Hydrogen Production

This invention relates to the production of hydrogen from biomass, in particular to sorption enhanced production of hydrogen from polyols such as glycerol in which almost pure hydrogen can be obtained. The invention is further characterised by being autothermic making it a highly attractive process for hydrogen formation.

There is increasing World concern about climate change and hence carbon dioxide emissions. In Europe at least, measures are now being taken at the highest level to try to reduce carbon dioxide emissions in all walks of life. This means a heavier reliance on renewable energy sources such as solar and wind power, increasing taxation on energy inefficient products and more investment in harnessing the power of the sea. A further quickly developing field is biofuels, especially for vehicles. The biofuels directive of 2003 (Directive 2003/30) established an indicative target value of 5.75 % market share for bio-fuels at the end of 2010. The reference value is based on the energy content of the fuel, and each member state is to set national targets. In the Energy and Climate Package and the related Renewable Energy Roadmap published on 10 January 2007, the European Commission proposes a minimum binding 10% target of biofuels for vehicle use to be reached in the EU by 2020.

The use of flexifuel vehicles, which run on alcohol, in particular ethanol (and also methanol), are already well known. Ethanol can be produced from sugar cane and is frequently used in blends with gasoline to form a biofuel (E5, E85) . Methanol is however toxic and is not an ideal material for the mass market, its use currently being confined therefore to racing vehicles.

The industry has therefore being considering ways in which other biomasses can be incorporated into both gasoline and diesel.

To introduce a biodiesel element into a diesel fuel, it is therefore known to add to a conventional mineral oil a biomass component. Vegetable oils such as rape seed oils are currently the most common biomass added to biofuels. Animal fats and fish oils are also capable of providing oils for biofuels.

The most common method for making biodiesel from oils and fats is through a chemical process called transesterification whereby the triglycerides obtained from plants (oil seed rape, soy bean, palm oil and many others) and animal fats are reacted with methanol using an alkaline catalyst. The process leaves behind two products - methyl esters (the chemical name for biodiesel) and glycerol (also known as glycerin). The glycerol is a 10% by-product of biodiesel manufacture. In addition, glycerol is also a by product from the hydrolysis and transesterification of triglycerides to produce long chain acids. Due to the enormous quantities of biofuels now being made, the amount of glycerol available on the market has increased enormously in recent times making the price of this material very low.

As glycerol is effectively a byproduct of biofuel formation, it would be very useful of glycerol could itself be used as a biofuel. The inventors have now realised that glycerol can be converted into hydrogen and hence serve as a raw material for biofuel formation. Hydrogen created from this residual glycerol can be considered a biofuel as it derives from a renewable source such as a plant or animal. In this way, the whole of the triglyceride molecule can be used to provide energy.

Hydrogen is, of course, of massive economic value in view of fuel cell technology. Fuel cells have emerged as one of the most promising new technologies for meeting future global energy needs. In particular, fuel cells that consume hydrogen are proving to be environmentally clean, quiet, and highly efficient devices for power generation. However, while hydrogen fuel cells have a low impact on the environment, the current methods for producing hydrogen require high- temperature steam reforming of non-renewable hydrocarbon fuels.

Hydrogen fuel cells are highly efficient devices for the production of electrical power. However, realization of hydrogen society for sustainable society development depends significantly on sustainable hydrogen production to achieve CO ₂ neutrality. There are two ways to achieve CO ₂ neutral hydrogen production to achieve full environmental benefit, one is the existing technologies for production of H ₂ by steam reforming or partial oxidation of non renewable fossil fuels with carbon dioxide sequestration, and another way is to produce hydrogen from renewable resources, such as water (by the action of sunlight) or biomass (by catalytic conversion). Renewable biomass is therefore an attractive alternative to fossil

feedstocks because it has essentially zero carbon dioxide impact. However, the effectiveness of biomass as a hydrogen source depends critically on integrated processes to use the residuals of the biomass conversion process to make economic sense. Vegetable oils have a better potential for producing hydrogen than lignocellulosic feedstocks, but their high costs make the process economics unfavourable. Lignocellulosic biomass has too low a hydrogen content, and direct conversion such as gasification cannot compete with existing technology for hydrogen production, the well-developed technology for steam reforming natural gas. The present inventors have therefore sought hydrogen production from low value residuals from a biomass conversion process such as bio-oils from fast pyrolysis or carbohydrate derived materials from bio-oils, by-products in pulping industry and biodiesel production.

The present inventors have realised that glycerol offers an ideal biomass residual for conversion to hydrogen.

The conversion of glycerol into hydrogen is not new although no one has yet realised the enormous potential of this reaction as never before has glycerol been such an abundant and cheap chemical. Before the biofuel revolution, glycerol would have been a prohibitively expensive fuel source. US2003/0170171 describes the production of hydrogen from, inter alia, glycerol in a condensed phase at relative low temperature, e.g. below 400°C. The reaction is catalysed by known transition metal catalysts. The yields of hydrogen in this process are, however, disappointing and there remains a need to provide better processes for the production of hydrogen from biomass derived liquids such as glycerol in high purity and selectivity.

This problem is solved by the process of this invention where sorption enhanced reforming is used, ideally on biomass residuals, in particular glycerol, to give highly pure hydrogen.

Viewed from one aspect therefore the invention provides a sorption enhanced reforming process for the conversion of at least one polyol to hydrogen comprising contacting water (e.g. in the form of steam) and a polyol in the presence of a transition metal catalyst and a carbon dioxide acceptor.

Viewed from another aspect the invention provides a reforming process for the conversion of at least one polyol to hydrogen comprising contacting water (e.g. in the form of steam) and a polyol in the presence of a transition metal catalyst and a carbon dioxide acceptor wherein the purity of hydrogen obtained is greater than 95%.

