A SOLID AMMONIA STORAGE AND DELIVERY MATERIAL

TITLE

A Solid Ammonia Storage and Delivery Material

BACKGROUND OF THE INVENTION 1. Field of the Invention

The present invention relates to the storage of ammonia in a solid form us¬ ing metal ammine complexes and delivery there from. The invention also relates to the use of ammonia stored in a solid form as the reducing agent in selective catalytic reduction (SCR) of NO_x in exhaust gases from com- bustion processes, to methods for producing such complexes and to an ammonia delivery device comprising such complexes.

2. Description of the Related Art

Current environmental regulations necessitate the use of catalysts in the exhaust gas from automotive vehicles, boilers and furnaces for control of NO_x emission leaving the system. Particularly, vehicles equipped with die- sel engines or other lean burn engines offer the benefit of improved fuel economy, but suffer from the drawback of increased formation of NO_x be¬ ing noxious and which must be eliminated from the exhaust gas. However, catalytic reduction of NO_x using conventional so-called three-way exhaust catalysts for automobiles is impossible because of the high content of oxy¬ gen in the exhaust gas. Instead, selective catalytic reduction (SCR) has proven useful for achieving the required low levels of NO_x in the exhaust gas both in stationary and mobile units. In such systems NO_x is con- tinuously removed from the exhaust gas by injection of a reducing agent into the exhaust gas prior to entering an SCR catalyst capable of achiev¬ ing a high conversion of NO_x. So far, ammonia has proven to be the most efficient reducing agent, which is usually introduced into the exhaust gas by controlled injection of gaseous ammonia, aqueous ammonia or aque- ous urea. In all cases, the amount of reducing agent being dosed has to be very precisely controlled. Injection of a too large amount of reducing agent will cause emission of ammonia with the exhaust gas whereas injec¬ tion of a too small amount of reducing agent causes a less than optimum conversion of NO_x.

In many mobile units powered by combustion engines, the preferred tech¬ nical solution is to use an aqueous solution of urea as the reducing agent since in this way potential hazards or safety issues relating to the transport and handling of liquid ammonia in high pressure containers are eliminated. However, there are several disadvantages related to the use of aqueous urea as the reducing agent. First of all, the use of urea solutions requires the carrying of a relatively large volume in order to provide sufficient amounts of ammonia to allow a vehicle to drive e.g. about 20,000 kilome¬ tres without having to substitute or refill the source of ammonia. In typical systems, an aqueous solution comprising about 30 wt % of urea is pre¬ ferred meaning that about 70 wt% of the content of a container holding the urea solution is used only to transport water. During the decomposition, one molecule of urea forms two molecules of NH₃ and one molecule of CO₂ and thus, ammonia only constitutes roughly 50wt% of the weight of the urea molecule and hence, the concentration of ammonia of the reduc¬ ing agent is very low. Similar concentrations of ammonia can be achieved using aqueous solutions of ammonia as reducing agents. Furthermore, when using solutions, specially designed spray nozzles combined with a precision liquid pump are required to ensure that a) the aqueous urea is delivered to the exhaust system at a desired (and dynamically changing) flow rate and b) the aqueous urea is efficiently dispersed in the gas phase before entering the catalyst. Finally, the aqueous solutions might freeze in cold weather conditions (below minus 11⁰C), or the urea solution may simply form precipitates, which will block the dosing system, e.g. the noz- zle. Altogether, these difficulties may limit the possibilities of using SCR technology in abatement of pollution from NO_x, particularly in connection with mobile units.

Transporting ammonia as a pressurized liquid is hazardous as the con- tainer may burst or a valve or tube might break in an accident giving a dis¬ charge of poisonous/lethal gaseous ammonia. In the case of the use of a solid storage medium containing absorbed/adsorbed ammonia, the safety issues are much less critical since a small amount of heat is required to release the ammonia from the storage medium.

International Patent Publication No. WO 99/01205 discloses a method and a device for selective catalytic NOx reduction in waste gases containing oxygen, using ammonia and a reduction catalyst. According to the method, gaseous ammonia is provided by heating a solid storage medium comprising one or more salts, especially a chloride and/or sulphate of one or more cations selected from alkaline earth metals, and/or one or more transition metals, preferably Mn, Fe, Co, Ni, Cu, and/or Zn, said storage medium being introduced into a container. In the preferred embodiments of WO 99/01205 the cation is Ca²⁺ or Sr²⁺. The inventive method and de- vice are stated to be particularly suitable for use in automobiles.

However, the use of the ammonia storage media known from WO 99/01205 suffers from various draw-backs hampering a wide-spread use in the automotive industry. The vapour pressure of ammonia above a solid salt phase is e.g. about 1 bar at room temperature and atmospheric pres¬ sure for calcium octa ammine chloride and strontium octa ammine chloride complexes rendering the use somewhat complicated due to the high pres¬ sures that must be taken into account in view of the toxicity of ammonia. Having such a high partial pressure of ammonia, the handling and trans- portation of the saturated storage material is still difficult and also danger- ous. Still further, the use of granulated storage materials - stated as pre¬ ferred embodiments in WO 99/01205 - requires measures for preventing the storage material from leaving the container during use in moving vehi¬ cles. Furthermore, a granulated material will have a considerable void be- tween the granules which will drastically reduce the bulk density of the storage material by a factor 1.5 - 2.

