NO20121234A1 - Integrated combustion reactor and methods for carrying out an endothermic reaction therein - Google Patents

Abstract

Integrated combustion reactors (ICRs) and methods of producing integrated combustion reactors are described, where combustion chambers (or channels) (12) are in direct heat contact with reaction chambers for an endothermic reaction (115). Very good results were obtained for combustion chambers which included a slot (18,19) for free flow through the chamber. Special reactor designs are also described. Methods for conducting reactions in integrated combustion reactors are described and results presented. Some of these methods are characterized by unexpected and very good results.

Description

Invent Ice area

The present invention relates to an integrated reactor for carrying out exothermic and endothermic reactions, methods for producing integrated reactors, and methods for carrying out reactions in integrated reactors.

Related patent applications

This application is a CIP application of US Patent Applications 09/375,614 and 09/640,903, which are incorporated herein as if reproduced in their entirety below. Pursuant to 35 U.S.C., Section 119 (e), this application claims priority from US Provisional Patent Application 60/269,628 of February 16, 2001.

Introduction

Currently, endothermic reactions carried out in micro-reactors are powered by utilizing heat from an external source, such as the effluent from an external combustion chamber. When this is done, the temperature in the gas flow that provides the heat is limited by requirements imposed by the construction materials. For example, a typical microreactor constructed of Inconel 625 could be limited in application for gas to temperatures of approx. 1050°C or lower. In practice, this means that the discharge from an external combustion chamber must be thinned with cooling gas (ie excess air) to bring the gas temperature down to meet material temperature requirements. This increases the total gas flow rate, which increases fan/compressor costs. Also, heating the gas stream causes external heat loss (associated with delivering the hot gas to the microreactor) and expensive high temperature materials to connect the combustion chamber to the microreactor.

On the other hand, an integrated combustion chamber can produce heat for the reaction in the immediate vicinity of the reaction zone, thereby reducing heat loss and increasing efficiency. Because traditional combustion catalysts are not stable at high temperatures (above about 1200°C) due to precious metal sintering, the integrated combustion chamber must remove heat at a rate sufficient to maintain local catalyst surface temperatures below this level or risk rapid catalyst deactivation.

Summary of the invention

In an integrated reactor, combustion/heat generation should take place in the immediate vicinity of the endothermic reaction. Preferably, an exothermic reaction takes place in microchannels which are interleaved with microchannels in which there is an endothermic reaction. Joint flow of endothermic and exothermic reaction streams is preferred, but crossflow or counterflow is also a possibility. The heat from an exothermic reaction is conducted from the exothermic reaction catalyst to the endothermic reaction catalyst, where it drives the endothermic reaction. This rapid heat removal from the combustion region makes it possible to use a very small fraction of excess air (for example near stoichiometric combustion, which would produce temperatures in excess of 1400°C in an adiabatic reaction). The use of a bypass catalyst design for one or both of the exothermic and endothermic microchannels can produce an advantageous capacity/pressure drop ratio. In a bypass catalyst design, gas flows in a 0.05-2mm gap up to a thin layer of engineered catalyst that is in contact with the microchannel wall. The catalyst can either be introduced into and up to the reactor wall or in one with the reactor wall. In the case of one with the reactor wall, a preferred method is wash coating of a catalyst on a wall or walls of the micro-channel. The catalyst may include the use of additional layers to increase surface area, such as a high surface area porous ceramic material, or layers to promote adhesion of a ceramic material to metal, such as amorphous titanium oxide deposited either by chemical vapor deposition (CVD) or from solution. The use of channels that have a minimum dimension of more than 2mm may be less efficient, as heat and mass transport limitations may be magnified. An integrated combustor can utilize the large surface area of microreactor channels to remove heat as it is generated and thereby keep the microreactor components from exceeding material temperature requirements when combusting with much less excess air (or thinner) than would be required for an external combustor.

According to one aspect, according to the invention, an integrated reactor has been produced which comprises: a first reaction chamber having a width of 2 mm or less, there being an open channel through the first reaction chamber having an internal volume which is made up of from 5 to 95% by volume of porous catalyst and from 5 to 95 vol% open space. The integrated reactor also comprises a second reaction chamber having a width of 2 mm or less, being an open channel through the second reaction chamber having an internal volume constituted by a catalyst and at least 5 vol% open space, and a reaction chamber wall separating the first chamber and the second chamber. This integrated reactor has heat flux characteristics of at least 1 W/cm<3>measured according to the heat flux measurement test.

The invention also relates to methods for carrying out exothermic and endothermic reactions in the reactor. An exothermic reaction substance is a chemical substance which, under selected conditions, will react with the generation of heat. Typically, a catalyst will catalyze the reaction.

In another aspect, according to the invention, an integrated reactor has been produced which comprises a first reaction chamber having a width of 2mm or less, there being an open channel through the first reaction chamber which has an internal volume made up of from 5 to 95 vol% porous catalyst and from 5 to 95 vol% open space, and a second reaction chamber having a width of 2mm or less, there being an open channel through the second reaction chamber having an internal volume constituted by a catalyst and at least 5 vol% open space. A reaction chamber wall separates the first chamber and the second chamber, and the integrated reactor has NOx performance characteristics of less than 100ppm as measured by the standard NOx measurement test.

According to the invention, a method for manufacturing an integrated reactor has also been developed, where the method is characterized by the fact that it comprises: provision of a single block of heat-conducting material, design of at least one first microchannel in the block, design of at least one second microchannel in the block, placement of at least one catalyst capable of catalyzing an exothermic reaction in said at least one first microchannel, as well as placement of at least one catalyst capable of catalyzing an endothermic reaction in said at least one second microchannel, whereby in an integrated reactor the first microchannel and the second microchannel are separated by less than 1 cm.

Furthermore, according to the invention, a method has been developed for carrying out an endothermic reaction in an integrated combustion reactor, where the method is characterized by an exothermic reacting substance being led into at least one exothermic reaction chamber which is equipped with at least one exothermic reaction chamber wall which is up to at least one exothermic reaction chamber , the reaction chamber containing an exothermic reaction catalyst in contact with at least said at least one exothermic reaction chamber wall located up to at least one endothermic reaction chamber, the exothermic reaction catalyst having an exposed surface inside the exothermic reaction chamber and the exothermic reaction catalyst's exposed surface and a second surface inside the exothermic reaction chamber, an open channel defines inside the exothermic reaction chamber, the gap having a thickness, in a direction perpendicular to the net flow, where the direction of net flow is the direction in which the gas will flow through the combustion chamber in operation, of 2mm or less, as the exothermic reaction substance reacts in the exothermic reaction chamber and generates heat, and that an endothermic reaction mixture is led into said at least one endothermic reaction chamber, and where the method has a volumetric heat flux of at least 1 W/cm <3>.

In another aspect, according to the invention, a method for carrying out an endothermic reaction in an integrated combustion chamber has been developed, where the method is characterized by the fact that it includes: that an endothermic reaction substance is led into at least one endothermic reaction chamber, that an exothermic reaction substance is led into at least one exothermic reaction chamber, where the exothermic reaction chamber is equipped with at least one exothermic reaction chamber wall located up to at least one endothermic reaction chamber, the endothermic reaction chamber containing an endothermic reaction catalyst in contact with said at least one endothermic reaction chamber wall located up to at least one exothermic reaction chamber, while the endothermic reaction catalyst comprises an exposed surface inside the endothermic reaction chamber and the endothermic reaction catalyst's exposed surface and a second surface inside the endothermic reaction chamber define a gap inside the endothermic reaction chamber, where the gap has a thickness, in a direction perpendicular to net flow, where the direction of net flow is the direction in which the gas will move through the endothermic chamber in operation, of 2 mm or less, and that the method is regulated so that the heat flux between said at least one exothermic chamber and said at least one endothermic reaction chamber is 1 W/cm<3> or more.

According to the invention, a method has also been developed for carrying out an endothermic reaction in an integrated combustion reactor, where the method is characterized by: that an endothermic reaction substance is led into at least one endothermic reaction chamber, that an exothermic reaction substance is led into at least one exothermic reaction chamber, where exothermic reaction chambers are equipped with at least one exothermic reaction chamber wall located up to at least one endothermic reaction chamber, where the endothermic reaction chamber contains an endothermic reaction catalyst in contact with at least said at least one endothermic reaction chamber wall located up to at least one exothermic reaction chamber, the endothermic reaction catalyst comprises an exposed surface inside the endothermic reaction chamber and the exposed surface of the endothermic reaction catalyst and a second surface inside the endothermic reaction chamber defines a gap inside the endothermic reaction chamber, the gap having a thickness, in a direction perpendicular to net flow, where the direction of net flow is the direction through which the gas will flow through the endothermic reaction chamber in use, of 2mm or less, whereby the exothermic reactant comprises air and a fuel, and where the exothermic reactant is converted to products and the products contain less than lOOppm N0X.

In another aspect (or in combination with one of the aspects described above), according to the invention, an integrated reactor has been produced which is characterized in that it comprises: at least one endothermic reaction chamber and/or at least one exothermic reaction chamber, where at least one reaction chamber comprises at least one porous catalyst material and at least one open channel, each of said at least one (exothermic or endothermic) reaction chambers having an internal volume bounded by reaction chamber walls in the height and width directions and by catalyst length in the longitudinal direction. The internal volume has the dimensions chamber height, chamber width and chamber length. At least one exothermic reaction chamber and at least one endothermic reaction chamber (located up to the exothermic reaction chamber) have a chamber height or chamber width of approx. 2mm or less. At a point where the chamber height or chamber width is approx. 2mm or less, the chamber height and chamber width form a cross-sectional area. The cross-sectional area in at least one reaction chamber consists of a porous catalyst material and an open area where the porous catalyst material occupies from 5 to 95% of the cross-sectional area and where the open area occupies from 5 to 95% of the cross-sectional area. The open area in the cross-sectional area occupies a continuous area of from 5 x 10~<8> to 1 x 10~<2>m<2>, and the porous catalyst material has a pore volume of from 5 to 98%, and more than 20% of the pore volume is made up of pores that have sizes from 0.1 to 300 µm.

The invention also relates to devices that have any of the unique constructional features or designs described here. For example, the invention comprises apparatus which comprises a fuel-air mixing manifold which is shown in fig. 19. The invention also relates to methods where any of the constructional features or designs are used or which are characterized by any of the properties or results described here.

Various embodiments of the invention can have such advantages as: low pressure drop, low need for excess air, high combustion stability, short contact time, low CO/NOx formation, operation with close to stoichiometric air supply, greater safety and high heat cycle stability. Operation at or near stoichiometric air supply reduces the total load on the system's air fan or air compressor, which will lead to significant cost savings.

A further advantage of lowering the combustion temperature (or the temperature of the exothermic reaction) necessary to drive the endothermic reaction is the use of alternative metals or alternative metallurgy, so that a lower cost of materials or a longer lifetime of the devices can be achieved.

Although combustion can have both homogeneous and heterogeneous contributions, catalytic combustion in a microchannel (or a channel with a smallest opening dimension smaller than the cooling diameter) will reduce the contribution from homogeneous reactions and favor heterogeneous (catalytic) combustion at the wall. This will also further increase safety by inhibiting the gas phase reactions which would otherwise take the combustion mixture well above the safe operating temperature limit for the material. Inhibition becomes stronger with decreasing minimum channel dimension and with increasing catalytic surface area on the channel walls.

In connection with other features of the invention, the use of a by-flow design where a reaction chamber has such a gap that gases can flow past (rather than through) a catalyst makes possible a significant improvement in performance compared to prior art. This improved performance was demonstrated in the integrated combustion reactor (ICR) tests at much higher heat flux (eg 29 W/cm3 on an area basis or 118 W/cm3 on a volumetric basis) than any reported in the literature at a minimal pressure drop ( for example < 0.3 bar) at a reactor length of 2.5 cm. Because the integrated combustion reactor can achieve such high heat fluxes without causing excessive pressure drops, endothermic reaction contact times are feasible, which are much shorter than those of flow-through catalytic devices or monolithic devices. Shorter contact times enable a higher productivity or throughput through unit volume of reactor.

Introduction of laterally distributed (across a channel) combustion fuel and air in joint flow with endothermic reactant flow concentrates the heat transport at the endothermic reactor inlet, where the concentration gradient (and therefore the reaction rate) is the highest, resulting in better results than for systems that distribute the combustion fuel over the total of the combustion catalyst surface is achieved. Although the distributed combustion examples still exhibit excellent heat flux compared to conventional steam reformers.

