HYDROFORMYLATION PROCESS

FIELD or THE INVENTION

The present invention refers to the use of an epoxide in order to reduce the formation of heavy ends in a continuous hydroformylation process, where an olefin or an olefin mixture is reacted with carbon monoxide and hydrogen in the presence of a rhodium complex catalyst, comprising at least one organobisphosphite ligand, in order to produce an aldehyde. Said epoxide is added to the reaction mixture, reducing the formation of heavy ends by 10-80 %.

BACKGROUND ART

Hydroformylation, also called the oxo process, is an important industrial process that generates aldehydes by reacting olefins with carbon monoxide and hydrogen in the presence of transition metal catalyst complexes. It is well known that preferred processes involve continuous hydroformylation and recycling of a catalyst solution containing a metal-organophosphoras ligand complex catalyst wherein the metal is a Group VIII transition metal, preferably rhodium. Such art is exemplified by for instance US 4,148,830, US 4,717,775, and US 4,769,498, Yielded aldehydes have a wide range of utility, for example, as intermediates for hydrogenation to aliphatic alcohols, so called oxo alcohols, for animation to aliphatic amines, for oxidation to aliphatic acids, and for aldol condensation to produce components of plastic izers.

Rhodium catalysts need ligands for stabilizing and/or activating the rhodium in rhodium- catalyzed low-pressure hydroformylation. Rhodium complexed with phosphorus ligands, wherein the phosphorus ligands typically are organophosphines and/or organophosphites, are well known in the art, Rhodium-bisphosphite catalysts are known to be used in the

hydroformylation of linear olefins, like propenes, butenes and hexenes and they have an effect of increased activity and selectivity to unbranched reaction products

Hydroformylation is often accompanied by a series of parallel and secondary reactions. Due to their high reactivity the formed aldehydes undergo aldol condensation reactions, Tischenko reactions and acetalization reactions and produce high-boiling by-products, such as dimers, timers and letramers. These high-boiling by-products produced during fee hydroformylation process are called heavy ends and they can cause a lot of problems to the process, since their accumulation can lead to a forced bleeding of rhodium and ligand containing solution. Heavy ends formation can take place in reactors, as well as in hot places like distillation bottoms.

The condensation reaction producing heavy ends is generating water and when heavy ends are formed to a large extent, there is consequently a large amount of water produced in the hydroformylation process. Water formed in the aldol condensation reaction in the reaction mixture causes the organobisphosphite ligand to undergo hydrolysis. When the phosphite ligand undergo hydrolysis, acidic decomposition products are formed, e.g. hydroxyalkylphosphonie acids, font catalyzes many of the reactions forming heavy ends, which in turn, produces additional water, causing further hydrolysis of the ligand. This autocatalytic reaction leads to a substantial amount of high boiling compounds in the process and a significant loss of phosphite ligand resulting in unnecessary high ligand costs,

A far more challenging problem than accelerated ligand decomposition is the accumulation of high-boiling compounds in the process. The concentration of heavy ends in the reaction mixture increases with time and heavy ends have to actively be removed from the process. Accumulation of heavy ends is troublesome since they take up space in the process and can affect both the yield and the selectivity of the hydroformylation products. There has to be a balance between the rate of removal and the rate of formation of heavy ends, If the heavy ends increase too much the useful capacity and effectiveness of the hydroformylation system is reduced and a catalyst change is required. Replacing the catalyst is a costly procedure, not only because of the very expensive catalyst and free and complexed ligand, but also because the reaction has to be stopped. The accumulation of heavy ends may also cause clogging in the reactor system, if not inhibited in time, causing a lot of process problems and reaction stops.

Accordingly, it is important to limit the amount of heavy ends accumulating in the reaction mixture and to prevent the formation of heavy ends during the hydroformylation reaction. The amount of heavy ends can to a certain extent be reduced via optimized reaction parameters, catalyst: concentration and the use of solvents. But heavy ends still have to actively be removed from the reaction mixture in order to avoid accumulation in the reactor; It is not possible to separate the more high-boiling heavy ends from the reaction solution by distillation, since the phosphite ligand thermally decomposes at high temperatures, It is generally required to bleed the heavy ends from the system. The bleed does unfortunately not only contain heavy ends, but also e.g., smaller amount of the aldehyde product, the rhodium-phosphite complex catalyst, free phosphite ligand, organic solvent, unreacted olefin and dissolved carbon monoxide and hydrogen, Since the phosphite ligand decompose at high temperatures, it is impossible to separate the rhodium-phosphite: complex catalyst from the heavy ends and recycle back to the hydroformylation reaction without losing the valuable ligand.

