.2 1

Stabilizer (H ₂ SO ₄ ): 1-10-2 mol / l

Promoter (NaBr): 8 mg / l

Solvent: MeOH

Temperature: 25 ° C

Pressure on the liquid: a) 67 bar (100% O ₂ ) b) 64 bar (9% H ₂ , 91% N ₂ ) Pressure in the membrane: a) 69 bar (100% H ₂ ) b) 66 bar (100% O ₂ )

Stirrer speed: 800 rpm

Fig. 6 shows a laboratory autoclave for investigating the kinetic properties of the catalytic membrane for the direct synthesis of hydrogen peroxide. The mode of operation is either a) H ₂ in the membrane and O ₂ in the liquid or b) O ₂ in the membrane and H ₂ / N ₂ in the liquid, where between the modes a) H ₂ addition via the membrane and O ₂ feed dissolved in the liquid, and b) O ₂ addition across the membrane and H ₂ feed to be distinguished.

In operation with supply of pure hydrogen into the membrane and pure oxygen into the headspace above the liquid, a productivity of about 54 g H ₂ O ₂ per g palladium and hour and a mean selectivity to H ₂ O ₂ of 20% was observed (cf. Fig. 7). This means that 80% of the hydrogen consumed led to the formation of water.

The productivity reached at dosage of pure oxygen into the membrane and 9% hydrogen and 91% nitrogen in the gas space in the reactor with 45 g H ₂ O ₂ per g palladium and hour a comparable order of magnitude, but at an average selectivity to H ₂ O ₂ of 90% (see Fig. 8). In this case, only 10% of the hydrogen used was converted to water.

The direct comparison impressively proves that only with optimized addition of the two reactants oxygen and hydrogen, a high selectivity to hydrogen peroxide is achieved, which, in addition to sufficiently high productivity and inherent safety is a mandatory requirement for an economic process for hydrogen peroxide synthesis. LIST OF REFERENCE NUMBERS

Saturator dosing pump membrane module expansion valve tank

Cooler splitter circulation pump

Claims

claims

1. An inherently safe process for carrying out the direct synthesis of hydrogen peroxide in a wettable, catalytic membrane contactor with high productivity and selectivity to hydrogen peroxide, comprising the following step:

A membrane contactor having a hydrophilic asymmetric membrane which has a coarse-pored carrier side and a fine-pored cover layer continuously coated with catalytically active components is provided.

characterized by the further step:

An oxygen-containing gas is fed to the carrier side, while on the side of the cover layer (reaction zone) is a partially saturated with hydrogen at the side of the gas pressure saturated reaction solution, which penetrates due to the capillary forces in the hydrophilic cover layer.

2. The method according to claim 1, characterized in that the hydrogen concentration in the reaction solution is lowered by inerting with an inert gas to about half of the expected at the gas / liquid phase boundary within the membrane dissolved oxygen concentration.

3. The method according to claim 1 or 2, characterized in that the reaction solution is reacted continuously, wherein the flow rate of the reaction solution is adjusted by a suitable membrane module such that in a module upstream of the saturator inserted brought hydrogen amount after passing through the module is almost completely implemented, so that even in a downstream flash tank no explosive gas mixture is formed.

4. The method according to claim 3, characterized in that after the passage of the reaction solution through the membrane module, the unreacted oxygen and optionally dissolved inert gases are removed from the reaction solution, and that a part of the reaction solution is recompressed and returned to the saturator.

5. The method according to claim 4, characterized in that the concentration of the produced hydrogen peroxide solution is adjusted in a simple manner by varying the return ratio to the product side requirements.

6. The method according to any one of the preceding claims, characterized in that the concentration range of the hydrogen peroxide solution prepared is 1 to 35 wt.%.

7. The method according to any one of the preceding claims, characterized in that nickel, palladium, platinum or a palladium compound or a platinum compound on α-alumina, γ-alumina, zirconia, titania or ceria and on carbon or with carbon coated, said ceramic materials can be used as a catalyst / carrier system.

8. The method according to claim 7, characterized in that the palladium or platinum compound contains silver, gold, rhodium, ruthenium, iridium or copper or tin.

9. The method according to claims 1 to 8, dadruch characterized in that asymmetric membranes are used which consist of a fine-pored cover layer with pore diameters in the range of 1 - 200 nm, a thickness of 1-50 microns and a porosity of 10 - 60%, a coarse-pored carrier of 0.2 to 3 mm thickness, 0.5 to 5 μm pore diameter and a porosity of 10 to 40%, and optionally one or more intermediate layers of 1 to 50 μm thickness, 10 to 50% porosity and 1 to 100 nm pore diameter.

10. The method according to claims 1 to 9, characterized in that tube bundle, capillary bundles, hollow fiber bundles, multi-channel elements, rotating disc membranes or other, today common in membrane technology module principles are used.

11. The method according to claims 1 to 10, characterized in that instead of pure oxygen, an oxygen-containing gas, in particular air or oxygen-enriched air, for. B. from a membrane separation process, is used as the oxidant.

12. The method according to claims 1 to 11, characterized in that instead of pure hydrogen, a hydrogen-containing gas, for. B. from the reforming or partial oxidation of hydrocarbons or biomass is used.

13. The method according to claims 1 to 12, characterized in that are used as reaction solution, water, alcohols, organic solvents or mixtures thereof with or without additions of acids, stabilizers, inhibitors or other auxiliary components.

14. The method according to claims 1 to 13, characterized in that the gas pressure in the membrane is modulated at a sufficiently high frequency periodically, whereby a permanent renewal of the gas / liquid phase interface within the porous membrane and thus an increase in the mass transfer rate is triggered.

15. A device for carrying out a method according to one of claims 1 to 14, characterized in that the device comprises: a saturator (1) for supplying hydrogen to a reaction solution, a membrane module (3), a gas / liquid separator (5) for Separating excess oxygen, a splinter

(7) for branching off a hydrogen peroxide product stream and a circulation pump (8) for recycling the portion of the reaction solution remaining after branching off the hydrogen peroxide product stream into the saturator (1).

16. The device according to claim 15, characterized in that the membrane module (3) has at least one asymmetric membrane consisting of a fine-pored cover layer with pore diameters in the range of 1 - 200 ntα, a thickness of 1-50 microns and a porosity of 10 - 60%, a coarse-pored carrier of 0.2 to 3 mm thickness, 0.5 to 5 μm pore diameter and a porosity of 10 to 40%, and optionally one or more intermediate layers of 1 to 50 μm thickness, 10 to 50% porosity and 1 to 100 nm Pore diameter exists.