It has now been found that these drawbacks may be overcome using a magnesium salt complex according to the present invention providing a compact, light-weight, cheap and more safe storage for ammonia having a very low vapour pressure of ammonia below 0.1 bar at room temperature to be used in the automotive industry.

SUMMARY OF THE INVENTION The present invention relates to a solid ammonia storage and delivery ma¬ terial comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NH₃)_nX_z, wherein M is one or more cations selected from alkaline earth metals, and/or one or more transition metals, such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination num¬ ber of 2 to 12, wherein M is Mg.

In a second aspect the invention relates to a method of producing a solid ammonia storage and delivery material comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula: M₃(NHs)_nXz, wherein M is one or more cations selected from alkaline earth metals, and/or one or more transition metals, such as Mn₁ Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination num- ber of 2 to 12, said method comprising the steps of

1 ) providing the solid salt,

2) saturating the salt with ammonia, and

3) compressing the ammonia salt complex.

In a third aspect the invention relates to a method for selective catalytic NO_x reduction in waste gases containing oxygen, using ammonia and a reduction catalyst wherein gaseous ammonia is provided by heating a solid storage medium comprising one or more ionic ammonia absorbing salts of the general formula: Ma(NHa)_nXz, wherein M is one or more cations selected from alkaline earth metals, and transition metals such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions such as chloride or sulphate ions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the co- ordination number of 2 to 12, wherein M is Mg and wherein the release rate of ammonia is controlled in consideration of the content of NO_x in the waste gases.

In a fourth aspect the invention relates to the use of a solid ammonia stor- age and delivery material comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NHs)_nX₂, wherein M is one or more cations selected from alkaline earth metals, and transition metals such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions such as chloride or sulphate ions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the co¬ ordination number of 2 to 12, wherein M is Mg as a source of ammonia as the reducing agent in selective catalytic reduction (SCR) of NO_x in exhaust gases from combustion processes.

In a fifth aspect the invention relates to an ammonia delivery device com¬ prising a container comprising an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

In a sixth aspect the invention relates to the use of an ammonia delivery device comprising a container comprising an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

_x in exhaust gases from combustion processes.

In a seventh aspect the invention relates to a solid ammonia storage and delivery material comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NHa)_nXz, wherein M is one or more cations selected from alkali metals, alkaline earth metals, and transition metals such as Li, Na, K₁ Cs, Mg, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, or Zn or combinations thereof such as NaAI, KAI, K₂Zn, CsCu, or K₂Fe, X is one or more anions selected from fluoride, chlo¬ ride, bromide, iodide, nitrate, thiocyanate, sulphate, molybdate, and phos¬ phate ions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination number of 2 to 12, said storage and delivery material having a density of 0.9 to 1.3 g/cm³.

Brief Description of the Drawings

The invention is disclosed more in detail with reference to the drawings in which figure 1 shows a photograph of an ammonia delivery device of the inven- tion, figure 2 shows schematically a top view of the location of thermocouples in the device of figure 1 , figure 3 shows a diagram of the experimental setup, figure 4 shows a graphical representation of the desorption of the ammo- nia, figure 5 is a plot of the buffer pressure during desorption of the ammonia, figure 6 shows the pressure and flow in a single pressure oscillation period during desorption, figure 7 schematically shows a device for compression of an ammonia de- livery material, figure 8 schematically shows an embodiment of an ammonia delivery de¬ vice of the invention, figure 9 schematically shows another embodiment of an ammonia delivery device of the invention, and figure 10 shows an equilibrium phase-diagram for the MgCI₂-NH₃ sys¬ tem.

Detailed Description of the Present Invention

The present invention is related to the use of metal-ammine salts as safe and efficient solid storage media for storage and controlled delivery of ammonia, which in turn is used as the reduction agent in selective catalytic reduction to reduce NO_x emissions, especially from automotive vehicles, boilers and furnaces.

The present invention relates to a solid ammonia storage and delivery ma¬ terial comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NHs)_nXz, wherein M is one or more cations selected from alkaline earth metals, and/or one or more transition metals, such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination num¬ ber of 2 to 12, wherein M is Mg. Suitable anions to be used according to the present invention may be se¬ lected from fluoride, chloride, bromide, iodide, nitrate, thiocyanate, sul¬ phate, molybdate, and phosphate ions. The anion is preferably the chlo¬ ride.

A preferred solid ammonia delivery material according to the invention is Mg(NHa)₆CI₂.

It has been found that although Mg(NHs)₆Cb has a very low partial pres- sure of ammonia above a salt phase, below 0.1 bar at room temperature, it is very suitable for use as a source of ammonia in SCR technology for abatement of pollution from NO_x.