Rhodium is a highly expensive catalyst that is recycled in the hydroformylation reaction. The loss of rhodium due to heavy ends removal is very costly. Heavy ends accumulation is therefore a serious process problem, often leading to a forced rhodium bleed. Reducing the formation of heavy ends, is a way to reduce the rhodium loss, because of the subsequently reduced amount of heavy ends removed from foe process.

There is also another problem with rhodium catalysts comprising phosphite ligands, they lose catalytic activity over time during continuous use ip a hydroformylation reaction. Some decomposition products of the phosphite ligands can coordinate with rhodium and form complexes that ate less reactive than rhodium-phosphite complex catalysts; These phosphite ligand decomposition products act as catalyst poisons or inhibitors, and lower the catalyst activity of rhodium-phosphite complex catalysts. If the ligand degradation caused by water from heavy ends formation were reduced, the hydroformylation process could be run for an extended period of time before tike catalyst solution had to be exchanged.

Due to foe tendency of phosphite ligands to undergo hydrolysis, alcoholysis and/or

transesterification, under customary hydrofonnylation conditions, there is a lot of prior art relating to methods of stabilization of phosphite ligands in hydroformylation. There is however not many publications focusing on the amount of heavy ends and how to effectively reduce the amount of heavy ends formed in a hydroformylation reaction. There are several publications on methods for stabilizing catalyst and/or organophosphite ligands using different kinds of amines. For instance EP3126319, disclosing a method where a water soluble amine is used instead of aqueous buffer solution to neutralize any phosphoric acidic compound. US4567306 discloses a method to decrease degradation of cyclic phosphite ligands by adding a tertiary amine. Tertiary amines reduce ligand destruction by neutralizing the acidic materials and forming ammonium salts. However, many amines also catalyze the undesirable condensation of the aldehyde products, thus leading to increased formation of undesirable by- products, i.e. heavy ends.

WO2019083700 discloses a process to reduce heavy end formation in a hydroformylation reaction by the addition of a certain class o f nitrogen compounds. Many nitrogen compounds are however associated with environmental and health concerns. WO2010003073 discloses a two- stage process of hydroformylation and product-catalyst separation for controlling heavy ends in a catalyst recycle stream. This method does only partly controls the amount of heavy ends and is associated with addition equipment, process redesign and a risk of ligand loss.

EP455261 discloses a process for producing a 1,3-diol and/or a 3 -hydroxyaldehyde by hydroformylation of an epoxide using a rhodium-containing catalyst. In this European patent application the epoxide is used a reactant, not as an additive, and there is no disclosure of decreasing the amount of heavy ends by using art epoxide, EP0590611, is focusing on stabilizing phosphite ligands against degradation by the addition of an epoxide. The patent is however silent On heavy ends complications and what effect an epoxide may have on the amount of heavy ends in a continuous hydroformylation process.

Accordingly, there is a need for an efficient way to reduce the amount of heavy ends formed in a continuous hydroformylation process and consequently increase the stability of the

organophosphite ligand by decreased hydrolysis, without the need of additional equipment-

DESCRIPTION OF THE INVENTION

It has surprisingly been found that by adding small amounts of an epoxide to a hydroformylation mixture, the amount of heavy ends produced during the hydroformylation process can be drastically reduced. This efficient way of limiting the formation of heavy ends and thus preventing heavy ends from accumulating in the reaction mixture, has also a stabilizing effect on the organobisphosphite ligand of a rhodium complex catalyst in a hydroformylation reaction.

The present invention refers to the use of an epoxide in order to reduce the formation of heavy ends in a continuous hydroformylation process, where an olefin or an olefin mixture is reacted with carbon monoxide and hydrogen in the presence of a rhodium complex catalyst , comprising at least one organobisphosphite ligand, in order to produce an aldehyde. Said epoxide is added to the reaction mixture in an amount of 0.01-1.5 wt%, reducing the formation of heavy ends by 10- 80 %. ln preferred embodiments of the present invention the formation of heavy ends is reduced by at least 50 %, That is a huge benefit for the economy of the hydroformylation process. The present invention is a great way of using an epoxide, as an heavy end inhibitor. Another benefit of the use according to the present invention is that the addition of epoxide increases the purity of the aldehyde product, without any other adjustments to the process.

The amount of epoxide used in accordance with the present invention is an amount sufficient to interact with the acidic ligand decomposition products which cause formation of heavy ends upon reaction with the aldehyde product. Preferably, the epoxide is added in an amount of 0.1-1 wt%, and most preferably 0.2-0.5 wt% of the reaction mixture . The epoxide is suitably a cycloaliphatic epoxide and preferably selected from cyclohexene oxide, 1,2-cyclo-dodecene oxide, 1,2-cyclododecene oxide and 3,4-epoxycyclohexylmethyl-3,4-epoxycyclo- hexanecarboxylate. Most preferably the epoxide is 3,4-epoxyeyclohexyImethyl-3,4-epoxycyclo- hexanecarboxylate.

