The trend in commodities today is for materials and processes which are "environmentally friendly". One such material is hydrogen peroxide. Hydrogen peroxide has many potential appiications in, for example, chemical oxidation processes. One especially large field of use is as a bleaching agent in the pulp and paper industry. The demand for hydrogen peroxide is expected to grow at a rapid rate for many years. However, the current commercially practiced processes for synthesis of hydrogen peroxide are inefficient and have many disadvantages. As such it would be advantageous to develop a more efficient process for production of this commodity.

Most hydrogen peroxide is manufactured by a well-known anthraquinone process through successive reduction and oxidation reactions. Among the disadvantages of this method are that it requires the addition of numerous organic solvents, forms many unwanted by-products, and requires various separation steps. Another method for forming hydrogen peroxide is by catalytic reaction of hydrogen and oxygen with supported or homogeneous platinum group metal catalysts suspended or dissolved in aqueous solutions. However, this method requires bringing hydrogen and oxygen into a dangerous, potentially explosive, mixture together at high pressures (for optimum performance, usually greater than 7000 kPa), constituting a serious safety hazard. One method for avoiding excessive mixing of hydrogen and oxygen is cathodic reduction of oxygen in an alkali metal hydroxide solution. However, this process requires input of significant amounts of electrical energy and use of corrosion resistant equipment. A second method for avoiding excessive mixing of hydrogen and oxygen is to use fuel and reactor cells such as Proton Exchange Membrane (PEM) fuel cells. However, since such cells typically require formation and conductance of electrons and ions across the fuel cell by means of an electrochemical potential, they typically require complex catalytic, ionic, and electrical equipment. This equipment is generally inappropriate for large scale manufacturing operations. Similarly, a method for producing hydrogen peroxide without excessive mixing of the hydrogen and oxygen using a palladium metal membrane is unsuitable. For example, such a membrane is expensive, susceptible to poisoning, requires relatively higher temperatures for satisfactory hydrogen fluxes, and may be lifetime limited due to hydrogen embrittlement.

One method for synthesis of hydrogen peroxide has been patented by The Dow Chemical Company (U.S. Pat. No. 5,512,263). Such method employs a composite membrane having a multi- phase conductive path to conduct both electrons and cations through the membrane to react with oxygen on the other side of the membrane. It would, however, be desirable to have alternative methods and devices for production of hydrogen peroxide.

It would be an advantage to have a device, and method for its use, wherein hydrogen and oxygen may be controllably, but directly, reacted to form hydrogen peroxide outside of the explosive

range, without the use of organic solvents or complex equipment for ionic and electrical transport.

The invention disclosed herein includes a membrane which is useful for synthesis of hydrogen peroxide. The membrane has an oxygen contact side and a hydrogen contact side and comprises a porous hydrophobic catalyst layer facing the oxygen contact side and a gas flux control layer facing the hydrogen contact side. The gas flux control layer is positioned between the hydrogen contact side and the porous hydrophobic catalyst layer such that the flux of hydrogen may be controllably delivered to the porous hydrophobic catalyst layer. It is a further aspect of this invention to provide a method for synthesizing hydrogen peroxide (H2O2) by reacting hydrogen and oxygen using the membrane described herein.

An advantage of this method is that hydrogen peroxide may be produced at attractive rates using a considerably simplified process when compared to existing technology. Furthermore, the porous hydrophobic catalyst layer advantageously permits oxygen to be transported to a reactive interface with hydrogen while inhibiting the flooding of water through the porous hydrophobic catalyst layer from the oxygen side of the membrane. Another advantage of this invention is that it provides an effective mechanism for controlling the flux of hydrogen as it is transported to a reactive interface with oxygen.

FIGURE 1 is one embodiment of this invention. It iilustrates a membrane having a Hydrogen Contact Side 1 and an Oxygen Contact Side 2. A Gas Flux Control Layer 3 faces the Hydrogen Contact Side 1 and is positioned between the Hydrogen Contact Side 1 and Porous Hydrophobic Catalyst Layer 5, more specifically between the Hydrogen Contact Side 1 and the Macroporous Support 4. The Macroporous Support 4 is positioned substantially between the Porous Hydrophobic Catalyst Layer 5 and the Gas Flux Control Layer 3. Covering the Oxygen Contact Side 2 surface of the Porous Hydrophobic Catalyst Layer 5 is an Erosion Control Layer 6.

FIGURE 2 is another embodiment of the invention. It illustrates oxygen and water entering the reactor through an inlet into an Oxygen Supply Chamber 7, and hydrogen entering the reactor through an inlet into a Hydrogen Supply Chamber 8. The Membrane 9 (see FIGURE 1), having the Hydrogen Contact Side 1 and the Oxygen Contact Side 2, separates the two chambers. The Hydrogen Contact Side 1 of the membrane faces the Hydrogen Supply Chamber 8 and the Oxygen Contact Side 2 faces the Oxygen Supply Chamber 7. Two outlets 10 for withdrawal of product and excess gas are located at opposite ends of the reactor from the two inlets.

FIGURE 3(a) depicts a cross section of a reactor containing "serpentine" channels 11 (depicted in FIG. 3(b)) for this invention. Serpentine channels refer, simply, to a series of Oxygen Supply Chambers 7 on the Oxygen Contact Side 2 of the Membrane 9 and a corresponding series of Hydrogen Supply Chambers 8 on the Hydrogen Contact Side 1 of the Membrane 9.

The membrane of this invention requires a porous hydrophobic catalyst layer and a gas flux control layer. For optimal synthesis of hydrogen peroxide, the porous hydrophobic catalyst layer faces an oxygen contact side of the membrane and the gas flux control layer faces a hydrogen

contact side of the membrane. The oxygen contact side of the membrane may simply be a surface of one side of the porous hydrophobic catalyst layer. Similarly, the hydrogen contact side of the membrane may simply be a surface of one side of the gas flux control layer. The gas flux control layer is positioned between the hydrogen contact side and the porous hydrophobic catalyst layer such that the flux of hydrogen may be controllably delivered to the porous hydrophobic catalyst layer. This membrane is useful for synthesis of hydrogen peroxide by placing hydrogen in contact with the hydrogen contact side of the membrane, placing oxygen in contact with the oxygen contact side of the membrane, and allowing the hydrogen and oxygen to be transported into the membrane to an interface at the porous hydrophobic catalyst layer. When the oxygen and hydrogen are contacted at the catalyst layer, reaction may take place, producing hydrogen peroxide.

For the purposes of this invention, the word "hydrophobic", when used in conjunction with the catalyst layer, is meant to mean that the layer substantially inhibits the spontaneous intrusion of water from the oxygen contact side of the membrane into the catalyst layer. "Spontaneous intrusion of water" can occur if the catalyst layer is substantially hydrophilic such that, water contacting the layer will wet the pores of the catalyst layer and be retained by capillary action unless the differential pressure across the layer exceeds the "bubble-point" pressure. Also for the purposes of this invention, the word "porous", when used in conjunction with the catalyst layer, is meant to mean that the layer permits oxygen to access catalyst sites in the catalyst layer without being significantly impeded by pore size and/or spontaneous intrusion of water into the pores. Therefore, by having the catalyst layer be both hydrophobic and porous, maximum interaction of hydrogen and oxygen is permitted at the catalyst layer.