A solid ammonia storage and delivery material comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

An ammonia saturated material of the invention may be prepared by ex¬ posing the dry salt to gaseous ammonia in a manner known per se. The ammonia saturated delivery material as prepared is powdery and rather "fluffy" and difficult to handle or transport and may be - during trans¬ port or use - transformed into small particle fragments thereby potentially blocking the dosing system of a device or give rise to hazardous dust into the environment. Furthermore, the powder has a low bulk density.

In a preferred embodiment of the invention the solid delivery material has a density of 0.9 to 1.3 g/cm³, more preferred a density of 1.1 to 1.3 g/cm³ giving a very high-density storage and delivery material.

The compacted material shows a very low release rate of ammonia at room temperature and atmospheric pressure and may be handled without special precautionary measures for protection against the action of am¬ monia. The compacted material can easily be handled during transport and during and after the final application.

It has surprisingly been found that a powdered ammonia delivery material of the present invention having a very low vapour pressure of ammonia at room temperature may be compacted to a very high density using several different methods for shaping of the material into a desired form and still be capable of delivery of ammonia at a sufficient rate to be suitable for use as a source of ammonia for a SCR reduction of NOx in e.g. automotive vehicles, boilers and furnaces. Such methods are e.g. pressing, extrusion, and injection moulding. In the case of pressing, a pressure might be ap¬ plied in several different ways in a manner known per se. In one embodi¬ ment, the material is compressed to shapes like dense blocks or tablets by placing the saturated salt in a groove/dent/hole/pit in a metal block (e.g. in a cylindrical hole) and applying pressure to compress the material using a corresponding piston.

The saturated material is preferably compressed to a bulk density above

70%, more preferred above 80% and even more preferred above 85%, of the skeleton density. In a preferred embodiment the present invention is related to the compaction and shaping of the saturated ammonia delivery material.

When an ammine complex of a salt is compressed to such a high extent, desorption from such a compacted material would be expected to be ex¬ tremely slow, mostly due to diffusion hindrance. In most dense materials desorption of ammonia would involve solid state diffusion which is known to be a slow process for virtually all materials. This has surprisingly been found not to be the case for the dense materials according to the present invention. It has been found that when ammonia desorbs, a progressing nano-porous structure is formed as the "reaction front" proceeds, leaving open paths for additional ammonia to leave from the central parts of the body of storage material. This is in contrast to e.g. classical heterogene¬ ous catalysis where conversion of reactants is only possible, if reactants are able to diffuse into the catalyst pore structure and the products are able to diffuse out of the catalyst pore structure.

The metal-ammine salts constitute a solid storage medium for ammonia, which represent a safe, practical and compact option for storage and transportation of ammonia (a single-crystalline compound of Mg(NHs)₆CI₂ has an ammonia density of 38 kmole NH₃/m³, whereas that of liquid am¬ monia is only slightly higher (40 kmole NH₃/m³)). Ammonia may be re¬ leased from the metal ammine salt by heating the salt to temperatures in the range from 10°C to the melting point of the metal salt ammine com- plex, preferably to a temperature from 100 to 700 ⁰C, more preferred to a temperature from 150 to 500 ⁰C.

During release of ammonia the metal-ammine salt of the formula M_a(NH₃)_nXz wherein M, X, a, n, and z has the meaning stated above, is gradually transformed into a salt of the formula M_a(NH₃)_mX₂ wherein 0≤m<n. When the desired amount of ammonia has been released, the re¬ sulting salt of formula M_a (NH₃)_mXz can usually be converted back into the salt of the formula M_a(NH3)_nX_z by treatment with a gas containing ammo- nia.

Anhydrous MgCI₂ absorbs up to six moles of NH3 (Gmelins Handbuch, 1939; Liu, 2004) according to reactions 1 to 3:

MgCI₂(S) + NH₃fø) <→ Mg(NH₃)CI₂(S) (1 )

Mg(NH₃)CI₂(S) + NH₃(g) <→ Mg(NH₃)₂CI₂(s) (2)

Mg(NH₃)₂CI₂(s) + 4NH₃fø) <→ Mg(NH₃)₆CI₂(s) (3)

Typical ammonia contents of the metal ammine complexes are in the range of 20-60 wt%, and preferred complexes comprise above 30 wt% ammonia, more preferred above 40 wt% ammonia. The inexpensive com¬ pound Mg(NH₃)₆CI₂ contains 51.7 wt% ammonia.

The present invention offers ammonia storage at significantly higher densi¬ ties (both on a volume and a weight basis) than both aqueous ammonia and aqueous urea. For several metal ammine salts it is possible to re¬ lease all ammonia and then transform the resulting material back into the original metal ammine salt in a large number of cycles. Additionally, the ammonia is directly delivered into the exhaust pipe as a gas, which is an advantage in itself - both for the simplicity of the flow control system and for an efficient mixing of reducing agent, ammonia, in the exhaust gas - but it also eliminates possible difficulties related to blocking of the dosing system because of precipitation in the liquid-based system. Some metal ammine complexes offer a further advantage in that the va¬ pour pressure of ammonia above a solid salt phase is relatively low. It is preferred that the vapour pressure is below 0.1 bar at room temperature, preferably below 0.01 bar. Specifically for Mg(NH₃)₆CI₂ the vapour pres- sure is as low as 0.002 bar at room temperature and atmospheric pres¬ sure which in practice eliminates any noxious effect of the ammonia as the release of ammonia is as low or lower than the release from common cleaning materials containing ammonia.