Viewed from another aspect the invention provides a sorption enhanced reforming process for the conversion of glycerol to hydrogen comprising contacting water (e.g. in the form of steam) and glycerol in the presence of a transition metal catalyst under conditions of at least 450 ⁰ C. Viewed from another aspect the invention provides a sorption enhanced reforming process for the conversion of glycerol to hydrogen comprising contacting steam and glycerol in the vapour phase in the presence of a transition metal catalyst.

Viewed from another aspect the invention provides the use of at least one polyol in a sorption enhanced reforming process for the conversion of at least one polyol to hydrogen.

Viewed from another aspect the invention provides the use of a byproduct of triglyceride hydrolysis and transesterification, preferably glycerol, as a substrate for sorption enhanced reforming into hydrogen.

The term reforming is used herein to mean the process of reacting the substrate (here the polyol) with water (in the form of steam) to give hydrogen and carbon dioxide.

Sorption enhanced steam reforming occurs where the carbon dioxide produced during reforming is absorbed by a carbon dioxide acceptor to force the equilibrium of reforming to the product hydrogen side. By polyol is meant a compound comprising at least two hydroxyl groups.

Preferably, polyols which are reformed in this invention will contain atoms of carbon, hydrogen and oxygen only.

The polyol to be reformed according to the invention can be a single polyol or a mixture of polyols. Preferably the polyols will have fewer than 12 carbon atoms, preferably fewer than 6 carbon atoms, e.g. 2 to 4 carbon atoms. Highly preferred polyols are mono or disaccharides such as glucose, sucrose or sorbitol, ethylene glycol and glycerol. In the most preferred embodiment this invention

concerns the reforming of glycerol, especially glycerol which is formed as a byproduct of triglyceride hydrolysis and transesterification. Triglyceride hydrolysis and transesterification is a crucial step in the formation fatty acids from biosources and to make biofuels. The transition metal catalyst used to catalyse the reforming reaction, e.g. reforming reaction, preferably comprises at least one group 6 to 10 transition metal, especially a group 8 to 10 transition metal such as Fe, Ru, Os, Co, Rh, Ir, Ni, Pd or Pt. Most preferably the transition metal is from group 9 or 10. Rh and Pt are perhaps the most preferred catalysts in terms of selectivity but these metals are rare and expensive. The inventors have found that Ni and Co are also able to provide excellent results and are comparatively cheap. Most especially, therefore the catalyst is Ni or Co. Mixtures of any of these transition metals could also be employed, in particular a Ni/Co mixture.

Where a mixture of two metals is employed they may be combined in any ratio. Preferred weight ratios include 1 :10 to 10:1, e.g. 1 :5 to 5:1, preferably 1 :3 to 3:1, e.g. about 1 :1.

As will become apparent below, it is also preferred if the catalyst to be used in the process of the invention contains at least one noble metal. According to the invention, noble metals are those that are resistant to oxidation and which also act as a catalyst for steam reforming. Examples include Rh, Pt, Au, Pd, Ag, Tc, Os, Ru, and Re..

The amount of noble metal can be less than 10 wt% of the total amount of metal catalyst present (in terms of metal weight), preferably less than 5 wt%.

It is preferred if the transition metal catalyst is supported. The support should be one that provides a stable platform for the chosen catalyst and the reaction conditions. Supports include silica, alumina, zirconia, titania, magnesia, carbon, lanthanum (III) oxide and mixed supports such as silica-alumina, silica nitride and boron nitride. Furthermore, nanoporous supports such as zeolites, carbon nanotubes, or carbon fullerene may be utilized. In a highly preferred embodiment, the support is a hydrotalcite. Hydrotalcite itself is an anionic clay mineral which, in its natural form, is a hydroxy carbonate of magnesium and aluminium. These clays contain exchangeable anions and whilst

hydrotalcite does occur naturally, it is simple to prepare synthetically. In these synthetic hydrotalcites the magnesium can be replaced by other M ²⁺ ions such as Co or Ni. In this way therefore, catalyst material can be loaded onto the hydrotalcite support in potentially very high loadings. The art is replete with methods of synthesisng these materials.

Preferred catalyst carrying hydrotalcites are of formula (I)

[Ni _x Co _y M _gw Al _z (OH) _q ] 2 ^+ , CO ₃ . H ₂ O (I)

wherein w, z and q are positive numbers; at least one of x and y is a positive number; and w, x, y, z and q add up so that the cation has an overall charge of 2+. In this formula, the number of waters of crystallisation is unspecified, i.e. the formula covers all amounts of water of crystallisation. Preferably w, x, y and z have a value up to 5, more preferably up to 4. The subscript q may have a value up to 20. Preferably q is 16. The subscript z is preferably 2. Preferably the total of x, y, w and z is 8. In a further embodiment z/(x+y+w+z) is preferred to be 0.25

A catalyst of this structure is new and forms a further aspect of the invention. Thus, viewed from another aspect the invention provides a catalyst for reforming of at least one polyol, e.g. glycerol comprising a compound of formula (I)

[Ni _x Co _y M _gw Al _z (OH) _q ] ²⁺ CO ₃ . H ₂ O (I)

wherein w, z and q are positive numbers; at least one of x and y is a positive number; and w, x, y, z and q add up so that the cation has an overall charge of 2+. In a further embodiment z/(x+y+w+z) is between greater than 0 and 0.3. In one embodiment, z/(x+y+w+z) is 0.25. Preferably the content of Ni and/or Co will be no more than 60 wt% of the content of total catalyst ((Ni+Co)/(Ni+Co+MgO+ Al ₂ O ₃ ). More preferably the content of Ni and/or Co will be no more than 50 wt%, e.g. no more than 40 wt%,

e.g. 40 wt%. .These percentages are given as wt% in a reduced catalyst base, which is common in catalysis.

