NL9202282A - Method for desulfurizing a gas and absorbent suitable for this method. - Google Patents

Description

Title: Method for desulfurizing a gas and absorbent suitable for this method.

The present invention relates to a process for desulfurizing gases using absorbents based on molybdenum and / or tungsten oxide, as well as such absorbents.

The need to purify gases that are further processed in chemical processes, delivered to customers or released to the atmosphere from sulfur compounds, in particular hydrogen sulfide, is well known.

In the gasification of coal or heavy petroleum fractions, the problem arises that hydrogen sulfide is formed. Due to its corrosive nature and toxicity to catalysts used to convert such gases into useful compounds, hydrogen sulfide should be removed. In the case where the sulfur-containing gas is used as a fuel gas, for example in power stations, the hydrogen sulfide content should be preferably less than 100 ppm for environmental reasons and because of the corrosive effect. In the case where the gas is used as a feed for catalytic conversions, the aim is to keep the content of H 2 S below 25 ppm, preferably below 10 ppm. A number of methods are therefore known which are aimed at the removal of hydrogen sulphide from gas.

Numerous methods are known for removing hydrogen sulfide from gas streams by ad- or absorption on solids. In the purification of gas flows from, for example, gasification of coal or oil, which are particularly large quantities that have a high temperature, metal oxides are often used for this purpose. So-called bulk metal oxides are very common, but their disintegration during absorption and regeneration results in dust formation, which makes these materials unattractive. The problem of disintegration can be at least partially overcome by the use of highly dispersed metal oxide on a stable support.

It is known for such purposes to use a silica as iron oxide support. In the literature (EP-A-71,983) an iron oxide absorbent on silica has been described which has a good degree of conversion and stability.

In the publication of Abel, Shultz and Langdon, Bur. Mines RI 7947 (1974) have described the results of a study of the utility of a large number of solid absorbents for the removal of hydrogen sulfide from hot gases. This publication shows that a large number of compounds, mainly metal compounds, exhibit an absorption of hydrogen sulfide. Examples thereof are compounds of iron, magnesium, calcium, boron, copper, cerium, molybdenum, yttrium, lanthanum, cobalt and zinc. Virtually all of these compounds have been tested as oxides, hydroxides or carbonates, in the form of fly ash sintered granules. Best results were obtained with sintered granules of 75% fly ash with 25% iron oxide. An absorption capacity of about 40 g sulfur per 100 g absorbent was thereby obtained at about 800 ° C. These absorbents exhibit the same drawback as described above with respect to bulk metal oxides, namely disintegration upon absorption and upon regeneration.

In the article by Furimsky et al, Solid adsorbents for removal of hydrogen sulfide from hot gases, Wissenschaft & Technik, Bd. 39, Heft 4, April 1986, pages 163-172, an overview of absorbents described in the literature for the absorption of hydrogen sulfide from hot gases is given. Examples of absorbents described therein include compounds of calcium, magnesium, iron, manganese, zinc, molybdenum, tungsten, nickel, cobalt, barium, vanadium and copper. According to the authors, deactivated cracking catalysts based on molybdenum, cobalt and nickel on aluminum oxide, after burning the coke, are suitable as absorbents.

The absorbents described in this article also exhibit the already indicated disadvantages of disintegration and the associated dust formation. In said deactivated cracking catalysts, this is caused by the reaction of the alumina to sulfate. In addition, a large number of the materials described have the disadvantage of a low temperature resistance, so that after some time sintering and therefore loss of activity occurs. A number of materials also exhibit the property of volatilization at a higher temperature, which is undesirable from the viewpoint of the environment and degree of loading.

Finally, a number of prior art materials have the problem that they are not or only very difficult to regenerate.

The object of the invention is to provide a method of the type described in the preamble, in which no problems arise with regard to the formation of metal sulphates which are difficult or not regenerable.

A further object of the present invention is to provide a method for removing H 2 S from gas streams, wherein the degree of loading is improved over the conventional materials.

The present invention therefore relates to a process for desulfurizing gases at elevated temperature, wherein the gas to be desulfurized is contacted with an absorbent based on one or more metal oxides, which method is characterized in that the gas to be desulfurized is contacted with an absorbent based on one or more molybdenum and / or tungsten oxides applied to a support.

In this process, a very good absorption of hydrogen sulphide is obtained, also at low temperature, while no or almost no formation of SO2, or of metal sulphates which are difficult to reduce, occurs.