For Mg(NH₃)₆CI₂ the partial pressure of ammonia at room temperature is 0.002 bar. Even though a partial pressure of ammonia of 0.002 bar at am¬ bient temperature in itself could cause health problems, the compacted material according to the invention saturated with ammonia releases am¬ monia at a very slow rate at ambient temperature and an equilibrium pres- sure of 0.002 bar will only be obtained after a considerable span of time, even if the material is placed in a very confined space. However, when raising the temperature e.g. in the delivery device, a quite quick desorption of ammonia is observed as discussed above.

For mobile units, it is particularly useful to use an ammonia delivery device comprising a container containing the metal ammine complex as such a container may easily be separated from mobile unit and replaced by a fresh at suitable intervals. In preferred embodiments, the metal ammine containers are recycled and recharged with ammonia in a separate re- charging unit.

Due to the slow release of ammonia in open atmosphere at ambient tem¬ peratures for the compressed materials of the present invention, the han¬ dling of the materials does not require extensive safety precautions and substitution of exhausted storage and delivery material with fresh material does not require an encapsulation of the material facilitating the handling as compared to the handling of the materials used in the methods of the state of the art.

In another aspect the invention relates to a method of producing a solid ammonia storage and delivery material comprising an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

■ M₃(NHs)_nXz, wherein M is one or more cations selected from alkaline earth metals, and/or one or more transition metals, such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination num¬ ber of 2 to 12, said method comprising the steps of 1 ) providing the solid salt,

2) saturating the salt with ammonia, and

3) compressing the ammonia salt complex to a dense, shaped body.

In a preferred embodiment of the method of the invention, the ammonia salt complex is compressed to a density of 0.9 to 1.3 g/cm³, more pre¬ ferred to a density of 1.1 to 1.3 g/cm³

In a further aspect the invention relates to a method of selective catalytic NO_x reduction in waste gases containing oxygen, using ammonia and a reduction catalyst wherein gaseous ammonia is provided by heating a solid storage and delivery medium comprising one or more ionic ammonia absorbing salts of the general formula:

Ma(NHs)_nXz, wherein M is one or more cations selected from alkaline earth metals, and transition metals such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination number of 2 to 12, wherein M is Mg and wherein the release rate of ammonia is controlled in consideration of the content of NO_x in the waste gases.

In a yet further aspect the invention relates to the use of a solid ammonia storage and delivery material comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula: M_a(NH₃)nX_z, wherein M is one or more cations selected from alkaline earth metals, and transition metals such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination number of 2 to 12, wherein M is Mg as a source of ammonia as the reducing agent in selec¬ tive catalytic reduction (SCR) of NO_x in exhaust gases from combustion processes.

In yet another aspect the invention relates to an ammonia delivery device comprising a container comprising an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NH₃)_nXz, wherein M is one or more cations selected from alkaline earth metals, and/or one or more transition metals, such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination num¬ ber of 2 to 12, wherein M is Mg, said container being provided with one or more closable outlet opening(s) connected to a pipe and further being ing provided with means for heating the container and the ammonia ab¬ sorbing salt for release of gaseous ammonia.

In the ammonia delivery device according to the invention, the closure of the closable outlet opening(s) may be in the form of one or more valve(s) known per se for use in connection with ammonia. Heating means may be in the form of an electrical resistive heating device known per se.

The heating means may alternatively be provided as heat produced by chemical reactions.

The salt is normally heated to temperatures in the range from 10°C to the melting point of the metal salt ammine complex, preferably to a tempera- ture from 100 to 700⁰C, more preferred to a temperature from 150 to 500 ⁰C.

In a preferred embodiment of the invention the release rate of ammonia is controlled by accurate control of the heating of the container and the am- monia absorbing salt for release of gaseous ammonia. The release of ammonia is preferably further controlled by reduction valves, flow control¬ lers or similar equipment or units. The release of ammonia from a con¬ tainer is preferably controlled by interaction with an electronic engine con¬ trol system for delivery of an optimum amount of ammonia in a specific ra- tio (e.g. NH₃:NOx = 1 :1 ) of the changing emission of NOx from the engine.

In still another aspect the invention relates to the use of an ammonia de¬ livery device comprising a container comprising an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

M₃(NHs)_nXz, wherein M is one or more cations selected from alkaline earth metals, and/or one or more transition metals, such as Mn, Fe, Co, Ni, Cu, and/or Zn, X is one or more anions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination num¬ ber of 2 to 12, wherein M is Mg, said container being provided with one or more closable outlet opening(s) connected to a pipe and further being provided with means for heating the container and the ammonia absorbing salt for release of gaseous ammonia as a source for ammonia in selective catalytic reduction of NO_x in exhaust gases from combustion processes.

