CZ20002004A3 - A process for the catalytic performance of multiphase reactions, in particular hydroformylation - Google Patents

Abstract

The method of carrying out multiphase reactions in a tubular reactor consists in that the catalyst is contained in the continuous phase and at least one starting material is contained in the dispersed phase and the loading factor B of the tubular reactor is equal to or greater than 0.8. The catalyst may be a metal of the VIII subgroup of the periodic table of elements, e.g. rhodium. The starting material may be olefins with a number of carbon atoms of 2 to 25.

Description

Method for catalytically performing multiphase reactions, especially hydroformylation

Technical field

The invention relates to a method for catalytically carrying out multiphase reactions in a tubular reactor, in particular for the production of aldehydes by reacting olefinically unsaturated compounds with hydrogen and carbon monoxide in the presence of a catalyst.

State of the art

Aldehydes are used in the synthesis of many organic compounds. The products of their direct conversion are alcohols and carboxylic acids, which have technical applications. The alcohols produced from these aldehydes are used, among other things, as solvents and as precursors for the production of plasticizers and detergents.

It is known that the reaction of olefins with carbon monoxide and hydrogen can produce aldehydes and alcohols. The reaction is catalyzed by hydridocarbonyls of metals, especially metals of group VIII of the periodic table. In addition to cobalt, which has a large technical use as a catalyst metal, rhodium has recently gained increasing importance. In contrast to cobalt, rhodium allows the reaction to be carried out at low pressures. The hydrogenation of olefins to saturated hydrocarbons is clearly lower when rhodium catalysts are used than when cobalt catalysts are used.

In the methods used in the art, a rhodium catalyst is formed during the process from a catalyst precursor, synthesis gas and optionally other modifying ligands. When using modified catalysts, the modifying ligands may be present in excess in the reaction mixture. Tertiary phosphines or phosphites have proven themselves as ligands. Their use makes it possible to reduce the reaction pressure to values below 30 MPa.

However, problems with this method arise from the separation of the reaction products and the regeneration of the catalysts homogeneously dissolved in the reaction product. Generally, for this purpose, the reaction product is distilled from the reaction mixture. In practice, due to the sensitivity of the catalyst or the products formed to heat, this can only be done in the hydroformylation of lower olefins with up to about 5 carbon atoms in the molecule.

Industrially, _C4 and _C5 aldehydes are produced, for example, according to DE 3234701 or DE 2715685.

In the latter method, the catalyst is dissolved in an organic phase consisting of the product and high-boiling components (derived from this product). Olefin and synthesis gas are introduced into this mixture. The product with synthesis gas is withdrawn from the reactor, in a more recent variant it is pumped over as a liquid. Since the catalyst slowly reduces its capacity, a part of it together with the high-boiling components must be continuously withdrawn and replaced by an equivalent amount. Due to the high price of rhodium, the regeneration of rhodium from the withdrawn stream is necessary. This regeneration process is complex and therefore burdens the method.

• 9 ·♦ ·· ··· ·· • ·

According to DE 2324701, this disadvantage is overcome by dissolving the catalyst in water. The solubility of the rhodium catalyst used in water is achieved by using trisulfonated triarylphosphines as ligands. Olefin and synthesis gas are introduced into the aqueous phase. The product formed by the reaction forms a second liquid phase. The liquid phases are separated outside the reactor and the separated catalyst phase is fed back into the reactor.

Technically, both of these processes are multiphase reactions.

The term multiphase reaction is used in the following text to refer to reactions that take place in the presence of two or more immiscible or only partially miscible liquid phases. This includes, for example, reactions between a gas and a liquid phase (gl), between two liquid phases that are immiscible or have an immiscibility region (11), and reactions that involve two liquid immiscible or only partially miscible phases and one gas phase (gll).

In addition, however, it is also possible to use other fluid phases, for example supercritical phases. Such a supercritical phase can be present as an alternative to those mentioned, but also in addition.

Examples of technically important gas-liquid (gl) reactions besides the hydroformylation of liquid olefins using a catalyst dissolved in the organic phase are the reaction of acetylene with carboxylic acids or hydrogenation using homogeneously dissolved catalysts or oxidation with air or oxygen.

·· ······ ·· φ • · · ·· · · ft · · • · · · · ftft · · · · ······· ftftftft · ••ft® · · · ft ftft · ·· ·· ftft ftft ftft ftftft

The cases mentioned have in common the problem of substance transfer, since these reactants exist in different phases. In the case of hydroformylation in a three-phase system, the process is particularly difficult to implement, since the reactants exist in three separate phases. Not only the olefin, but also the synthesis gas must be transported to the aqueous catalyst phase in order to come into contact with the catalyst. Finally, there must be a return transport from the aqueous phase. Since the transport processes are often slower than the reactions themselves, such reactions are determined by the substance transfer rate, i.e., they are transport-inhibited reactions. Multiphase reactions are associated with a number of problems that make their technical implementation considerably more difficult compared to simple homogeneous reactions. Some typical problems are listed below.

In all cases, the substances must be brought into as close contact as possible with each other in order to minimize the problem of substance transfer: the largest possible surface area for substance transfer and the interphase _interface must be created. On the other hand, it must be possible to easily separate the phases again after the reaction has taken place: too much mixing can lead to problems here. In the presence of two liquid phases, an emulsion can form, in the case of a gas-liquid process, foaming. In the case of the three-phase processes mentioned, all of the problems can even occur simultaneously.

In addition to the large surface area of the substance transition _{a s} , the highest possible substance transition coefficient ki should be achieved in all multiphase reactions. In total, the so-called KLA value, i.e. the product of ki and _{a s} in the substance transition equation j = ki*a _s * (C* - C) where • · ·

·· ·« ·· • · · · • · · · · · j [mol/s], mass flow of reacting components penetrating the phase interface, ki [m/s], mass transfer coefficient, a _s [m ² ], area of the phase interface in the reactor,

C* [mol/dm ³ ] , maximum solubility of the starting substance in the second phase and

C [mol/dm ³ ] , the actual concentration of the starting substance, which is in turn linked to the reaction rate, to be maximum.