It is also recognized that the present invention will be able to use alternative exothermic reactions, such as oxidation reactions, including partial oxidation reactions, to drive an endothermic reaction.

Dictionary

Shims refer to generally planar plates or sheets which may be of arbitrary width and height and preferably have a thickness (smallest dimension) of 2mm or less and in some preferred embodiments between 50 and 500um.

"Unit Operation" means chemical reaction, evaporation, compression, chemical separation, distillation, condensation, heating or cooling. "Unit operation" does not simply mean mixing or fluid transport, although mixing and transport often occur together with unit operations.

A microchannel has at least one dimension of 2mm or less.

An "open channel" is a gap of at least 0.05mm that extends all the way through a reaction chamber so that gases can flow through the reaction chamber with. relatively low pressure drop.

In operation, a reactant enters a combustion or reaction chamber in a mass flow path flowing past and in contact with a "porous material" or "porous catalyst". A portion of the reactant molecularly diffuses transversely into the porous catalyst and reacts to form a product or products, and then the product(s) diffuse transversely into the mass flow path and out of the reactor.

The term "mass flow path" refers to open areas or open channels inside the reaction chamber. A continuous, mass flow area enables rapid gas flow through the reaction chamber without large pressure drops. In preferred embodiments, there is a laminar flow in that mass flow region. Mass flow regions inside each reaction chamber preferably have a cross-sectional area of from 5 x 10"<8> to 1 x 10"2m<2>, more preferably from 5 x 10"<7> to 1 x 10"<4>m<2>. The combined mass flow areas preferably comprise at least 5%, more preferably 30-80%, of either 1) the internal volume of the reaction chamber or 2) the cross-section of the reaction chamber.

"Equilibrium conversion" is defined in the classical way, where the greatest achievable conversion is a function of the reactor temperature, pressure and the composition of the added material. In hydrocarbon steam reforming reactions, the equilibrium conversion increases with increasing temperature and decreases with increasing pressure.

Brief description of the drawings

Fig. 1 shows an integrated reactor.

Fig. 2a and 2b show reaction chamber designs.

Fig. 3 and 4a show cross-sections of integrated reactors.

Fig. 4b shows an extended view of an integrated reactor. Figures 5 and 6 show integrated steam reforming reactors. Fig. 7 shows designs of disks for the production of a reactor. Fig. 8 shows a schematic view of a reactor used in the examples.

Fig. 9-12 shows data from the examples.

Fig. 13 shows a schematic view of a reactor used in the examples.

Fig. 14-16 shows data from the examples.

Fig. 17-20 show designs of reactors described in the examples.

Fig. 21-24 shows data from the examples.

Fig. 25 shows a design of a reactor described in the examples.

Fig. 26-27 show data from the examples.

Detailed description of the invention

An integrated reactor according to the present invention comprises a first reaction chamber containing a catalyst capable of catalyzing an exothermic reaction, and an adjacent second reaction chamber containing a catalyst capable of catalyzing an endothermic reaction. A reaction chamber wall separates the first and second reaction chambers.

A diagram of part of a preferred type of reactor apparatus is shown in fig. 1. An exothermic reaction chamber 12 contains exothermic reaction catalysts 14, 16 and an open channel 18. An endothermic reaction chamber 15 comprises an endothermic reaction catalyst 17 and an open channel 19.

In the present invention, the exothermic (and/or endothermic) reaction chamber preferably has a width (the smallest dimension perpendicular to flow) of 2mm or less, preferably 1mm or less and in some embodiments 0.5mm or less. The dimensions of a reaction chamber are the internal dimensions and include the catalyst, but do not include the chamber walls. A reaction chamber wall (separating the exothermic and endothermic reaction chambers) should be thermally conductive and preferably have a thickness (the distance between the reaction chambers) of 5mm or less, more preferably 2mm or less and in some embodiments a width of from 1 to 2mm. A short heat transfer distance is desired for good performance. It has been observed that these short heat transport distances combined with preferred reactor designs can produce surprisingly high volumetric productivity and low pressure drop.

As described in more detail below, preferred methods according to the invention can be described with the help of the designs of the apparatus and/or with the help of measurable characteristics, such as heat flux, volumetric productivity and/or pressure drop (which could also be described in connection with process conditions, such as flow rate, temperature etc.).

Fig. 1 shows reaction chambers that have the shape of parallelepipeds, but it should be understood that other shapes will also be possible, such as cylinders (for example adjacent cylinders or cylinders with an exothermic catalyst partially enclosed by an arch containing an endothermic reaction catalyst, or vice versa), or prisms (preferably close-packed prisms to reduce the heat transport distance and maximize the surface area for heat transport). Such shapes could, for example, be produced by drilling through a block or by laminating a stack of disks with shaped, flush openings, so that the stacked and connected disks form a passage which has borders formed by the edges of the openings. In order to increase surface area, in some embodiments the reaction chamber (either exothermic, endothermic or both) may have protrusions or a set of microchannels as shown in fig. 2a and 2b. Projections 22 or microchannel walls 24 can be coated with a catalyst so that a catalyst coating (not shown) is formed, such as by wash coating, chemical vapor coating, etc.

In another preferred embodiment (see Fig. 3), a reaction chamber wall 31 has fins 32. The fins can have any shape and can project partially or completely over the width of a reaction chamber. Preferably, one or more catalysts (not shown) are deposited over the reaction chamber walls to form exothermic or endothermic reaction chambers. In the device shown, flow is in operation into or out of the drawing plane. Preferably, the reaction chambers are stacked in alternating layers or exothermic 34 and endothermic 36 reaction chambers separated by a heat-conducting chamber wall 38.

Alternating layers of exothermic and endothermic reaction chambers is a general feature of the invention, and preferably there are at least two, more preferably at least five, layers of endothermic reaction chambers which alternate with at least one, more preferably at least four, layers of exothermic reaction chambers. Preferably, the apparatus and methods are designed so that the outer layers have half the reactant flow compared to inner layers of the same type. For example, in a device that has two exothermic reaction chambers interleaved between three endothermic reaction chambers, the outer two endothermic reaction chambers have half of the flow in the inner endothermic reaction chamber.

In another embodiment, see fig. 4a, is an integrated reaction chamber formed by adjacent reaction chambers where one reaction chamber contains a porous catalyst insert 42 (in fig. 4a the insert fills the entire reaction chamber so that the reactants will flow through the catalyst, a less preferred embodiment which can be used when relatively large pressure drops are acceptable or catalyst reaction -the kinetics do not support very fast reactions of tens of milliseconds). Reactor walls 43, 45 separate a chamber 42 from an adjacent reaction chamber 44, which has open channels which may contain a catalyst (not shown) placed in each channel. Fig. 4b shows an exploded view with the pieces that can be joined to form the connected device.

Fig. 5 shows an embodiment which, for example, could be for a steam reforming device in which hydrogen (or alternatively fuel) is burned to supply heat. One

reformate material flows through an open channel 52. Hydrogen (or alternative combustion fuel) flows through an open channel 54. Air (or other oxygen-containing substance) in channels 56 flows through openings 58 and into channel 54 where the oxygen reacts with hydrogen, mainly on a combustion catalyst 53. Preferably, all dimensions in the hydrogen channel are less than or close to the quench distance for hydrogen (or alternatively fuel, which may include hydrogen and other hydrocarbons) for the selected oxidant at the required temperature, although this may is not possible, especially for a hydrogen fuel. Combustion exhaust 55 preheats the oxygen-containing substance. Combustion heat is supplied to a reforming catalyst 57 through a reaction chamber wall 59. It is desirable that the dimensions of the open channel 54 are regulated to suppress flame formation. In this embodiment, the oxygen-containing substance is introduced into the front (ie near the hydrogen intake) of the reactor, and the combustion gas and the reformate stream are in a common flow arrangement. This design has two significant advantages: (1) both the exothermic and the endothermic reactor have the highest reaction of reactants and therefore the fastest reaction in the same area, and (2) the hydrogen is mixed with oxygen in the reaction chamber, thereby avoiding premixing and reaction outside the reaction chamber.

Fig. 6a shows a similar arrangement with the exception that the wall separating the hydrogen channel and the air or oxygen channel has several openings 62, which distribute oxygen through the hydrogen channel. Fig. 6b shows a view from above of plates that can be used to construct hydrogen channel(s) 61 and a separating plate 64 (any of the two plates can be used). In a variant of the process shown in fig. 6, hydrogen can be distributed in an air (or oxygen) chamber which is up to the endothermic reaction chamber.

Of course, it should be understood that in any of the devices described here, alternative reactants can be used instead of the reactants mentioned. For example, methane or other fuel could be used instead of hydrogen. Correspondingly, conditions can be varied, for example flow could be cross flow or counter flow. Although in some figures the apparatus is shown with only one exothermic reaction chamber, it should be understood that several alternating layers could be used, or two exothermic reaction chambers could form a sandwich construction with an endothermic reaction chamber.

A reaction chamber has the dimensions height, width and length. The height and/or width is preferably 2mm or less,

more preferably lmm or less (whereby the reaction chamber falls within the classic definition of a microchannel). The length of the reaction chamber is typically longer. Preferably, the length of the reaction chamber is over limb, more preferably i

the range from l-50cm. Typically, the sides of the reaction chamber are delimited by reaction chamber walls. These walls are preferably made of a hard material, such as a ceramic material, an iron-based alloy, such as steel, or monel, or high-temperature nickel-based superalloys, such as Inconel 625, Inconel 617 or Haynes 230. More preferably, the reaction chamber walls consist of stainless steel or Inconel, which is durable and has good thermal conductivity.

In addition to heat transport between adjacent reaction chambers, in some embodiments a reaction chamber may be in thermal contact with a microchannel heat exchanger. This combination of reaction chamber(s) and heat exchanger(s) can result in high heat transfer rates. Examples of and more detailed description of the use of microchannel heat exchangers are provided in US Patent Application 09/492,246 dated January 27, 2000, incorporated herein by reference. In some embodiments, the reaction chamber(s) and heat exchangers have a heat flux of at least 0.6 W per cm<3>reactor volume.

In some preferred embodiments, the reaction chamber has an inlet and an outlet with a continuous mass flow path from the inlet to the outlet. In these preferred embodiments, the pressure drop from inlet to outlet is preferably less than 20%, more preferably less than 10% of the system's inlet pressure. The pressure drop is preferably less than 350 kPa, and more preferably the pressure drop is less than 70 kPa. A low pressure drop is desirable to reduce the size and cost of other system equipment, such as pumps and compressors. In other, less preferred embodiments, the reaction chamber may comprise a section, such as a porous plug, which inhibits mass flow.

The integrated reactor works best with specific manifold chamber designs that 1) prevent

combustion reaction upstream of the microchannel catalyst and 2) distributes one of the combustion reactants evenly across the microchannel cross-section.

Preferably, the width of the mass flow path (or the open channel gap) inside the reaction chamber is less than or equal to 1mm, and the length (the direction of net flow) is preferably less than or equal to 50cm. The width of the porous catalyst can vary, but is preferably at least 20%, more preferably 50%, of the circumference of the mass flow path.

The present invention can also be used for reactions in the liquid phase. In liquid-phase reactions, the critical channel dimension will probably be smaller than the dimension for gas-phase reactions to accommodate the reduced mass diffusion rate that brings the reactants to the catalytic surface.

The "porous catalyst material" described here relates to a porous material having a pore volume of from 5 to 98%, more preferably from 30 to 95%, of the total volume of the porous material. At least 20% (more preferably at least 50%) of the material's pore volume consists of pores in the size range (diameter range) from 0.1 to 300 µm, more preferably from 0.3 to 200 µm and even more preferably from 1 to 100 µm. Pore volume and pore size distribution are measured by mercury pore symmetry (assuming cylindrical geometry for the pores) and nitrogen adsorption. As is known, mercury porosity symmetry and nitrogen absorption are complementary techniques where mercury porosity symmetry is more accurate for measuring large pore sizes (greater than 30nm) and nitrogen absorption is more accurate for small pores (less than 50nm). Pore sizes in the range from approx. 0.1 to 300um allows molecules to molecularly diffuse through the materials under most gas phase catalysis conditions. The porous material can itself be a catalyst, but preferably the porous material comprises a metal, ceramic or composite carrier on which one or more layers of the catalyst material have been deposited. The porosity may be geometrically regular as in a honeycomb or parallel pore structure, or the porosity may be geometrically tortuous or random. Preferably, the support is a foamed metal, a foamed ceramic fabric, a metal felt (ie mat-shaped, non-woven fibers) or a metal cloth.