In yet a further aspect, the invention relates to a solid ammonia storage and delivery material comprising: an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NHa)_nXz, wherein M is one or more cations selected from alkali metals, alkaline earth metals, and transition metals such as Li, Na, K, Cs, Mg, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, or Zn or combinations thereof such as NaAI, KAI, K₂Zn, CsCu, or K₂Fe, X is one or more anions selected from fluoride, chlo¬ ride, bromide, iodide, nitrate, thiocyanate, sulphate, molybdate, and phos- phate ions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination number of 2 to 12, said storage and delivery material having a density of 0.9 to 1.3 g/cm³. Still further, the invention relates to a method of producing a solid ammo¬ nia storage and delivery material comprising an ammonia absorbing salt, wherein the ammonia absorbing salt is an ionic salt of the general formula:

Ma(NHs)_nX₂, wherein M is one or more cations selected from alkali metals, alkaline earth metals, and transition metals such as Li, Na, K, Cs, Mg, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, or Zn or combinations thereof such as NaAI, KAI, K₂Zn, CsCu, or K₂Fe, X is one or more anions selected from fluoride, chlo¬ ride, bromide, iodide, nitrate, thiocyanate, sulphate, molybdate, and phos- phate ions, a is the number of cations per salt molecule, z is the number of anions per salt molecule, and n is the coordination number of 2 to 12, said storage and delivery material having a density of 0.9 to 1.3 g/cm³, said method comprising the steps of 1 ) providing the solid salt, 2) saturating the salt with ammonia, and

3) compressing the ammonia salt complex to a density of 0.9 to 1.3 g/cm³.

The present invention is especially suitable for use in reduction of emis¬ sion of NO_x in exhaust gases from stationary and mobile combustion en- gines or power plants fuelled by diesel oil, petrol, natural gas or any other fossil fuels.

Thus, the present invention is also especially suitable as a source for pro¬ viding ammonia in selective catalytic reduction in exhaust gasses for re- duction of emission from stationary and mobile combustion engines or power plants fuelled by methanol, ethanol, hydrogen, methane, ethane or any other synthetic fuels.

Mobile combustion engines for which the invention is suitable are may e.g. be automobiles, trucks, trains, ships or any other motorised vehicle. The invention is particularly suitable for use in connection with reduction of NO_x in combustion gases from automobiles and trucks.

Stationary power plants for which the invention is suitable are preferably power plants generating electricity.

The present invention is also especially suitable for use in reduction of emission of NO_x in exhaust gases from solid oxide fuel cells (SOFC). SOFCs are operated at a high temperature, where there may be a small production of NOx, which can be removed by adding ammonia using the present invention.

Description of the Preferred Embodiments

The invention is now explained more in detail with reference to the draw- ings showing preferred embodiments of the invention.

MATERIALS AND METHODS

MgCI₂ powder: anhydrous, >98% purity, Merck Schuchardt.

Glove-bag: Aldrich premium AtmosBag from Aldrich Chemical Company, Inc., 1001 West Saint Paul Ave., Milwaukee, Wl 53233, USA.

Ammonia Gas: Ammonia Gas 99.9 % from Hede Nielsen, lndustriparken 27-31 , 2750 Ballerup, Denmark.

Preparation of ammonia saturated MgCI? powder.

The ammonia carrier, Mg(NH₃)6Cl2, was prepared by placing MgCl₂ pow¬ der for several days in a glove-bag containing ammonia gas at atmos- pheric pressure (1 bar) and at room temperature. The degree of saturation was checked by temperature programmed desorption (TPD) and verified to be near 100% of the theoretical amount. The absorption/desorption was found to be fully reversible.

The rate of absorption is dramatically increased at higher NH₃ pressures (minutes rather than days) (Touzain and Moundamga-lniamy, 1994).

An equilibrium phase-diagram of the MgCI₂-NH₃ system is shown in Figure 10 also showing the process of saturation when carried out at 5 bars.

Starting from point B' (pure MgCI₂) the pressure was increased to 5 bars at point A corresponding to an isothermal absorption at T=298.15K and resulting in the formation of fully saturated Mg(NH₃)₆CI₂. The rate of forma¬ tion depends on the equilibrium pressure drop; however absorption is generally quite fast at ammonia pressures of more than 4 bars (Touzain and Moundamga-lniamy, 1994). As mentioned, the desorption reaction only proceeds at elevated temperatures, which is also indicated on the path from A to B (isobaric desorption at P=5 bar).