The invention is based on the surprising insight that molybdenum and / or tungsten oxide on a support absorbs hydrogen sulphide very well with a high degree of loading, that is to say a degree of loading corresponding to a total conversion of the oxide into the sulphide, while no or virtually no no sulfate is formed. In addition, the absorbents are easy to regenerate to almost their original capacity.

With the absorbents it appears to be important that the distribution of the active component on the carrier material is good. In particular, well-dispersed oxides appear to have a very high capacity, while the rate of absorption is also high, which is of great importance for the prevention of H2S breakdown.

As the active component, it is preferred to use an oxide selected from the group consisting of oxides of molybdenum, tungsten and combinations thereof with oxides of iron, vanadium, copper, manganese, cobalt, zinc, nickel and / or chromium. More particularly, the active component is selected from the group consisting of molybdenum (VI) oxide, iron molybdate, mixtures of iron (II) and / or iron (III) oxide with molybdenum (VI) oxide and mixtures of iron molybdate with molybdenum (VI) oxide and / or iron (II) and / or iron (III) oxide.

As the support material, any suitable solid can be used which has sufficient thermal and chemical stability under the conditions of absorption and regeneration. Examples of carrier materials to be used are silicon dioxide, aluminum dioxide, silica-alumina, zeolites and aluminum phosphate. More particular preference is given to silicon dioxide and especially aluminum phosphate.

When an aluminum phosphate is used as a support material, no problems appear with regard to metal sulphate formation, even if an excess of aluminum relative to the phosphate is present in the support. The absorbent is therefore particularly easy to regenerate. It is not necessary to use a multi-step regeneration method. It also has the advantage of good regenerability and good stability at high temperature, which improves the applicability of the absorbent.

Another advantage of the present invention lies in the fact that the aluminum phosphate carrier is particularly stable at the temperatures used at high temperature desulfurization and will retain all or almost all of the original structure, even with longer use and frequent regeneration.

The structure of the amorphous aluminum phosphate support can be adjusted very well by varying preparation conditions, such as pH, temperature and ratio of aluminum to phosphate. It is possible, for example, to work with undersized aluminum or with excess aluminum with respect to the stoichiometric quantity required. It is also possible to work with substantially stoichiometric amounts of aluminum and phosphate.

Surprisingly, it has been found that it is no problem in the invention if an aluminum phosphate is used with amounts of aluminum and phosphate that differ from the stoichiometric amount. Although it would be expected that the use of an excess of aluminum in particular would give rise to problems with sulphate and sulphide formation, this has in no way been the case.

The aluminum phosphate support material can be prepared in various ways. In general, the preparation will start from separate aqueous solutions of a soluble phosphate and a soluble aluminum compound. From the solubility point of view, preference is given to aluminum nitrate and ammonium phosphate. The solutions are combined with aluminum phosphate to precipitate.

This can be done by simultaneously injecting both solutions, preferably at a constant pH, into a reactor which, if desired, may already contain water or an aqueous solution.

It is also possible to put either solution into the reactor and inject the other solution. In general this will not be done at a constant pH.

The distribution of the metal component over the support is important, since a poor distribution is detrimental to the absorption capacity. The active component is preferably present as homogeneously as possible on the carrier. This can be achieved by carefully selecting the application method and / or by selecting the degree of loading.

The specific surface area is preferably 5 to 500 m2 / g, while the average pore diameter, based on cylindrical pores, is between 1 and 100 nm.

For the purposes of the invention, specific surface is understood to mean the BET surface, as defined by S. Brunauer et al, in J.A.C.S. 60f 309 (1938). Nitrogen adsorption at 77 K was used according to the so-called three-point measurement. For the calculation, the surface area for a nitrogen molecule was 16.2 A2.

The average pore radius is determined from a cylindrical pore model using the following formula:. ,. , ... 20,000 x pore volume (cm3 / q)

Average pore radius (A = -—-— c -: -; -, 9 BET surface (m2 / g)

The pore volume used herein is determined gravimetrically by impregnation with water under vacuum. The specific pore volume can also be determined by mercury porosimetry up to a pressure of 2000 bar. The values obtained by both methods are in good agreement.

The content of carrier material in the absorbent used according to the invention is preferably between 25 and 99% by weight.