Preferred active principal constituents of the catalysts include: elements in IUPAC groups IIA, IVA, VA, VIA, VHA, VIIIA, IB, IIB, IVB, the lanthanide series and the actinide series. The layers of catalyst, if present, are preferably also porous. The average pore size (volume average) of the catalyst layer(s) is preferably smaller than the average pore size of the support. The pore size in the catalyst layer(s) placed on the support is preferably in the range from 10~<9>m to 10~<7>m measured by N2 adsorption by the BET method. More preferably, at least 50 vol% of the total pore volume consists of pores in the size range from 10~<9>m to 10~<7>m in diameter. Diffusion inside these small pores in the catalyst layer(s) is typically Knudsen in nature for gas phase systems, whereby the molecules collide with the walls of the pores more often than with other gas phase molecules.

In preferred embodiments, the catalysts are in the form of inserts that can be conveniently introduced into and removed from a reaction chamber. Reaction chambers (either of the same type or of different types) can be combined in series with several types of catalysts. For example, reactants can be passed through a first reaction chamber containing a first type of catalyst, and the products from this chamber can be passed into a subsequent reaction chamber (or a subsequent step in the same reaction chamber) containing a second type of catalyst where the product (or more correctly termed, the intermediate product) is converted into a more desired product. If desired, further reactant(s) can be added to the subsequent reaction chamber.

The catalyst will also be able to be applied by other methods, such as wash coating. On metal surfaces, it is preferred to first apply a buffer layer by chemical vapor deposition, thermal oxidation, etc., which improves the adhesion of the subsequent washing coatings.

Preferred reactors and methods for carrying out reactions in integrated reactors can be characterized by their properties. Unless otherwise stated, these properties are measured by using the test conditions described in the example section. The invention can be characterized arbitrarily by the properties individually or in any combination. The average volumetric heat flux is preferably at least 1 W/cm<3>, or, in other preferred embodiments, at least 5 or 10 or 20 or 50 or 100 or approx. 120 W/cm<3>, and in some embodiments between 10 and approx. 120 W/cm<3>. The devices can be characterized by the low NOx output measured by the standard NOx test measurement described in the example section. NOx output is preferably less than 100ppm, more preferably less than 50ppm, even more preferably less than 10ppm and even more preferably less than 5ppm, and in some embodiments the NOx output is in the range of about 5-20 ppm. The method according to the invention comprises combustion where preferably less than 100% excess air (or equivalent excess oxygen) is used, more preferably less than 75%, even more preferably less than 50%, even more preferably less than 25%, or 10% or 5 % surplus air. For the characterization of devices, excess air is measured under the conditions specified in the heat flux measurement test (when characterizing together with NOx output) measured under the conditions specified in the standard NOx test measurement. Pressure drop through the exothermic and/or endothermic reaction chambers is preferably less than the following (in order of preference, based on the length of the reaction chamber) 295,000, 250,000, 125,000, 50,000, 25,000, 12,500, 2500 or 1500 Pa/cm. For devices, the pressure drop is measured under the conditions specified in the heat flux measurement test (when characterizing together with the NOx output) measured under the conditions specified during the standard NOx test measurement.

The devices can be made from such materials as plastic, metal, ceramic substances and composites, depending on the desired properties. Walls separating the device from the surroundings can be heat-insulating, but the walls up to the exothermic and endothermic reaction chambers should be heat-conducting.

The devices can be manufactured by forming the chambers inside a single block of material, by joining several components and by stacking binding discs.

A preferred integrated reactor body can be made from a single block of metal. Its channels will be able to be formed using an EDM string in the main body, and associated wiring will be able to be manufactured separately and welded on, increasing design flexibility.

EDM string is used to form slits or holes in a block of metal which are microchannels through which current passes and a unit operation takes place. Punching EDM, laser machining and, in some larger ducts, conventional milling, can also be used to fabricate ducts from a single block of metal.

The wafers provided with openings can be formed by processes including: conventional machining, string EDM, laser cutting, photochemical machining, electrochemical machining, casting, water jet, pressing, etching (for example, chemical, photochemical and plasma etching) as well as combinations thereof . For low cost, pressing is particularly desirable. The sheets can be joined by diffusion bonding methods, such as a frame press or a HIP chamber. They can also be joined using reactive metal bonding or other methods that form a surface seal. Alternatively, laser welding intermediate plates will be able to join the devices or plates so that seals are formed between flow paths. Devices will alternatively be able to be joined by using adhesives. In preferred embodiments, the devices are laminated in a single step, in less preferred embodiments, a first set of wafers is bonded together and then bonded to a second (or more) set of wafers. In some preferred embodiments, a set of wafers is bonded together in a single step, and then the resulting bonded article is cut into multiple devices.

Catalytic processes according to the present invention include: acetylation, addition reactions, alkylation, dealkylation, hydrodealkylation, reductive alkylation, amination, aromatization, arylation, autothermal reforming, carbonylation, decarbonylation, reductive carbonylation, carboxylation, reductive carboxylation, reductive coupling, condensation, cracking, hydrocracking , cyclization, cyclooligomerization, dehalogenation, dimerization, epoxidation, esterification, substitution, Fischer-Tropsch, halogenation, hydrohalogenation, homologation, hydration, dehydration, hydrogenation, dehydrogenation, hydrocarboxylation, hydroformylation, hydrogenolysis, hydrometallization, hydrosilation, hydrolysis , water treatment (HDS/HDN), isomerization, methylation, demethylation, metathesis, nitration, oxidation, partial oxidation, polymerization, reduction, reforming, reverse water gas shift, sulfonation, telomerization, transesterification, trimerization and water gas shift.

Another advantage of the present invention is that good yields (or other measures of good performance) can be achieved with short contact times. In preferred methods, the contact time is less than 100 milliseconds, more preferably less than 50ms and even more preferably between 1 and 25ms for gas phase reactions. Liquid phase reactions can be expected to be at least three orders of magnitude slower. Contact time can be lowered by reducing the diffusion distance between the mass flow and the porous catalyst with a simultaneous reduction in channel length. At these contact times, in a preferred embodiment of hydrocarbon steam reforming, at least 70%, more preferably at least 90%, of the equilibrium conversion of the hydrocarbon entering at the beginning of said at least one reaction chamber is converted to hydrogen, carbon monoxide and/or carbon dioxide. Similar improvements can be achieved in other processes.

Some process characteristics of some preferred processes according to the invention include the following: 1. Safely operates at a fuel:oxygen ratio close to stoichiometric (less than 100% excess air) for use in combustion as the exothermic reaction. This reduces the air requirement, which improves the overall system heating efficiency and reduces the necessary function of the external air fan or compressor. 2. Drives steam reforming at short contact times or vice versa at high gas space rates per hour. This is necessary to produce a compact device. 3. Works at a high heat flux. This is necessary for working with short contact times. 4. Works at a low pressure drop per unit length of reactor. This enables a higher productivity per volume unit. 5. Optionally: cools/inhibits gas phase reactions. When the channel dimension approaches the quench diameter ("quench diameter") or falls below, the contribution of undesired, homogeneous gas-phase combustion reaction is reduced.

In its broader aspects, the invention relates to any pair (or several, i.e. different substances can be driven through different reaction chambers containing different catalysts) of endothermic and exothermic reactions.

In steam reforming, gas space velocity per hour is preferably above 10,000, more preferably above 50,000 and can be in the range from approx. lOO.OOOh-1 to over 10<6>h_<1>, which corresponds to a contact time of the order of 100 to 1 millisecond. Operating pressure for methane steam reforming can be in the range from 1 atmosphere to 50 atmospheres. A range of from 1 to 30 atmospheres is preferred. The ratio between steam and carbon can be in the range from 1 to 10. A range of from 1 to 3 is preferred.

Various hydrocarbons can be reformed to produce hydrogen, including methane, ethane, propane, alkanes in general, alkenes, alcohols, ethers, ketones and the like, including mixtures such as gasoline, diesel, petroleum and others.

In addition, the invention can be used to intensify endothermic reactions other than steam reforming. As an example, the present invention can be used to intensify a dehydrogenation reaction by supplying heat via an integrated combustion reaction.

Examples

Preferred catalysts for use in the apparatus described in the examples were prepared by the following methods: The catalyst in the reformer channels contained a catalyst of 13.8%-Rh/6%-MgO/aL 2 O 3 on a metal felt of FeCrAlY alloy from Technetics, Deland, Florida. The reforming catalyst was prepared using a wash coating technique based on FeCrAlY felt with approx. 0.25mm thickness and 90% porosity. Before wash coating, the metal felt was pretreated by rapid heating to 900°C in air for two hours. To increase the adhesion between the metal upper surface and the catalyst, a dense and pinprick-free interface layer was first coated on the oxidized FeCrAlY felt by metal-organic chemical vapor deposition (MOCVD). This interface layer can be Al 2 O 3 , Al 2 O 3 +SiO 2 , or TiO 2 , etc. For example, when TiO 2 was coated, titanium isopropoxide (Strem Chemical, Newburyport, MA) was vapor deposited at a temperature of from 250 to 900°C at a pressure of from 0.1 to 100 torr. Titanium oxide coatings with excellent adhesion to the foam were obtained at a deposition temperature of 600°C and a reactor pressure of 3 torr. This layer not only increases the adhesion between the metal felt and the catalyst, it also protects the FeCrAlY from corrosion during the steam reforming reaction. 13.8wt%Rh6wt%MgO/Al2O3 powdered catalyst was prepared by 1) calcination of a high surface area gamma alumina at 500°C for 5 h, 2) impregnation of the gamma alumina with MgO using the incipient wetting method with an aqueous solution of magnesium nitrate obtaining a MgO-modified gamma alumina support, 3) drying of the modified support at 110°C for 4 hours followed by 4) a second calcination at 900°C for 2 hours, 5) impregnation of the modified support with Rh 2 O 3 by the initial moisture method from a rhodium nitrate solution, 6) followed by a final drying at 110°C for 4 hours and 7) final calcinations at 500°C for 3 hours to obtain a powder of the supported catalyst. Catalyst coating slurry was prepared by mixing the above powdered catalyst with deionized water in a ratio of 1:6. The mixture was milled in a ball mill for 24 hours to obtain a coating slurry containing catalyst particles smaller than lum. The heat treated and CVD coated felt was wash coated by dipping the felt in catalyst slurry. The wash coating process was repeated to achieve the desired weight gain. Between each coating, the felt coated with catalyst was dried in an oven at 100°C for 1 hour. The coating process was repeated to achieve the desired coating thickness or catalyst loading. After the final coating step, the catalyst was dried overnight in an oven at 100°C and calcined by slow heating in air at a rate of 2°C/min to a temperature in the range of 300 to 500°C. the amount of catalyst coated was measured to be 0.1 gram of catalyst per 6.5cm<2>felt. Prior to steam reforming testing, the engineered catalyst felt was subjected to an activation treatment, preferably reduction, at 300-400°C.

The integrated combustion catalyst may be a wash-coated catalyst that is placed directly on the inner Inconel walls of the ICR device. The Inconel surface is first cleaned, with ultrasound if possible, in hexane, nitric acid (20%) and acetone (or propanol). Preferably, the cleaning solutions are allowed to flow over the Inconel rafts. A natural chromium oxide layer is then formed on the Inconel surface by heating in air (flow if possible) at 3.5°C/min to 500°C and held at 500°C for 2 hours. The temperature is then increased by 3.5°C/min to 950°C and held at 950°C for 2 hours. The Inconel is then allowed to cool to room temperature at a rate not faster than 5°C/min. The active palladium component is then applied to the chromium oxide layer by immersing the required deposition area in a 10% by weight solution of palladium nitrate. This is achieved either by static immersion or by pumping the fluid into a device to a required liquid level. The solution is then allowed to contact the deposit surface for 2 minutes. The solution is then removed from contact with the Inconel surface, and the amount of palladium that remains is calculated using a differential measurement. During channel coating, nitrogen flows through the channel to ensure that no clogging occurs. The catalyst is then dried at 100°C for 1 hour, if possible in a vacuum. The catalyst is then calcined by heating at 3.5°C/min to 800°C and held at 850°C for 1 hour. The catalyst is then allowed to cool to room temperature at a rate not greater than 5°C/min.

For some examples, a felt form of the combustion catalyst was fabricated and then introduced into the combustion microchannel(s).