Generally, the porous hydrophobic catalyst layer comprises a catalyst and a substantially hydrophobic material. For example, the porous hydrophobic catalyst layer may comprise a catalytic material and a hydrophobic composite material wherein the composite comprises a first material that has been treated with a second material of low surface energy such that the composite exhibits a contact angle with water of greater than ninety degrees. The catalytic material may be intermixed with the hydrophobic material, or may be organized as a more discrete layer (within the porous hydrophobic catalyst layer) wherein the hydrophobic material may be part of a layer which is positioned primarily towards the oxygen contact side of the porous hydrophobic catalyst layer and the catalytic material is part of a layer which is positioned primarily towards the hydrogen contact side of the porous hydrophobic catalyst layer. Examples of materials which are substantially hydrophobic include: styrene divinylbenzene copolymers; polyethylene, polypropylene or ethylene- propylene copolymers; silica which has been rendered hydrophobic by treatment with a silane or with fluorine or a fluorinated compound; fluorinated polymers and copolymers; carbon which has been rendered hydrophobic by treatment with a silane or with fluorine or a fluoridated compound; and combinations thereof.

The catalyst is typically an oxygen reducing catalyst. Preferably, the oxygen reducing catalyst comprises a metal selected from the group consisting essentially of platinum, palladium, rhodium, rhenium, indium, gold, silver, copper, cobalt, iron, nickel, and combinations thereof. Those of skill in the art recognize catalysts which are beneficial for oxygen reduction, such as: silver, gold, bismuth, palladium, cobalt (see, for example, Putten et al., J. Chem. Soc., Chem. Commun. 477 (1986), incorporated herein by reference), niobium-titanium, lanthanum-manganese mixtures, indium-tin oxide mixtures, praseodymium-indium oxide mixtures, metal phthalocyanines (see, for example, Cook et al., 137 [No. 6] J. Electrochem. Soc. 2007 (1990), incorporated herein by reference), metal porphyrins (see, for example, Chan et al., 105 J. Am. Chem. Soc. 3713-14 (1983), incorporated herein by reference), and anthraquinone-based catalysts (see, for example, Degrand, 169 J. Electroanal. Chem. 259-68 (1984), incorporated herein by reference). Preferred oxygen reducing catalysts comprise at least palladium. Various additive metals may also be useful, in combination with the above-described oxygen reducing catalysts, such as lead, zinc, copper, gallium, tin, and bismuth. One embodiment for the catalyst is to utilize the catalytic material such that it is present in at least two different oxidation states.

Catalysts may be in the form of pure materials, or they may be supported on carriers. When carriers are utilized, a preferable carrier is selected from the group consisting of carbon, silica, titania, zirconia, alumina, lanthanum oxides, cerium oxides, zeolites, heteropolyacids, alkaline earth sulfates, alkaline earth phosphates, titanium silicates, vanadium silicates, and combinations thereof.

For example, one embodiment of this invention is to deposit ultrafine gold particies (for example particle having a radius of less than 10 nanometers) on one of the above carriers, such as a titanium silicate. The catalyst, or catalyst with carrier, may then be rendered hydrophobic by any appropriate method such as treating it, mixing it, or forming a composite, with the substantially hydrophobic material. As described in further detail herein below, one further embodiment of this invention is to also include at least one hydrogen peroxide selectivity increasing additive in the porous hydrophobic catalyst layer. Such additive may be doped into the porous hydrophobic catalyst layer using a pretreatment process. Methods for incorporating and depositing catalysts and other additives onto and into other materials are well known in the art. In light of the disclosure herein, one of skill in the art is capable of optimizing these deposition methods to form the porous hydrophobic catalyst layer.

Examples of membrane fabrication and catalyst deposition methods are disclosed in: A. B. Stiles, "Catalyst Manufacture, Laboratory and Commercial Preparations", Marcel Dekker, Inc., New York, ISBN 0-8247-7055-2, 1983; C. N. Satterfield, "Heterogeneous Catalysis in Practice", McGraw-Hill Book Company, New York, ISBN 0-07-054875-7, 1980; A. J. Appleby and F. R. Foulkes, "Fuel Cell Handbook", Van Nostrand Reinhold, New York, ISBN 00-442-31926-6, 1989; K. Kinoshita, "Electrochemical Oxygen Technology", John Wiley & Sons, Inc., New York, ISBN 0-471-57043-5, 1992; Nidola et al., U.S. Pat. No. 4,364,803 (1982); Watanabe et al, "New Preparation Method of a High Performance Gas Diffusion Electrode Working at 100% Utilization of Catalyst Clusters and

Analysis of the Reaction Layer", J. Electroanal. Chem., 197 (1986) 195-208; Watanabe et al, "Experimental Analysis of the Reaction Layer Structure in a Gas Diffusion Electrode", J. Electroanal.

Chem., 195 (1985) 81-93; and Takenaka et al., U.S. Pat. No. 4,328,086 (1982). The relevant teachings of all references are incorporated herein by reference.

It may also be desirable to incorporate an ion exchange material into the porous hydrophobic catalyst layer The ion exchange material is useful for enhancing the productivity of the catalyst. For example, in the case of an oxidatively or reductively reactive catalyst, which is reactively converted to a species that is soluble, an ion exchange material will capture the soluble catalyst and allow: (1) reaction back to the insoluble form by an added reactant (oxidant or reductant), and/or (2) regeneration of the insoluble catalyst in a separate step. Therefore, the ion exchange material inhibits catalyst loss without removing spent catalyst from the system nor introducing materials that may cause product decomposition. For example, when a membrane is used without the ion exchange membrane, conditions of operation typically oxidize and dissolve the catalyst into the product stream. However, when the ion exchange membrane is used, passage of hydrogen gas over the ionically-bound catalyst ion will reconvert it into the insoluble catalyst metal.

The ion exchange material will then be available for more incorporation of catalyst ions. Since, when using the ion exchange material, less catalyst metal will be lost, a longer lifetime can be expected.

Furthermore, proper choice of ion exchange material may add acidity or promotion to the catalyst.

In one embodiment, a polymer of perfluorosulfonic acid (PFSA) is incorporated into the porous hydrophobic catalyst layer which employs palladium as the catalytic material. For a discussion of PFSA polymers, and methods of preparing such polymers, see De Vellis et al., U.S. Patent No.

4,846,977, col. 5, lines 1-36. See also Kirk-Othmer, "Perfluorinated-lonomer Membranes," Encyclopedia of Chemical Technology, pp. 591 - 98 (1984) and A. Eisenberg and H. Yeager, "Perfluorinated lonomer Membranes", ACS Symposium Series No. 180 (1982). An example of a commercially available PFSA polymer is NAFIONTM (E.l. du Pont de Nemours and Company).