The invention also relates to an absorbent suitable for high temperature desulfurization of gases, which absorbent is based on at least one oxide of molybdenum and / or tungsten applied to an aluminum phosphate support material. The composition of such an absorbent, its preparation and its use substantially correspond to what has been stated above in the context of the gas desulfurization method.

Depending on the intended application, the degree of removal of the hydrogen sulphide will be adjusted. For less critical applications, the aim is often to achieve an H2S content that does not exceed 100 ppm. When using catalytic conversions of the treated gas with sulfur-sensitive catalysts, the hydrogen sulfide content must of course be significantly lower. For such applications the aim is to have an H2S content of at most 25 ppm, more in particular at most 10 ppm.

The content of H2S in the treated gas can be adjusted by varying the contact time of the gas with absorbent. This can be done on the one hand by varying the gas velocity and on the other hand by varying the amount and height of the bed in which the absorbent is applied. Suitable contact times are between 0.01 and 10 seconds.

During the absorption of hydrogen sulfide, most of the hydrogen sulfide will react with the metal oxide to form metal sulfide. The regeneration of the sulfur-loaded absorbent thus occurs by reacting the metal sulfide with oxygen to form metal oxide. Depending on the amount of oxygen used, the sulfur can be obtained in the form of elemental sulfur, of SO2, or of mixtures thereof.

When relatively little oxygen is used in the regeneration, mainly elemental sulfur in the gas is obtained. Since sulfur can easily cause blockages in the further equipment, it is preferable to condense this sulfur immediately after regeneration.

When a large excess of oxygen is used, all sulfur is released in the form of SO2. If desired, this can be separated and / or further processed in a known manner.

The advantages of the present invention are particularly evident in very high temperature desulfurization. These temperatures are generally above 150 ° C and more particularly between 400 and 625 ° C. At these temperatures it is therefore possible to regenerate the absorbent, the regeneration of which takes place by oxidation with oxygen and preferably at substantially comparable temperature as the absorption.

According to the invention, it is preferable if the contact takes place at a temperature of 150 to 400 ° C in case the absorbent is based on molybdenum (VI) oxide; and at a temperature of 150 to 625 ° C, preferably 400 to 625 ° C if the absorbent is based on iron (II) molybdate.

The absorbents are preferably made by homogeneous deposition precipitation of the metal component or precursor therefor on the finished support. This method is described in J.W. Geus, Prep, or Cat. Ill, ed. G. Poncelet et al. (1983) p. 1-25.

In general, the amount of metal oxide applied to the support is not critical and can vary within wide limits. The appropriate amounts are related to the specific surface area of the support and will be between 1 and 50% by weight. The aim is of course to have the highest possible content of metal component finely distributed on the support, since this allows the highest possible loading thereof. It is preferable to use a particle size of at most 0.1 µm, more in particular at most 0.05 µm. The lower limit of the particle size is not very critical and is preferably at least 1 nm.

The method according to the invention is especially suitable for treating large amounts of hot gases, which can for instance originate from the gasification of coal or oil. Such gases are often used as fuel gas in power plants and other gas-fired installations. It is also possible to catalytically convert such gases into other compounds, such as ammonia, methanol or higher hydrocarbons, for example via a Fischer Tropsch synthesis.

Depending on the intended application, the degree of removal of the hydrogen sulphide will be adjusted. For less critical applications, the aim is often to have a Ü2S content that does not exceed 100 ppm. When using catalytic conversions of the treated gas, the hydrogen sulfide content must of course be significantly lower. For such applications the aim is to have an H2S content of at most 25 ppm, more in particular at most 10 ppm.

The content of the treated gas can be adjusted by varying the contact time of the gas with absorbent. This can be done on the one hand by varying the gas velocity and on the other hand by varying the amount and height of the bed in which the absorbent is applied. Suitable contact times are between 0.01 and 10 seconds.

The invention is now illustrated by means of a few examples, without being limited thereto.

EXAMPLES

1 Experimental procedures 1.1 Preparation of the absorbents

All absorbents were prepared by homogeneous deposition precipitation as described in J.W. Geus, Prep. or Cat.Ill, ed. G. Poncelet et al. (1983) p. 1-25. The absorbents based on molybdenum on silica were prepared by injecting an aqueous solution of molybdenum (III) chloride under the surface of a vigorously stirred suspension of silicon dioxide (Aerosil 200, Degussa) at a constant pH of 5. The molybdenum (III) chloride was prepared by electrochemical reduction of an aqueous solution of ammonium hepta molybdate (Merck pa) at pH 1.