The support material will typically be calcined. Calcination involves heating the material at a temperature of from 300 to 800°C, preferably 400 to 800°C, more preferably 500 to 700 ⁰ C, especially 550 to 600°C. Calcination preferably takes place in air flow. It is to reduce the catalyst before reforming. This can be achieved using hydrogen.

Once the temperature is above 200 ⁰ C hydrotalcites dehydrate and at 450 ⁰ C the hydroxide dehydrates. The support material can be loaded with high levels of metal catalysts. The total metal loading of the catalyst may be in the range 10 to 60 wt%, e.g. 20 to 50 wt%, preferably 25 to 45 wt%. This means that the weight of loaded catalyst in terms of metal can be 10 to 60 % etc of the weight of the catalysts, i.e. if you have lOOg of reduced catalyst, 60 wt% loading means 60 g of metals as active catalysts for the reforming.

It has been found also that the particle size of the loaded catalyst is important. It is preferred if the catalyst particles are nanoparticulate. By nanoparticulate means that the particles of the catalyst are nanoparticles, i.e. less than 80 nm, preferably less than 30 nm, especially less than 15 nm in diameter. Most preferably, the particles are around 2 to 50 nm in diameter, e.g. 10 to 30 nm, especially 10 to 25 nm in diameter. Particles diameters can be measured using well known techniques such as electron microscopy or chemisorption. The nanoparticles are preferably crystalline. Nanoparticles have been found to minimise coking.

It has also been found that hydrotalcite supports also minimise coking due to its basic nature and the combination therefore of a nanoparticle catalyst loaded onto a hydrotalcite support is an ideal material to minimise coking. In particular the combination of a Ni or Co or Ni/Co catalyst loaded onto a hydrotalcite support is advantageous.

The support material itself is preferably mesoporous. Preferably, the mesoporous support material contains pores with diameters between 2 and 50 nm.

The synthesis of these materials is well known. In the present invention the use of a coprecipitation method can be used, as is well known in the art. The

method may involve the coprecipitation of solutions of bivalent and trivalent metals in the presence of an anion. A summary of coprecipitation synthesis can be found in "Catalysis Today" 11 (2), (1991), pages 173 to 301.

Where a noble metal is employed, the noble metal may be introduced into the calcinated hydrotalcite derived catalyst by an incipient wetness method. The pore volume of the catalyst can be measured first by water titration or N ₂ adsorption. The noble metal precursor may then be dissolved in a solution having the same volume as the pore volume of the hydrotalcite derived catalyst. After contact between the catalyst and solution is made, the samples can then be dried and calcinated.

Viewed from another aspect the invention provides a process for reforming a substrate to hydrogen comprising contacting water (e.g. in the form of steam) and a said substrate in the presence of a transition metal catalyst as hereinbefore defined. Useful substrates include polyols and hydrocarbons. The steam reforming reaction of a polyol will primarily be described with reference to glycerol although it will be appreciated that the chemical principles are the same irrespective of the polyol employed.

The ideal reaction for the reforming of glycerol to hydrogen is

C ₃ H ₈ O ₃ + 3H ₂ O → 3CO ₂ + 7H ₂

This reaction is highly endothermic. The actual reactions taking place are complex but can be represented by the unbalanced equations:

Reforming: C ₃ H ₈ O ₃ + H ₂ O → CO + H ₂ Water gas shift: CO + H ₂ O → CO ₂ + H ₂

All these reactions are reversible. To maximise the yield of hydrogen, it is preferred to force the equilibrium of this reaction to the right (products side). This is achieved in the invention by absorbing the carbon dioxide which is formed hence the process of the invention preferably involves sorption enhanced reforming of polyols.

The process of the invention therefore concerns sorption enhanced steam reforming and therefore utilises a carbon dioxide acceptor to drive the equilibrium of the above reaction towards hydrogen formation.

By carbon dioxide acceptor is meant a compound which reacts reversibly with carbon dioxide thereby extracting it from the atmosphere in the process of the invention and driving the equilibrium of reaction towards hydrogen formation.

Any carbon dioxide acceptor can be used in the invention. Known carbon dioxide acceptors include metal oxides, in particular transition metal oxides such as calcium oxide and mixed metal oxides of general formula XYO ₂ , XYO ₃ , XYO ₄ , XY ₂ O ₄ , or X ₂ YO ₄ where the first metal ion X is preferably selected from groups I or II of the periodic table, i.e. is an alkali metal or alkaline earth metal, or a transition metal in the I ⁺ , 2 ⁺ or 3 ⁺ oxidation state. The second metal ion Y is preferably from the transition metal or lanthanide series of metals or is an ion of Al, Si, Ga, Ge, In, Sn, Tl, Pb or Bi. Metal ions X and Y must be different. Preferably, the metal ion X is an ion of Li, Na, Mg, K, Ca, Sr or Ba. Most preferably, the metal ion X is lithium, sodium or calcium, especially calcium. Most preferably the Y metal ion is a transition metal of groups 4 to 6, especially aluminium and titanium or most especially zirconium. These materials may also be doped, e.g. with alkali or alkaline earth metal ions, preferably Na ⁺ or K ⁺ . A description of these types of acceptor can be found in WO2006/111343.

Preferably however, the carbon dioxide acceptor will comprise calcium oxide as the main active component, preferably as part of a mixed oxide. This improves stability. Calcium oxide can be readily obtained from the calcination of the mineral dolomite. The carbon dioxide acceptor of the invention is preferably nanoparticulate. By nanoparticulate means that the particles of acceptor formed by the process of the invention are nanoparticles, i.e. less than 500 nm, preferably less than 300 ran, especially less than 100 nm in diameter. Most preferably, the particles are around 2 to 80 nm in diameter, e.g. 10 to 50 nm, especially 10 to 25 nm in diameter. Particle diameters can be measured using well known techniques such as electron microscopy or XRD. The nanoparticles are preferably crystalline.