Another problem in multiphase reactions is the heat removal in exothermic reactions. If the reaction rate is increased by improving the mass transfer, more heat must naturally be removed, which can lead to an undesirable increase in temperature until the reaction is complete.

This issue of multiphase reactions can be technically solved, for example, by using a stirred tank reactor.

For the hydroformylation of olefins, the use of a high-pressure stirrer is disadvantageous because it requires the design of a hermetically sealed shaft, which is prone to failure. In addition, multiple stirrers must be used in this process to achieve sufficient phase mixing.

The large heats of reaction generated during hydroformylation can generally only be controlled by using heat exchangers located inside the reactor.

The use of a stirrer vessel always leads to back-mixing, thereby reducing the effective concentration of the reaction substrates, which • · • · · · · ·

This disadvantage has to be compensated by the investment of expensive reaction space.

In the hydroformylation of olefins, the space-time yield decreases sharply with increasing number of carbon atoms in the olefin. The connection between molecular size and reaction rate is known (B. Cornils, W. Hermann, Aqueous-Phase Organometallic Catalysis, Concepts and Application, Verlag Wiley-VCH, pp. 308-310).

For example, the reaction rate ratio of the hydroformylation of 1-pentene and 1-hexene is 2.6:1. The conversion of higher olefins is therefore becoming increasingly uneconomical.

In view of the foregoing explanations, there is a need for a method that avoids the above-mentioned disadvantages and is at the same time simple to implement technically.

The essence of the invention

The object of the present invention is therefore to develop a method for carrying out multiphase reactions which is particularly suitable for the production of aldehydes by catalytic hydroformylation of olefins.

Technically, the new method should meet the following requirements for a multi-phase method:

• creation of a high and stable substance transition between the phases involved, • simple to implement, if possible with conventional technical equipment, • simple and reliable heat dissipation, • high operational safety,

• simple and reliable scaling.

Due to the production of aldehydes to be carried out, they specifically approach it:

• high selectivity, suppression of especially high-boiling by-products, • no or only slight catalyst removal, • high space-time yield, small reactors, • high product purity.

Using the method according to the invention, a surprisingly simple method for carrying out multiphase reactions has been found, which can be carried out in a tubular reactor, optionally filled with packing bodies or mounted elements, and which is suitable for the hydroformylation of olefins with synthesis gas at high space-time yields and selectivities.

The subject of the present invention is therefore a method for catalytically carrying out multiphase reactions in a tubular reactor, wherein the catalyst is contained in the continuous phase and at least one starting material in the dispersed phase, and the reactor loading factor B is equal to or greater than 0.8.

In a special embodiment of the present invention, olefins are hydroformylated by multiphase reactions, i.e. converted to aldehydes using synthesis gas.

The tubular reactor used according to the invention may contain packing bodies or mounted elements. Packing bodies in the sense of the present invention are, for example, Raschig rings, saddles, Pall rings, plates, wire mesh coils, wire mesh screens. Examples of mounted

The elements are filter plates, flow-closing barriers, column trays, perforated plates or other mixing devices. However, several narrow tubes connected in parallel are also possible as mounted elements in the sense of the present invention, resulting in a multi-tube reactor. Structured mixing packings or demister packings are particularly preferred.

The crucial importance in the method according to the invention is that the minimum reactor cross-sectional load is observed or exceeded. When the reactor is operated upwards (flow direction from bottom to top), the point of flooding of the packing should be exceeded. The reactor is therefore operated above the point at which a bubble column is usually operated. When the reactor is operated downwards (flow direction from top to bottom), the reactor cross-sectional load is set so that the reactor is completely flooded. The operation is therefore above the point at which it can still be called a trickle bed.

To more accurately determine the minimum reactor loading to be observed, the loading factor B of the tubular reactor is calculated as a dimensionless pressure loss using

B = PD/PS , where PD [Pa/m] means the pressure loss related to the length of the reactor under operating conditions and PS [Pa/m] means the theoretical quantity with the dimension of pressure related to the length, defined as the ratio of the mass flow M [kg/s] of all components in the reactor to the volume flow V [m ³ /s] of all components under operating conditions multiplied by g = 9.81 m/s ² , i.e. PS = (M/V)*g.

·· ·4 4 4 · 4 · 4 ·· • · 4 4 4 4 4 4 4 4 •44·· 44 · ·« • ·· 444 4 44· 4

44 »4 · 4

Illustratively, PS would be the static pressure per meter of a multiphase mixture in a vertically positioned tube if all phases flowed at the same rate. PS is a theoretical value that results from the amount of flows fed into the reactor and can be given independently of the direction of reactor flow, the flow rate of all phases, or the state of flooding of the reactor.

The pressure loss PD [Pa/m] is used as a theoretical quantity to determine the process conditions and can be calculated using the usual methods for single-phase or multi-phase flows. Common methods for calculating the pressure loss PD in pipes, installed elements or filter media from packing bodies, etc. can be found, for example, in the VDI-Wärmeatlas, 7th expanded edition, VDI-Verlag GmbH, Dusseldorf 1994, paragraphs Lal to Lgb7 and also in the standard work by Heinz Brauer, Grundlagen der Einphasen- und Mehrphasenströmungen, Verlag Sauerländer, Aarau and Frankfurt am Main, 1971.

The pressure loss PD for single-phase flow through an empty pipe is given by

PD = Cw* p/2*w ² /D , s

p [kg/m ³ ], density of the flowing medium under operating conditions, w [m/s], flow velocity (volumetric flow/cross-sectional area,

D [m], pipe diameter and

Cw [-], coefficient of hydraulic resistance of the flowing pipe.

Μ ···· • · · · · · • * ··· 9 9 9 • 9 9 9 9 9 9 9 • 9 9 9 9 9 9 9 • · ·· ·· ··

When flowing through packing bodies, filter media or mounted elements, the velocity w can be replaced by the effective velocity (w/ψ) and the pipe diameter D can also be replaced by the hydraulic channel diameter d _H of the packing bodies or mounted elements, so that

PD = Cw* p/2* (w/ψ) ² *l/d _H , s

Ůh [m], hydraulic diameter of the channel,

Ψ [~] _r empty tube fraction,

Cw [-], coefficient of hydraulic resistance of the flow-through device with filling.