The preparation of well-dispersed silica-based molybdenum catalysts using molybdenum precursors of higher valency than three is not possible due to the low interaction of these materials with the support during precipitation.

Iron molybdate on silica was achieved in a number of ways. One method was the coprecipitation of an iron and molybdenum precursor. The use of a solution of iron (III) chloride was not possible because of the redox reaction between Fe3 + and Mo3 + ions. This resulted in reduction of the iron (III) ions and oxidation of the molybdenum (III) ions. Therefore, a mixture of iron (II) and molybdenum (III) chloride was used to prevent the formation of higher valence molybdenum. The aqueous solution of iron (II) was obtained by dissolving iron powder (Baker 99.99%) in concentrated hydrochloric acid (Merck p.a.). A number of absorbents with varying iron / molybdenum ratios were prepared. Also, the absorbents were prepared by deposition of Fe2 + and Mo3 + precursors one after the other on the silica support. The order in which the precipitation was performed was found to have a strong influence on the action of the final absorbent.

Using aluminum phosphate as a support, we only used the coprecipitation method. Fe3 + and Mo ^ + were used as precursor herein. The preparation of aluminum phosphate with an aluminum / phosphate ratio of two was carried out at a pH of 5.

After the homogeneous deposition precipitation, the absorbents were filtered and dried overnight at 120 ° C. The absorbents were tableted and broken again. A sieve fraction between 150 and 500 µl was separated, and calcined in air at 500 ° C with a heating rate of 5 ° C / minute. After reaching 500 ° C, they were kept at that temperature for one hour.

The various absorbents are listed in the following table.

Γ \ Λ Λ Λ Λ A Λ

1) Precipitation of iron (II) followed by precipitation of molybdenum (III) 2) Precipitation of molybdenum (III) followed by precipitation of iron (II) 3) Coprecipitation of iron (II) and molybdenum (III) 4) Coprecipitation of iron (III) and molybdenum (IV)

The test procedure followed when testing the various absorbents was done with an apparatus as described in Fig. 1. This apparatus consists of a gas flow control system with a mass controlled supply of various gases, solenoid valves, and saturator and reactor furnace and analysis equipment. The gas was analyzed using a thermal conductivity detector (TCD) and a UV-VIS spectrophotometer, Varian Cary 1. The latter measured the hydrogen sulfide concentration at 232 nm and the sulfur dioxide concentration at 262 nm.

The experiments consisted of a number of absorption and regeneration cycles, temperature-programmed reductions and oxidations, during which the hydrogen sulfide and sulfur dioxide concentrations were monitored.

During the absorption procedure, use was made of a sample size of 1.5 ml, a reaction mixture of 1% by volume hydrogen sulfide and 10% by volume hydrogen in argon with a total amount of 100 ml / minute, a space velocity of 4000 hours-1 and a temperature of 400 ° C. The absorption step was generally preceded by a reduction step where any iron sulphates formed were reduced in 10% hydrogen at the same temperature as the absorption step. After the absorption procedure, the absorbent was regenerated with a reaction mixture of 2% by volume of oxygen in argon under otherwise equal conditions.

10% by volume of hydrogen in argon was used during the temperature programmed reduction, while the temperature was increased by 4 ° C / minute. For the temperature-programmed oxidation, a gas mixture of 2% by volume of oxygen in argon was used. During two experiments, the reactor was rinsed with argon for 10 minutes.

The various absorbents described in the table were tested in the manner described above. In the various figures appended to this application, the absorption of hydrogen sulfide and possible formation of SO2 by the absorbent is indicated.

Figure 2 shows the absorption curve of 20 wt.% Molybdenum oxide on silica, SM020 /. Sulfur dioxide production does not take place at the beginning of the absorption. Also, reduction treatments after regeneration do not produce sulfur dioxide. This indicates that no sulfate formation occurs during oxygen regeneration if the absorbent is based only on molybdenum oxide.

Nevertheless, some SO2 formation does take place during the absorption step itself. This is caused by reduction of M0O3 with H2S to molybdenum (IV) oxide and SO2. However, to prevent sulfur dioxide formation during the absorption step, the hydrogen concentration can be increased, which is indicated in FIG. Figure 3 shows an absorption curve with 50% by volume of hydrogen in the feed. No SO2 is found in the waste gas.