Engineered combustion catalyst was also produced based on the FeCrAlY felt from Technetics. Similar to the preparation of engineered steam reforming catalysts, the FeCrAlY felt support was first fired at 900°C in air for 2 h in a muffle furnace (rate of rise = 200°C/min). After the burning process, the felt was cooled to room temperature. It was then dip coated in a colloidal Al 2 O 3 solution (PQ corporation) containing gamma Al 2 O 3 particles of um size. This step was performed by immersing the felt in the solution and then removing excess solution on the felt on an absorbent sheet followed by drying in vacuum at 110°C overnight. The sample was heated to 500°C for a period of 3 hours before the addition of Pd. Pd was then added by soaking the support, which now contained an Al2O3 layer, in a 20% by weight Pd (Nett solution (Engelhard). When removing excess Pd(N03)2 solution, the sample was dried in vacuum at 110°C for at least 4 hours. Final calcination was performed by heating at 2°C/min to 350°C and holding isoternally at this temperature for 3 hours. The prepared Pd/Al2O3 catalyst has a nominal loading of 47 wt% Pd over Al2O3 and 0.126 grams of catalyst /g FeCrAlY.

Example 1

Integrated Combustion Reactor (ICR).

This integrated catalytic combustion reactor consisted of

a single methane steam reforming channel that shared a wall with a single catalytic combustion channel. Heat was transported through this common wall from the hot (combustion) side to the cold (reforming reaction) side to drive the endothermic reaction. One design (version 1) of the integrated combustion reactor had no branching and collecting chamber for the combustion flow, but instead introduced the unmixed combustion gases directly onto the catalyst. This was done to ensure that homogeneous combustion did not occur upstream of the catalyst (ie in the manifold). A second design of the integrated combustor (version 2) was produced which included 6.4mm x 10.2mm chambers on the combustor side, which were similar to those on the reformer side (see Fig. 8). This was done to reduce the probability of channel formation on the combustion chamber side and distribute the flow more evenly (and thereby the heat of reaction) over the entire width of the reactor and thereby extend the life of the catalyst and increase conversion in the reformer. The integrated combustion reactor's overall dimensions were 5.8mm x 12.7mm x 43.2mm, and each catalyst's dimensions were 2.5mm x 12.7mm x 25.4mm. For the purposes of heat flux calculations for this example only, 10 mm catalyst width was included when the remaining catalyst volume was sealed to reactant flow. The channel catalysts in both the reformer and the combustion chamber worked with flow of reactants through the catalyst (and therefore very high pressure

fall), which had no bypass gap after assembly. The test conditions that were used are shown in table 1. Combustion fuel and air were not preheated, and SMR reactants were preheated to approx. 600-650°C. Although outlet pressures are not indicated, they are close to ambient pressure (approx. 1 bar absolute).

The integrated combustion reactors (version 1 and version 2) were fabricated from a 0.25mm heat transfer disc which was welded along the circumference between two 2.9mm x 17.8mm x 43.2mm plates. One plate held the combustion catalyst against the heat transport disc while the other plate held the SMR catalyst against the opposite side of the disc. Each plate provided access for reactants (combustion or SMR) to a cavity (machined into the inner surface) through an inconel tube with an outer diameter of 3.2mm and an inner diameter of 1.8mm at one end and discharged products through a corresponding tube at the other . With the exception of the combustion plate in version 1, a flow path with a length of 6.4 mm between the catalyst and the inlet or outlet pipe was included so that the collection areas at the leading edge and trailing edge would contribute to distributing the flow over the cross-section. Each catalyst was held in intimate thermal contact with the heat transfer disk by means of 1.3 mm x 25.4 mm rails machined to the side of the flow path. Fuel was introduced into the device through a 1.6mm tube (1.1mm internal diameter) which was fixed concentrically inside the air inlet tube, so that air and fuel were not allowed to mix with each other until the end of the 1.6mm the tube (approx. 1-2mm from the heat transfer disc).

The heat transport area between the two catalysts was 10.2mm x 25.4mm. The reactor core volume (which was used to calculate the average volumetric heat flux) was assumed to be the heating area times the sum of heat transport disc thickness (0.25mm), SMR channel thickness (0.25mm) and combustion channel thickness (0.25mm).

After two heating cycles and six hours of operation in an application of the first design, the flow pattern was changed from cocurrent to countercurrent in an attempt to further improve heat transport to the steam reformer. Only hydrogen/air mixtures were used in the operation of the first design to eliminate the possibility of coking in the combustion chamber.

Version 2 was tested continuously (in a common current form) burning a hydrogen/air mixture and for four hours burning a methane/air mixture. During operation of the version 2 integrated combustion reactor, the fuel equivalence ratio was not allowed to exceed 0.7, thereby maintaining maximum adiabatic flame temperatures for H.2/air mixtures of 1740°C (as opposed to about 2110°C at stoichiometric) and 1565° C for CH4/air mixtures.

CR = combustion reaction

SR = steam reforming of methane

Results

Version 1

A hydrogen conversion of 99.9% was achieved in the integrated combustion reactor with 5% excess air (fuel equivalence ratio of 0.95), giving a peak combustion chamber outlet temperature of 1050°C or lower. In contrast, an external hydrogen combustor would require 186% excess air (fuel equivalence ratio of 0.35) to maintain the adiabatic combustion product temperature at 1050°C or lower. A typical pressure drop through the combustion channel was approx. 275.8 KPa at 1.75ms contact time. Pressure drop in the integrated combustor microreactor was much higher on the combustion side due primarily to much higher flow rates (much shorter contact times) and secondarily due to the way the air was supplied

(through a very narrow annulus).

The ICR steam reformer (at a steam to carbon ratio of about 3) was able to extract heat from the integrated combustor in the co-flow design, maintain steam reforming temperatures of 780-800°C and methane conversions as high as 75% at contact times in the reformer of 20-60ms (see fig. 9-10, contact times calculated based on total channel volume including catalyst and the total gas flow rate at standard temperature-pressure).

It was found that in this device, for total channel contact times in the combustion chamber of less than 3.5ms (5% excess air) the hydrogen combustion zone appeared to extend beyond the outlet of the catalytic combustion region. A decreasing combustor outlet gas temperature and an increasing gas temperature immediately downstream of the outlet with increasing combustor flow rate were shown therein. High fuel flow rates in the combustor (ie, less than 5ms contact times) were necessary to achieve reformer temperatures of 750°C and higher. This is probably due to the channeling of the combustion flow down the middle 30% of the combustion catalyst. If channeling and catalyst deactivation could have been eliminated, it is believed that hydrogen/air contact times of less than 1.2ms could have been achieved. Although exit temperatures from the combustion chamber were well below adiabatic flame temperatures, examination of the combustion catalyst showed evidence of precious metal sintering or vaporization and redeposition. This suggests that there was an area of unacceptably high temperature in the combustion chamber (above 1200°C).

Version 2 of combustion chamber (CR)

The inclusion of a manifold in the combustor of version 2 of the integrated combustor greatly improved the efficiency of operation of both the combustor and the reformer. Combustion chamber test results for version 2 of the integrated combustion reactor are shown in Table 2. The highest pressure drop across the combustion chamber side of version 2 of the integrated combustion reactor was 5.4 bar (78 psid) for ClU/air at 2.8ms total channel contact time .

During operation of version 2 of the device, two combustion chamber operating regimes were observed. In the first regime, the highest temperature observed was on the steam reforming side of the manifold, directly opposite the combustion chamber inlet (typically over 100°C above the highest bulk temperature). This is a strong sign that in this regime there is a homogeneous combustion zone in the branching chamber of the combustion chamber. In the second combustion regime, the highest temperature in the reformer is the main part temperature (in the thermocouple well in the middle of the side plate, sunk 1.78mm deep into the plate wall) rather than in the branching chamber. In addition, in the second regime the temperature in the combustion chamber inlet drops below the self-ignition temperature for the fuel/air mixture. This is a sign that regime 2 does not involve combustion in the manifold. In other words, in regime 2, a flame will not attach stably to the fuel inlet pipe in the manifold. Table 2 describes the conditions under which each regime was observed when operating version 2. The transition between regime 1/regime 2 (identified by a sudden, discontinuous change in the reformer's branching chamber temperature) was observed twice during testing, each time approx. 3 to 4 minutes after a change was made to the process conditions.

In this device, under the conditions shown in Table 2, hydrogen combustion works only in regime 1, probably due to the unusually high flame speed of hydrogen/air mixtures (an order of magnitude higher than methane/air mixtures). On the other hand, methane combustion was observed in both regimes, where the transition occurred somewhere between approx. 600 and lOOOscm<3>total reactant flow. Both combustion fuels/regimes produced similar results

(expressed in steam reformer performance) at identical total flows and fuel equivalence ratios, although reformer temperatures were much more uniform (across flow path 2) for regime 2. No reduction in combustor fuel formation was observed during 7 hours of operation.

Version 2, steam reformer (SR)

Due to the ability of the Version 2 combustor to function more efficiently and smoothly, the steam reformer was able to achieve much higher conversions than the Version 1 reformer. The results for the Version 2 integrated combustion reactor are shown in Fig.

11-12. As expected, conversion increases with increasing temperature (see fig. 11), but the temperature inside the reformer is difficult to find out in such a small device, as the temperatures in headers and in the main part are probably much lower than the inlet gas temperature. Conversion of methane in the steam reformer decreased with decreasing contact time between 40 and 20ms (see fig. 12), although the change is exaggerated as the temperature also fell with increasing flow (by at least 25°C). Selectivity towards CO increases significantly with increasing temperature (fig. 11), but not with increasing contact time (fig. 12). The highest steam reforming conversions were at the highest condition (65.4 W thermal methane burned in 43% excess air), where 98.6% methane conversion and 66% selectivity to CO were achieved at a feed pressure of approx. 3 40 KPa (48 psig), steam:carbon of 3:1 and contact time (in steam reformer) of 38.Oms. After 1.5 hours of operation at the highest combustor temperature, conversion dropped to 98%, a sign of weak deactivation. The lowest methane conversion by steam reforming was 93.6% at a contact time in the steam reforming device of 19.00 rpm. The steam reformer was on-line for a total of 6.5 hours. The carbon balance ended within +- 4% for all exceptions except one.

CONCLUSIONS

The following findings are significant in example 1:

1. Integrated hydrogen/air combustion (bypass design) looks very promising for well-mixed, evenly distributed flows, little excess air and total contact times as low as 3.5ms if a header is included in the design. 100% H2 conversion was achieved under these combustion chamber (CR) conditions. Integrated methane/air combustion (by-flow design) looks extremely promising for well-mixed, evenly distributed CR streams with little excess air and total contact times as low as 2.7ms if a header is included in the design. An incredible CH4 conversion of 99.6% with 100% selectivity for CO2 was achieved under these combustion chamber conditions without air preheating. Such performance would have required at least five times longer contact time (and about 400°C air preheating) for methane combustion in an external catalytic combustion chamber. 2. An integrated catalytic combustion channel can generate the heat necessary to sustain an endothermic catalytic steam reforming reaction in an adjacent channel of the same size under the "worst case" heat loss conditions present in the Version 2 ICR device. Steam reforming methane conversions as high as 98.6% with 66% selectivity to CO was observed. 3. Inadequate distribution of the combustion reactants across the cross-section of the combustion catalyst (ie channel formation) results in greatly reduced conversions in both the combustion chamber and in the reformer. 4. The design of these ICR devices (versions 1 and 2) caused an unexpectedly large pressure drop in both the reformer (about 3.3 bar or 48 psid) and in the combustor (about 5.4 bar or 78 psid) when operating at temperature of approx. 850-900°C at an SR contact time of 4 Oms.

Example 2

This example describes the design, manufacture and test results from a highly efficient, small microchannel reactor with high throughput, where heat-producing (exothermic) and heat-consuming (endothermic) reaction channels are immediately adjacent to each other (integrated). Combustion of hydrogen in air was used as the exothermic reaction, while steam reforming of methane at a ratio between steam and carbon of 3:1 was used as the endothermic reaction. A new combustion reactor integral design (bypass flow) was used, which enabled much higher throughput at minimal (ie less than about 0.8kg/cm<2>) pressure drop by allowing each reactant stream to flow in a narrow slot ( 0.125mm) up to the porous engineered catalyst. The new design included a central combustion zone (of two microchannels) flanked by a reformer channel on either side.

In the integrated combustion reactor (see Fig. 13) a shortened combustion catalyst layer (approx. 0.5 cm) was used to enable a combination of catalytic and homogeneous hydrogen combustion. Hydrogen was only distributed across the channel width and not across the length of the catalyst's nozzle. The entire hydrogen fuel flow was fed into the combustion channel through eight approx. 0.02cm holes 0.08cm upstream of the catalyst bed.