The data d _H and ψ specific to the filler bodies are often part of the specifications provided for the filler bodies. For a number of filler bodies, the data is given in the above-mentioned VDI-Wármeatlas.

The empty tube fraction ψ can also be determined experimentally by, for example, determining the volume of the reactor before and after filling with packing bodies. The hydraulic diameter of the channel can be calculated, if not known, from the specific surface area F [m ² /m ³ ] of the packing bodies or mounted elements (usually known or determinable experimentally) according to the simple relationship d _H = 4vg/F.

The coefficient of hydraulic resistance of pipes, mounted elements and fillers is usually described in terms of the Reynolds number Re, which provides information about the flow state under selected conditions. For mounted elements and fillers, etc., the following relationship can almost always be used:

Cw = Ki/Re ⁿ + K ₂ /Re ^m , where n = 1, m = 0 is often used (compilation according to S. Ergun, Chem. Engng. Progr. 48, (1948), 89) or n = 1,.

m = 0.1 (compilation according to Brauer et al.). Κχ, K ₂ are constants specific to the packing bodies, which are known from the data provided or from the literature (examples can be found in the VDI-Wármeatlas and in Brauer et al.). However, they can also be determined experimentally by operating the reactor with packing bodies with liquid at different speeds and determining Cw as a function of Re from the known data and the measured pressure losses.

The dimensionless Reynolds number Re is finally defined as

Re = w*(p/r|)*D for an empty tube, or

Re = (w/ψ) * (ρ/η)*d _H for a pipe with mounted elements or filler bodies, η [Pa*s] denotes the viscosity and p [kg/m ³ ] the density of the flowing medium.

The pressure loss in two-phase flows (here gas-liquid for synthesis gas/catalyst solution) increases excessively. Usually, according to Lockhart-Martinnelli (in Brauer et al.), the pressure loss of a two-phase flow P _lg is related to the pressure loss of one of the two phases, for example the pressure loss of the pure flowing liquid phase Ρχ, and is substituted into the relation for the pressure loss ratio of the two phases Ρχ and P _g considered as flowing separately.

·· ···· ·· ·· • · · · · · ·

For the calculation of pressure losses in two-phase flows, dimensionless pressures according to φ ² = Pig/Pi and X ² = Pi/Pg are often used. Another dependence φ ² = function (X ² ) is often investigated. Examples can be found in the following places in the literature:

Y. Sáto, T. Hirose, F. Takahashi, M. Toda: Pressure Loss and Liquid Hold Up in Packed Bed Reactor with Cocurrent Gas-Liquid Down Flow; J. Chem. Eng. Of Japan, s/. 6 (Z.. 2), 1973, 147-152;

D, Sweeney: A Correlation for Pressure Drop in Two-Phase Concurrent Flow in Packed Beds; AIChE-Joumal, vol.13, 7/1967, 663-669,

V.W. Weekman, J.E. Myers: Fluid-Flow Characteristics of Concurrent Gas-Liquid Flow in Packed Beds, AIChE-Joumal, s/. 10 (c. 6), 11/1964, 951-957,

R.P. Larkins, R.P. White, D.W. Jeřfrey: Two-Phase Concurrent Flow in Packed Beds; AIChE-Joumal, 2), 6/1961, 231-239 or

N. Midoux, M. Favier, J.-C. Charpentier; Flow Pattern, Pressure Loss and Liquid Iloldup Data in Gas-Liquid Down-Flow Packed Beds with Foanring and Nonfoamins Liquids; J. Chem. Eng. Of Japan, 5), 1976, 350-356.

The relationship proposed by Midoux, which has been tested for many gas-liquid systems, is often used for calculation. For example, for non-foaming systems, φ ² = 1 + 1/X + 1.14/X ⁰ ' ⁵⁴ .

This relationship, named after Lockart-Martinelli, is shown graphically in many works, and detailed discussions of it can be found in many textbooks on manufacturing process technology, for example also in Brauer el al.

«φ ·* ···· · · ........

;···♦·· Φ·*·

-, φφφφ φφφ· · · ·♦ ·· ·· ·φ φ·

The pressure loss of the two-phase flow P _gl results from the experimentally determined pressure loss of the pure flowing liquid phase Pi or from the pressure loss of the pure flowing liquid phase P _lA estimated as explained above using

Pgi = Φ ² * Pi ·

In the special case of aldehyde production by hydroformylation of olefins, the calculation of the pressure drop is even more complex. In addition to the synthesis gas phase and the liquid catalyst phase, the presence of an organic liquid phase must be taken into account. This issue can be computationally challenging by determining an additional dimensionless pressure <t> ² org = Pgii/Pig, so the pressure drop can be calculated as follows:

Pgii = Φ ² * Φ ² ογ ₉ * Pi

Generally, for reactor length L [m],

PD = Pgi/L, or PD = P _gll /L.

The pressure loss of the multiphase flow can thus be calculated using the usual means of chemical process technology. This applies analogously to the previously defined dimensionless pressure loss B, i.e. the loading factor of the multiphase reactor.

The magnitude of the dimensionless load factor B represents a necessary basic condition of the method according to the invention; B should be greater than or equal to 0.8, preferably greater than or equal to 0.9 or particularly preferably greater than or equal to 1.

4* 444« ·· • 4 4 • · 4 4· • · 4 4 • 4 4 4

4· • 4 4

4 4 • · 4 4 * · 4 9

4· 44

In the range of B greater than or equal to 0.8, the reactor operated from the top down begins to flood. It should be emphasized that, if these conditions are met, the advantages of the method according to the invention are also achieved when the reactor is operated from the bottom up or in another direction.

Higher reactor cross-section loadings (B >> 1), observable on the increasing pressure difference across the reactor, are possible at any time and even desirable, as long as the increasing space-time yields justify a correspondingly increasing energy consumption. The upper limit is therefore only given by practical considerations, such as energy consumption or the difficulties in separating the phases after the reaction has taken place. It is therefore clear that in addition to the volume flow of the individual phases, or the velocities in the empty tube w = ν/(πϋ ² /4) derived from this, the dimensions of the reactor (length L, diameter D) and also in particular the data of the packing bodies used (hydraulic diameter d _H , empty tube proportion ψ) play an important role. By choosing these parameters correctly, the process can be easily adapted to a variety of requirements, it is only important to comply with the requirement B >= 0.8, preferably B >= 0.9 and particularly preferably Β >= 1.