The sulfur dioxide formation depends not only on hydrogen concentration but also on the dispersion of the active compound over the carrier. Figure 4 shows an absorption curve of a 10 wt% molybdenum oxide on silica, SM010 The sulfur dioxide formation is less than with the 20 wt% molybdenum oxide absorbent. The most striking difference, however, is the absorption capacity for hydrogen sulfide. The 10 wt% molybdenum oxide absorbent shows complete conversion to Mo4 sulfide. This is not the case with the SM020. The difference in behavior is undoubtedly due to the difference in dispersion between the two absorbents. Due to the high degree of dispersion of molybdenum oxide, the absolute hydrogen sulfide removal capacity is higher for the 10 wt% product than for the 20 wt% product.

In Figure 5, a number of absorption curves for molybdenum oxide on silica (SM010) at different temperatures are given. H2S is released at 235 ° C after complete absorption. This is probably due to the reaction of molybdenum (VI) sulfide to molybdenum (IV) sulfide. This can also be concluded from figure 6, which shows the hydrogen sulfide uptake as a function of temperature.

The absorption capacity of the molybdenum oxide absorbent decreases at a temperature of about 525 ° C. There are three reasons for this decrease in capacity. First, there is a loss of molybdenum oxide due to evaporation. The second reason is the formation of larger molybdenum oxide particles which, as shown in previous figures, have a lower capacity. The third reason for the reduced capacity is the formation of molybdenum (VI) particles, which occurs at a high temperature by reduction of molybdenum (VI) to molybdenum (IV).

Figure 7 shows the absorption curve of an SFe5.8Moio.6 absorbent. Sulfur dioxide formation is again observed in the reduction of molybdenum (VI) to molybdenum (IV) by hydrogen sulfide. However, this can be suppressed by increasing the hydrogen partial pressure.

Figure 8 shows the capacity of H2S per gram of absorbent for an iron-free and an iron-containing absorbent. It is clear that the absorption of iron in the absorbent increases the capacity.

The iron molybdate absorbent mentioned above still shows some sulfur dioxide production during reduction. This indicates that iron sulfate has formed during regeneration. However, experiments at various temperatures show that the iron sulfates formed from iron molybdate are less stable than those formed with pure iron oxide. The decomposition temperatures of iron sulfate obtained from iron molybdate in oxygen appear to be a function of the iron molybdenum ratio. Table 2 below shows decomposition temperatures for various iron molybdenum ratios.

In figure 9 the capacities of three absorbents are indicated with increasing iron oxide content after a number of absorption regeneration cycles at 400 and 600 ° C. The measurements were taken at these temperatures where no sulfate formation was observed. The results indicate that the addition of iron oxide may improve the capacity of molybdenum oxide based absorbents. The optimum working temperature increases with the iron content. Since the iron molybdate-based absorbents are easier and cheaper to obtain than on molybdenum oxide, the additional costs may outweigh this. Another particularly advantageous property of the iron molybdates is the higher thermal stability, which makes the absorbents less vulnerable to unwanted temperature increases during regeneration.

Experiments have also been conducted to determine the influence of the order of precipitation of iron and molybdenum. Figures 10 and 11 show the absorption curves for a material obtained by precipitating first iron (II) followed by molybdenum (III), respectively. first precipitation of molybdenum (III) followed by precipitation of iron (II).

The first figure has the same characteristic shape as the absorbent obtained by coprecipitation. When the order of precipitation is reversed, the results are completely different. The relevant figure shows that iron sulfate formation does not take place at 400 ° C. This is a particularly remarkable result. Although it is particularly likely that iron oxide will be exposed to food, iron sulfate is not formed. The reason for this is still unclear.

Since the preparation of iron molybdate on silicon dioxide is quite difficult, tests have been carried out with an absorbent on aluminum phosphate. Due to the positive interaction of the metal precursors with the phosphate groups or the surface of the support, the use of iron (III) and molybdenum (VI) ions is possible. Figure 12 shows a hydrogen sulfide absorption curve of 12 wt% iron oxide and 10 wt% molybdenum oxide on aluminum phosphate absorbent PFei2Moio. The same characteristic curve is obtained as with absorbents based on silicon dioxide. However, the capacity is slightly lower due to interaction of some of the active compounds with the phosphate groups. Namely, the metal phosphates are not active in the absorption of hydrogen sulfide.