The dimensions of the catalysts in the steam reformer were 0.25mm x 1.24cm x 2.54cm, and the combustion catalysts were 0.25mm x 1.27cm x 0.51cm, although only 1.02cm of the width was exposed to reactant flow, being 0.12cm on edge was used to hold the catalyst in place on each side. For the purposes of heat flux calculations in this example, only approx. 1.0 cm catalyst width included, with the rest of the catalyst volume closed for reactant flow. A thickness of 0.30mm was allowed for each catalyst with bypass channel thicknesses of approx. 0.13mm and approx. 0.43mm in the channels in the steam reformer and the combustion reactor. The integrated combustion reactor was operated with common flow. The experimental setup used for testing was the same as that used in example 1. The integrated combustion reactor was tested down both a stoichiometric amount of hydrogen in air and with 40% excess air in the combustion reactor channels while methane steam reforming reaction was carried out in the channels of the steam reforming device.

Main part temperatures were measured inside approx. 0.16 cm deep heat wells at 3 positions across the length of the integrated combustor at the rear of each reformer channel at positions corresponding to the top, middle, and bottom of the catalyst bed in the steam reformer. The combustor was tested for a range of contact times and air equivalence ratios using hydrogen in air supplied by Brooks mass flow controllers (MFCs). The reformer was tested over a wide range of reformer contact times at an average temperature of approx. 750-800°C. Methane was supplied by a Brooks mass flow controller, and water was supplied by an HPLC or a syringe pump. Feed compositions were determined both for portions from the combustor and from the reformer using an MTI gas chromatograph. All tests were performed at ambient pressure using a 3:1 molar ratio of steam to carbon.

A summary of the results of the tests carried out is shown in Table 3. The tests show a large drop (36-80°C) in temperatures in the steam reformer along the length of the reactor, suggesting that the main part of the combustion takes place at the inlet to the combustion chamber (in cataly -sator zone). It is worth noting that the ideal area for supplying heat to the reforming reaction is at the catalyst inlet where the SMR reactant concentration gradient (and therefore the reaction rate) is highest. The co-flow design in the channels of the integrated combustion reactor and steam methane reforming device coupled with the fast kinetics of hydrogen combustion delivers the bulk of the heat right at the catalyst bed inlet and thereby maximizes the speed of the steam methane reforming reaction. c> Thermocouples are placed down the length of the reactor with the highest temperatures near the inlet for the common reforming and combustion stream.

The device showed differences of over 50°C in steam reformer temperatures at the same locations on each channel in the reformer, suggesting that the fuel or air is not evenly distributed between the two combustion chamber half-channels. The integrated combustion reactor transported up to 54.8% of the heat generated by the combustion into the endothermic reaction (it is estimated that about 40% of the energy generated was used to heat the combustion products to 740°C). When operating at peak capacity, heat losses in the device were thus as low as 10% of the energy available at 740°C, which showed an incredibly high efficiency for such a small device. The calculated largest average heat flux was 13.2 W/cm<2>. This flux is higher than anything reported in the literature for driving an endothermic reaction. It follows that the integrated combustion reactor bypass concept can be used to produce high throughput reactors, which are much more compact and better behaved than devices using monolithic or flow-through catalyst designs.

Results from tests with the second type of integrated combustion reactor (ICR) show much higher efficiency than with an external combustion chamber (as high as 56.7% of combustion fuel energy was transferred to the steam reforming reaction of the highest possible 60.4% that can be delivered at 740°C at 92% combustion fuel formation and 47% excess air). The heat flux was 13 W/cm<2>, which is higher than any reported in the literature. For the conditions where this heat flux occurred, the pressure drops measured across the combustion chamber and reformer were 0.46 bar (6.7 psid) and 0.76 bar (11.0 psid) respectively. These results confirm that the bypass catalyst design enables catalytic combustion heat transport rates and efficiencies far in excess of those of existing ICR-like devices. In contrast, the flow-through device (Example 1) incurred pressure drops across the combustion chamber and reformer of 5.3 bar (77 psid) and 5.0 bar (72 psid) respectively at a heat flux of 3.0 W/cm<2>.

Results that are typical of the reformer's behavior are shown in fig. 14. Values of equilibrium selectivity towards CO calculated for measured reactor temperatures (assumed to be equal to the average of all six main part temperatures) were in good agreement with measured selectivity values. From the data in fig. 14, it is clear that the integrated combustion reactor is capable of achieving equilibrium conversion for contact times in the reformer of 15ms and near equilibrium conversion at contact times of 7.5ms.

Fig. 15-16 supports the hypothesis that either the fuel or the air was not evenly distributed between the two half-channels. The data in fig. 15 shows that for the range of contact times that were tested, no increase in conversion is seen when contact time is shortened, which suggests that the limitation of conversion is not due to catalytic activity.

A strong trend in hydrogen conversion is seen with decreasing excess air, suggesting that below 150% excess air, one of the combustion half-channels may be operated fuel-rich, limiting the highest achievable total conversion.

The integrated combustion reactor was operated for a total of 21 hours over 6 days with no significant deactivation observed for either the catalyst in the combustor or in the reformer. During testing, the measured main part temperatures never exceeded 860°C and were typically 750-800°C.

CONCLUSIONS

1. An integrated combustion reactor (ICR) was designed, fabricated and proved capable of achieving a heat transfer rate of 13.2 W/cm<2> with pressure drops of less than 0.76 bar and 0.46 bar in the channels in the reformer and the combustion chamber, respectively. 2. Equilibrium conversion and selectivity were achieved in this integrated combustion reactor for total channel reformer contact times as low as 15ms, even though the combustion catalyst bed extended only 1/5 of the length of the catalyst bed in the reformer. 3. After 21 hours of operation at 750-800°C, a single channel integrated combustion reactor showed no significant signs of catalyst deactivation, although the combustor was normally operated with only 41% excess air (with some occurrences of less than 3% excess air). Due to the unusually high capacity of this "single channel" integrated combustion reactor, heat losses as low as 10% of the total available energy at 740°C were achievable.

Example 3

This example describes the design and testing of a microchannel reactor with several interleaved channels, where heat-producing (exothermic) and heat-consuming (endothermic) reaction channels are interleaved (integrated) with each other. Combustion of hydrogen in air was used as the exothermic reaction, while steam reforming of methane at a ratio between steam and carbon of 2:1 was used as the endothermic reaction. The pressure for both reactions was close to atmospheric pressure and operated at a pressure necessary to overcome the pressure drop in the system. For the steam methane reforming side, the typical operating pressure was 139 kPa, and for the combustion side, the typical operating pressure was also 137 kPa. A bypass design was used, which enabled much higher throughput at minimal (less than 0.2 bar) pressure drop by allowing the reactant stream in each channel (either exothermic or endothermic) to flow in a narrow (approx. 0.2mm) gap between two layers of porous, engineered catalysts, which are in intimate thermal contact with the solid channel wall or heat transfer surface. The reactants mostly diffuse from the gap into the porous catalyst. The reactants then continue to diffuse and react within the porous catalyst located up to the heat transfer surface. The design included nine combustion channels interleaved between ten reformer channels in a common flow arrangement. The test results from the integrated combustion reactor (ICR) show that the concept of the integrated combustion reactor can be scaled up for use in a multi-channel device.

A main body for an integrated combustion reactor was fabricated from Inconel 625 using wire EDM to form grooves in a block of solid metal. The device was made from a 5.33cm long, 5.23cm high and 2.54cm wide block. Nine combustion channels were machined 5.3cm through the length of the block by string EDM, with each channel having a total width of 1.78cm and 0.081cm height. Each combustion channel was centered in the block's width of 2.5 cm. Interleaved between the nine combustor channels were eight steam methane reformer channels, and these interleaved channels were joined by two smaller SMR channels which provided heat sinks on both sides of each combustor channel. All steam methane reformer channels in the block were 4.57 cm in length, whereby the channel began 0.25 cm from the side where the combustion reactants enter and ended 0.51 cm from the exit of the combustion channels. The eight steam methane reformer ducts had heights of 0.81 mm while two smaller, converging SMR ducts had heights of 0.43 mm. The separation between all the channels in the steam-methane reformer and the combustor was maintained using 0.15 cm thick heat transfer mesh, and the interleaved channels were centered within the device height of 5.23 cm. The general channel orientation and the interfoliated nature of the channels are both shown in fig. 17.

Fluid access to the SMR channels is through the sides where the channels are machined through a direction perpendicular to the combustion channels. In order for the SMR flow in each channel to move simultaneously with the combustion channels, two 4.0 cm x 0.38 cm side pieces were produced for each SMR channel with heights that were adapted to the heights in the SMR channel (press fit). For each SMR channel, each of these side pieces was placed 0.38cm into the channel and positioned so that a 0.51cm opening was left upstream of the catalyst for reactant gas to pass through from the manifold, while another was placed on the opposite side and left a corresponding opening downstream of the catalyst for gases to leave the collection chamber. These openings start 0.25cm from the combustion inlet side of the block. The bypassed SMR catalysts are then introduced into the SMR channels, of length 3.57 cm, width 1.78 cm and thickness 0.03 cm. Two SMR catalysts enter each of the eight central channels up against both heat transfer surfaces, resulting in a gap of 0.2mm. A single SMR catalyst enters the two mating SMR channels, positioned against the heat transfer surface it shares with the combustion channel, resulting in a gap of 0.12mm. The catalysts are centered inside the 4.57cm long and 2.54cm wide SMR channel.

In order to maintain the above-described bypass gap and keep the catalysts against the channel wall, two 3.57 cm long, 0.13 cm wide strips of "Inconel" 625 were placed between the 2 catalysts, whereby the thickness of the 2 strips was equal to the gap height . These 2 strips run with their lengths in accordance with the catalyst lengths, and were placed on the edges of the channels that the 2 catalysts form. The final step involves placing the second side piece 0.38 cm into the width of each SMR channel on the opposite side of the first piece. That arrangement forms a 0.51 cm long opening for the SMR products to exit the channel, and this opening is 0.51 cm in length where the edge from which the combustion products exit the device. The overall effect of the addition of the 2 side pieces is to build up a small z-shaped manifold for flow through the channel formed by the SMR catalyst gap. A drawing showing the shape of the flow path for the SMR channels can be found in Fig. 18.

The SMR manifold that distributes the SMR streams from a single inlet tube to the 10 channels is fabricated from a single block of material, 5.08 cm long, 4.71 cm high and 0.95 wide. In this block, individual, circular cylinder channels lead from the openings mentioned in the above paragraph to a half-tube welded over the opening. The SMR branch pipe which takes the outlets from each stream into a single outlet pipe has a similar design with its connecting cylinders adapted to the outlet channels.

Each of the 9 combustion channels had 2 combustion catalysts of the same dimensions as the SMR catalysts and used the same dimensions for the strips used for separation in the SMR channels. The combustion catalysts are pushed into the length of each channel, 0.76 cm from the side where the combustion reactants enter, and doing so ensures that the surfaces of the SMR and combustion catalysts flush for optimal heat transfer.

The combustion gas manifold that brings in the fuel (hydrogen) and the air must distribute air and fuel equally to the 9 channels, while the 2 combustible streams are kept separate until they meet in the combustion channel inlet. This is accomplished by taking a 1.27 cm x 5.23 cm x 2.54 cm wide block of "Inconel" 625 and using a string EDM to machine 1.27 cm x 1.78 cm x 0.3 cm through channels, one for the air and the fuel for each channel. A single air duct joins the inlet of each combustion duct, while a single fuel duct is a height of 1.93mm below it. The 2 channels are connected at the combustion channel inlet using a 0.38mm long channel machined in the plane that forms the interface with the combustion channel. This enables the fuel to make a 90° turn and move into the machined pocket at the interface with the main body to meet the incoming air flow where mixing takes place, as shown in Fig. 18e. The fuel and air channels produced in this way with string EDM are closed with welds on the opposite surface of the interface with the reactor body. The branch pipe for delivering air from a single inlet pipe in the channels produced by string EDM is formed by machining 0.25 cm diameter pipes that protrude from the side of the block. These protrude 2.16 cm of the width from the side, which forms an individual branch pipe for air across the combustion channel. These individual channel branch pipes are connected to the inlet by a single pipe. The branch pipe for the fuel has a similar design. The total manifold is shown in figures 17 and 18.