For example, in the case of a slow reaction, the hydraulic diameter of the packings can be chosen so small, or their specific surface area so large, that the required conditions for B can be achieved even at low flow rates. In this way, sufficient residence times are achieved over the length of a technically reasonable reactor. For very fast reactions, the opposite procedure is recommended.

Another criterion in carrying out the process according to the invention is the ratio of the mass flow of the liquid phase M _x , containing the catalyst, to the mass flow of the dispersed phase or the dispersed phases M ₂ . In the case of hydroformylation, the mass flow of the catalyst phase Mi is significantly greater than the mass flow of the dispersed phases, i.e. the organic olefin phase M _2a and the synthesis gas phase M _2b . In the process according to the invention, the mass ratio Μχ/Μ ₂ of the continuous phase (Μχ) to the dispersed phases (M ₂ ) may be greater than 2, preferably Μχ/Μ ₂ > 10. Flow ratios with Mi/M ₂ > 100 are quite possible and often even advantageous. Under the condition Μχ/Μ ₂ > 2, the catalyst phase is a continuous phase, while the dispersion phases are divided into fine bubbles or fine droplets. In the process according to the invention, it is possible for at least one starting material to be dispersed by means of energy introduced into the reactor by the continuous phase. This leads to the division of at least one starting material into bubbles or droplets in the continuous catalyst phase.

This can also be estimated using conventional engineering methods. Suitable for this are formulations with dimensionless coefficients, such as ds/du - k * Re _g i(gii) ^m * We _g i( _g ii) ⁿ s

d _s diameter of the drop or bubble according to Sauter (in Brauer et al.), d _H hydraulic diameter of the filling bodies,

Re _g i(gii) ^m Reynolds number of multiphase flow = = Wgugii,* (ρι/ηι) * (dn/ψ) ,

We _g i ₍ gii) Weber number of multiphase flow = = WgKgll) ² * (Pi/Ogx) * (dn/ψ) ² , k, m, n empirical constants (known or determinable by experiments, velocity in empty tube [m/s] = V/(kD ² /4),

• · • · • 9

V volume flow at operating conditions [m ³ /s] , p density at operating conditions [kg/m ³ ], η viscosity at operating conditions [Pa*s] and γ interfacial tension at operating conditions [N/m] and with the symbols 1 (liquid phase), g (gas phase), gl (two-phase gas-liquid flow) and gll (three-phase gas-liquid-liquid flow).

With structured packings, such as Sulzer-SMV, or narrow tubes as assembled elements, it seems likely that a calculated bubble or droplet diameter d _s greater than the channel diameter is not meaningful. However, this does not apply to permeable packings and packing bodies, such as wire mesh coils or wire mesh nets (so-called demister packings or condensate drains). In the method according to the invention, calculated droplet diameters can be used that are at least equal to or smaller than the hydraulic channel diameter:

d _s /d _H <= 1, preferably < 0.9.

From the calculated droplet diameter, the transition area of the substance can finally be calculated according to

A _s = 6 (p _g d ₃ [m ² /m ³ ] .

For the phase fraction (p _g of the dispersed phase (in the case of hydroformylation, the synthesis gas and the organic phase are dispersed) it can be substituted with the phase velocities in the empty tube <Pg ~ Wg/Wgl • · • · · · 9 9

0444 4 ·

9 9 9 9 9 9

4 49 4 9

The residence time τ of the phases flowing through the reactor can be approximately calculated from τ ~L*\[//w _lg . The residence time τ is usually much shorter than one hour in the process according to the invention and can be in the range of minutes or even shorter. However, in this completely unusual process - high catalyst throughput in the reactor, comparatively very low proportion of starting material in the reaction mixture, and thus very short residence time - surprisingly high space-time yields are achieved in many multiphase reactions. Alternatively, it is possible to operate at significantly lower temperatures than usual for the same space-time yields, because this economically allows an increase in the reaction rate, which can, for example, lead to a minimization of subsequent reactions and thus improved selectivity.

The method according to the invention can be adapted very flexibly to the most diverse requirements. The following embodiments of the method according to the invention are offered for special requirements.

If the operating purpose requires a very long mixing area or if stagnant areas are desired, for example for the loss of substance flows, then a cascade arrangement of tubular reactors with mounted elements and packing bodies is offered.

A cascade arrangement of tubular reactors or an alternative arrangement of empty tube sections or sections with filling may be recommended when a particularly low pressure loss is required.

Furthermore, parallel arrangement of tube reactors or the use of a multi-tube reactor is possible, whereby the tube can take over the function of the mounted elements. In addition, the reactors can be provided with multiple gas feeds over the length • · • φ • · · φ φφφφ * φ · *

φ φ φ φ φ · φ φφφ φ φ φ φ β· · reactor, when the gas consumption is so high that unfavorable gas-to-liquid phase ratios occur upstream of the reactor when both phases are fed together.

The special conditions of the process according to the invention allow for further embodiments of the process. Thus, the required high circulation of the catalyst phase or the continuous phase can be used in addition to the operation of the injection nozzle, which is arranged upstream of the tubular reactor itself as a compressor of the gas and liquid stream. This can be used for thorough premixing of the two phases and also for compression of the gas phase, which allows the process at higher pre-pressures in the reactor. Finally, if suction is used instead of gas compression, it is even possible to conduct the gas in a cycle with simultaneous premixing of the phases. The energy supplied to the tubular reactor by the continuous phase containing the catalyst can therefore be used for dispersing the starting material phase or at least one starting material.

In the process according to the invention, heat removal is also not critical in strongly exothermic reactions, such as hydroformylation. The high throughput of the catalyst cycle acts as a heat carrier, so that even in the adiabatic operation of the reactor, only small temperature differences occur, resulting in a homogeneous temperature distribution in the reactor without temperature peaks. The heat generated can then be conveniently removed using a conventional heat exchanger arranged arbitrarily in the outer catalyst cycle or used for energy generation. For better heat removal, it may be advantageous, depending on the circumstances, to run the catalyst cycle even higher (i.e. at a higher value of B) than is required according to the experimental results, since the catalyst cycle can set a lower temperature gradient in the reactor.