40

It is believed that the use of nanoparticulate acceptors improves the stability and the kinetic ability of the acceptor to capture carbon dioxide and improves the ease of regeneration of the acceptor.

The nanoparticles may coagulate to form larger porous particles normally with relatively uniform size between 1-2 μm.

The carbon dioxide acceptor reacts reversibly with carbon dioxide. The following description of the carbon dioxide acceptance procedure is based on the use of calcium oxide however the chemical principles apply to any acceptor.

The reversible reaction of calcium oxide and carbon dioxide is represented by the equation:

CaOO) + CO ₂ (g) → CaCO ₃ (S)

This reaction is highly exothermic. Overall, however, when the reactions involved in sorption enhanced steam reforming are considered:

C ₃ H ₈ O ₃ + 3H ₂ O → 3CO ₂ + 7H ₂ AH ₂₉₈ =I 27 kJ/mol

CaOO) + CO ₂ (g) → CaCO ₃ O) AH ₂₉₈ =- 178kJ/mol the overall energy change is slightly exothermic. This means that the heat of carbon dioxide absorption can be used to allow the reforming step to take place without any external heat being provided. Capture of carbon dioxide and the steam reforming process preferably take place between 400 and 700°C, e.g. 450°C to 700 ⁰ C. The process of the invention preferably takes place at temperatures in the range of 500 to 700°C, more preferably in the range 550-650 ⁰ C.

The carbon dioxide acceptors of the invention are able to capture at least 8 wt%, preferably at least 10 wt%, more preferably at least 12 wt%, especially at least 15wt% of their own weight of carbon dioxide, highly preferably at least 20 wt% of their own weight of carbon dioxide. Typically, an acceptor will be able to absorb carbon dioxide for at least 15 mins, e.g. 15 mins to 1 hour before regeneration is required. It will be appreciated therefore that under these capture conditions, the process of the invention involves steam reforming of polyols.

41-

The ratio of carbon dioxide acceptor to catalyst is important. The theoretical ideal ratio can generally be estimated by calculating the balance of the heat of Ni and Co oxidation with the heat of the CO ₂ acceptor regeneration, so that there is no requirement of external heating for acceptor regeneration. The ratio depends, inter alia, on issues such as catalyst metal loading, catalyst composition, capacity of the CO ₂ acceptors, reforming conditions and process heat integration. This ratio will be an important parameter for process design to optimize the process. The ratio of catalyst and CO ₂ acceptors can be reduced by good heat integration.

It is generally preferred, however, if the amount of catalyst (in terms of the weight of catalyst metal, e.g. Ni or Co) exceeds the weight of acceptor, i.e. the weight ratio is greater than 1 :1. Preferably the weight ratio is 10 parts catalyst to 1 part acceptor to 1 :1, e.g. 5:1 to 1 : 1 , more preferably 3 : 1 to 1 :1, especially 2:1 to 1 :1. Once a carbon dioxide acceptor has absorbed all the carbon dioxide which it is capable of absorbing, it can no longer work as an acceptor. At this point therefore, the equilibrium of the reforming reaction would establish itself and other products would start to form such as carbon monoxide. Methane is also a possible by product at this point as it can form via a methanation. This is obviously undesirable.

Once saturation of the acceptor does occur therefore, the reforming reaction can be switched to another reactor where there is fresh acceptor. The saturated acceptor can then be regenerated ready to be used again when the receptor in the second reactor becomes saturated. It will be appreciated therefore that any number of reactors can be set up in parallel to run the process of the invention.

By regeneration is meant reversing the carbon dioxide acceptance reaction to give carbon dioxide and carbon dioxide acceptor. Thus, it is envisaged that the exhaust gas containing carbon dioxide passes over the carbon dioxide acceptor in one reactor, e.g. a fluidised bed reactor. Once the acceptor had taken its full amount of carbon dioxide the acceptor can be regenerated and the released carbon dioxide captured and stored. Meantime, the reforming reaction and hence exhaust gas can be transferred to a second reactor to continue the reforming process. Once the second acceptor has taken its full amount of carbon dioxide, it too can be regenerated whilst the reforming reaction returns to

42

the first reactor and so on. It will be clear that the principles of this process can be expanded to any number of reactors.

Regeneration of the acceptor can be achieved in any convenient fashion. For example, regeneration of the acceptor can be carried out using an inert gas or using high temperature steam. In such a process, the acceptor is exposed to steam at the temperatures above (e.g. 500 to 800°C), especially 550 to 650 ⁰ C. During the regeneration process carbon dioxide is released and can be stored or vented. Thus, it is possible for the capture process to be stopped, the acceptor regenerated and capture to be restarted without having to remove the acceptor from its location or to significantly change the temperature of reaction. As regeneration can occur in the same time frame or faster than absorption, preferably less than 0.5 times of the time of sorption enhanced reforming this allows for successive capture and regeneration steps to be carried out.

The present inventors have found a new and greatly beneficial method of regenerating the acceptor however. The regeneration reaction is highly endothermic and requires an input of significant amounts of heat. Whilst heat could be generated by burning hydrogen, the present inventors have realised that this heat can be provided by allowing an oxygen containing gas to enter the reactor, typically air. The reactor obviously contains the catalyst on which there is metal which, under the high temperatures already present in the reactor, will oxidise rapidly generating much heat. For example,

Ni (s) + Y ₁ 0 ₂ → NiO (δH° ₂₉₈ = -240 kJ/mol)

3Co(s) + 2O ₂ (g) → Co ₃ O ₄ (δH° ₂₉₈ = -910 kJ/mol) CaCO ₃ 0) → CaOO) + CO ₂ (g) (δH° ₂₉₈ = 178 kJ/mol)

The heat required for the regeneration process can therefore be supplied by oxidation of the catalyst. The oxidation of 1 mol Co can, for example regenerate about 1.8 mol CaCO ₃ . This forms a further aspect of the invention.