The use of aluminum phosphate as a carrier therefore considerably simplifies the preparation of iron molybdates. The interaction of the metal precursors with the surface phosphate groups makes it possible to use iron (III) and molybdenum (VI) as precursor during precipitation. This gives a high degree of dispersion and thermal stability.

Claims (26)

Method for desulfurizing gases at elevated temperature, wherein the gas to be desulfurized is contacted with an absorbent based on one or more metal oxides, characterized in that the gas to be desulfurized is contacted with an absorbent based on one of one or more molybdenum and / or tungsten oxides applied to a support. 2. Process according to claim 1, characterized in that a molybdenum oxide is used which is selected from the group consisting of oxides of molybdenum, tungsten and combinations thereof with oxides of iron, vanadium, copper, manganese, cobalt, zinc, nickel and / or chrome. Process according to claim 1 or 2, characterized in that a molybdenum oxide is selected which is selected from the group consisting of molybdenum (VI) oxide, iron (II) molybdate, mixtures of iron (II) - and / or iron ( III) oxide with molybdenum (VI) oxide and mixtures of iron (II) molybdate with molybdenum (VI) oxide and / or iron (II) and / or iron (III) oxide. A method according to claims 1-3, characterized in that the support material is selected from the group consisting of silicon dioxide, aluminum dioxide, silica-alumina, zeolites and aluminum phosphate. Method according to claims 1-4, characterized in that the content of carrier material is between 25 and 99% by weight. A method according to claim 4 or 5, characterized in that an excess of aluminum relative to phosphate groups is present in the aluminum phosphate support material. Process according to claim 4 or 5, characterized in that the aluminum phosphate has an aluminum base relative to phosphate groups. Process according to claim 4 or 5, characterized in that the aluminum phosphate contains a substantially stoichiometric amount of aluminum relative to phosphate groups. Method according to claims 1-8, characterized in that the gases are brought into contact with the absorbent at a temperature of 150 to 800 ° C, preferably of 150 to 625 ° C. Method according to claim 9, characterized in that the contact takes place at a temperature of 150 to 400 ° C if the absorbent is based on molybdenum or tungsten oxide; and the contact is made at a temperature of 150 to 625 ° C, preferably 400 to 625 ° C if the absorbent is based on a combination of molybdenum and / or tungsten oxide with at least one iron oxide. A method according to claims 1-10, characterized in that the absorbent has a specific surface area of 5 to 500 m2 / g. Method according to claims 1-11, characterized in that the absorbent has an average pore diameter of 1 to 100 nm. Process according to claims 3-12, characterized in that the Fe / Mo molar ratio is 0 to 20. A method according to claims 1-13, wherein the absorbent contains 5 to 25% by weight of iron oxide. A method according to claims 1-14, wherein the gas to be desulfurized is selected from the group consisting of gases from coal and gasification: A method according to claims 1-15, wherein the treatment with the absorbent takes place under conditions that the content of H 2 S in the treated gas is at most 100 ppm, more in particular at most 25 ppm. The method of claims 1-16, wherein the contact time between gas and absorbent is between 0.01 and 10 s. Absorbent for sulfur compounds suitable for high temperature desulfurization of gases according to claims 1-17, comprising an aluminum phosphate carrier material with one or more molybdenum oxides applied thereon. Absorbent according to claim 18, characterized in that a molybdenum oxide used is selected from the group consisting of molybdenum (VI) oxide, iron (II) molybdate, mixtures of iron (II) - and / or iron (III) - oxide with molybdenum (VI) oxide and mixtures of iron (II) molybdate with molybdenum (VI) oxide and / or iron (II) and / or iron (III) oxide. Absorbent according to claim 18 or 19, characterized in that the content of carrier material is between 25 and 99% by weight. Absorbent according to claims 18-20, characterized in that an excess of aluminum relative to phosphate groups is present in the aluminum phosphate. Absorbent according to claims 18-20, characterized in that in the aluminum phosphate there is an undersize of aluminum over phosphate groups. Absorbent according to claims 18-20, characterized in that the aluminum phosphate contains a substantially stoichiometric amount of aluminum relative to phosphate groups. Absorbent according to claims 18-23, characterized in that the absorbent has a specific surface area of 5 to 500 m2 / g. Absorbent according to claims 18-24, characterized in that the absorbent has an average pore diameter between 1 and 100 nm. Absorbent according to claims 18-25, characterized in that the Fe / Mo molar ratio is 0 to 20.