Example 1 An ammonia delivery device was made from stainless steel in the form of a cylindrical container, having the dimensions 2RO = H = 10 cm, where RO is the inner radius and H is the height of the reactor. A photograph of the device is shown in figure 1. The device was provided with five wells for in¬ sertion of thermocouples placed perpendicular to the tangent of the reac- tor circumference for determining the radial temperature distribution. Fur¬ thermore, a thermocouple was placed on the outer wall to measure the ac¬ tual temperature of the heated steel wall. Ammonia-saturated MgCI₂ pow¬ der (258.8 grams) was then placed and compacted slightly by manual pressure to a bed density of approximately 331 kg/m³. A thin sheet of quartz cotton was placed on top of the salt to prevent any grains from be- ing carried out of the reactor. As desorption of ammonia from the complex requires elevated temperature, a heating wire was wrapped around the reactor and thermal insulation (Rockwool) was placed on top of this. Using a PC with labview interface, the power to the heating tape surrounding the storage container was turned on/off when the pressure in the buffer was below/above the pressure set-point. The desorbed gaseous ammonia flowed into a small buffer container, that was placed after the desorption unit and having a volume of approximately 200 millilitres. The pressure in¬ side the buffer was measured using a digital Kobold SEN-87 pressure me¬ ter. The position of the thermocouples (denoted T1 to T5) are shown in the below table 1 and illustrated in figure 2.

Table 1

Radial position of five thermocouples

A mass-flow controller (Brooks Smart Mass Flow 5850S calibrated to NH₃) dosed the ammonia from the buffer container into a "tailpipe" conveying a stream of carrier gas of air (1000 litres/minute) corresponding roughly to the current of exhaust gas from a small car. In addition to dosing ammo¬ nia, the same device also measured the actual flow through the valve in millilitres/minute (at T=298.15 K and P=1 bar). According to the stoichiometry of the SCR reaction, the mixture ratio between NH₃ and NOx should be approximately 1 :1 (e.g. (Koebel and Kleeman, 2000; Fang and DaCosta, 2003)). The transient NOx concentration (and by that the tran¬ sient NH3 concentration) in the exhaust gas from a car is very complicated to describe a continuous mathematical function and as an approximation, a sinusoidal function was used. The amplitude was set to 210 millili- tres/minute and the period was set to 120 seconds. Due to limitations in the dynamics of the mass-flow controller, however, the resulting outlet flow was not completely sinusoidal. Heating of the desorption unit was con- trolled using the ON-OFF controller programmed along with the data- collection in Labview. The control parameter was chosen as the buffer pressure and the set-point pressure was set to 5 bars. At such high pres¬ sures it is fairly safe to assume that there is no pressure gradient across the porous bed (Lu and Spinner, 1996), and therefore the reactor pressure should be equal to the buffer pressure. It is possible to reduce the set- point pressure, which will reduce the temperature required for desorption. However, the main reason for the choice of set-point pressure was to maintain a well-defined flow through the mass-flow controller. With proper flow-mapping the flow through the valve should in principle be independ- ent of the back-pressure.

The result of the experiment appears from Figure 4 which shows a graphi¬ cal representation of recorded temperature at positions T1 and T5 (the melting point for MgCb is 714 ⁰C) during desorption of ammonia from the ammonia-saturated MgCb powder, and figure 5 shows the recorded pres¬ sure in the buffer container during the experiment. Phases I-IV and an in¬ termediate phase (phase (H-III)) have been indicated in figures 4 and 5 and are explained more in detail below. During the experiment, a controller increased the temperature of the unit slowly to sustain a desired pressure of ammonia in the buffer. In the case of the experiment in figures 4 and 5, the buffer was kept at a pressure close to 5 bars. Other experiments were done at lower buffer pressures. This resulted in a lower operating temperature due to the thermodynamics of the storage material. A higher ammonia pressure gives a higher desorp- tion temperature.

Phase (I) - Warm-up During this phase the pressure was allowed to build up to the set-point, no ammonia was extracted during this initial warm-up phase.

Phase (II) - Desorption of first four moles of ammonia

When reaching the set-point buffer pressure, heating was turned off and the mass-flow controller began to dose ammonia. This eventually caused the pressure to drop and heating was turned on again. Switching between on and off for dosing ammonia caused the pressure to oscillate around the set-point. The small oscillations in the measured out-flow (nearly sinusoi¬ dal flow) can also be seen in the large oscillations around the set-point, as it is shown in figure 6.

The large (and slow) oscillations are also seen in the temperature, and due to the low effective thermal conductivity of the porous solid matrix this is most pronounced in the vicinity of the source of heat (T1 ). During phase II, the oscillations of the pressure and temperature continued to increase in amplitude and decrease in frequency. Combined with the large tem¬ perature gradients observed from figure 4 this indicates, that the reaction proceeds along an inward moving reaction front. Such a moving front will increase the way of heat-transfer, which will result in increasing amplitude and decreasing frequency of the pressure/temperature oscillations. Phase H-III - Transition phase

As the front moves toward the centre of the reactor, less and less ammo¬ nia is available for desorption. Eventually, a new front builds up near the heat-source, in which the fifth mole of coordinated ammonia is desorbed. This again reduces the way needed for heat-transfer and thus reduces the amplitude and increases the frequency of the oscillations. The apparently lower amplitude of the oscillations in the transition phase as compared to phase III, indicates that there still is ammonia bound as Mg(NH3)6Cl2 left for desorption near the centre.