The olefui in the hydroformylation reaction of the present invention is preferably a C2-C6 olefin, such as ethene, propene and/or butene. Most preferably the olefin is 1-butene and/or cis- or trans-2-butene, or a mixture thereof. The butene feed in the hydroformylation reaction of the present invention comprises a maximum of 2 wt % isobutene. Hydroformylation of mixed butene feeds containing 1-butene and cis- and trans- 2-butene, will generate a mixture of n-valeraldehyde and 2-methyl butyraidehyde. The n/iso ratio is affected by the choice of ligand. The ligand used m combination with rhodium in the present ihvention is an organobisphosphite ligand of Formula (I)

wherein -R ¹ , -R ² , -R ³ , -R ⁴ , R ³ , R ⁶ , R ⁷ and R ⁸ individually is hydrogen or a linear or branched alkyl group and -Ar is a substituted of unsubstituted aryl group, such as a group of Formula (II) or (III),

In preferred embodiments of tike present invention, -Ar is a group ofFormula.(II) and -R ¹ , -R ³ , - R ⁶ and -R ⁸ are n-bufyl, iso-butyl or tert-butyl. In a further preferred embodiment, -R ⁴ and -R ⁵ are methyl. Most preferably the ligand has the structure of formula (IV), having CAS no. 198979- 98-5, hereinafter designated A4N3-

Said Formula (I) ligand is suitably used in an amount of 0,5-15 wt%, such as 1-10 wt% or 1 -5 wt% calculated on total reaction mixture.

Due to the reduced amount of heavy ends formed in the hydroformylation process when using an epoxide according to the present invention, a reduced amount of water is present in the reaction mixture and the organobisphosphite ligand is hydrolyzed to a less extent. This allows the time between partial or complete catalyst changes to be dramatically increased.

The continuous hydroformylation process ofthe present invention is preferably conducted at a temperature in the range of 80° C to 130° C , and at a total pressure of not more than about 20 bar, A continuous or periodic addition of epoxide to the reaction mixture will maintain: a low level of epoxide in the reactor that will keep the heavy ends concentration at a manageable, steady level without autocatalytic degradation of the organobisphosphite ligand. An accurate amount of epoxide addition can be achieved by statistical process control. A device measuring the amount of unreacted epoxide, the amount of formed heavy ends, and/or the amount of acidic ligand decomposition products in the reaction mixture, can be used to calculate the necessary amount of epoxide addition in order to keep a concentration of about 0.5 wt%, preferably 0.1 wt%, epoxide in the reaction mixture. In circumstances where the organobisphosphite ligand of Formula (I) risk exposure to oxygen, small amounts of one or more less efficient and/or cheaper tertiary phosphines as anti-oxidants can be added in order to avoid or substantially reduce oxidation of the ligand. The anti-oxidant is preferably added in an amount of 0:01-5, such as 0.05-2 wt% of the reaction mixture. Said anti- oxidant is preferably a tertiary phosphine, such as triarylphosphine, dicycloalkylarylphosphine and/or cycloalkyldiarylphosphine. Especially preferred anti-oxidants include, phenyldi(o-, m- or p-tolyl)phosphine, diphenyl(o, m- or p-tolyl)phosphine, tri(o-, m~ or p-tolyl)phosphine, phenyldibenyzylphosphine, diphenylbenzylphosphine, tribenzylphosphine, phenyldinaphthyl- phosphine, diphenylnaphthyiphosphine, trinaphthylphosphine, dicyclohexyl-benzylphosphine and/or cyclohexyldibenzylphosphine.

Rhodium is suitably and advantageously present in the reaction mixture of the present invention in an amount of 20- 1000, such as 50-550, ppm by weight of the reaction mixture and preferably charged in form of a precursor selected from the group consisting of a hydride, a halide, a nitrate, a carbonyl compound, an acetate and a dicarbonyl-acetylacetonate. In especially preferred embodiments of the present invention, said precursor is selected from rhodium(lII)nitrate, rhodium(III)acetate, rhodium(I)acetate, acetylacetonatedicarbonyl rhodium(l),

di(rhodium)tetracarbonyl dichloride, dodecancarbonyitetrarhodium and/or hexadecane carbonylhexarhodium.

Solvents may be present in the reaction mixtures of the present invention and are typically saturated hydrocarbons, aromatic hydrocarbons, ethers, aldehydes, ketones, nitriles and aldehyde condensation products. Suitable solvents are for instance pentanes, cyclohexane, benzene, xylene, toluene, diethyl ether, butyraldehyde, valeraildehyde, acetophenone, cyclohexanone and/or benzomtrile.