As mentioned, the membrane also includes a gas flux control layer which faces the hydrogen contact side of the membrane. The gas flux control layer is positioned between the hydrogen contact side and the porous hydrophobic catalyst layer such that the flux of hydrogen may be controllably delivered to the porous hydrophobic catalyst layer. In order for this invention to function desirably, the gas flux control layer must allow hydrogen to be delivered from the hydrogen contact side to the oxygen contact side at a rate at least equal to that required for maintaining an acceptable minimum rate of reaction with oxygen. However, the gas flux control layer must also prevent excessive flow of hydrogen to the oxygen contact side, since such excessive flow can create a danger of an explosion, or necessitate venting or recycle of large volumes of undesirably mixed gases, adding complicated and costly steps to the synthesis reaction. The gas flux control layer also preferably inhibits excessive oxygen transport across the membrane. "Flux" as used herein shall mean the flow rate of a permeating species per unit cross-sectional area of the gas flux

control layer (that is (standard cc)/(cm2.sec), wherein "standard" is equal to 0° C and 760 mmHg pressure). For this invention, the gas flux control layer is selected from a material which is effective for at least one of the following means for transporting hydrogen from the hydrogen contact side of the membrane: solution-diffusion transport; viscous flow; Knudsen flow; and any combination thereof.

Gas transport through dense polymer typically occurs by a solution-diffusion (S-D) mechanism. In this mode of transport, gas dissolves into and diffuses through the membrane (for example, see Vieth, W.R. , Diffusion in and Through Polymers, Oxford University Press, New York 1991) "Equation I" describes this mechanism of transport: <BR> <BR> <BR> <BR> <BR> <BR> ~ APj (Equation 1) <BR> <BR> <BR> Nss and <BR> <BR> <BR> <BR> <BR> <BR> pA ~ N.,xA <BR> <BR> <BR> B - - (Equation 2) where p is permeability, Nss is the steady state flux, I is the thickness, apj is the partial pressure driving force across the membrane and a is the selectivity.

Defects in the a solution-diffusion membrane can lead to gas transport which can be described by either a viscous and/or Knudsen flow model. While both of these mechanisms describe transport of gas through a pore or capillary in an otherwise dense material, the two flow mechanisms are quite different.

In viscous flow, flux is dominated by interactions between the flowing molecules. Viscous flow is well described by the "Hagen-Pouiseuille Equation" (see Bird, R.B., Stewart, W.E. and Lighffoot, E.N., Transport Phenomena, John Wiley & Sons, Inc. 1965, (Equation 3).

(Equation 3) wherein S is the cross sectional area, r is the radius, p is pressure, u is viscosity, R is the gas constant, I is the membrane thickness, and T is temperature. The values of S and r are characteristic of the pore structure for a given membrane and U is a physical property of the gas.

Viscous flow differs from solution-diffusion flow in that the driving force is the pressure difference rather than the partial pressure difference across the membrane. Another major difference is that viscous flow increases with system pressure as well as differential pressure. Because of this, the viscous permeance (Nv/Dp) and permeability are an increasing function of pressure.

For a given membrane flux obeying the above equation, the selectivity of the membrane is the reciprocal of the ratio of the viscosity of the two gases:

(Equation 4) The final flow mechanism is intermediate to the previous two mechanisms. Like viscous flow, Knudsen flow describes flow in a small channel or capillary (see Hwang, S. and Kammermeyer, K. Membranes in Separations, John Wiley & Sons 1975). However, in Knudsen flow, the diameter of the channel is less than the mean free path of the molecule and the molecule has more collisions with the wall of the capillary than with other molecules. Because of this, the driving force for Knudsen flow is the partial pressure difference across the membrane. Knudsen flow can be described by the following equation: 8rJ.r3 Ap NK 2HMwRT)112 1 I/2 (Equation 5) (2IIMWRT) where NK iS the flux, r is the radius of the capillary, R is the gas constant, Mw is the molecular weight, T is the temperature, 1 is the membrane thickness and Ap is the partial pressure difference across the flow channel.

For Knudsen flow, the selectivity of the membrane is the square root of the ratio of the reciprocal molecular weight of the two gases: M2 6) aK l/2 (Equation 6) MWB The Knudsen selectivity is intermediate to viscous and solution-diffusion selectivity and much closer to viscous flow selectivity. A major difference between Knudsen flow and viscous flow is that the Knudsen flow permeability coefficient is independent of pressure. Like S-D flux, Knudsen flux increases linearly with partial pressure.

Solution-Diffusion membranes are typically used for gas separations. This is primarily due to the high selectivity achievable by this mechanism. In gas separations, a large differential pressure is imposed on the membrane to maximize productivity and promote the enhanced flux of one component of the mixture. Membranes for membrane reactor applications have different requirements. The purpose of the membrane reactor is to separate the pure bulk phases of the two potentially dangerous reactive species and to controllably react (mix) the reactants. For a typical membrane-type reactor, two essentially pure gases are flowing on opposing sides of the membrane.

Although the pressure differential across the membrane can be relatively small, the partial pressure difference across the membrane is high for both gases. In a solution diffusion membrane, this causes diffusion of the low pressure gas into the high pressure gas. Flammability limits can be reached by back-diffusion. High selectivity is desired in order to prevent the back-diffusion.

A viscous flow membrane would have very desirable selectivity since flow is driven by the absolute pressure gradient. Gas substantially flows from the high pressure side of the membrane to the low pressure side. The difficulty with gas delivery by a viscous flow mechanism is the degree of uniformity required in the membrane. For uniform delivery of gas, the discreet pores need to be uniformly distributed over the membrane. Equally important is the pore size distribution. Flow through a pore increases as the fourth power of the radius. Small changes in the pore size can lead to large changes in gas delivery.

For purposes of the invention described herein, a flux control layer exhibiting all three flow mechanisms is a desirable means for gas delivery. The three flow mechanisms lead to different methods to regulate the gas flux across the membrane. One method of control is absolute pressure.

Increasing the pressure of the system at a fixed differential pressure increases the solution diffusion, Knudsen and viscous contributions to an equivalent extent. As pressure is increased, oxygen flux increases due to the increased driving force. Since there is substantially no reaction occurring on the hydrogen side of the membrane, helping to diminish the influx of oxygen, it is possible to reach flammability limits on the hydrogen side of the reactor. A decrease in oxygen pressure could help to alleviate the problem. Oxygen flux will decrease linearly with the pressure decrease. For example, at a typical pressure used in a reactor employing the membrane of this invention of 200-400 pounds per square inch (psig), a 10% decrease in oxygen flux would require the differential pressure to increase by 20 to 40 psi. With a membrane containing viscous flow defects, increasing the differential pressure will simultaneously increase hydrogen flux and decrease the oxygen flux thereby requiring smaller changes in differential pressure. This increase in hydrogen flux with a decrease in oxygen flux is identical to replacing the membrane with a membrane that has higher selectivity. In essence, selectivity of the flux control layer can be regulated by differential pressure for a membrane that incorporates all three flow mechanisms.

The composition of the gas flux control layer may be any which functions as set forth herein.

Since there are many materials and combinations of materials (mixtures, blends, multilayers) that have acceptable flux rates, selection of a gas flux control layer for obtaining the described fluxes should be within the skill in the art when combined with the teachings provided herein. It may be either organic, inorganic, or a combination thereof. One composition for the gas flux control layer comprises a composite of polytetrafluoroethylene and carbon. Catalyst may be incorporated predominantly into one side of the composite in order to form the catalyst layer. Preferably, the gas flux control layer comprises an organic, polymeric material selected from the group consisting essentially of polycarbonates, polyester, polyestercarbonates, polysulfones, polyolefins, polyphenylene oxides, polyethers, polyimides, polystyrenes, polyetherimides, polyamideimides, and polyethersulfones. A preferable material is tetrabromobisphenol A polycarbonate (TBBA-PC). More preferably, the composition of the gas flux control layer is halogenated. Commercial embodiments of such materials include sulfonated styrene grafts on a polytetrafluoroethylene backbone

(commercially available from RAI Research Corporation as RAIPORETM membranes) and crosslinked sulfonated copolymers of vinyl compounds (commercially available from lonics, Inc., as TYPE CRTM membranes).