The combustion product manifold to a single outlet pipe was fabricated from a block of "Inconel" 625, which was 0.95 cm long, 4.71 cm high, and 2.29 cm wide. The branch pipe is a large open space with a length of approx. 0.51 cm for current to collect in, with a 0.46 cm diameter outlet on the underside of the surface opposite the interface with the reactor.

Gaseous reactants were metered from pressure cylinders via Brooks mass flow controllers, while water was delivered via a high pressure liquid (HPLC) pump. An external combustion chamber and a heat exchanger network were used to continuously evaporate the supplied water, preheat the steam reforming reactant to 500-600°C, and preheat the combustion air to 500°C. Device body temperatures were measured in thermowells (drilled into selected steps between channels) using 0.81 mm type K thermocouples, and recorded using an "Iotec" data logging system. Inlet and outlet gas pressures were measured for the combustion and reforming streams via calibrated pressure transducers. Gas temperatures were also measured with a thermocouple at the combustion gas inlet and outlet, and at the reformer gas inlet and outlet. The product gases were cooled in water-cooled condensers and the dry outlet flow measured in a dry test meter. Dry product gas (H 2 , CO, CO 2 , CH 4 , O 2 , and N 2 ) was measured using an online gas chromatograph. All tests for the integrated combustion reactor were performed using a steam-to-carbon molar ratio of 2:1.

Catalysts were reduced prior to the first day of testing at 125°C 1 Ml hour. Start-up of the integrated combustion reactor was carried out in several stages. First, an inert stream (nitrogen, 1-5 SLPM) was initiated in both the combustion and reformer channels. The external combustion chamber was then ignited to ensure water evaporation and preheating in external, microchannel-based heat exchangers. Then the nitrogen flow to the combustion chamber was shut off, and the air flow initiated. Hydrogen was then introduced into the combustion channels as approx. 3% of the total current, which caused catalytic ignition of the combustion reactants. Then the hydrogen content was slowly increased to bring the reactor temperature to approx. 700°C. A small stream of hydrogen was then introduced into the reforming channels to prevent catalyst oxidation by steam, water was fed through the evaporator thereby supplying steam to the reforming channels. Finally, the nitrogen to the reformer was shut off, and methane was introduced into the reformer, and the hydrogen to the reformer was shut off. After the reforming reaction started, additional hydrogen was added to the combustion channels to maintain the reactor temperature at between 750 and 850°C. Changes in process streams carried out during steady state operation were carried out in a balanced manner (changing both the reforming and combustion streams simultaneously) to prevent large transient deviations of the integrated combustion reactor temperature from the desired operating temperature.

A summary of the results of tests carried out using the integrated combustion reactor at near ambient pressure is shown in Table 4. Note that in Table 4, as a result of the combustion air being preheated to 500°C, and a low % excess air (48 -75%) was used, adiabatic flame temperatures inside the combustion channels theoretically exceed 2100°. However, the highest measured temperatures inside the integrated combustion reactor were well below 940°C, demonstrating the effectiveness of the integrated combustion reactor concept in transporting heat away from the combustion channels and into the endothermic reaction. Note also in Table 4 that high heat fluxes (close to 3 W/cm<2>) and high conversions can be achieved even with very short contact times (as low as 33 ms). Contact time is defined here as the total channel volume (which accommodates the catalyst and the bypass area) divided by the total volumetric inlet flow amount (converted to standard conditions, 0°C and 1 atm). Even for very fast contact times of the data shown in Table 4, pressure drop is minimal (about 0.14-0.21 bar), showing the distinct advantage of the bypass arrangement in the integrated combustion reactor.

The integrated combustion reactor was designed to provide proof of principle of the concept of the integrated combustion reactor in a multi-channel device and is not an optimized prototype. Optimizing the inlet and outlet branch pipes, as well as channel geometry, is expected to enable increased capacity and/or reduce the pressure drop. Measurements of methane conversion in the integrated combustion reactor (channels in the reformer) were performed for a range of combustion channel air equivalence ratios. Air equivalence ratio is defined as the ratio between available combustion oxygen and that required for stoichiometric combustion. It was observed that methane conversion gradually decreased with increasing air equivalence ratio from approx. 83% conversion at an equivalence ratio of 1.75 to approx. 63% conversion at an air equivalence ratio of 7.0. Although some coking of the reforming catalyst was observed when the reforming reactant preheat was allowed to drop below 500°C, activity was fully restored after burning off the coke with air and re-reducing the catalyst. The integrated combustion reactor was operated for over 10 h, but included 3 heating cycles.

In summary, an integrated multi-channel combustion reactor was designed, fabricated and shown to be capable of achieving a heat transfer rate of 2.8 W/cm2 and a volumetric heat flux of 13 W/cm<2>at pressure drops of less than 0.23 bar and 0.19 bar in the channels in the reformer and the combustion chamber, respectively. High conversion (70-82%) was achieved in this integrated combustion reactor for total channel reformer contact times of 33-60 ms. After over 10 hours of operation at 750-850°C, the integrated combustor showed no significant signs of catalyst deactivation, even though the combustor was operated with 75% excess air or less for more than 1½ hours.

Example 4

An integrated combustion microchannel reactor capable of producing exceptionally high heat fluxes (up to 29 W/cm<2> and 118 W/cm<3>) to a highly exothermic reaction with minimal pressure drop was designed, fabricated and demonstrated. This report describes the design and testing of this multi-channel interfoliated microchannel reactor in which heat-producing (exothermic) and heat-consuming (endothermic) reaction channels were interfoliated (integrated) with each other. Combustion of hydrogen in air was used as the exothermic reaction, while steam reforming of methane (SMR) with a steam to carbon ratio of 2:1 was used as the endothermic reaction. A bypass design was used, which enabled much higher throughput with minimal (<0.34 bar) pressure drop. By allowing the reactant stream in each channel (either exothermic or endothermic) to flow in a narrow (approx. 0.3 mm) gap between 2 layers of porous, engineered catalyst that was in intimate thermal contact with the fixed channel wall or heat transfer surface. The design included 4 combustion channels interleaved between 6 reforming device channels in a common flow arrangement. The test results show that integrated microchannel combustion reactors can be used to deliver much higher volumetric heat fluxes (i.e. 118 W/cm<3>) than is possible in conventional reformers, and these heat fluxes can be achieved with minimal pressure drop (0.28-0.34 bars). High conversions were also possible at very high gas space velocities per hour (eg IO<6>h"<1>)

Gas space velocity per hour is defined as the number of SMR reactor core gas volumes (defined at 0°C and 1 atm) that will pass through SMR catalyst containing channels every hour. The space velocity can be calculated by dividing 3.6 x IO<6> by the contact time in milliseconds, both of which are defined at STP 0°C and 1 atm. The reactor volume used in all volumetric heat flux calculations includes the entire core reactor volume which has included all reformer channels, including catalysts, all metal steps between reformer and combustor channels, as well as all combustion channels used to supply heat. The external packing is not included, as this volume does not contribute to the transport of heat between the 2 fluid streams.

Experimental method

A multichannel reactor with high heat flux and with integrated combustion was designed, manufactured and tested. The reactor body was fabricated from "Inconel" 625 using string EDM to form grooves in a 2.5 cm x 3.48 cm x 4.3 cm solid block of metal. External cover plates fitted with inlet and outlet gas manifolds (ie manifold and header) were independently machined and welded to the body. Each branching chamber was designed to distribute the flow evenly from channel to channel, and within each channel. A drawing of the exterior of this device is shown in Fig. 19a and comprises an air inlet 102, a hydrogen inlet 104, a reformer feed inlet 106, a

reformer product outlet 108, as well as a combustion chamber outlet 110. For test purposes, temperature monitoring was performed with thermocouples 112. All the small tubes 112 in Fig. 19 are thermocouple tubes. Fig. 19b and 19c show exploded views of the integrated reforming device. Air is fed into a branching chamber 114 and distributed by uniform pressure into slits 116 with uniform pressure through each slit. Correspondingly, hydrogen is fed in through the inlet 104 and with uniform pressure into each tube into slotted tubes 118. The tubes 118 protrude from the hydrogen inlet into a

branch line 120. Schematic views of a gas mixer 125 are shown in Figs. 19d and 19e (which are not drawn to scale). Projection 127 defines a gap 129 for air flow. Hydrogen flowing out of slits 131 passes through a channel 133 and mixes with air in the combustion chamber 122 inside the reactor body 124. As can be seen, this design mixes air and fuel inside the combustion chamber and very close to a combustion catalyst 137, thereby avoiding heat loss from combustion outside the combustion chamber. Feedstock for the steam reformer is fed into an inlet distribution chamber 130 and passes at uniform pressure into the front of the steam reformer chambers. An outline of the reformate flow through an individual chamber 132 is shown in Fig. 20. Plates 134 have supports 136 which press the reforming catalyst against the top and bottom of the reforming chamber.

The main body of the device in Ex. 4 was completely similar to the main body of the device in Ex. 3 except that there were 4 fewer SMR channels and 4 fewer combustion channels, and the length of the catalyst in the direction of flow was only 2.5 cm instead of 3.5 cm. The dimensions of the SMR slits were 0.89 mm thickness, 17.8 mm width and 25.4 mm length. There were 6 SMR slots. The dimensions of the combustion slits were 0.97 mm thickness, 17.8 mm width and 25.4 mm length. There were 5 combustion channels. The channels were arranged so that the outermost channel was a half-reformer channel up to a combustion channel, then up to a full reformer channel, etc., in the same way as in the device in Ex. 3. The reformer half channel was 0.46 mm thick, 17.8 mm wide and 25.4 mm long. The half-channel contained only one catalyst insert, which was placed up to the wall shared with the combustion channel. The full reformer channels contained 2 engineered catalysts that were positioned up to each channel wall.

The flows in the exothermic and endothermic reaction channels are in the same direction (joint flow), although in this design the flow in the reformer enters and leaves in a direction perpendicular to the direction of flow during the reaction to occupy the branch connections on a other surface of the device, than the combustion flow branching, as was done with the device in Ex. 3. Note that the catalytic combustion was most likely accompanied by some homogeneous combustion in the by-flow gap. During testing, the device was insulated with ceramic fiber insulation. 2 versions of the reactor were manufactured and tested, whereby a combined time with flow of over 300 h was achieved.

Both the endothermic (combustion) and exothermic (SMR) reactions were catalyzed by catalyst inserts 134, 14, which measured 0.028 cm x 1.8 cm x 2.5 cm and were held against the channel walls by 1 .3 cm wide strips of "Inconel" 625 metal inserted in the bypass gap along each edge and down towards the center [134 and 136 in Fig. 19b are respectively the catalyst and the spacers in the steam methane reformer. 14 is the combustion catalyst. There are equivalent spacers between combustion catalysts (not shown)]. The by-flow gaps in each combustion channel were approx. 0.41 mm high, while the bypass gaps in the channels in the steam methane reformer were approx. 0.33mm high and approx. 0.2mm high for the full and half capacity channels respectively. The outermost channels in the steam reformer were of such a size that they had half the flow of the flow in the inner channels, the outer channels only receiving half the heat. The thickness of these outermost channels was approx.

0.46mm, and they contained a 0.28mm catalytic converter. Catalysts 8 and 11 (Fig. 19b) were introduced with a press fit along each side of each channel in the steam methane reformer to direct the flow into and out of the inlet distribution chamber areas and the outlet collection chamber areas and to prevent bypassing of the catalyst on the outside of the bypass gap. Each of the outer channels had catalyst only against the innermost wall and was designed to allow only half the flow to flow through a full channel at a given pressure drop.