Thus, viewed from another aspect the invention provides a process for the regeneration of a carbon dioxide acceptor that has absorbed carbon dioxide in the presence of a transition metal catalyst comprising exposing said catalyst to an

43-

oxygen containing gas, preferably air, causing oxidation of at least a part of the transition metal of said catalyst; using the heat generated by said oxidation reaction to regenerate said carbon dioxide acceptor by reversing the carbon dioxide acceptance reaction to form carbon dioxide and carbon dioxide acceptor.

Viewed from another aspect therefore the invention provides a sorption enhanced reforming process for the conversion of at least one polyol to hydrogen comprising contacting water (e.g. in the form of steam) and a polyol in the presence of a transition metal catalyst and a carbon dioxide acceptor; allowing said carbon dioxide acceptor to absorb carbon dioxide formed during said reforming process; contacting said transition metal catalyst with an oxygen containing gas to oxidise said transition meal generating heat; using the heat of oxidation to regenerate said carbon dioxide acceptor. The oxygen containing gas, preferably air, used for regeneration plays a multi-functional role as it not only provides heat for regeneration by oxidation of the catalyst but also provides an atmosphere having a very low carbon dioxide content. This encourages therefore the equilibrium of CaCO ₃ (S) — > CaO(s) + CO ₂ (g) reaction to move to the right. In contrast, during reforming, the atmosphere in the reactor has a very high level of carbon dioxide forcing the equilibrium of the towards left (i.e. towards calcium carbonate in the equation above). Using air as an oxygen containing gas lowers the regeneration temperature, as nitrogen in the air lowers the partial pressure of CO ₂ in the regeneration reactors. This factor combining with the sufficient heat supply by in-situ oxidation makes it possible to use almost identical temperature in both reforming and regeneration reactors, which makes continuous operation viable.

Regeneration is preferably carried out at the same temperature as the reforming reaction e.g. 400 and 700°C, preferably 450 ⁰ C to 700 ⁰ C. The regeneration process of the invention preferably takes place at temperatures in the range of 500 to 700 ⁰ C, more preferably in the range 550-650°C.

The problem with this technique for regeneration is of course that once regeneration has occurred, the metal ions of the catalyst are in the form of an oxide

44

and are no longer capable of catalysing the steam reforming reaction. It is necessary therefore to reduce the oxides back down to the metal. Remarkably, the inventors have realised that this can be achieved using hydrogen. As the whole process produces hydrogen, there is an abundance of this gas available to reduce the oxidised catalyst. This therefore forms a still yet further aspect of the invention.

Viewed from another aspect therefore the invention provides a sorption enhanced reforming process for the conversion of at least one polyol to hydrogen comprising contacting water (e.g. in the form of steam) and a polyol in the presence of a transition metal catalyst and a carbon dioxide acceptor; allowing said carbon dioxide acceptor to absorb carbon dioxide formed during said reforming process; contacting said transition metal catalyst with an oxygen containing gas to oxidise said transition metal generating heat; using the heat of oxidation to regenerate said carbon dioxide acceptor; reducing said oxidised transition metal catalyst using hydrogen.

The whole process can then restart.

The reduction of the oxidised transition metal catalyst can take place in the reactor in a hydrogen atmosphere where said hydrogen is recycled from the hydrogen produced by the reforming reaction. It is also possible simply to restart the reforming reaction in the presence of the oxidised catalyst. Polyol will partially reduce the oxidised metal to reform enough polyol to generate hydrogen which will then reduce the rest of the oxidised transition metal catalyst.

The inventors have realised however that there exists a problem reducing in this fashion. After regeneration has been completed and the switch between supplying the oxygen containing gas to the reactor and returning to supplying steam and polyol has taken place to restart the reforming process the catalyst is oxidised. As the catalyst has been oxidised, there is a short induction time whilst the catalyst is reduced back down to the metal by hydrogen generated by initially limited reforming. During this induction time, there is insufficient active catalyst for complete reforming and the product gases will become contaminated with unreacted water and perhaps carbon monoxide and methane.

45

The inventors have realised that an induction time can be avoided if the catalyst is doped with small amounts of noble catalyst metals such as Pt and Pd. These noble metals are too expensive to use as the main elements of the catalyst but can be used in small amounts to prevent induction time. These metals will begin the reforming process whilst the bulk of the oxidised catalyst is reduced. Moreover, these metals are noble and hence will not themselves undergo oxidation in the presence of oxygen during the regeneration step.

This forms a further aspect of the invention. Thus, viewed from a further aspect the invention provides a catalyst for reforming at least one polyol comprising at least one transition metal Ni and/or Co and at least one noble metal supported on hydrotalcite.

Viewed from another aspect the invention provides a sorption enhanced reforming process for the conversion of at least one polyol to hydrogen comprising contacting water (e.g. in the form of steam) and a polyol in the presence of a transition metal and noble metal catalyst and a carbon dioxide acceptor; allowing said carbon dioxide acceptor to absorb carbon dioxide formed during said reforming process; contacting said transition metal catalyst with an oxygen containing gas to oxidise said transition meal generating heat; using the heat of oxidation to regenerate said carbon dioxide acceptor; restarting said reforming process for the conversion of at least one polyol to hydrogen by contacting water (e.g. in the form of steam) and a polyol in the presence of said oxidised transition metal and noble metal catalyst and regenerated carbon dioxide acceptor wherein said oxidised transition metal catalyst is reduced by hydrogen and the reforming process initiated by the non-oxidised noble metal catalyst.

The feed to the reactor where the steam reforming occurs needs to contain polyol, e.g. glycerol and water, normally in the form of steam. The molar ratio of water to polyol is preferably as near to stoichiometric as possible, e.g. around 1 :3 polyol to steam, ideally 1 :3.1. It may be however that excess water is required to further improve hydrogen purity although preferably the molar ratio should not exceed 1 :10, preferably no more than 1 :5.