In many embodiments of this invention it is preferable that the membrane further comprise a macroporous support. The macroporous support may be positioned anywhere within the membrane as long as it provides sufficient mechanical strength to withstand any differential pressure across the membrane. Alternatively, the macroporous support layer may be omitted and another layer in the membrane, such as the erosion control layer (described hereinbelow), may be constructed of sufficient mechanical strength as to additionally serve the same purpose as the macroporous support layer. Typically, when a macroporous support layer is used, it is positioned substantially between the porous hydrophobic catalyst layer and the gas flux control layer. By "macroporous" it is meant that the support is of a pore size sufficient enough to provide negligible resistance to gas flow compared to the flux reduction layer. By stating "substantially between" the layers, it is intended to cover embodiments in which the macroporous support and either the porous hydrophobic catalyst layer and/or the gas flux layer overlap and are commingled. Examples of porous substrates/supports are carbon paper (for example Toray TGPH-120w) and supports for reverse osmosis membranes such as disclosed in U. S. Patent 4,277,344, assigned to FilmTec Corporation.

Additionally, the support may desirably be rendered hydrophobic by incorporating a hydrophobic material such as poly[l -trimethylsilyl-l -propyne] ("PTMSP") into the porous support to enhance overall hydrogen flux, provide a surface with enhanced capability to impede the transportation of undesirable fluid components through the composite structure, or to minimize the effect of pinhole leaks in the membrane.

The porous supporting layer is characterized in that it does not greatly impede the transport of molecular hydrogen when the hydrogen is placed in contact with the macroporous supporting layer. Generally, the selectivity of the support layer doesn't matter and is typically very low.

However, it cannot be chosen indiscriminately since it also does have some impact on both the flux and selectivity of the membrane. In one embodiment, the macroporous support is a porous polymer membrane. Illustrative of such polymeric supporting layers are celluiose ester and porous polysulfone membranes. Such membranes are commercially available under the trade names MILLIPORETM, PELLICONTM and DIAFLOWTM. Where such supporting membranes are thin or highly deformable, a frame may also be necessary to adequately support the semi-permeable membrane.

One preferred embodiment is to utilize a support having a "dual porosity" structure. By "dual porosity", it is meant that the support structure has a "coarse" pore layer facing the hydrogen contact side of the membrane, and a "fine" pore layer facing the oxygen contact side of the membrane. The "coarse" pores (for example radius of 10-30 micrometers) are larger than the "fine" pores (for example radius of 0.8-1.0 micrometers), which means that the capillary pressure of any liquid (for

example water or product on the oxygen contact side of the membrane) is larger in the fine pore layer than it is in the coarse pore layer. Liquid penetrates the fine pore layer, and would continue to penetrate the coarse pore layer. Interface control is achieved by increasing the pressure of the gas to a point between the capillary pressure of the liquid in each pore size region. Thus, the gas pressure is greater than the capillary pressure in the coarse pore layer, and forces the liquid back.

Since it is less than the capillary pressure in the fine pore layer, liquid is further inhibited from intrusion towards the hydrogen contact side of the membrane. In conjunction with electrolytic process, a method for utilization of a dual porosity mechanism has been disclosed and patented by The Dow Chemical Company. See U.S. Pat. No.'s 4,341,606 and 4,260,469, incorporated herein by reference. Such a structure may be fabricated in many different ways, but one typical method would be to form a composite of two layers, one layer having the coarse pores and the other layer having the fine pores.

A preferred embodiment of this invention is for the membrane to further comprise a catalyst erosion control layer. The catalyst erosion control layer is preferably located on the surface of the porous hydrophobic catalyst layer, more preferably, the oxygen contact side of the porous hydrophobic catalyst layer. The function of the catalyst erosion control layer is to inhibit the erosion of catalyst out of the porous hydrophobic catalyst layer. Typically, this occurs by the catalyst being carried away by either water and/or hydrogen peroxide synthesis products which are formed and removed from the oxygen contact side of the membrane. The catalyst erosion control layer may comprise any material useful for performing this function. One embodiment comprises porous carbon paper. A preferred embodiment comprises ion exchange resin, as described previously for incorporation into the porous hydrophobic layer itself. As described previously, the ion exchange resin (for example NAFION) will capture soluble catalyst and allow: (1) reaction back to the insoluble form by an added reactant (oxidant or reductant), and/or (2) regeneration of the insoluble catalyst in a separate step. Therefore, the ion exchange material inhibits catalyst loss without removing spent catalyst from the system nor introducing materials that may cause product decomposition. Another preferred embodiment is to utilize a catalyst erosion control layer which comprises both porous carbon paper and ion exchange resin. As described previously with regards to the "macroporous support", it may also be advantageous to fabricate the catalyst erosion control layer such that it has sufficient mechanical strength to withstand differential pressure across the membrane.

In light of the disclosure herein, and membrane preparation techniques which are already known to those of skill in the art (see citations set forth herein above), fabrication of the membranes of this invention may be accomplished. The membranes of this invention may further be subjected to treatments with heat or by stretching in order to modify properties of the membranes. Optionally, the membranes may be subjected to other treatments known to one of skill in the art such as solvent annealing, etching, irradiating, cross-linking, fluorinating, sulfonating, plasma treating, and the like.

In one preferred embodiment, the membrane is heat annealed before use. The membrane is exposed to temperatures above the beta transition and below the glass transition temperature of the membrane for a period of time to partially densify the polymer. This procedure can optionally be performed under vacuum.

The membrane may be utilized in many different structural forms. One embodiment of this invention is to use a hollow fiber form of the membrane, however, typically the membrane is in the form of a substantially flat sheet. Since the membrane may be relatively thin or highly deformable, it may be desirable to provide added support to the membrane. There are well known ways to produce membranes to do this. In one embodiment, the peripheral area of the membrane is affixed to a framing structure which supports the outer edge of the membrane. The membrane can be affixed to the framing structure by a clamping mechanism, adhesive, chemical bonding, or other techniques known by one of skill in the art. The membrane affixed to the frame can then be sealingly engaged in the conventional manner in a vessel so that the membrane surface inside the framing support separates two otherwise non-communicating compartments in the vessel. The skilled artisan will recognize that the structure which supports the membrane can be an integral part of the vessel or even the outer edge of the membrane.

Another embodiment of this invention is a chemical reactor which incorporates the membrane of this invention. Such a chemical reactor typically comprises: the membrane, as set forth herein; a means for supplying hydrogen gas to the hydrogen contact side of the membrane; a means for supplying oxygen gas to the oxygen contact side of the membrane; and a means for removing product from the oxygen contact side of the membrane. The hydrogen contact side of the membrane is positioned such that it faces, and operatively connects to, a hydrogen supply chamber.