Gaseous reactants were metered from pressure cylinders via Brook's mass flow controllers, while water was delivered via high-pressure liquid pumps. An external combustion chamber and an external heat exchanger network were used to continuously vaporize the feed water, preheat the steam reforming reactants to approx. 800-845°C, as well as preheating the combustion air to approx. 500-650°C. Temperatures in the device were measured in thermowells (drilled in selected struts between the channels) using 0.08 cm thermocouples of type K and recorded with "Labview" with a data logging system. The inlet and outlet gas pressures were measured for the combustion and reforming streams via calibrated pressure transducers. Gas temperatures were also measured by thermocouple at the combustion gas inlet and outlet, as well as at the reformer gas inlet and outlet. The product gases were cooled in water-cooled condensers, and the dry volumetric outlet flow was measured in a dry test meter. Dry product gases (H 2 , CO, CO 2 , CH 4 , O 2 and N 2 ) were measured using an online MTI gas chromatograph. The behavior of the steam reformer was eluted using a 2:2 steam to carbon molar ratio. Gas space velocity per t (GHSV) was calculated based on the total channel volume in which flow was exposed to catalyst (including the catalyst, spacer strips and bypass volume) based on volumetric flows defined at 0°C and 1 atm. Heat flux was calculated by determining the amount of heat that was transported into the endothermic steam reforming reaction. For a known molar flow rate of reforming reactant, a known amount of conversion is measured by means of gas chromatography analysis of the composition of the outflowing material and the flow rate of the starting product. From the total number of moles that are converted, the total heat required is calculated. Heat flux values were calculated as an average over the entire wall area in contact with the SMR catalyst. Calculated average heat flux values were based only on the required SMR heat of reaction for the measured SMR conversion and selectivity. Calculated volumetric heat flux values were based on the entire volume containing both reforming and combustion catalysts and respective slots, including the walls between the channels, but not including any peripheral metal. This volume includes the total volume through which heat is transported between the 2 fluids.

Catalysts were reduced on the first day of testing at 125°C oxygen. Start-up of the integrated combustion reactor was carried out in several stages. First, an inert stream (nitrogen, 1-5 SLPM) was initiated in the channels in both the combustor and the reformer. Next, an external combustion chamber was initiated to produce a hot stream of combustion products sufficient to transport heat in this external heat exchanger to vaporize water necessary for the reforming reaction and provide necessary reactant preheating. Then, the nitrogen flow to the combustion chamber was shut off, and air flow was initiated. Hydrogen was then introduced into the combustion channels as approx. 3% of the total current, which caused catalytic ignition of the combustion reactants. Then the hydrogen content was slowly increased to bring the reactor temperature to approx. 700°C. A small stream of hydrogen was then introduced into the channels of the reformer to prevent catalyst oxidation by steam, and water was fed through the evaporator to thereby supply steam to the channels of the reformer. Finally, nitrogen to the reformer was shut off, methane was introduced to the reformer, and hydrogen to the reformer was shut off. After the reforming reaction started, additional hydrogen was added to the combustion channels to maintain the reactor temperature between 750 and 850°C. Changes in process streams carried out during steady state operation were carried out in a balanced way (simultaneous change of both the reforming and combustion streams) to prevent large transient deviations of the temperature in the integrated combustion reactor from the desired operating temperature. Combustion was carried out in FDR with excess air values as low as 13%.

This device was operated near atmospheric pressure in both the reforming and combustion reactions. The typical operating pressure on the reforming side was approx. 0.07 kg/cm<2> (10 psig). The typical operating pressure for the combustion side was approx. 0.07 kg/cm<2>. Air and fuel inlet pressures were similar.

Results and discussion

Results of testing using the high heat flux integrated combustion microchannel reactor are shown in Table 5 and Figures 21-24. Methane conversion approaches equilibrium conversion and selectivity towards carbon monoxide and hydrogen even at very high space velocities (up to 10<6>hr"\ see Fig. 21).

NOx measurements were made of the dry combustion effluent stream with testing of the device in Ex. 4 under the conditions shown under Test 2 in Table 5. The concentration of NOX which was measured under this condition (4.7

SLPM H2, 20 SLPM air, 850° core temperature) in the dry effluent was 10-12 ppm. This is comparable to NOx levels exceeding 100 ppm in conventional methane stream reformers. This measurement is called here the "standard NOx test measurement".

Results from the tests where the amount of excess air on the combustion side was varied show that the integrated combustion reactor with high heat flux is able to function efficiently with 20% excess air (see Fig. 22). This means that the combustion air requirement and associated air compressor costs and air recovery function can be greatly reduced compared to traditional catalytic combustion chambers.

The volumetric heat flux (ie the energy transported to the endothermic reaction per unit volume of reactor) reflects the compactness or degree of intensification of the process. The high volumetric heat flux values shown in Fig. 24 demonstrate the ability of this device to drive heat (circles, left axis) into the endothermic reaction with a much smaller reactor (and catalyst) volume and minimal pressure drop (triangles, right axis). The highest volumetric heat fluxes observed in the high heat flux integrated reactor (Fig. 24) are 200-1000 times higher than those typically seen in conventional reformers. The much lower reactor volume suggests the potential for substantial savings in reactor materials and catalyst costs per unit volume of synthesis gas produced.

Conclusions

An integrated combustion microchannel reactor was designed, manufactured and tested which can achieve high methane conversions at heat fluxes as high as 29 W/cm<2> and volumetric heat fluxes of 118 W/cm<3> at pressure drops below approx. 0.3 kg/cm<2> and approx. 0.35kg/cm<2> in the reformer and combustion chamber sides, respectively, with a pressure drop of less than approx. 0.11 kg/cm<2>and 0.14 kg/cm<2>in the reforming device side and the combustion chamber side, respectively, per cm reactor current length. Close to equilibrium methane conversions (73-98%) were achieved in this microchannel device at very high gas space velocities per hour (1.3 x IO<5>- 1.0 x IO<6>) at steam:C, approx. 1 atm and 850°C average wall temperature. The high heat flux microchannel reactor proved capable of operating with as little as 20% excess combustion air.

Example 5

An integrated combustion microchannel reactor with distributed fuel introduction and a cross-flow reforming and combustion flow orientation was designed, manufactured and demonstrated. In this example, the design and testing of this microchannel reactor is described. Combustion of hydrogen in air was used as the exothermic reaction, while steam methane reforming (SMR) at a steam to carbon ratio of 3:1 was used as the endothermic reaction. The device comprised a single SMR by-flow channel with 3 combustion by-flow channels in a cross-flow orientation relative to the SMR channel. The SMR reactants flowed in a 0.13 mm gap between the wall and a layer of porous engineered catalyst, which was in intimate thermal contact with the fixed channel wall or heat transport surface. Combustion air flowed in 3 parallel, 2.5 mm cylindrical channels, each with combustion catalyst coated on the wall. At 3 points that were distributed evenly along the flow length of the combustion channels, hydrogen fuel was injected into the combustion channel. Each combustion channel also contained a static mixer made from 0.5 mm "Inconel" sheet material which was provided with notches and twists. The test results show that this integrated cross-flow combustion device can achieve high heat fluxes (approx. 15 W/cm2). High conversions were also possible at very high space velocities.

Experimental method

A cross-flow microchannel device with integrated combustion was designed, manufactured and tested. The device body (see Fig. 25) was fabricated from Inconel 625 using string EDM and conventional machining to fabricate both an SMR plate 202 (approximately 0.6 cm thick) and a combustion plate 200 (1, 1 cm thick), each of which was approx. 5.8 cm x 3.0 cm. Engineered SMR was then placed inside the SMR channel between the plates and the 2 pieces were welded together to seal the device. 4 raised areas, each 1.3 cm wide and 1.1 cm long were distributed evenly over the SMR channel to hold the catalyst against the wall (heat transfer surface). The SMR catalyst measured 3.8 cm x 0.03 cm thick and was placed in a slot of the same dimension except that the slot's thickness was approx. 0.04 cm. External cover plates containing the inlet and outlet gas manifolds (ie the inlet manifolds, and the outlet manifolds) were machined independently and welded to the body with the exception of the combustion fuel inlets, which were fed through 3 separate inlet tubes. The SMR inlet manifold was designed to distribute and flow uniformly across the 3.8 cm catalyst width prior to flow into the 0.13 cm bypass gap. Combustion air flowed in the 0.3 cm diameter, cylindrical bypass channels in a cross-flow orientation relative to the SMR channel. Combustion air was mixed with hydrogen fuel and introduced from the opposite wall into 3 parallel 2.5 cm cylindrical channels, each of which had combustion catalyst coated on the wall. At 3 points spaced every 0.13 cm along the flow length in each channel, hydrogen fuel was introduced into the combustion channel through 0.003 cm holes. A static mixer made of 0.5 mm "Inconel" sheet material equipped with notches and twists was introduced into each combustion channel, which increased mixing of the air and fuel inside each channel. Combustion catalyst was coated on the static mixer and on the wall of the combustion channels to promote catalytic combustion of the fuel/air mixture. During testing, the device was insulated with ceramic fiber insulation.

Gaseous reactants were metered from pressure cylinders via Brook's mass flow controllers, while water was supplied by a high pressure washer (HPLC) pump. An external combustion chamber and heat exchanger network was used to continuously evaporate the feed water, preheat steam reforming reactants and combustion air to the values shown in the table below. Device body temperatures were measured in thermowells (drilled into the outside of the combustion plate) using 0.81 mm type K thermocouples and recorded using a Labview with a data logging system. Inlet and outlet gas were measured for the combustion and reforming streams via calibrated pressure transducers. Gas temperatures were also measured via thermocouples in the combustion inlet and outlet and in the reformer gas inlet and outlet. The product gases were cooled in water-cooled condensers, and the dry volumetric outlet flow was measured in a dry test meter. Dry product gases (H 2 , CO, CO 2 , CH 4 , O 2 and N 2 ) were measured using an online MTI gas chromatograph. The SMR behavior was evaluated using a 2:1 molar ratio of steam to carbon. Gas space velocity per hour (GHSV) was calculated based on the total channel volume in which flow was exposed to catalyst (including the catalyst, spacer strips and bypass volume) based on volumetric flows defined at 0°C and 1 atm. The average heat flux was calculated based on the entire wall area in contact with the SMR catalyst and up to the plane below the combustion cylinders and therefore defines the plane through which the heat is transported between fluids.

Catalysts were reduced on the first day of testing at 125°C oxygen. Startup of the small super channel was carried out in several steps. First, nitrogen flow was started in the SMR side, and air flow was initiated in the combustion side. Hydrogen was then introduced into the combustion channels as approx. 3% of the total current, which caused catalytic ignition of the combustion reactants. Then the hydrogen content was slowly increased to bring the reactor temperature to approx. 400°C. A small stream of hydrogen was then introduced into the reformer channels to prevent catalyst oxidation by steam, and water was fed through the evaporator, thereby producing steam to the reformer channels. Finally, the reformer nitrogen was shut off, methane was introduced into the reformer, and hydrogen to the reformer was shut off. After the reforming reaction started, additional hydrogen was supplied to the combustion channels to maintain the reactor body temperature between 725 and approx. 875°C. Changes in process flows were carried out during steady-state operation in a balanced way (changing both reforming and combustion flows at the same time) to prevent large transient deviations of the temperature from the desired operating temperature. To assist with the analysis of hydrogen conversion, a cooling vapor of approx.

400°C sometimes introduced directly into the combustion chamber. This was not very effective for cooling the combustion in the device's outlet collecting chamber, and increased the heat losses significantly when it was used.

Results and discussion

Results of tests using a cross-flow integrated combustion microchannel reactor are shown in the table below and Fig. 26. Important features shown include: catalyst applied directly to the wall, static mixer introduced into the reaction chamber, distributed fuel, as well as cross-flow orientation of combustion fluids and reformate.

In all tests, the conversion of hydrogen in the combustion reaction chamber was 100%.

Example 6

DESIGN AND OPERATION

An integrated reforming and combustion reactor was evaluated in a cross-flow orientation. A single SMR slot was up to 3 combustion cylinders.

The SMR catalyst was placed against the wall shared by the combustion channels, which supplies the energy that sustains the endothermic reforming reaction. The SMR channel gap was 0.38 mm, and a 0.03 mm porous SMR catalyst was introduced, giving a gap of approx. 0.076 mm. The length of this reforming channel was 11.4 mm and its width was 12.7 mm. The combustion section was formed by 3 channels of diameter 0.254 cm in a plane parallel to the SMR channels, but with its flow path aligned at 90° to the flow - but in the SMR channel, in cross flow. It was an approx. 1.27 mm metal step that separated the reforming channel and the top of the cylindrical combustion channels. The flow length of the combustion cylinders was 12.7 mm, which corresponded to the reformer channel width of 12.7 mm for this cross-flow orientation. The combustion catalyst was made from a 0.025 cm thick FeCrAlY felt wrapped around a 0.15 cm wide and 0.051 cm thick static mixer insert. The mixer had 2 functions:

1) pressing the felt against the chamber walls, and

2) to mix the fuel as the air passed through

the inner diameter of the felt.

The static mixer was also prepared on the following

manner:

1. The approx. The 1.5 mm wide side was cut into 2 length sections of 12.7 mm separated by a 0.75 mm width that protruded 0.96 mm in length. 2. The first length of 12.7 mm was twisted 90° from its initial orientation. 3. The beginning of the next (and last) 12.7 mm length section was turned 90° towards the end of the first 12.7 mm length section. 4. The end of the second 12.7 mm length was twisted 90° in the opposite direction of the first 12.7 mm length section.