46

The production of hydrogen according to the process of the invention allows the isolation of hydrogen in incredible purity without any separation procedure. Purity greater than 95%, e.g. at least 97%, more preferably at least 98%, especially at least 99% can be achieved (in terms of dry gas purity). A purity of greater than 95% means therefore that the content of hydrogen in the dry exhaust gas from the process of the invention is greater than 95% without any separation procedures taking place. As all the carbon dioxide is absorbed and CO converted to carbon dioxide, hydrogen is, in theory the only gas which should be present. Note this value is based on dry gas (and excludes therefore steam). It is a remarkable feature of the invention that the reactors described herein are almost perfectly balanced thermally. This means that the entire reforming process can be run without an external heat source. All the heat required can be generated by the chemical reactions described above.

The present invention, in view of its energy balance, lends itself therefore to deployment not just in power stations but also in homes and offices. Small scale reformers are a key technology for the early stages of a hydrogen economy.

It is also a feature of the invention that the reforming process produces a product, hydrogen, which can be used as a fuel to provide heat if necessary. If a need arises to generate heat over and above that occurring during the processes described above, hydrogen can simply be oxidised to generate heat and steam in situ.

The hydrogen produced by the process of the invention may be used in fuel cell technology or for another other use of hydrogen gas. A CO content lower than 1%, e.g. less than 0.5%, more preferably less than 0.1%, especially less than 0.05% can be achieved (in terms of dry gas purity). The hydrogen produced by the process can be directly used as feed in high temperature fuel cell such as high temperature PEM fuel cell and solid fuel cell. The relatively pure hydrogen will increase the efficiency of the fuel cell and reduce the size of the fuel cell. Many fuel cell technologies will benefit from the hydrogen production technology. Viewed from another aspect the invention provides process comprising

(I) a sorption enhanced reforming process for the conversion of at least one polyol derived from biomass to hydrogen;

47

(II) transferring said hydrogen into a fuel cell.

The fuel cell is preferably a high temperature fuel cell, e.g. one operating at least 120 ⁰ C, preferably at least 150°C. Since the hydrogen produced by the sorption enhanced reforming process of the invention is so pure, it makes an ideal fuel for a high temperature fuel cell, especially in view of its low CO and CO ₂ concentration.

The whole process of the invention can then be autothermal.

The process of the invention has been described in relation to polyol derived biomass but there is no reason for these biomasses to be mixed with other biomass. The equipment required to carry out the process of the invention is known.

Steam reforming of methane is a well known process and the reactors used for that process are suitable for use in the present invention. If desired the reactor can be positioned within a furnace and reactants introduced into the reactor via a pump. Reactants can be conveyed in an inert gas if necessary. Mass flow regulators can be used as is non in the art to control flow rates and partial pressures within the reactor. The product hydrogen gas stream can be cooled using a heat exchanger and can be sent to a separator to remove any minor non hydrogen contaminants if desired.

Whilst the invention has been described primarily with reference to the reforming of polyols it will be appreciated that the sorption enhanced steam reforming reaction described herein has utility with a wide variety of starting materials, in particular bioderived materials including all the liquids and gases from biomass, such as bioliquids, biogas and bio-synthesis gas from biomass gasification. Other substrates on which the reforming reaction of the invention could be carried out therefore include hydrocarbons, in particular those which are non gaseous at ambient temperature and have a boiling point of less than 36O ⁰ C and which are therefore components of diesel and gasoline. The reforming reaction of the invention could also be carried out on other biomaterials such as oxygenated biomaterials like acids and aldehydes and monoalcohols such as ethanol. The use of other substrates is particularly preferred where the sorption process of the invention involves an oxygen containing gas regeneration step. Mixtures of any of these feeds can also be used.

48-

Thus viewed from another aspect the invention provides a sorption enhanced reforming process for the conversion of at least one substrate to hydrogen comprising contacting water (e.g. in the form of steam) and the substrate in the presence of a transition metal catalyst and a carbon dioxide acceptor; allowing said carbon dioxide acceptor to absorb carbon dioxide formed during said reforming process; contacting said transition metal catalyst with an oxygen containing gas to oxidise said transition metal generating heat; using the heat of oxidation to regenerate said carbon dioxide acceptor. The substrate used here may be a hydrocarbon, alcohol, polyol, oxygenated hydrocarbon. Suitable hydrocarbons may have up to 15 carbon atoms, preferably up to 12 carbon atoms.

Viewed from another aspect the invention provides a sorption enhanced reforming process for the conversion of a substrate to hydrogen comprising contacting water (e.g. in the form of steam) and a substrate in the presence of a transition metal catalyst and a carbon dioxide acceptor; allowing said carbon dioxide acceptor to absorb carbon dioxide formed during said reforming process; contacting said transition metal catalyst with an oxygen containing gas to oxidise said transition metal generating heat; using the heat of oxidation to regenerate said carbon dioxide acceptor; and reducing said oxidised transition metal catalyst using hydrogen.

Viewed from another aspect the invention provides a sorption enhanced reforming process for the conversion of a substrate to hydrogen comprising contacting water (e.g. in the form of steam) and a substrate in the presence of a transition metal and noble metal catalyst and a carbon dioxide acceptor; allowing said carbon dioxide acceptor to absorb carbon dioxide formed during said reforming process; contacting said transition metal catalyst with an oxygen containing gas to oxidise said transition meal generating heat; using the heat of oxidation to regenerate said carbon dioxide acceptor;

restarting said reforming process for the conversion of a substrate to hydrogen by contacting water (e.g. in the form of steam) and a substrate in the presence of said oxidised transition metal and noble metal catalyst and regenerated carbon dioxide acceptor wherein said oxidised transition metal catalyst is reduced by hydrogen and the reforming process initiated by the non-oxidised noble metal catalyst.

