PROCESS FOR THE HYDROGENATION OF ALKYLARYL KETONES

The present invention relates to a process for the hydrogenation of alkylaryl ketones such as acetophenone . Background of the invention

Processes for hydrogenation of alkylaryl ketones such as acetophenone to the corresponding alcohols are well known in the art. Processes generally comprise hydrogenation of acetophenone by contacting feed comprising acetophenone with hydrogen at elevated pressure and temperature in the presence of a heterogeneous catalyst. Such processes have been described in prior art such as EP-A-0714877.

US-A-6, 046, 369 teaches hydrogenation of acetophenone to 1-phenylethanol at a liquid hold-up ratio of from 30 to 90%, preferably of from 40 to 70% in order to be able to apply relatively mild process conditions such as a temperature of from 60 to 150 ⁰ C and a pressure of from 1 MPa to 5 MPa wherein the amount of ethylbenzene produced as by-product is controlled at sufficiently low level. The use of downflow operation is discouraged in US-A-6, 046,369 (column 4, lines 8-24). Comparative Example 1 of US-A-6, 046 , 369 exemplifies downflow operation at a liquid hold-up ratio of 25%. In this comparative example, the conversion of acetophenone was lower and the amount of ethylbenzene produced was higher than those in Examples 1-3 of US-A-6, 046, 369.

WO 2004/085354 relates to hydrogenation of alkylaryl ketones, such as acetophenone, in the presence of 0.5 to 30 %wt of phenolic compounds. This document contains no information on the liquid hold-up either recommended or applied for this process.

Summary of the invention

US-A-6, 046, 369 warns against the use of low liquid hold-up, that is to say less than 30%, in the hydrogenation of acetophenone . However, it has now been found that a relatively low liquid hold-up can still be applied in the hydrogenation of alkylaryl ketones such as acetophenone into the corresponding alcohols at high conversion and yield and with little formation of byproducts, if some specific requirements are met. A low liquid hold-up has the advantage that it reduces the pressure drop over a reactor.

The present invention relates to a process for the hydrogenation of an alkylaryl ketone which process comprises contacting a feed containing an alkylaryl ketone with hydrogen in trickle-flow mode in the presence of a heterogeneous hydrogenation catalyst, in which process the dynamic liquid hold-up is of from 10 to less than 30% and the external area of the catalyst particles is effectively wetted by the liquid flowing down. Detailed description of the invention

Hydrogenation within the context of the present application is understood as the chemical reaction of the alkylaryl ketones with molecular hydrogen in the presence of a suitable catalyst, as for instance described in Ullmann's Encyclopedia of Industrial Chemistry,

5th edition, Volume A13, pages 407-410. Hydrogen is added in this reaction to the carbon-oxygen double bond of the alkylaryl ketones, which thereby are converted to the corresponding alkylaryl alcohols. The term alkylaryl alcohol describes α- and/or β-aryl alkanol, and mixtures thereof .

The process according to the present invention is carried out in trickle-flow mode. Trickle-flow mode means

that liquid phase flows downward through a fixed bed of catalyst particles . Gas phase may flow either upward or downward. The liquid and gas flow velocities are so low that the liquid flows as a continuous film over the catalyst. The gas phase is the continuous phase in the catalyst bed. It is well known to someone skilled in the art how to attain a trickle-flow mode. It is known that important features are the gas and liquid velocities, the particle size and the physical properties of the liquid. Criteria for obtaining trickle flow, by choosing superficial velocities of gas and liquid of particular density, viscosity and surface tension, are for example given in Ullmann's Encyclopedia of Industrial Chemistry, 5th edition, Volume B4, pages 311-312. The process of the present invention is carried out such that the liquid is in downflow. Liquid downflow operation is advantageous in that such operation gives lower axial dispersion of the reactants, less flow non- idealities such as channelling in the catalyst bed and lower pressure drop.

The liquid hold-up is the liquid volume contained in a unit bed volume, or the liquid volume fraction in the total bed volume. The dynamic liquid hold-up in a trickle-bed process concerns the liquid which drains out of the bed after the gas and liquid flows have been stopped. The dynamic liquid is contrary to the static liquid. Static liquid remains around the catalyst contacting points due to capillary forces. The dynamic liquid hold-up can be measured by a sudden shut-down of the liquid and gas feeds and closing of the gas/liquid outlet from the bottom of the reactor. The dynamic liquid hold-up then concerns the liquid which has drained out of the bed and collects at the bottom of the reactor. The