The oxygen contact side of the membrane is positioned such that it faces, and operatively connects to, an oxygen supply chamber. "Operatively connects" means that each chamber is positioned with respect to the membrane such that a relevant composition (for example, hydrogen or oxygen) can be placed in contact with its respective contact side of the membrane. "Chamber" includes any vessel, space, zone, or the like, capable of substantially containing and facilitating contact between any relevant composition and an appropriate surface of the membrane. Thus, a hydrogen supply chamber provides an effective environment for introducing, containing, and placing hydrogen, or a hydrogen containing mixture, in contact with the hydrogen contact side of the membrane. Similarly, the oxygen supply chamber provides an effective environment for introducing, containing, and placing oxygen, or an oxygen containing mixture, in contact with the oxygen contact side of the membrane. In addition, each chamber desirably has at least one opening for supply and/or removal of relevant composition(s), reaction products, or both. It may also be useful to utilize a "Serpentine" channel arrangement wherein the hydrogen and/or oxygen chambers on respective sides of the membrane consist of a series of connected parallel channels with alternating flow directions. With such a serpentine channel, it is preferred that the channels on one side of the membrane are aligned

in parallel with respective channels on the other side of the membrane. Furthermore, more than one opening per chamber may also be provided wherein one opening is an inlet for introducing a relevant composition into its respective chamber and one opening is an outlet for removing reaction products and/or unreacted relevant compositions.

Another aspect of this invention is a method of using the membrane of this invention for synthesis of hydrogen peroxide. The method comprises placing hydrogen in contact with the hydrogen contact side of the membrane, placing oxygen in contact with the oxygen contact side of the membrane, and contacting the hydrogen and oxygen at an interface in the hydrophobic catalyst layer. Conditions, such as temperature and pressure should also be provided which are sufficient to react the hydrogen and oxygen to form the hydrogen peroxide.

When the hydrogen is contacted with the hydrogen contact side, the hydrogen is transported through at least the flux control layer to the porous hydrophobic catalyst layer. When the oxygen is provided at the oxygen contact side of the membrane, it is transported to the porous hydrophobic catalyst layer and is placed in contact with the hydrogen at the catalyst layer. The hydrogen then reacts with the oxygen to form a reaction product comprising hydrogen peroxide. As described above, preferably the oxygen contact side comprises an oxygen reducing catalyst (for example palladium). The oxygen reducing catalyst at the oxygen contact side is chosen to enhance the rate of the reaction between hydrogen and oxygen, and to provide high selectivity to produce hydrogen peroxide rather than water. For purposes of this invention, hydrogen peroxide "selectivity" is defined as the moles of hydrogen peroxide produced divided by the total of the moles of water produced plus the moles of hydrogen peroxide produced (that is H202 selectivity = moles H202 produced i (moles H2O produced + moles H202 produced)).

Although the oxygen is preferably provided to the oxygen contact side of the membrane as a stream of pure oxygen gas, a typical method of introducing oxygen to the oxygen contact side is as a component in a mixture such as air. It is also preferable for the oxygen to be introduced in a mixture with water. The water helps dilute the hydrogen peroxide product, thereby reducing its potential decomposition. The water may also assist in the removal of the heat of reaction. It is desirable that, when the oxygen is introduced in a mixture with water, the concentration of oxygen is high enough such that at least bubbles, or even pockets, of oxygen are present in the water (in contrast to substantially all of the oxygen being dissolved in the water). Hydrogen peroxide stabilizers may also be included in the oxygen feed stream. Typical hydrogen peroxide stabilizers include amino-tri(methylene phosphonic acid), 1-hydroxyethylidene-1,1-diphosphonic acid, ethylene diamine tetra(methylene phosphonic acid), pyrophosphoric acid, salts thereof, and combinations thereof.

It is most preferred to further include additives to the membrane which are optimized for increasing hydrogen peroxide selectivity. Such additives may be provided to the membrane in any number of ways depending upon the membrane reactor design utilized. For example, such additives

may be doped into the porous hydrophobic catalyst layer using a pretreatment process, or they may be supplied by means of one or both of the oxygen- or hydrogen-containing feed streams. Such additives, for example, may include H2SO4, HCN, HNO3, H3PO4, HCI, HBr, HI, (COOH)2, CH2COOH, HCOOH, salts thereof, and combinations thereof. Many such additives are known in the art for increasing selectivity to hydrogen peroxide. See T.Z. Pospelova et al., "Palladium-Catalysed Synthesis of Hydrogen Peroxide from the Elements," Russian Journal of Physical Chemistry 35(2), 143 (1961) (incorporated herein by reference). Additionally, gaseous species such a HX, wherein X is a halogen, CO, CO2, or H2S could be added to the hydrogen-containing feed stream to increase selectivity to hydrogen peroxide. Another class of additives that may be added to the oxygen- containing feed stream are surfactants or phase transfer agents, such as t-BuNX, wherein X is hydroxyl or a halogen. These materials can act separately or in conjunction with other additives to bring about increased selectivity.

This method of chemical synthesis may, if desired, be conducted at an elevated temperature. Generally, the temperature should not exceed a temperature at which any one of the materials of the chemical reactor (for example the membrane), or product (for example hydrogen peroxide) significantly decompose or degrade. This temperature, and the significance of chemical reactor degradation, vary according to the specific composition of the membrane. Generally, the temperature is maintained at less than 75" C, however, one of skill in the art is capable of selecting an appropriate temperature as other conditions in a chemical reactor may be varied.

In addition, the method of the invention is typically conducted at a pressure of from ambient (taken as 100 kPa) to 14,000 kPa (2030 psi). It is preferred that a pressure differential between each side of the composite membrane does not exceed 700 kPa (100 psi) to avoid damage to the membrane. Robust membranes, however, may allow the use of higher differential pressures from increased hydrogen pressures. Elevated hydrogen pressures can enhance the flux of hydrogen relative to that of oxygen. Generally, increased pressure provides an increased mass transfer rate of the reactants. A particularly preferred pressure is from 750 kPa (109 psi) to 6,800 kPa (986 psi).

It is typically preferable that the pressure of hydrogen on the hydrogen contact side of the membrane is greater than that of the oxygen on the oxygen contact side of the membrane. However, those of skill in the art will recognize that it is also important to also maintain a flux of hydrogen from the hydrogen contact side which results in a concentration of hydrogen in the oxygen supply chamber which is outside of the flammability range.

Finally, it is preferable to remove any reaction products from the oxygen contact side of the membrane. This isolates desirable reaction products and minimizes undesirable side reactions and decomposition of hydrogen peroxide. This may be accomplished using any means known by those of skill in the art. A simple method for reaction product removal is for the product to be swept up at the surface of the oxygen contact side of the membrane into a continually flowing oxygen feed stream containing liquid water.

Examples The invention will be further clarified by a consideration of the following examples, which are intended to be purely exemplary of the use of the invention.

As described in the examples set forth below, unless stated otherwise, the carbon paper was purchased from E-TEK, Inc., Natick, MA. Specifically, Toray TGPH120TM, from E-TEK was utilized. Toray TGPH-120TM is a porous carbon paper having a nominal thickness of 12-14 mils and density of approximately 0.55 grams/cm3. The gas permeability of the Toray carbon paper is approximately 54,000 X 10-4 cc / cm2x sec. x cm. Hg.