A major benefit of using biofuels to produce hydrogen which is then used for fuel production is that the resulting fuel will be as environmentally friendly as E95 ethanol fuel. The present invention therefore allows the formation therefore of large amounts of bio-neutral hydrogen. Hydrogen is a very important raw material in the chemical and petroleum industries. Large quantities are used in the manufacture of ammonia and methanol and in a variety of petroleum hydrotreating processes. Hydrogen management has become a priority in current refinery operations and when planning to produce lower sulphur gasoline and diesel fuels. In many refineries, hydroprocessing capacity and the associated hydrogen network is limiting refinery throughput and operating margins. Furthermore, higher hydrogen purities within the refinery network are becoming more important to boost hydrotreater capacity, achieve product value improvements and lengthen catalyst life cycles. The process of the invention maximises hydrogen purity and provides environmentally neutral hydrogen.

Moreover, the process of the invention is capable of reforming feeds from all manner of different sources and these can be mixed together. Methane reforming is already well known by the possibility of using reformers to reform not only methane but any reformable feed is new. This allows more flexible on-purpose hydrogen production in refinery. A mixture of fossil fuels and biofuels can be used as the feedstocks for hydrogen production. The composition of the mixture can be adjusted based on the availability of different fuels and the requirement of the hydrogen in the refinery. The sorption enhanced reforming can use the products from biomass gasification or pyrolysis, however, more preferred feedstocks are the bio products of biorefinery such as glycerol.

Light alkanes or mixed alkanes and alkenes in the refinery off gas can be also used as feedstock for hydrogen production. Another possibility is the gasification of refinery residues following a sorption enhanced reforming. These sources and this process are summarised in Figure 8. The invention will now be described with reference to the following non limiting figures and examples.

Brief Description of the Figures

Figure 1 shows two parallel reforming reactors containing catalyst and carbon dioxide acceptor. Glycerol and steam are supplied to reactor 1 allowing the formation of hydrogen in the product gas stream in very high purity. Carbon dioxide is absorbed by the acceptor in the reactor (not shown). Once the carbon dioxide acceptor is saturated, the glycerol and steam feed is switched into reactor 2 which also contains catalyst and acceptor. The reforming process continues.

Once the switch to reactor 2 occurs, air is allowed to enter reactor 1 generating heat through the oxidation of the catalyst particles and hence allowing the regeneration of the carbon dioxide acceptor. As there is neglible carbon dioxide in the feed to reactor (1) at this point, the equilibrium of the carbon dioxide acceptance reaction is forced to the left releasing carbon dioxide. This may be stored or vented.

Once regeneration is complete, the glycerol and steam is switched back to reactor 1 and regeneration begins in reactor 2. At this point, hydrogen can be supplied to reactor 1, e.g. via a recycle from the product gas stream, to allow reduction of the oxidised catalyst particles. Alternatively, the reforming process can simply be restarted without a hydrogen recycle although a short induction time might occur as reduction of the oxidised catalyst takes place in situ. Induction time can be reduced or avoided by the use of a catalyst containing noble metal. Reforming therefore continues during the reduction process as the catalyst is doped by noble metals which act as reforming catalysts and resist oxidation. Figure 2 shows the sorption enhanced steam reforming of glycerol in the presence of calcium oxide using a 30% Ni- 10% Co catalyst supported on

hydrotalcite. After approximately 25 mins, the carbon dioxide acceptor becomes saturated and conversion to hydrogen reduces as the natural equilibrium establishes.

Figure 3 shows acceptor regeneration in air. The regeneration process occurs rapidly over a period of around 10 minutes. It will be seen that the oxygen content in the reactor is very low until 10 mins indicating that it is all reacting. At 10 mins, the oxygen content begins to increase indicating oxidation is complete. After 20 min, inert gas Ar used to remove oxygen in the reactor.

Figure 4 shows the sorption enhanced steam reforming of glycerol after air regeneration of the acceptor and 2 minute reduction of the catalyst. It will be seen that there is a short induction time whilst the reduction of the oxidised catalyst occurs. The overall hydrogen purity is slightly less than in Figure 2 possibly due to incomplete reduction of the catalyst.

Figure 5 shows the results of repeated sorption and regeneration cycles. The material is shown to be stable over many cycles. Figure 6 shows changes in Hydrogen, CO, CO ₂ , and CH ₄ mole fraction with time on stream during sorption enhanced reforming of sorbitol on 30%Ni-10%Co catalyst at 575°C

Figure 7 shows changes in Hydrogen, CO, CO ₂ , and CH ₄ mole fraction with time on stream during sorption enhanced reforming of glucose on 30%Ni-10%Co catalyst at 550°C.

Figure 8 shows the various feedstock sources available in a refinery.

Examples

Catalyst Preparation A homologous series of hydrotalcite-derived Ni-Co catalysts were prepared by co-precipitation. The total metal loading was fixed at 40 wt% of total amount of reduced catalyst and Ni-Co composition varied from 40-0, to 30-10, 20-20, 10-30 and 0-40. It was found that the 40% Ni sample had the largest surface area. The samples were calcined at 600 ⁰ C at a heating rate of 5 °C/min and kept at the target temperature for a period of 6 h. After the calcination the catalyst was crushed and sieved to a particle size of 250 - 500 μm.

Table 1 : Prepared hydrotalcite catalyst and their general formula

Table 2: BET area of hydrotalcite derived materials after calcination

The calcined catalyst was reduced before the reaction in a mixture of Ar and hydrogen at 670 ⁰ C for 1O h, using a heating rate of 2 ⁰ C /min from ambient temperature to 670 ⁰ C. After reduction, the temperature was decreased to 575°C and the reactive mixture was introduced in the reactor.