dynamic liquid hold-up in the present invention is of from 10 to less than 30%, more specifically of from 10 to less than 25%, even more specifically from 15 to less than 25%, and most specifically of from 15 to 20%. The fraction of the external area of the catalyst particles effectively wetted by the liquid flowing down, is called the wetting efficiency. This wetting efficiency may, in theory, vary from 0 (0% wetting) to 1 (100% wetting) . If the liquid flows too slowly the liquid film can break up into rivulets leading to incomplete liquid coverage of the external catalyst surface. Proper design of the gas and liquid distribution and of the catalyst bed, can reduce or prevent dry zones. Typically, the wetting efficiency varies between 0.6 and 1. According to the present invention, the external area of the catalyst particles should be effectively wetted by the liquid flowing down. Preferably, the wetting efficiency is of from 0.8 to 1.0, more specifically of from 0.9 to 1.0. More preferably, the external area of the catalyst particles is essentially completely wetted, which in this specification means that the wetting efficiency is 0.99 or higher. Most preferably, the wetting efficiency is 1.0.

Correlations for wetting efficiency are, for example, given in Ullmann's Encyclopedia of Industrial Chemistry, 5th edition, Volume B4, page 317. Further, S. T. Sie has defined a wetting criterion in Revue de l'institut Frangais du Petrole, vol. 46, no. 4, July- August 1991, pages 501-515. This criterion was developed specifically in relation to the trickle-flow hydrotreating of heavy oil fractions. In trickle-flow hydrotreaters, liquid oil and gaseous hydrogen flow down cocurrently over a fixed bed of solid catalyst particles.

The heavy oil fractions generally processed in the trickle-flow mode, are composed of very many different molecules. The above article is silent to any applicability of said criterion in the present field of hydrogenation of alkylaryl ketones.

In the above criterion defined by Sie, the so-called "wetting number" W should exceed 4 * ICT ⁶ . W is dimensionless and is defined as:

W = ( η * u ) / (p * d _p ² * g) wherein η is the dynamic viscosity of the liquid (unit: N s πT ² or kg s ^"1 πT ¹ ) . u is the superficial velocity of the liquid (unit: m s ^"1 ), which is the quotient of the volumetric flow rate of the liquid (unit: m ³ s ^"1 ) divided by the cross- sectional flow area (unit: m ² ) . In the present invention, where a fixed bed of catalyst particles is used, the velocity calculated as if the reactor were empty (i.e. without any catalyst particles) is the "superficial" velocity. p is the specific gravity (density) of the liquid (unit : kg πT ³ ) . d _p is the catalyst particle diameter (unit: m) . g is the acceleration due to gravity (which amounts to 9.81 m s ^~2 ) .

The quotient of the dynamic viscosity divided by the density of the liquid (i.e. η / p) is the kinematic viscosity of the liquid (i.e. V; unit: m ² s ^"1 ) .

According to Sie, if W > 4 * 10 ^~6 , then there is question of a wetting efficiency of 1.0, implying that the external area of the catalyst particles is completely wetted by the liquid flowing down. Further, this implies that there is question of an even irrigation of the

completely wetted external area of the catalyst particles, as described by Sie in the above-mentioned article (see especially Fig. 10 on page 510) .

It has been found by the present inventors that the above criterion as defined by Sie is also indicative for good reaction results in terms of conversion and yield, in the hydrogenation of alkylaryl ketones as is demonstrated in the Examples below with reference to the conversion of acetophenone to 1-phenylethanol . Therefore, most specifically for the present invention the above criterion as defined by Sie is satisfied. Said criterion may be satisfied by recycling part of the liquid reaction mixture to the inlet of the reactor, thereby increasing the volumetric flow rate of the liquid reaction mixture flowing through the reactor. This will increase the superficial velocity of the liquid reaction mixture, and as a consequence the wetting number W is also increased. Alternative or additional measures to achieve this, is to increase the fresh liquid feed rate and to reduce the reactor diameter, as this would also increase the superficial velocity of the liquid reaction mixture.

Suitable catalysts for use in the present invention may contain as metal or metal compound at least one metal selected from the group consisting of groups IA, HB, VI and VIII of the Periodic System. Suitable catalysts comprise at least one of the metals or metal compounds selected from groups VI, VIII and IB, such as chromium, copper, zinc, nickel, palladium and platinum. Preferably, the hydrogenation catalyst comprises copper, chromium and/or palladium as metal or metal compound, as these catalysts usually are not apt to hydrogenate the aromatic ring system under the conditions typically used for this process. Most preferably, the hydrogenation catalyst

comprises both copper and chromium as metal or metal compound. Such catalysts have shown a high catalytic activity and selectivity over a long period of operation.