9 ·· ·» ···· • · · · · · · · « · · 9 9 · 9 9 9 9 • 9 9 9 9 9 9 9 V 9 *

The method according to the invention provides significant advantages compared to the prior art; the following can be mentioned:

• High space-time yields can be achieved at comparatively low temperatures.

• The formation of by-products is extremely low, values of 1 to 2 wt% and lower are possible.

• The catalyst is saved, deactivation is very small, and continuous removal is eliminated.

In the case of the production of aldehydes by hydroformylation of olefins by the method according to the invention, additional advantages are added:

• Due to the higher reaction rate, this process can also be economically used for the hydroformylation of higher olefins with more than 10 carbon atoms.

• For gaseous olefins, the proportion of starting materials remaining after partial conversion can be recycled by simple recirculation using an injection nozzle.

In the following text, a special embodiment of the present invention, a process for the hydroformylation of olefins with synthesis gas in a multiphase system, is to be described in more detail. The catalyst phase is used as the continuous phase and is pumped through a tubular reactor by means of a pump.

All solvents that meet the following conditions are suitable as solvents for the production of the catalyst solution or phase:

• · • · • « « · · · · ♦ ♦ ·

• The solvent is poorly soluble in the product phase, • The product is only slightly soluble in the catalyst phase, which consists of the catalyst and the solvent.

• The solvent has a sufficiently high solubility for the catalyst used.

The solvent may contain phase transfer agents, surface-active or amphiphilic agents or tensides as additives. Water is used as the preferred solvent.

Metals of the VIII. subgroup (Fe, Co, Ni, Ru, Rh, Pd, Os, Ir, Pt) can be used as hydroformylation catalysts. These metals or compounds of these metals should be soluble in the catalyst phase under the reaction conditions, but not in the product phase. If catalyst solutions are used, this requires water-soluble metal compounds. Rhodium or water-soluble rhodium compounds are used as the preferred catalyst. Suitable rhodium salts are, for example, rhodium sulfate, rhodium nitrate, rhodium carboxylates, such as rhodium acetate, rhodium propionate, rhodium butyrate or rhodium 2-ethylhexanoate.

The type of ligands depends on the metal used and the solvent of the catalyst solution, for these complexes it is true that they are catalytically effective and their catalytic effect is not lost in continuous operation. The condition for this is that the ligands do not change, for example by reaction with the solvent.

Triarylphosphines can be used as ligands for catalytically active metals. Suitable phosphines have one or two phosphorus atoms, which have three aryl radicals per phosphorus atom, the aryl radicals being the same or different and representing phenyl, naphthyl, biphenyl, phenyl21 • · 9 · •··· 9 9

9 9 9 9 9 naphthyl or binaphyl radical, in particular phenyl, biphenyl or binaphthyl radical. The aryl radicals can be linked to the phosphorus atom directly or via the group -(CH ₂ ) _X -, where x is an integer from 1 to 4, preferably from 1 to 2, particularly preferably 1. For water-soluble catalyst systems, the ligand should preferably contain three radicals -(SCDM, where M is the same or different and represents H, an alkali metal ion, an ammonium ion, a quaternary ammonium ion, an alkaline earth metal ion (theoretically half) or a zinc ion.

The -SO ₃ M radicals are usually substituents on aryl radicals and provide the triarylphosphines with the necessary water solubility. A preferred sulfonated triarylphosphine with one phosphorus atom is trisodium tri-(m-sulfophenyl)phosphine.

In addition to sulfonate units (-SO3M), other polar groups, such as carboxylate units, can also be used.

The aqueous phase can be used directly in the hydroformylation or the catalyst can first be preformed under reaction conditions and then used in the preformed state. However, the aqueous catalyst solution can also be prepared in a relatively simple manner by dissolving a water-soluble metal salt and water-soluble ligands in water and causing them to form complexes.

The concentration of the metal salt required in the process according to the invention can be adjusted within a wide range. The highest value is predetermined by the solubility. The reaction rate also depends on the concentration of the metal salt. As a rule, higher reaction rates are achieved at higher concentrations of the metal salt. On the other hand, higher concentrations of the metal salt mean higher • · · ·

9 • · · · « • · · * « 9 99 9 9

9 9 9 9 9

9.9 9 9

9 99 99 costs. Therefore, the optimum can be selected according to the reactivity of the starting material and other reaction conditions. The rhodium content in the catalyst phase is usually 20 ppm to 2,000 ppm, preferably 100 to 1,000 ppm.

In the catalyst system used, the molar ratio between metal and ligands can be varied to achieve the optimum for each individual conversion. This ratio is between 1:5 and 1:200, in particular between 1:10 and 1:60.

The pH of the catalyst solution can be optimized for the hydroformylation of each olefin with respect to the selectivity of aldehyde formation. It is between 2 and 8, preferably between 3.0 and

5.5.

Olefinic compounds having 2 to 25 carbon atoms, preferably 2 to 12 carbon atoms, can be used as starting materials for hydroformylation. Olefinic compounds may contain one or more carbon-carbon double bonds, which may be at the end or within the chain. Olefinic compounds with a carbon-carbon double bond at the end of the chain are preferred.

An olefin of uniform structure can be used. Mixtures of olefins can also be used. The mixture can consist of isomeric olefins with the same number of carbon atoms or of olefins with different numbers of carbon atoms or of a mixture which contains isomeric olefins as well as olefins with different numbers of carbon atoms. Furthermore, the olefins or olefin mixtures can contain substances which are inert under the reaction conditions, such as aliphatic hydrocarbons.

Olefins from a variety of sources can be used in the process of the invention. For example, olefins from cracking, dehydrogenation or Fischer-Tropsch processes can be used. φ · φ φ • Φ Φ 9

Φ ΦΦΦ» Φ • Φ Φ ΦΦΦ

ΦΦ · Φ Φ Φ Φ Φ Φ synthesis. Suitable starting materials are also olefins or mixtures of olefins which are formed by dimerization, oligomerization, codimerization, cooligomerization or metathesis of olefins. The olefins used can be (under normal conditions) gaseous, liquid or solid. Solid olefins are used as solutions. Inert liquids which are poorly soluble in the catalyst phase are used as solvents. Solvents which have a higher boiling point than the products produced are particularly preferred, since this facilitates distillation separation and recycling.