Phase III - Desorption of fifth mole of ammonia The sudden increase of the amplitude marks the end of the transition phase and the beginning of the third phase, in which only the fifth mole of ammonia desorbed. The decrease of the pressure overshoot as compared to phase Il is most obviously due to the fact that only one mole of ammo¬ nia was released as compared to the four moles released during phase Il (cf. reactions 2 and 3), therefore reducing the overall desorption rate. However, the increased loss of heat to the surroundings due to the higher temperature level at T1 could also help moderating the tempera- ture/pressure peaks. During this phase it is difficult to see any increase of amplitude and decrease of frequency of the pressure oscillation; however the peak temperature does increase throughout the phase.

Phase IV - Desorption of sixth mole of ammonia The amplitude of the pressure overshoot was even smaller during this phase than for phase III. One mole of ammonia was desorbed in both phase III and IV, but the increased loss of heat at T1 during phase IV helped moderating the temperature peaks and hence, also the overshoot of pressure. Since this experiment was run for several hours, including a night, a maximum allowable temperature of 500 ⁰C was set as a safety pa- rameter. From figures 4 and 5 it can be seen, that this temperature was reached, which resulted in the loss of buffer pressure. As this occurred during the night, it was not corrected until the following morning by setting the maximum allowable temperature to 650 ⁰C (64 ⁰C less than the melt- ing point temperature of MgCb). The temperature and pressure fluctuated a somewhat during this phase. This might indicate some sort of build-up of pressure and subsequent release due to mass-transfer hindrance. During phase IV, the temperature throughout the reactor became quite high. However, in an optimized system the set-point for pressure will certainly be lower than the chosen 5 bars, which will also reduce the required tem¬ perature for all three desorption reactions to occur.

Integration of the outlet flow curve in time yields the total accumulated vol¬ ume of NH3 released through the valve. This value can be converted to number of moles by means of the ideal gas law. In order for the process to be efficient, this value should be close to the theoretically predicted amount of NHb contained in the salt. This is easily calculated, since the ra¬ tio of MgCh and NH3 in a 100% saturated salt is 1 :6. The stoichiometric calculations showed that, theoretically, 7.88 moles of ammonia should be stored within the given mass of Mg(NH3)eCl2 salt (258.8 grams), while the integration of the calibrated signal from the mass flow controller revealed that 7.86 moles of ammonia was released through controller giving a very high storage efficiency of approximately 99.8%.

Example 2

Compression of solid ammonia storage medium

Figure 7 schematically shows a device compression of the solid ammonia delivery material for the preparation of cylindrical tablets (dimensions: 13mm in diameter; 10mm high). In this embodiment, the solid ammonia delivery material was compressed in a chamber by applying a pressure of 2 - 4 tons/cm² using a piston compressing the powdered saturated storage material. When the piston was removed, the delivery material was in the desired shape of e.g. a tablet, a cylinder or a rod, and had a density above 80% of the theoretical crystal density.

Figure 8 schematically shows an embodiment of an ammonia delivery de¬ vice of the invention for desorption of the compressed delivery material. In this embodiment, one or more tablets of solid ammonia delivery material 1 are placed in a container 2, which can be heated by a heating device 3. Desorbed ammonia leaves the container through a nozzle 4. Heat for the heating device 3 may be provided by e.g. resistive electric heating or chemical reactions. Such chemical reactions could be generated e.g. by combustion of a part of the released ammonia. If the delivery device is used in connection with SCR of NOx in exhaust gases, waste heat from the engine producing the gases can also be applied.

Figure 9 schematically shows another preferred embodiment where only a part of the stored solid delivery material 1 is heated at a time. The solid delivery material is stored in compressed form, and introduced into a hot chamber 2 one at the time at intervals corresponding to the requirement for gaseous ammonia. The hot chamber is heated by a heating device 3 operated after the same principles as described above 2. Gaseous am¬ monia leaves the hot chamber through a nozzle 4, and when all ammonia is desorbed from a tablet 5 of solid ammonia delivery material, it is dis- carded into a separate container 6.

In a similar type of embodiment, the entire storage material is separated into a number of compartments each having their own heating source so that it is possible to have complete desorption of a given fraction of the material without using any moving parts to replace saturated/unsaturated salt e.g. on-board the vehicle during use.

The bed-density of the delivery material used in Example 1 was quite low (331 kg/m³ when compacted gently by hand) when compared to the den¬ sity of MgCI₂ and Mg(NHa)₆CI₂ (1252 kg/m³, cf. the below table 2).