Usually, the catalyst is activated by reduction of the catalyst before start of operation. The reduction of the catalyst is preferably carried out in the presence of a liquid phase with the help of hydrogen. A specific preferred process has been described in GB-B-587, 181. Suitably, within the context of the present application, the alkylaryl ketones employed are ketones of alkylated benzenes in which the alkyl substituents are straight or branched alkyl substituents comprising from 2 to 10 carbon atoms, for example ethyl, n-propyl, n-butyl and/or sec-butyl substituents. More suitably, the alkylaryl ketone to be hydrogenated in the present process, is acetophenone . A commonly known process in which feeds containing acetophenone are produced is the preparation of propylene oxide, more specifically a process comprising: (i) contacting a feed comprising ethylbenzene with oxygen to obtain a feed comprising ethylbenzene hydroperoxide and acetophenone,

(ii) contacting the feed obtained in step (i) with propene in the presence of a catalyst to obtain a reaction mixture comprising propylene oxide, 1-phenylethanol and acetophenone, and

(iii) removing at least part of the propylene oxide and 1-phenylethanol from the reaction mixture obtained in step (ii) to obtain the feed containing acetophenone. The product obtained in step (ii) comprises 1- phenylethanol . This 1-phenylethanol can be dehydrated to obtain styrene. Therefore, the above process may further comprise (iv) dehydrating at least part of the

1-phenylethanol obtained in step (ii) to obtain styrene. The 1-phenylethanol can be dehydrated either when still being part of the reaction mixture as obtained in step (ii) or the 1-phenylethanol can have been separated off optionally in combination with other compounds .

Process conditions which can be used in the process of the present invention comprise a temperature of from 50 to 250 ⁰ C, more preferably of from 60 to 170 ⁰ C, even more preferably of from 70 to 160 ⁰ C, and most preferably of from 80 to 150 ⁰ C, and a pressure of from 0.1 to

100 * 10^ N/m ² (bar), more preferably of from 1 to 50 * 10 ⁵ N/m ² , most preferably of from 10 to

40 * 10 ⁵ N/m ² .

The process according to the present invention is further illustrated by the following examples wherein a feed comprising acetophenone is hydrogenated. In the following examples, the conversion of acetophenone is expressed as the molar flow of acetophenone converted divided by the molar flow of acetophenone supplied times 100%. The yield of 1-phenylethanol is expressed as the molar flow of 1-phenylethanol produced divided by the molar flow of acetophenone supplied times 100%. Comparative Example 1

The following experiment was carried out in downflow in a bench scale unit comprising a reactor connected to a heating/cooling system, a high pressure feed pump, a high pressure pump for recycling product to the feed, and two vessels (for incoming and outgoing feed streams), and a gas inlet connected to sources of hydrogen and nitrogen. 560 g of a copper-chromium catalyst (3 mm * 3 mm tablets) were added to the reactor. The bed inside the reactor comprised said catalyst particles and glass balls, which were added to fill the remaining empty space above the

catalyst particles. The glass balls had a diameter of 3 mm and were used to provide adequate fluid distribution. The volume ratio of catalyst particles to glass balls was about 95:5. The section of the reactor containing the catalyst particles and glass balls had a cylindrical shape. The diameter of this cylindrical section was 32 mm. The reactor was first purged with nitrogen at a pressure of 2.3 * 10^ N/m ² , then the reactor temperature was raised to 130 ⁰ C. Hydrogen was introduced at a concentration of 1% volume, then the hydrogen concentration was gradually increased to 100% volume at a rate such that the reactor temperature did not exceed 170 ⁰ C. The temperature was then raised to 175 ⁰ C which temperature was maintained for 4 hours. A fresh liquid feed comprising acetophenone was then fed to the reactor at a rate of 400 g/hour, together with a hydrogen gas stream which was fed at a rate of 60 liter/hour (normal conditions). The reactor temperature was decreased by controlling the inlet temperature of the feed at 102 ⁰ C. The outlet temperature was 113 ⁰ C. The hydrogen pressure in the reactor was increased to

24 * 10 ⁵ N/m ² .

The fresh liquid feed comprised 49.3 %wt of acetophenone, 16.5 %wt of 1-phenylethanol, 15.3 %wt of 2-phenylethanol, 8.4 %wt of ethylbenzene, 2.4 %wt of phenylacetaldehyde, 0.7 %wt of styrene, 0.4 %wt of benzylalcohol, 0.1 %wt of benzaldehyde, the remainder comprising other aromatic compounds (but no phenol in an amount of 0.1 %wt or higher) . Part of the liquid reaction product was recycled to the inlet of the reactor at a rate of 400 g/hour. Therefore, the total rate of liquid feed to the inlet of the reactor was 800 g/hour, that is 1.43 g liquid/g

catalyst/hours. The process was carried out in trickle- flow downflow mode, meaning that both liquid phase and gas phase flowed concurrently downward through the catalyst bed. At 2.7, 3.7, 4.7 and 5.7 hours after the start of the feed of the liquid to the reactor, samples were taken from the reaction product. For each of the samples, the conversion of acetophenone and the yield of 1-phenylethanol were measured. The average conversion of acetophenone was 60.2% and the average yield of 1-phenylethanol was 55.4%. Further, ethylbenzene formation from acetophenone was at an average level of 1.7 %wt on total feed.