Preferably, α-olefinic compounds are used in the process according to the invention. Examples of such α-olefinic compounds are 1-alkenes, alkyl alkenoates, alkylene alkanoates, alkenyl alkyl ethers and alkenols, such as propene, butene, pentene, butadiene, pentadiene, 1-hexene, 1-heptene, 1-octene,

1- nonene, 1-decene, 1-undecene, 1-dodecene, 1-hexadecene,

2-ethyl-1-hexene, 1,4-hexadiene, 1,7-octadiene, 3-cyclohexyl-1-butene, styrene, 4-vinylcyclohexene, allyl acetate, vinyl formate, vinyl acetate, vinyl propionate, allyl methyl ether, vinyl methyl ether, vinyl ethyl ether, allyl alcohol, 3-phenyl-1-propene, hex-1-en-4-ol, oct-1-en-4-ol, 3-butenyl acetate, allyl propionate, allyl butyrate, n-propyl-7-octenoate, 7-octenoic acid, 5-hexenamide, 1-methoxy-2,7-octadiene, 3-methoxy-1,7-octadiene, in particular propene, 1-butene, technically available mixtures which essentially contain 1-butene, 2-butene and isobutene, and/or 1-pentene.

The products of hydroformylation of olefins are aldehydes one carbon atom longer and optionally alcohols. The aldehydes produced by the process according to the invention can be used for the production of alcohols by hydrogenation. The alcohols thus obtained are in turn precursors for plasticizers, for example as phthalic acid diesters or detergents.

Furthermore, the aldehydes according to the process of the invention are used in aldol condensations and in the production of carboxylic acids by oxidation.

Mixtures of hydrogen and carbon monoxide (synthesis gas) or other mixtures of hydrogen, carbon monoxide and substances inert under the reaction conditions are used as hydroformylation agents.

When using liquid olefins or solid olefins in solution, it is advantageous to use the hydroformylating agent in excess in order to achieve as complete a conversion as possible. This reduces the processing costs. When using gaseous olefins, on the other hand, it may be advantageous to use the hydroformylating agent in deficit, since the excess gaseous olefin is separated from the liquid product and can be fed back into the process.

The molar ratio of olefin to hydrogen and olefin to carbon monoxide may be greater than or equal to 1.

The process according to the invention, in the case of hydroformylation using a gaseous olefin, is initially a two-phase reaction, whereby a liquid product phase is formed during the reaction, thereby creating a three-phase system. When using a liquid olefin, a three-phase system exists from the start.

The process according to the invention can be carried out in one or more tubular reactors with mounted elements corresponding to the above description.

In the process according to the invention, the catalyst phase is a continuous phase; therefore, a weight ratio between the catalyst phase and the dispersed phase or phases is advantageously in the range of 2:1 to 3500:1, preferably in the range of 40:1 to 2500:1.

When the process for hydroformylating olefins according to the invention is used, the weight ratio between the catalyst phase and the olefin phase at the reactor inlet may be in the range of 5,000:1 to 4:1, preferably in the range of 2,000:1 to 50:1. The weight ratio between the catalyst phase and the hydroformylating agent (usually synthesis gas) is 4:1 to 10,000:1, preferably 200:1 to 4,000:1.

The reactants can be fed preheated, i.e. in the reaction temperature range, or they can be fed cold. Due to the high phase ratio with the catalyst phase, preheating can also be carried out using process heat.

If the process for hydroformylating olefins according to the invention is used, the conversion takes place in a temperature range of 20°C to 250°C, preferably in the range of 90°C to 150°C, with the total pressure being in the range of between 1 MPa and 30 MPa, preferably between 2 MPa and 15 MPa.

The reaction tube can be filled in one direction from top to bottom or vice versa. For safety reasons, filling from the top is preferred.

The reaction heat can be removed by various heat exchangers. The heat exchangers do not have to be located near the reaction space, but can also be located anywhere outside the reactor. The individual heat flows depend on the specific reaction heat and also on the required temperatures in the reactor and in the processing equipment.

Φ · «φ · * • · · · · • · φ · · φ ♦ • Φ · · · · 9 • Φ Φ 9 Φ Φ

Φ Φ Φ Φ · Φ

Φ Φ

Φ · · Φ

Φ Φ ·

ΦΦΦ Φ • Φ · « · Φ Φ Φ

The removed heat can therefore be used very simply, for example in the process itself, to heat the distillation equipment or to produce steam.

The mixture leaving the reactor can be degassed in the case of using gaseous olefins or in the case of complete conversion in a gas-liquid separation tank. The separation of gas from liquid can be carried out at the same pressure as at the reactor outlet. This is particularly advantageous if at least part of the gas after depressurization is recirculated to the reactor. Otherwise, depressurization can also be carried out at lower pressures (up to 0.1 MPa).

The separated gas stream can be completely or partially fed back into the reactor.

This recirculation can be achieved in a known manner, for example by means of an injection or mixing nozzle which is placed in the flow of the catalyst cycle upstream of the reactor, or by means of a rotary compressor. For energy reasons, an injection or mixing nozzle which is placed in the flow of the catalyst cycle upstream of the reactor is preferably used.

The remaining or possibly the total amount of gas can be cooled or fed uncooled to the waste gas recovery system. When using a cooler, the gas condensate formed in the cooler can be piped to a tank for separating gas from liquid.

The degassed liquid mixture is mechanically separated in a liquid-liquid separation tank into a catalyst phase and a product phase. This can occur in settling tanks ♦ · • · φ φ • φ φ • φφφ»

Φ 8 8 8 8

ΦΦ · Φ

Φ Φ Φ « ΦΦΦ

ΦΦΦ 8 of various types of construction or in centrifuges. For price reasons, settling tanks are preferred.

The residence times in the separation devices are not fundamentally critical, but they are preferably kept short. This has the following advantages: The separation device is small and its price is correspondingly low. With short residence times, practically no side reactions occur in the separation tank. This means that the phase separation occurs quickly, but the differences in the densities of the two phases must be sufficiently large and their viscosities small. All four quantities are a function of temperature and can be easily determined by exploratory experiments.