TM Also set forth in the examples below are references to TeflonTM-coated Toray paper.

Unless stated otherwise, this is intended to mean a Toray TGPH-120 carbon paper which has been coated with a dispersion of (poly)tetrafluoroethylene (PTFE), specifically DuPont's TeflonT- 30TM. The TeflonTM-coated paper is dried in an oven under vacuum at approximately 325"C to melt the PTFE and disperse it over the whole surface of the carbon paper. The resulting treated carbon paper is hydrophobic (as demonstrated by water repulsion). The paper remains porous, but now has a density of approximately 1.1 grams/cm3. The gas permeability of the TeflonTM-coated Toray <BR> <BR> <BR> carbon paper is approximately 38,000 X 10 4 CC / cm2x sec. x cm. Hg. The coated carbon paper has a volume of dispersion containing the equivalent of approximately 10-18 mg of PTFE per centimeter squared (cm2) of carbon paper.

EXAMPLE 1 (a) Carbon Paper and Catalyst Composite Preparation Catalyst was prepared by mixing 1.25 gram (g) of 20 weight percent (wt%) Pd on carbon (obtained from E-TEK, Inc.) with 5g glycerol, 1.25 g water, and 8.3 g of a 5 wt% solution of NAFIONTM in an alcohol/water solution (obtained from Aldrich Chemical, Milwaukee, WI) to obtain a catalyst paint. The catalyst paint was applied to TeflonTM-coated carbon paper using an artist's brush. Only a thin coating was applied each time. It was then dried in the oven under vacuum at 135"C, cooled to room temperature (25"C), and the painting continued until all the paint was transferred onto the carbon paper to obtain a carbon paper composite with the required Pd loading.

The composite was then dried at 135"C for 45 minutes under vacuum. The carbon paper/catalyst composite was then heat pressed at 2.25 MPa, and a plate temperature of 140"C to form a carbon paper composite having an approximate size of 36 x 11 cm.

(b) TBBA-PC Asymmetric Membrane Preparation A TBBA-PC (tetrabromobisphenol A polycarbonate) asymmetric membrane was prepared by forming an 32 wt% solution of TBBA-PC in N-methylpyrrolidone at 58"C. The solution was degassed in a vacuum oven and was cast on a PYREXTM glass plate heated to 75"C. The resulting cast film and the plate was immediately immersed in water having a temperature between 10"C to 25"C and left in the water for between 1 to 2 hours. The resulting asymmetric membrane was air

dried and then dried in vacuum at 60"C. The thickness of the membrane was 0.2 mm and the membrane had one side that was comparatively more dense than its opposite side which was comparatively more microporous. Permeation rates of hydrogen and oxygen through the membrane were measured using standard techniques, as described and referenced in J. Comyn, Editor, "Polymer Permeability", Elsevier Applied Science Publishers, London/New York, 1985, ISBN 0- 85334-322-5. The hydrogen flux through the membrane ranged from 1x10-5 to 1X104 (standard cc)/(cm2.sec). This hydrogen flux was as high as 15 times greater than the oxygen flux through the membrane.

EXAMPLE 2 H202 Synthesis Using a Polymeric-based Membrane (a) A TBBA-PC asymmetric membrane (prepared as described in Example 1 (b)) and a carbon paper/catalyst composite (prepared as described in Example 1(a)), and having a Pd catalyst loading of 0.6 mg/cm2, was sandwiched together in a reactor similar to that depicted in "FIG. 2". The reactor consisted of metal plates having milled flow channels arranged in a "serpentine" pattern, the flow channels being used to supply hydrogen to one side of the membrane, and oxygen and water to the other side of the membrane. The area of membrane exposed to the serpentine channels on each face of the membrane was 15 in2 (100 cm2). The dense side (as opposed to the microporous, hydrogen contact, side) of the asymmetric TBBA-PC film was placed against the exposed carbon paper side of the carbon paper/catalyst composite. The catalyst side of the composite thus faced the oxygen/water flow channel. The aqueous "water" solution also contained hydrogen peroxide selectivity increasing additives of 0.01 M H2SO4 and 0.008M HBr. This aqueous solution was added to the oxygen feed stream at a rate of 0.25 ml-/min., while the molecular oxygen was added at a rate of 0.5 Umin. Flows of molecular hydrogen and the oxygen/water feeds were maintained and the pressure in the reactor was increased to 400 psig (2758 kPa) on the hydrogen contact side and 350 psig (2413 kPa) on the oxygen contact side to maintain a pressure differential of 50 psig (345 kPa) between H2 and 02. This allowed for the diffusion of H2 through the membrane. The oufflowing liquid on the oxygen contact side of the polymeric membrane was analyzed for hydrogen peroxide. The concentration of hydrogen peroxide was 1.35 wt%.

(b) A second experiment using a membrane containing a catalyst layer with a Pd loading of 0.13 mg/cm2 (made using 5% Pd on Vulcan XC-72 carbon using similar procedures as in "1 (a)") produced 0.58 wt% hydrogen peroxide at an aqueous solution feed flow rate of 1 mUmin. (oxygen flow of 0.250 Umin), and 1.85 wt% hydrogen peroxide at an aqueous solution feed flow rate of 0.125 ml/min. (oxygen flow of 0.250 Umin). The experiment was conducted under the same pressure and differential pressure as used in "2(a)", above.

(c) A third experiment using a membrane containing a catalyst layer with a Pd loading of 6 mg/cm2 (made from 20 wt% Pd on Vulcan XC-72 carbon using similar procedures as in "1(a)") produced 0.65 wt% hydrogen peroxide under the same conditions used in "2(b)", above.

(d) Another experiment was done using the membrane and conditions of "2(a)", above, except that the reactor was operated at a pressure of 150 psig (1034 kPa) for H2 and 100 psig (690 kPa) for O2 with a differential pressure of 50 psig (345 kPa). The concentration of peroxide was measured at 0.55 wt% on a continuous basis. The concentration reached that level over a period of 10 hours and maintained it at that level.

In each of the examples "2(a)" through "2(d)", above, the H2 diffusion through the membrane ranged from 7 to 30 standard cc /min for the 15 in2 (100 cm2) of area exposed to the gas supply channels. Analysis of the exiting gas on the oxygen side showed that 80% of the hydrogen diffusing through the membrane completely reacted to form water or hydrogen peroxide.

EXAMPLE 3 Polymeric Gas Flux Control Layer Preparation 1 2-dimethylcyclohexane (DMC) (98% purity, cis and trans isomers) was obtained from Aldrich Chemical (Milwaukee, WI). Methylene chloride (MeCI), HPLC-GC/MS grade, was obtained from Fisher Scientific (Pittsburgh, Ca). TBBA-PC was obtained from The Dow Chemical Company (Midland, Ml). Methylene chloride and the 1,2 dimethylcyclohexane were mixed to a ratio of 2.06 :1 w/w methylene chloride to dimethylcyclohexane. To make the casting solution, 14.8 wt% TBBA-PC was added to the MeCI/DMC solution.

A phosphoric acid fuel cell electrode (FCE) on TeflonTM-coated carbon paper was purchased from E-TEK Inc. ("Gas Diffusion Electrodes and Catalyst Materials", 1995 Catalogue, E-TEK Inc.