The dynamic liquid hold-up was measured by first shutting down the liquid and gas feeds and closing the gas/liquid outlet from the bottom of the reactor. The liquid which drained out of the bed and collected at the bottom of the reactor was measured, and amounted to 20 cm ³ . Then, after removing the drained-out liquid, the bed volume was measured by measuring the amount of ethylbenzene needed to completely fill the bed whilst the gas/liquid outlet was closed. This bed volume amounted to 151 cm ³ , and it does not include the volume occupied by the catalyst particles. The dynamic liquid hold-up could finally be calculated, and amounted to 13%.

Further, the wetting number W was 2 * 10 ^~6 . W was calculated using the above-mentioned formula as given by Sie . The dynamic viscosity (η) and density (p) of the liquid reaction mixture were 6.3 * 10 ^~4 kg s ^"1 m ^"1 and 906 kg πT ³ , respectively. The kinematic viscosity (V; i.e. η/p) of the liquid reaction mixture was therefore 7.0 * 10 ^~7 m ² s ^"1 . Further, the particle diameter (d _p ) was

3 * ICT ³ m and the acceleration due to gravity (g) is 9.81 m s ^"2 .

As discussed above, the superficial velocity (u) of the liquid reaction mixture is defined as follows: (volumetric flow rate of the liquid) / (cross-sectional flow area) . Since the section of the reactor containing the bed had a cylindrical shape, the cross-sectional flow area is π r ² , wherein r is the radius of the circular cross-section which is half the diameter which was 32 mm. Therefore, the cross-sectional flow area was

8.0 * 10 ^~4 m ² . The volumetric flow rate is the volume of liquid reaction mixture flowing through the bed per time unit. Therefore, the volumetric flow rate is the quotient of the sum of the fresh liquid feed rate (400 g/hr) and the recycle feed rate (400 g/hr) divided by the density of the liquid reaction mixture (906 kg πT ³ ) . Therefore, in this Comparative Example, said superficial velocity was 3.1 * 10 ^~4 m s ^"1 . Example 1 Comparative Example 1 was repeated except that the inlet temperature of the feed was controlled at 90 ⁰ C and the outlet temperature was 109 ⁰ C. Further, part of the liquid reaction product was recycled to the inlet of the reactor at a rate of 1600 g/hour instead of 400 g/hour . Therefore, the total rate of liquid feed to the inlet of the reactor was 2000 g/hour, that is to say 2.5 times higher than in Comparative Example 1, that is 3.57 g liquid/g catalyst/hour.

At 5, 6, 7 and 8 hours after the start of the feed of the liquid to the reactor, samples were taken from the reaction product. For each of the samples, the conversion of acetophenone and the yield of 1-phenylethanol were measured. The average conversion of acetophenone was

62.3% and the average yield of 1-phenylethanol was 58.5%. Further, ethylbenzene formation from acetophenone was at an average level of 1.3 %wt on total feed.

The dynamic liquid hold-up was measured in the same way as in Comparative Example 1 and amounted to 19% (29 cm ³ of liquid drained out of the bed; and a bed volume of 151 cm ³ ) .

Further, the wetting number W was 6 * ICT ⁶ . This wetting number W was calculated in the same way as in Comparative Example 1. The only difference was that in Example 1 the recycle feed rate was 1600 g/hr (not 400 g/hr as in Comparative Example 1) . This resulted in a different superficial velocity for Example 1, which was 7.7 * 10 ^~4 m s ^"1 . Thus, in Example 1 where the above-mentioned criterion for wetting efficiency (W > 4 * 10 ^~6 ) was satisfied, the acetophenone conversion and the 1-phenylethanol yield were higher than in Comparative Example 1, despite the lower average reaction temperature in Example 1. That is to say, if the reaction temperature in Example 1 would be the same as that in Comparative Example 1, even better results would be expected for Example 1. In addition, in Example 1, ethylbenzene formation from acetophenone was at a lower level than in Comparative Example 1.

The above Examples demonstrate that even when the dynamic liquid hold-up is relatively low, for example less than the lower limit of 30% as taught in US-A-6, 046,369 (13% in above Comparative Example 1; 19% in above Example 1), still good reaction results can be obtained when the external area of the catalyst particles is effectively wetted by the liquid flowing down. Suitably, the above-mentioned criterion for wetting

efficiency (W > 4 * 10 ⁶ ) is satisfied in order to obtain such good reaction results.