In addition, the density and viscosity of the catalyst solution can be varied by the choice of solvent and catalyst concentration. Alternatively, the density and viscosity of the product phase can be varied by the addition of solvent.

The phase separation can be carried out in a wide temperature range. The separation temperature can also be higher than the temperature of the reaction product at the reactor outlet. However, for energy reasons, it is not advantageous to use a temperature higher than the temperature of the liquid in the gas separator. The freezing point of one of the two liquid phases can be considered the lowest possible temperature. However, due to the short separation times, as mentioned above, temperatures that are too low are not chosen.

The product stream is separated by known methods, for example by distillation.

The separated catalyst solution is returned to the reactor, optionally after removal of a small portion and corresponding replacement with fresh catalyst solution.

·· · · 444« • 44 4 4 · 4 · · φ · 444 4 4 4 4 ·

4 4 444 · 44« 4 • 4 44

The following examples are intended to illustrate the invention without limiting the scope of its application, which is apparent from the claims.

Examples of embodiments of the invention

Hydroformylation of olefins

The hydroformylation of olefins was carried out in an experimental apparatus which is schematically shown in Fig. 1. In this, an aqueous catalyst solution is pumped in a cycle by a pump _1. Olefin 3 and synthesis gas _4 are mixed with the catalyst phase _2. The multiphase mixture 5 is pumped through a tubular reactor 6 which is provided with static mixing elements. The resulting mixture T_, consisting of product, unreacted starting material and catalyst, is degassed in a gas separator 8. The gas 9, consisting of synthesis gas, or olefin gas and enriched inert substances, is mostly reintroduced into the reactor 6 via a gas recirculation line 10 using a mixing nozzle or an injection nozzle 11. A small part of the gas flow 9 is discharged via a line 12. By suitable cooling 13 and reflux, olefin losses can be minimized. The discharge 14 is reduced to enriched inerts and small amounts of unreacted synthesis gas. The liquid stream 15 resulting from degassing _8 is led to a phase separation tank 16. Here, the aqueous catalyst phase 2 is separated and fed back into the cycle. The reaction heat can be removed by heat exchangers 17, 19, 20 located outside the reactor. The product is withdrawn via line 18 and optionally further purified.

Examples 1 to 7 describe the hydroformylation of propene by the process of the invention.

Examples 1 to 5 • · ♦ · • · · « · · • · · * ♦ « · · · · · · · • 4 4 9 4 9 9

4 49

A reactor 6 with a length of 3 m and a diameter of 17.3 mm was used, which contained static mixing elements from Sulzer with a hydraulic diameter of 2 mm.

For this example, the gas recirculation line 10 was closed. Water was used as the catalyst solvent. The pH value was 7. The reactor was run at a catalyst load of 400 kg/h at a temperature of 120°C. The reaction pressure was 5 MPa. Rhodium was used as the catalyst with a concentration of 800 ppm relative to the solvent phase. Tri(m-sulfophenyl)phosphine (TSTPP) in the form of its sodium salt was used as the ligand, the P/Rh ratio was 60. For experiment 2, the reaction conditions were set as in experiment 1, with the difference that the reaction temperature was 130°C. For experiment 3, the reaction conditions were set as in experiment 1, with the difference that the reaction pressure was 7 MPa. For experiment 4, the reaction conditions were set as in experiment 1, with the difference that the pH value of the catalyst solution was set to 4. For experiment 5, the reaction conditions were set as in experiment 1, with the difference that the catalyst loading of the reactor was 300 kg/h. The feed flows of the starting materials and also the products are given in the table in mol/h.

όο ♦ tt tttt tttt «tttttt ·· • tttt tt* tt tttttt tttttt·· tttt tt tttt « tttt tttttt · tttttt · tttt tttt tttt tttt tttt tttttt

Example 1 2 3 4 5 RZA [t/(m ^J .h)J 0.98 1.39 1.23 1.00 0.71 B 13.16 13.34 13.16 13.16 8.06 Starting substance WHAT 22.79 16.99 28.24 23.92 12.73 h ₂ 21.2S 15.87 26.38 22.87 11.89 _n2 0.12 0.09 0.15 0.10 0.07 Propene 51.49 35.57 31.13 36.01 33.45 Propane 0.17 0.12 0.10 0.11 0.11 Product WHAT 2.15 2.09 2,OS 0.66 2,H h ₂ . 0.S9 0.93 0.9S 0.37 0.97 Propene 3.35 6.73 3.04 3.45 4.00 i-Butanal 0.24 0.51 0.53 0.39 0.20 n-Butanal 5.46 11.03 9,S2 6.19 4.66 i-Butanol <0.01 <0.01 <0.01 <0.01 <0.01 n-Butanol 0.07 0.07 0.07 0.06 0.07 2-Ethylhexanal 0.09 0.03 0.08 <0.01 0.02 2-Ethylhexenal 0.26 0.19 0.1S <0.01 0.09 Waste gas WHAT 9.50 3.06 13.23 13.72 6.81 h ₂ S,45 1.46 9.97 10.82 3.59 _n2 0.29 0.10 0.09 0.10 0.16 Propene 37.37 13.56 17.72 19.47 36.93 Propane 0.17 0.10 0.17 0.10 0.09 i-Butanal 0.24 0.16 0.15 0.18 0.14 n-Butanal 4.59 3.13 2.72 2.62 2.56

4* 44 ·44· 4· • · · · · · ♦ · •444 44 4 44

444 4 444 4 «· 44 44 44 4 4 4

Examples 6 and 7:

These examples illustrate the use of the gas recirculation line 10. This minimizes the loss of valuable substances through the waste gas line. Conversions of >90% can already be easily achieved in these examples. By means of the cooler 13, it is possible to almost completely prevent the removal of propene in the waste gas, so that propene now only leaves with the liquid product. The product 18 removed from the phase separation tank, which still contains propene, is processed according to the prior art by, for example, separating it with synthesis gas or by distillation. Propene conversions of more than 99% are thus possible without loss of space-time yield.