Doping with noble metals

Noble metals are introduced to the calcinated hydrotalcite derived catalysts by an incipient wetness method. The pore volume of the catalyst is measure by water titration or N ₂ adsorption. The noble metal precursor is then dissolved in a solution of the same volume of the pore of the hydrotalcite derived catalyst. The solution is brought into contact with the hydrotalcite derived catalyst. The sample is then dried in a vacuum oven at 100°C for 1O h, calcinated at 600 ⁰ C at a heating rate of 5 °C/min and kept at the target temperature for a period of 6 h.

Example 1

Sorption Enhanced Glycerol Reforming

5g calcinated hydrotalcite derived 30%Ni-10%Co catalysts and 2.5g calcinated dolomite (Arctic dolomite (CaMg(CO ₃ ) ₂ ), both with a particle size range of 250 - 500 μm were installed together in a fixed bed reactor. The calcined catalyst was reduced before the reaction in a mixture of Ar and hydrogen at 670 ⁰ C for 1O h, using a heating rate of 2 K/min from ambient temperature to 670 ⁰ C. After reduction, the temperature was decreased to 575°C and the reactive mixture was introduced in the reactor.

A mixture of glycerol and water with a molar ratio of (1 :9) was fed into an evaporator with a flow rate of 5 g/h by a liquid mass flow controller. Inert gas of Ar (40 ml/min) was introduced into the line to carry the liquid into the evaporator. The liquids were vaporized at 400°C and introduced into the reactor.

The reaction was controlled at 575°C. The composition in the effluent obtained by GC analysis are plotted as function of time on stream in Fig. 2. More than 99.5% hydrogen was produced. It decreases when the solid acceptors become saturated.

Example 2 Regeneration

The carbon dioxide acceptor of example 1 was regenerated in situ. Regeneration was performed at 575°C using air 300 mL/min. Fig. 3 indicates that all the oxygen in air introduced into the reactor was initially consumed by the oxidation of Ni-Co metallic nanoparticles to supply the heat for the regeneration of CaCO ₃ . Carbon dioxide is then released by the acceptor.

Example 3

Glycerol Reforming after air regeneration

Sorption enhanced reforming of glycerol was performed after regeneration with air, under the conditions identified in example 1. A pre-reduction in hydrogen at 575°C was performed for two minutes before the reaction is started. Fig. 4 indicate a slightly lower purity of hydrogen compared to the example 1, most likely due to incomplete reduction of the metal oxides.

Example 4

l.lg 20%Ni-20%Co hydrotalcite and 2.6g synthesized CaO based carbon dioxide acceptor were installed together in the fixed bed reactor. The flow rate of glycerol/water mixture was 5 g/hr and the ratio of glycerol and water was 1 :3.1. The cycles 1-3 were performed at 575 ⁰ C, cycles 4-8 were performed at 525°C, and cycles 9-22 were performed at 550°C. Regeneration of sorbent was performed with oxidation of Co and Ni at an identical temperature to the sorption enhanced reforming reaction. Results are presented in Fig 5.

The hydrogen purity is larger than 95% at all the temperatures and the hydrogen purity is slightly lower at 525°C compared to 550 and 575°C. The conversion of glycerol is not 100% at 525°C. Temperature is not sensitive to the hydrogen purity in the range of 550 and 575°C , but the kinetics of solid acceptor for CO ₂ capture is larger at 550°C, and no oxygenates were found in condensed water. The temperature of 55O ⁰ C is preferred for hydrogen production with a high CO ₂ capture capacity of solid acceptors. Both catalysts and solid acceptors were found to be stable during the cyclic tests more than 10 days.

The results indicate that hydrogen can be produced at a low ratio of glycerol and water (1 :3.1), close to the stoichiometric number, which provides a lot of advantages including high thermal efficiency, enhanced reaction and CO ₂ capture due to high partial pressures of glycerol and CO ₂ generated in the reaction. It enhanced also the material stability due to low concentration of steam.

Example 5

Sorption Enhanced Sorbitol Reforming

2 g calcinated hydrotalcite derived 30%Ni-10%Co catalysts and 17.6 g calcinated dolomite (Arctic dolomite (CaMg(CO ₃ ) ₂ ), both with a particle size range of 250 -

500 μm were installed together in a fixed bed reactor. The catalyst pre-treatment was performed as described in Example 1.

A mixture of sorbitol and water with a molar ratio of (1 : 11 ), corresponding to carbon/steam ratio of (6:11) was fed into an evaporator with a flow rate of 3.47 g/h by a liquid mass flow controller. Inert gas of Ar (20 ml/min) and N ₂ (20 ml/min) was introduced into the line to carry the liquid into the evaporator. The liquids were vaporized at 400°C and introduced into the reactor.

The reaction was controlled at 575°C. The compositions in effluent obtained by GC analysis are plotted as function of time on stream in Fig. 6. More than 99% hydrogen can be produced. It decreases when the solid acceptors become saturated.

Example 6

Sorption Enhanced Glucose Reforming

2 g calcinated hydrotalcite derived 20%Ni-20%Co catalysts and 1O g calcinated dolomite (Arctic dolomite (CaMg(CO ₃ ) ₂ ), both with a particle size range of 250 - 500 μm were installed together in a fixed bed reactor. The catalyst pre-treatment was performed as described in example 1.

A mixture of glucose and water with a carbon to steam ratio of 1 :6 is fed into an evaporator with a flow rate of 16.9 g/h by a liquid mass flow controller. Inert gases of Ar (37.5 ml/min) and N ₂ (21 ml/min) are introduced into the line to carry the liquid into the evaporator. The liquids are vaporized at 400°C and introduced into the reactor.

The reaction is controlled at 550°C. The compositions in effluent obtained by GC analysis are plotted as function of time on stream in Fig. 6. More than 99% hydrogen can be produced. It decreases when the solid acceptors become saturated.