Natick, MA). Using 80 grit sandpaper, the non-catalytic side of the electrode was roughened. The electrode was placed in a fume hood with the catalyst side down. The surface of the electrode was saturated with methylene chloride to fill the pores of the macroporous carbon paper. Immediately after the surface liquid disappeared (evaporated), the casting solution was poured onto one end of the FCE. A 38 mil doctor blade was drawn down the length of the FCE uniformly covering the FCE with the casting solution. The solvents were allowed to evaporate leaving the gas flux control layer.

EXAMPLE 4 Control Membrane having Cu/Pd catalyst with no other promoter A fuel cell electrode utilizing a catalyst containing 1.2 wt% copper/20 wt% palladium was purchased from E-TEK, Inc. The metal loading of the membrane was 1.5 mg/cm2. A TBBA-PC gas flux control layer was applied to the electrode to form the membrane using the method of Example 3.

The membrane was loaded into the reactor. Pressure in the hydrogen chamber of the reactor was increased to 250 psi with hydrogen. The membrane was conditioned with hydrogen at 250 psi.

Nitrogen was substituted for hydrogen and the reactor returned to 100 psi. The reactor was brought to operating conditions at 200 psi of H2 in the hydrogen supply chamber and 190 psi in the oxygen supply chamber. A 10 psi differential pressure existed across the membrane. Water (Optima, Fisher Scientific, Fairlawn, NJ) was fed with the oxygen at 0.2 cc/min. Hydrogen flow was maintained at 0.74 standard liters per minute at 0°C, 1 atmosphere (hereinafter SLPM). Oxygen flow was held at 0.38 SLPM. Productivity (that is the combined water and hydrogen peroxide generated by the reaction) of the membrane was 0.33 Ib/ft2hr and the peroxide component of the product stream was 0.02wt%.

Control Membrane -Promoter Solution Fed With Oxygen-containina Feed Stream A phosphoric acid fuel cell electrode comprised of 20wt% palladium on Vulcan XC carbon with 0.5 mg/cm2 total metal loading was purchased from E-TEK Inc. The electrode was subsequently treated with a formalin solution to deposit copper onto the catalyst. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor. Hydrogen pressure was increased to 250 psi to condition the catalyst.

Hydrogen was fed to the reactor on the a 0.43 SLPM. Oxygen was fed to the oxygen contact side of the membrane at 0.65 SLPM. Hydrogen pressure was controlled at 100 psi. A 20 psi differential was maintained across the membrane. Temperature was maintained at 7 "C. A promoter solution containing 0.01 M H2SO4 and 0.001 M HBr was fed with the oxygen stream at a rate of 4 ml/minute. Under these conditions, the productivity of the reaction was 0.15 Ib/ft2hr and the hydrogen peroxide concentration of the product was 50 ppm.

Hydrogen pressure was increased to 310 psi. Differential pressure was increased to 25 psi.

Hydrogen was fed to the reactor on the a 0.52 SLPM. Oxygen was fed to the catalyst side of the membrane at 0.69 SLPM. Promoter flow rate was decreased to 0.5ml/minute. Temperature increased to 13 "C. Productivity increased to 0.36 Ib/ft2hr. The hydrogen peroxide concentration increased to 0.28wt%.

Control Membrane increased Acidity Promoter A phosphoric acid fuel cell electrode comprised of 20wt% palladium on Vulcan XC carbon with 0.5mg/cm2 total metal loading was purchased from E-TEK Inc. The electrode was subsequently treated with a formalin solution to deposit copper onto the catalyst. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor. Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi.

Hydrogen was fed to the reactor on the a 0.39 SLPM. Oxygen was fed to the oxygen contact side of the membrane at 0.28 SLPM. Hydrogen pressure was controlled at 100 psi. A 10 psi differential was maintained across the membrane. Temperature was maintained at 24 "C. A

promoter solution containing 0.05 M H2SO4 and 0.01 M HBr was fed with the oxygen stream at a rate of 0.1 ml/minute. Under these conditions, the productivity of the reaction was 0.11 Ib/ft2hr and the hydrogen peroxide concentration of the product was 0.23wt%.

Hydrogen pressure was increased to 150 psi. All other conditions were held constant.

Productivity increased to 0.14 Ib/ft2hr. The hydrogen peroxide concentration was essentially unchanged (0.22 wt%).

Promoter Doped Membrane -No Copper Treatment A phosphoric acid fuel cell electrode comprised of 20 wt% palladium on Vulcan XC carbon with 0.5mg/cm2 total metal loading was purchased from E-TEK Inc. Using a paint brush, a solution comprised of 100 ul 48wt% HBr, 450 ul of 98wt% H2SO4, and 40ml methanol was painted on both sides of the phosphoric acid fuel cell electrode. The methanol wet the TeflonTM-coated carbon paper and the catalyst layer carrying the promoter inside of the electrode. The methanol was allowed to evaporate leaving the promoter within the membrane. This doping procedure was repeated. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor. Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi.

Hydrogen was fed to the reactor at 0.5 SLPM. Oxygen was fed to the catalyst side of the membrane at 0.34 SLPM. Hydrogen pressure was controlled at 100 psi. A 10 psi differential was maintained across the membrane. A promoter solution containing 0.05 M H2SO4 and 0.01 M HBr was fed with the oxygen stream at a rate of 1 ml/minute. Under these conditions, the productivity of the reaction was 0.2 Ib/ft2hr and the hydrogen peroxide concentration of the product was 0.47wt%.

Pressure in the reactor was increased to 200 psi. Hydrogen flow was increased to 0.6 SLPM and oxygen flow was increased to 0.41 SLPM. All other conditions were held constant.

Productivity increased to 0.31 Ib/ft2hr. Hydrogen peroxide concentration was 0.48 wt%.

Pressure was further increased to 250 psi and differential pressure decreased to 5 psi. A 90 wt% hydrogen/10 wt% nitrogen feedstream was used to prevent flammable condition on the oxygen side of the membrane. Hydrogen flow was reduced to 0.45 SLPM. Temperature was reduced to 11 "C. All other variables were held constant. Productivity was 0.2 Ib/ft2hr and the concentration of hydrogen peroxide was 0.7 wt%.

Acid-Doped. Cooper Treated. polymeric gas flux control layer A phosphoric acid fuel cell electrode comprised of 20wt% palladium on Vulcan XC carbon with 0.5mg/cm2 total metal loading was purchased from E-TEK Inc. (Natick, Ma.). The electrode was subsequently treated with a formalin solution to deposit copper onto the catalyst. Using a paint brush, a solution comprised of 100 ul 48wt% HBr an 450 ul of H2SO4 and 40ml methanol was painted on both sides of the phosphoric acid fuel cell electrode. The methanol wet the TefionTM-coated

carbon paper and the catalyst layer carrying the promoter inside of the electrode. The methanol was allowed to evaporate leaving the promoter within the membrane. A single doping procedure was used. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor. Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi. Promoter concentration was 0.04 M H2SO4 and 0.004 M HBr.