Table 2

A low density means that the entire storage system would require more space. This problem was solved by compressing Mg(NH3)βCl2 into solid rods having a density of 1219 kg/m³ (97% of the solid density). TPD ex¬ periments confirmed that it was possible to desorp all ammonia from this tablet, thus increasing the potential storage capacity by a factor of 3.7 (on a molar basis) to approximately 93% of the volumetric ammonia storage capacity of liquid ammonia. A nearly quantitative desorption of ammonia from the dense material is possible because the front of desorption leaves behind a porous layer of anhydrous MgCl2. This automatically generates the required pore system needed for mass-transfer through the structure. This is considered an ideal combination of a) an initially very compact structure having almost no void and being easy to handle, b) a high capac¬ ity for containing and delivering ammonia, c) a low external surface area, and d) a high degree of safety. Example 3

Comparison of the use of ammonia delivery devices of the invention and urea-technology

By calculating the amount of NOx (assumed to be pure NO) generated per kilometre in a model fuel (taken as pure n-octane, p = 696.8 kg/m3), the corresponding driving distance for a given amount of ammonia or urea can be found.

Based on the EURO 3 standards (The European Parliament, 1998) as well as by the values used by some researchers (Hyundai Motor Co.: Choi et al., 2001 ), the assumed NOx concentrations and the fuel economies are: 150 ppm and 10 km/litre for stoichiometric combustion (typically gasoline) and 300 ppm and 15 km/litre for lean burn combustion (λ=1.5, typically diesel).

Taking 1 litre (or 696.8 g) of fuel as a basis of calculations the generated NOx per. kilometre would be 5.87-10^"2 mole/km or 0.18 g/km for stoichiometric combustion, while for lean burn combustion (λ=1.5), the NOx emission is 1.15-10^'2 mole/km or 0.34 g/km.

In the below table 3 is shown the required mass/volume of the high- density Mg(NH₃)6Cl2 needed to drive 20,000 km (excluding the mass/volume of any equipment) compared to that of the 32.5 % urea solu- tion. The comparison is based on the combustion of pure n-octane and the above assumptions. The fuel is assumed to be pure n-octane and the ex¬ haust is assumed to have an average molar fraction of 300 ppm NOx. The engine is assumed to run lean with a fuel economy of 15 km/litre fuel. The density of a 32.5 % wt/wt urea solution is 1090 kg/m³ and the density of the Mg(NH₃)₆CI₂ is assumed to be 1219 kg/m³. Table 3

The results clearly show the superior ammonia storage capacity of Mg(NH₃)₆CI₂ over that of the urea solution. Combined with a high effi¬ ciency of approximately 99% this ammonia storage compound is very ap¬ propriate for the purpose. Due to the nature of the compaction of the mate¬ rials, the present invention is also superior to the disclosures of WO 99/01205, where the preferred embodiment of granulated material of ei¬ ther Ca(NH₃)₈Cl₂ or Sr(NH₃)SCI₂ will results in a reduction in volumetric ca¬ pacity of the theoretical salt densities by a factor of 1.5-2.

References

Fang, H. L., DaCosta, H. F., 2003. Urea thermolysis and NOx reduction with and without SCR catalysts. Applied Catalysis B: Environmental 46,

Goetz, V., Marty, A., 1992. A model for reversible solid-gas reactions sub¬ mitted to temperature and pressure constraints: Simulation of the rate of reaction in solid-gas reactor used as chemical heat pump. Chem. Eng. Sci.47 (17-18), 4445-4454. H. Van Vlack, L., 1989. Elements of Materials Science and Engineering, sixth Edition. Addison-Wesley Publishing Company Inc. Hyundai Motor Co.: Choi, S.-m., Yoon, Y.-k., Kim, S.-j., Yeo, G.-k., Hee- sung Engelhard Corp.: Han, H.-s., 2001. Development of UREA-SCR sys¬ tem for light-duty diesel passenger car. SAE Technical Paper Series. Koebel, M.; Elsener, M., Kleeman, M., 2000. Urea-scr: A promising tech¬ nique to reduce NOx emissions from automotive diesel engines. Catalysis Today 59, 335-345.

Liu, Chun Yi & Aika, K.-i., 2004. Ammonia absorption on alkaline earth hal- ides as ammonia separation and storage procedure. Bull. Chem. Soc. Jpn. 77 (1 ), 123-131.

Lu, Hui-Bo; Mazet, N., Spinner, B., 1996. Modelling of gas-solid reaction - coupling of heat and mass transfer with chemical reaction. Chem. Eng. Sci.51 (15), 3829-3845. Matsumoto, S., 1997. Recent advances in automobile exhaust catalyst. Catalysis Surveys from Japan 1 , 111-117.

Olovsson, I., 1965. Packing principles in the structures of metal ammine salts. Acta Cryst. 18, 889-893. CRC Handbook, 2004. Handbook of chemistry and physics (web edition). Gmelins Handbuch, 1939. Magnesium teil b: Die verbindungen des mag¬ nesiums (27).

The European Parliament, 1998. Directive 98/69/ec. The European Parliament, 1999. Directive 1999/96/ec. The European Parliament, 2003. Directive 2003/17/ec. Touzain, P., Moundamga-lniamy, 1994. Thermochemical heat- transformation:

Study of the ammonia/magnesium chloride-GIC pair in a laboratory pilot. MoI. Crys. Liq. Cryst. 245, 243-248.