In examples 6 and 7, the reactor was run at a catalyst load of 400 kg/h and a temperature of 120 °C. The reaction pressure was 5 MPa. The rhodium concentration was 800 ppm relative to the solvent phase. TSTPP in the form of the sodium salt was used as the ligand, the P/Rh ratio was 60. The feed flows of the starting materials and also the products are given in the table in mol/h.

99 • 9 9 999 ·· ··<··

Example 6 7 RZA[t/(nÁh)] 1.06 1.04 B 12.12 12.12 Propene conversion 93.70% 92.33% Syngas conversion 92.09% 90.27% Entry Starting substance. WHAT 10.41 10.23 h ₂ 9.56 9.63 _n2 0.07 0.07 Propene 12.42 12.43 Propane 0.07 0.07 Product WHAT 1.27 1.65 h ₂ 0.24 0.39 Propene 0.5S 0.73 i-Butanal 0.86 0.77 n-Butanal 8.04 8.24 i-Butanol <0.001 <0.001 n-Butanol 0.13 0.11 2-Ethylhexanal 0.15 0.12 2-Ethylhexenal 0.29 0.18 Waste gas WHAT 0.1S 0.09 h ₂ 0.09 0.05 _n2 0.01 0.01 Propene 0.10 0.09 Propane 0.003 0.002

4 94 4 4 • 4 4 • 9»99 • · 9 9 • 4 · 9

9 9 9 ► 9 « » 9 4 1

4 4

Hydroformylation of 1-hexene

Example 8

This example describes the use of the process according to the invention for the continuous hydroformylation of 1-hexene. The conversion rate in a longer reactor was determined by continuously reusing the crude product obtained. This gave values similar to those obtained from a reactor several meters long, in which gas was re-dosed every 1 m. Water was used as the solvent for the rhodium catalyst. A reactor with a volume of 235 ml (1 m long) was run through with a catalyst load of 400 kg/h at a temperature of 130 °C. The reaction pressure was 3 MPa. The rhodium concentration was 800 ppm relative to the solvent phase. TSTPP in the form of the sodium salt was used as the ligand, the P/Rh ratio was 60. The synthesis gas load was on average 132.8 1/h (under normal conditions). The inlet flow of the liquid starting material was set at 1.3 1/h. After 8 passes (= 8m reactor length) almost 90% conversion was achieved, but the RZA was still 0.44 t/(m ³ .h), calculated over the entire reactor length. This value is much higher than the technically known data. The overall selectivity after 8 passes was almost 93%.

Number of transitions 1 2 4 5 6 7 8 Total Conversion [%] 19.1 36.9 49.8 59.8 68.5 76.5 S2.9 86.8 High-boiling components 0 0.37 0.28 2.1 2.65 4.29 4.59 4.62 Overall selectivity [%] 85.1 89.1 91.7 92.1 92.4 92,S 93.1 92.9 Total RZA[t/(m’*h)] 0.748 0.743 0.689 0.619 0.568 0.521 0.486 0.444 B 14.26 14.26 14.26 14.26 14.26 14.26 14.26 14.26

• · · · · · • · · · · · · · · • ft · ·· · · · 9 · • ftft ftftft ft ··· · • ftftft ft··· ftft ft· ftft ftft ftft ·* ft

Example 9

Hydroformylation of 1-decene

A reactor with a length of 6 m and a diameter of 17.3 mm was used, which contained static mixing elements from Sulzer with a hydraulic diameter of 2 mm.

For this example, the gas recirculation line 10 was closed. Water was used as the catalyst solvent. The pH value was 4.5. The reactor was run at a catalyst load of 400 kg/h at a temperature of 125 °C. The reaction pressure was 7 MPa. The rhodium concentration was 800 ppm relative to the solvent phase. TSTPP in the form of its sodium salt was used as the ligand, the P/Rh ratio was 5. The feed flows of the starting materials and also the products are given in the table in mol/h.

-Example 9 RZA [t/(m ³ -h)] 0.05 B 13.96 Starting substance WHAT 11.3S h ₂ 10.52 N; 0.05 1-Dec 7.12 Product 1-Dec 6.67 Undecanal 0.34 2-Methyldecanal 0.11 Waste gas WHAT 10.36 h ₂ 9.89 _n2 0.05

♦· » · ·

I Φ · ·· » φ φ » φ φ φφ φφφ» ·· * · φ · φ φ φφφ φ φ

ΦΦΦ φ · · Φ

Φ· ·· ···

T>V DLOO-doo/f

Claims (14)

PATENT CLAIMS CLAIMS 1. A process for the catalytic performance of multiphase reactions in a tubular reactor, characterized in that a catalyst is present in the continuous phase and at least one starting material is present in the dispersed phase and the tubular reactor load factor B is equal to or greater than 0.8. Process according to claim 1, characterized in that the olefins are hydroformylated by the multiphase reaction. 3. A process according to claim 2 wherein the olefins contain from 2 to 25 carbon atoms. Process according to one of Claims 1 to 3, characterized in that metal VIII is used as the catalyst. secondary groups. Process according to one of Claims 1 to 4, characterized in that rhodium is used as the catalyst. Process according to one of Claims 1 to 5, characterized in that water-soluble rhodium compounds are used as catalyst. Process according to one of Claims 1 to 6, characterized in that water or a mixture of water and one organic solvent is used as the continuous phase. Method according to one of Claims 1 to 7, characterized in: 44 4444 • characterized in that the load factor B is greater than or equal to 0.9. Method according to one of claims 1 to 7, characterized in that the load factor B is greater than or equal to 1.0. Method according to one of Claims 1 to 9, characterized in that the ratio of the weights of the continuous phase to the dispersed phase or the dispersed phase is greater than 2. Method according to one of claims 1 to 10, characterized in that the continuous phase in front of the tubular reactor drives the injection nozzle. Process according to one of Claims 1 to 11, characterized in that the at least one starting material is dispersed by means of the energy supplied by the continuous phase to the tubular reactor. Use of aldehydes produced according to claims 1 to 12 for the production of alcohols. Use of aldehydes produced according to claims 1 to 12 in aldol condensations. Use of aldehydes produced according to claims 1 to 12 for the production of carboxylic acids.