The following data were collected: Productivity wt % H2O2 p(H2) (Ib/ft2hr) (psi) 0.1 2.3 100 0.12 3.7 150 0.16 7.4 250 0.17 8.1 275 0.23 12.7 400 Acid-Doped. Copper Treated. Polymeric Flux Control Layer-Sibunit Carbon A phosphoric acid fuel cell electrode comprised of 10 wt% palladium and 1.2 wt%Copper on Sibunit carbon with 0.5mg/cm2 total metal loading was purchased from E-TEK Inc. Using a paint brush, a solution comprised of 100 ul 48wt% HBr, 450 ul of 98wt% H2SO4, and 40ml methanol was painted on both sides of the phosphoric acid fuel cell electrode. The methanol wet the TeflonTM- coated carbon paper and the catalyst layer carrying the promoter inside of the electrode. The methanol was allowed to evaporate leaving the promoter within the membrane. A single doping procedure was used. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor.

Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi.

Hydrogen was fed to the reactor at 0.85 SLPM. Oxygen was fed to the catalyst side of the membrane at 0.50 SLPM. A 70 wt% hydrogen/30 wt% nitrogen feed was introduced to the hydrogen supply chamber and the pressure controlled at 400 psi. A 10 psi differential was maintained across the membrane. A promoter solution containing 0.05 M H2SO4 and 0.01 M HBr was fed with the oxygen stream at a rate of 0.3 ml/minute. Under these conditions, the productivity of the reaction was 0.27 Ib/ft2hr and the hydrogen peroxide concentration of the product was 5.0 wt%.

Acid-Doped. Copper Treated. polymeric gas flux control layer A phosphoric acid fuel cell electrode comprised of 20 wt% palladium and 1.2 wt% copper on Vulcan XC carbon with 0.5mg/cm2 total metal loading was purchased from E-TEK Inc. Using a paint brush, a solution comprised of 100 ul 48wt% HBr, 450 ul of 98wt% H2SO4, and 40ml methanol was

painted on both sides of the phosphoric acid fuel cell electrode. The methanol wet the TeflonTM- coated carbon paper and the catalyst layer carrying the promoter inside of the electrode. The methanol was allowed to evaporate leaving the promoter within the membrane. A single doping procedure was used. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor.

Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi.

Hydrogen was fed to the reactor at 0.85 SLPM. Oxygen was fed to the oxygen contact side of the membrane at 0.50 SLPM. Hydrogen pressure was controlled at 400 psi. A 65 psi differential was maintained across the membrane. A promoter solution containing 0.05 M H2SO4 and 0.01 M HBr was fed with the oxygen stream at a rate of 0.2 ml/minute. Under these conditions, the productivity of the reaction was 0.21 Ib/ft2hr and the hydrogen peroxide concentration of the product was 1.5 wt%.

Acid-Doped. Copper Treated. polymeric gas flux control layer A phosphoric acid fuel cell electrode comprised of 20 wt% palladium and 1.2 wt% copper on Vulcan XC carbon with 1.0 mg/cm2 total metal loading was purchased from E-TEK Inc. Using a paint brush, a solution comprised of 100 ul 48 wt% HBr, 450 ul of 98wt% H2SO4, and 40ml methanol was painted on both sides of the phosphoric acid fuel cell electrode. The methanol wet the TeflonTM- coated carbon paper and the catalyst layer carrying the promoter inside of the electrode. The methanol was allowed to evaporate leaving the promoter within the membrane. A single doping procedure was used. A TBBA-PC gas flux control layer was applied to the electrode using the method of Example 3. The resulting membrane was placed in a serpentine channel reactor.

Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi.

Hydrogen was fed to the reactor at 1.0 SLPM. Oxygen was fed to the catalyst side of the membrane at 0.60 SLPM. Hydrogen pressure was controlled at 310 psi. A 30 psi differential was maintained across the membrane. A promoter solution containing 0.05 M H2SO4 and 0.01 M HBr was fed with the oxygen stream at a rate of 0.1 ml/minute. Under these conditions, the productivity of the reaction was 0.82 Ib/ft2hr and the hydrogen peroxide concentration of the product was 1.6 wt %.

EXAMPLE 5 Further Teflon Modification of TeflonTM-Coated Carbon Paper An aqueous dispersion of PTFE, specifically DuPont's Teflon30BTM, was purchased from DuPont Fluoroproducts of Wilmington, DE. Teflon 30B is a negatively charged, hydrophobic colloid, containing approximately 60% [by total weight] of 0.05 to 0.5 mm PTFE resin particles suspended in water.

The TeflonTM-coated carbon paper was fully immersed in a bath of the Teflon 30B. The Teflon 30B bath containing the "coated" carbon paper was placed in a vacuum oven, and 29.5 in of vacuum applied. At intervals, the vacuum was released, whereupon the bubbles resulting from the vacuum treatment would collapse. This process was repeated 3 to 6 times until only a few bubbles occurred during the vacuum step. The carbon paper was then removed from the Teflon 30B bath.

This treated carbon paper was then dried in an oven at 90 to 1 200C for approximately 30 minutes.

The treated carbon paper was then subject to a devolatilzation heating step at 2880C for 30 minutes.

The final step was a sintering step where the treated carbon paper was heated at 3780 C for 30 minutes. The entire process of immersion in a Teflon 30B vacuum bath, drying, devolatilization, and sintering was repeated twice, but could have been repeated more times if desired. At the conclusion of this treatment, the modified carbon paper was 15.0 mils thick, and weighed <BR> <BR> <BR> 57.0 mg./cm2. The gas permeability of the modified paper was approximately 0.7 X 1 4 CC / cm2x sec. x cm. Hg. Hereinafter, this resulting Teflon modification of the TeflonTMcoated carbon paper shall be referred to as the "TeflonTM-modified carbon paper" (as opposed to the '7eflonTM-coated carbon paper").

The TeflonTM-modified carbon paper and catalyst (20 wt% palladium, 1.2 wt% copper on Vulcan XC-72) was supplied to E-TEK, Inc. A membrane was prepared by incorporating the catalyst onto one side of the TeflonTM-modified carbon paper such that the Teflon and carbon paper could serve as the gas flux control layer and the catalyst was on the oxygen contact side of the TeflonTM- modified carbon paper membrane. The membrane had a total metal loading of 0.6mg/cm2. Using a paint brush, a solution comprised of 100 ul 48wt% HBr, 450 ul of 98wt% H2SO4, and 40ml methanol was painted on the oxygen contact side of the membrane. The methanol wet the TeflonTM-modified carbon paper and the catalyst layer carrying the promoter inside of the membrane. The methanol was allowed to evaporate leaving the promoter within the membrane. The resulting membrane was placed in a serpentine channel reactor. Hydrogen pressure was increased to 250 psi to condition the catalyst. Nitrogen was substituted for hydrogen and the reactor returned to 100 psi.

Hydrogen was fed to the reactor at 0.39 SLPM. Oxygen was fed to the oxygen contact side of the membrane at 0.27 SLPM. Hydrogen pressure was controlled at 250 psi. A 5 psi differential was maintained across the membrane. A promoter solution containing 0.05 M H2SO4 and 0.01 M HBr was fed with the oxygen stream at a rate of 0.2 ml/minute. Under these conditions, the productivity of the reaction was 0.13 Ib/ft2hr and the hydrogen peroxide concentration of the product was 0.1 wt%.

Other embodiments of the invention will be apparent to the skilled in the art from a consideration of this specification or practice of the invention disclosed herein. It is intended that the specification and example be considered as exemplary only, with the true scope and spirit of the invention being indicated by the following claims.