The fuel was fed into each combustion channel at the inlet of the cylindrical channel from the bottom of the device, or on the opposite side of the plane separating the reforming channel. Combustion fuel enters 0.76 mm ahead of the catalyst/static mixer insert, which is brought inside the cylindrical combustion channel, each channel with its own fuel port.

The overall dimensions of the block assembly in the plane of the channels are 41.9 mm across in the SMR channel direction and 43.2 mm in the combustion channel direction, centered over the 11.4 mm x 12.7 mm integrated combustion reforming core where the reactions take place and heat is transported. The volume through which all heat is transported and used for the calculations of volumetric heat flux is 11.4 mm x 12.7 plus x the sum of 2.54 mm (combustion diameter) and 1.72 mm (metal step separating the channels) and 0.38 mm (full reforming channel), total 0.6 cm<3>(0.037 in<3>). The area or plane through which all heat is transported is 11.4 mm x 12.7 mm, or 145 cm<2>. The inlet manifold in the process was custom made to have a smooth transition from the inside diameter of a 4.57mm inside diameter inlet pipe to a 12.7mm wide by 0.36µm high slot to avoid coking and allow for additional preheating from device losses . An outlet collection chamber of the same design leads to the outlet pipe with an internal diameter of 4.57 mm.

This steam methane reforming microchannel reactor maintained a pressure differential of 11 atm absolute at 850°C between the reforming and combustion channels. The metal step separating the channels was 1.27 mm or 1270 µm. The device was operated for more than 13 days, or more than 300 h without any change in structural integrity. No play caches were found between the streams. The high pressure difference between the streams was supported over a thin metal step using microchannels, particularly circular microchannels, under a high-pressure reforming slot.

Behaviour

The device was demonstrated for the reforming reaction at 3:1 molar ratio of steam to carbon, 12 atmospheres absolute pressure and 850°C and 50 milliseconds. contact time. The device produced equilibrium conversion as seen in Fig. 27. Methane conversion and CO selectivity decreased continuously over time for the next 5 days. The conditions were changed to a 2:1 molar ratio of steam to carbon, while the temperature, pressure and contact time were kept constant, and the loss of activity continued. The device exhibited significant heat fluxes and average volumetric heat fluxes as shown in Table 5. The first 2 lines relate to the 3:1 ratio between steam and carbon at 5.0 milliseconds. contact times, whereby line no. 1 relates to the beginning of the experiment in Fig. 24 and no. 2 relates to the end of the experiment. Correspondingly, lines no. 3 and no. 4 in the table below relate to the start and end behavior for the molar ratio of 2:1 between steam and carbon at 5 milliseconds. contact time, also shown in Fig. 27. All data reflect an SMR reactant inlet temperature of from 830-840°C.

The corresponding settings on the combustion side to the setting in Table 5 are shown in Table 6. They reflect a constant air flow rate of 5.4 SLPM in 3 channels, and a fairly constant hydrogen flow rate of close to 0.5 SLPM.

In all tests, the conversion of hydrogen in the combustion reaction chamber was 100%.

Heat flux measurement test

The drive device for a methane steam reforming reaction at 850°C, 1.70 bar, 3:2 ratio of steam to carbon and a contact time of 100 ms. Contact time is defined as the total reaction volume divided by the total volumetric inlet flow rate of reactants at standard temperature and pressure (STP: 273K and 1 atm absolute). The total reaction volume includes the steam reforming catalyst and the reaction channel volume containing the catalyst.

E.g. if the cumulative volumetric sum of reaction chambers including reforming catalysts is 1 cm<3>, the total inlet flow rate of reactants will be 0.6 standard 1 per my. The inlet rate for methane will be 0.15 standard 1 per min., and the inlet flow rate for steam will be calculated to be 0.45 1 per my. at standard temperature and pressure. In this example, the molar inlet flow rate for methane will be approx. 0.00045 mol per Sec. These numbers are linear with the total reaction chamber volume. A 2 cm<3>reaction chamber volume will require 0.0009 mol per Sec.

Methane conversion is determined by measuring the feedstock composition and outlet flow rate of methane reforming reaction products and then calculating based on the following formula: Conversion %=100 x (mol methane in - mol methane out)/(mol methane in).

Mol methane in=inlet flow rate of methane at STP/(22.4 l/mol).

Mol methane out=[outlet flow rate for total dry product gas/(22.4 l/mol)] x% methane in dry gas by gas chromatography analysis.

Dry gas is defined as the product gas flow rate after condensation of the unreacted water or other condensable fluids.

Selectivity towards CO%=100x(mol XO/(mol C02+mol CO+mol C(s) if present)).

Selectivity towards CO2%=100x(mol C02/(mol C02+mol CO+mol C(s) if present)).

Heat load=(conversion%/100)x molar methane in x(reaction heat during steam reforming to carbon monoxide at 850°C (226800 J/mol)x selectivity to CO%+reaction heat during steam reforming of methane to carbon dioxide at 850°C (193200 J/ mol) x selectivity to CO2%)/100, units i

W.

Heat flux=heat load/reactor core volume, units W/cm3.

The reactor core volume includes all reaction chambers or channels, all associated combustion chambers or channels, as well as all separating metal steps through which heat is transported between fluids. In short, this volume includes the total volume through which heat is transported for the methane steam reforming reaction. This volume does not include peripheral metal, manifold volume or other associated equipment that depends on individual geometries of the device.

The following conditions must be met by the combustion reaction that adds heat to the heat flux measurement test: 1. The gas phase fuel that must be used is hydrogen. 2. The total air flow rate is given such a size that a mixture of the hydrogen and air flow rates into the reactor reaches an excess air percentage of 80%. The excess air is defined as the total molar flow rate of oxygen in the mixture of hydrogen and air divided by the molar flow rate of oxygen necessary to completely oxidize the hydrogen at its molar fuel flow rate. Since one mole of oxygen can completely oxidize 2 moles of hydrogen, 80% excess air corresponds to a molar ratio between air and hydrogen of 4.28:1. Air is taken as 21 mol% oxygen,

the rest nitrogen.

3. The hydrogen and air flow into the combustion reactor at 900°C. 4. The air and hydrogen must either be mixed in a manifold which is directly upstream of the combustion reactor, or in the reactor itself. 5. The standard volumetric flow rate of hydrogen through the combustion reactor per 0.15 SLPM methane flow rate through the methane steam reforming reactor is at least 0.140 SLPM and at most 0.204 SLPM. 6. The corresponding minimum and maximum air flow rates through the combustion reactor, based on the condition 80% excess air, per 0.15 SLPM methane flow rates through the methane steam reforming reactor are 0.600 SLPM and 0.875 respectively

SLPM.

7. The inlet pressure of the hydrogen and air streams should not

be higher than 2.38 bar (20 psig).

Pressure test- high temperature test for integrated combustion reactor

In preferred embodiments, any of the devices described herein are capable of withstanding internal pressure differences. E.g. some preferred embodiments satisfy the requirements of the following pressure test. For a microchannel unit operating device at least one critical channel dimension of less than approx. 2 mm, work with at least 2 inlet fluid streams. The first fluid flow must be on

850°C and 12.6 kg/cm2 (180 psig). The second fluid flow must

be at 800°C and 0.7 kg/cm<2>. Any flow rate can be used. The device is operated with gas flow for both flows for 300 h. After 300 h of operation, each fluid flow line is pressurized to approx. 3.5 kg/cm<2> which is held in 2

t. The pressure must remain constant, which indicates minimal leakage paths to the environment. The second fluid flow line is then pressurized to approx. 3.5 kg/cm<2>, while the first fluid flow line is allowed to be open to the atmosphere, and is held for 2 h. The pressure must remain constant, indicating minimal internal leakage paths. A minimal leakage path is defined as a leakage rate of less than 10~<6>standard cm3 per Sec. helium, when helium is used as fluid in the final leak test.

The invention also includes methods for carrying out unit operations in the device which has the pressure resistance characteristic described above.

Claims (10)

1. Procedure for carrying out an endothermic reaction in an integrated combustion reactor, characterized in that it comprises: that an exothermic reaction composition is led into at least one exothermic reaction chamber, where the exothermic reaction chamber comprises an exothermic reaction catalyst; where the integrated reactor comprises alternating layers of exothermic and endothermic reaction chambers comprising at least 2 layers of endothermic reaction chambers alternating with at least one layer comprising the exothermic reaction chamber; wherein the endothermic reaction chambers comprise an endothermic reaction catalyst; wherein the at least one metal exothermic reaction chamber comprises a wall adjacent to at least one of the endothermic reaction chambers; wherein one of the exothermic reaction catalysts or the endothermic reaction catalyst comprises a porous catalyst insert; wherein the exothermic reaction composition reacts in the exothermic reaction chamber and generates heat; and conveying an endothermic reacting composition into the at least one endothermic reaction chamber; and where the method has a volumetric heat flux of at least lW/cc. 2. Method in accordance with claim 1, characterized in that the porous catalyst insert fills the endothermic reaction chamber. 3. Method in accordance with claim one or more of the preceding claims, characterized in that an oxygen-containing composition passes through a plurality of slits and into a fuel flowing through the exothermic reaction chamber. 4. Method in accordance with claim one or more of the preceding claims, characterized in that the method has a volumetric heat flux of at least 20 W/cc. 5. Method in accordance with claim one or more of the preceding claims, characterized in that the endothermic ration includes hydrocarbon steam reforming and where at least 70% of the equilibrium conversion to the hydrocarbon that is taken in at the beginning of the endothermic reaction chambers is converted to hydrogen, carbon monoxide and carbon dioxide. 6. Method in accordance with claim one or more of the preceding claims, characterized in that the endothermic reaction comprises hydrocarbon steam reforming with a steam to carbon ratio in the range from 1 to 3. 7. Method in accordance with claim one or more of the preceding claims, characterized in that the endothermic reaction comprises hydrocarbon steam reforming with a contact time of less than 100 milliseconds. 8. An integrated reactor, characterized in that it comprises: alternating layers of exothermic and endothermic reaction chambers comprising at least 2 layers of endothermic reaction chambers alternating with at least one layer comprising an exothermic reaction chamber; wherein the exothermic reaction chamber comprises an exothermic reaction catalyst; wherein the endothermic reaction chamber comprises an endothermic reaction catalyst; where the exothermic reaction chamber comprises a wall adjacent to at least one of the endothermic reaction chambers; wherein one of the exothermic reaction catalysts or the endothermic reaction catalyst comprises a porous catalyst insert; and where the integrated reactor has a volumetric heat flux characterized by at least lW/cc when measured in accordance with "the Volumetric Heat Flux Measurement Test". 9. Method for carrying out an endothermic reaction in an integrated reactor, characterized in that it comprises: placing a fuel or an oxidant into at least one exothermic reaction chamber, where the exothermic reaction chamber comprises an exothermic reaction catalyst; where the integrated reactor comprises alternating layers of exothermic and endothermic reaction chambers comprising at least 2 layers of endothermic reaction chambers alternating with at least one layer comprising the exothermic reaction chamber; wherein the endothermic reaction chambers comprise an endothermic reaction catalyst; wherein the at least one exothermic reaction chamber comprises a first wall adjacent to at least one of the endothermic reaction chambers, and a second wall comprising slits; wherein an oxidant or a fuel is brought through the slots and into the at least one exothermic reaction chamber; where the oxidant and fuel react in the exothermic reaction chamber and generate heat; and imparting an endothermic reacting composition into the at least one endothermic reaction chamber; and where the method has a volumetric heat flux of at least lW/cc. 10. An integrated reactor, characterized in that it comprises: alternating layers of exothermic and endothermic reaction chambers comprising at least 2 layers of endothermic reaction chambers alternating with at least one layer comprising an exothermic reaction chamber; wherein the exothermic reaction chamber comprises an exothermic reaction catalyst; wherein the endothermic reaction chamber comprises an endothermic reaction catalyst; wherein the at least one exothermic reaction chamber comprises a wall adjacent to at least one of the endothermic reaction chambers; wherein the exothermic reaction chamber comprises a first wall adjacent to at least one of the endothermic reaction chambers, and a second wall comprising slits; where the second wall separates the exothermic reaction chamber from an oxidant or a fuel chamber; and where the integrated reactor has a volumetric heat flux characterized by at least lW/cc when measured in accordance with "the Volumetric Heat Flux Measurement Test".