The hydroformylation process of the present invention preferably involves a catalyst recycle procedure. Catalyst recycle procedures are well known in die art and usually a portion of the liquid reaction solution is continuously removed to a vaporizer/separator. The separated vaporized or distilled aldehyde product may then be condensed and recovered in any conventional manner. The remaining non-volatilized liquid residue which contains rhodium _" phosphite complex catalyst, solvent, free bisphosphite ligand and usually some remaining aldehyde product is then recycled back, with or without further treatment; along with by- products and non-volatilized gaseous reactants that might be dissolved in said recycled liquid residue, to the hydroformylation reactor. The reactant gases removed by distillation from the vaporizer may also be recycled back to the reactor.

The following Examples further illustrate the invention without limiting it in any way.

EXAMPLES

Example 1 illustrates a reduction: in heavy ends formation, when the use of the present invention is applied to a hydroformylation reaction mixture in an industrial production plant.

Example 2 illustrates an increase in purity of the aldehyde product, when the use of the present invention is applied to a hydroformylation reaction mixture in an industrial production plant. Example 3 illustrates the effect of the use of the present invention in a pilot scale, evidencing a reduction in the rate of heavy ends Formation.

Example 4 is a comparative lab experiment that illustrates the effect of different amounts of epoxide addition to a hydroformylation reaction mixture, evidencing a reduction in heavy ends formation rate and a reduction in ligand hydrolysis.

The added epoxide in all of the examples is 3,4-epoxycyclohexylmethyl 3,4- epoxycyclohexanecarboxylate:

Example 1

The amount of heavy ends removed, along with the aldehyde product, from an industrial hydroformylation process of the present invention was measured during a period of 8 months. During the first 5 months the amount of daily removed heavy ends gradually increased from: about 2 tons to 5 tons. After 5 months an epoxide (3,4-epoxycyclohexylmethyl 3,4- epoxycyclohexanecarboxylate) was added to the reactor in an amount of 1 wt% of the reaction mixture. The epoxide was thereafter periodically added to the reactor in order to keep a concentration of about 0.5 wt% epoxide in the reaction mixture.

The amount of heavy ends leaving with the product decreased drastically during the following month to a daily production level of 1 ton heavy ends, a 80 % decrease in heavy ends formation. The result, evidencing a substantial reduction in heavy ends formation upon epoxide addition, is given in attached Figure 1,

Example 2

The amount of aldehyde in the product stream from an industrial hydroformylation process of the present invention was measured during a period of 11 weeks. During the first 7 weeks the amount of valeraldehyde was below or about 98.5 % of the content of the product stream. After 7 weeks the above-described epoxide was added to the reactor in an amount of t wt% of the reaction mixture. The epoxide was thereafter periodically added to the reactor in order to keep a concentration of about 0.5 wt% epoxide in the reaction mixture.

The valeraldehyde content of the product stream steadily increased during the following 4 weeks to almost 99.5 %, a 1 % increase in purity, due to a dramatic decrease in more low-boiling heavy ends. The result, evidencing an increase in purity of the aldehyde product upon epoxide addition, is given in attached Figure 2,

Example 3

Hydraformylation reactions were conducted for a period of 35 days in a continuous pilot scale process with two stirred tank reactors in series and a liquid recycle. An aged catalyst solution was taken from an industrial process and divided in two. One with an addition of roughly 1 wt% epoxide and the other one without any added epoxide. Samples of the reaction mixture were taken regularly and analyzed with gas chromatography (GC) and high pressure liquid

chromatography (HPLC). The result of the analyses are given in attached Figure 3, evidencing a clear decline in the rate of heavy ends formation. A reduction in the rate of heavy ends formation of up to 65%. Example 4

Three parallel experiment in. lab scale were carried out to illustrate the present invention. An

aged catalyst solution was taken from an industrial process and divided in three parts. The

reactors were loaded with equal amounts of the reactor solution. In two of the reactors, epoxide

was added to the reactor solution in amounts of 0.5 wt% and 1 wt%, respectively. The reactors

were heated to 90°C and pressurized with nitrogen to 14 bar and left for 5 weeks under stirring.

Samples of the reaction mixture were taken at regular intervals and concentration of the

compounds in the reaction mixture were determined from analysis o f GC and HPLC. Obtained

result is given in attached Figure 4 and 5, illustrating that the addition of an epoxide reduces the

rate of heavy ends formation and reduces the ligand hydrolysis reaction, by scavenging the acidic

ligand decomposition products In closed systems with an aged catalyst solution, the addition of

the epoxide decreases the heavy ends formation by 50% and the positive effect on the ligand

stability is clearly visible.

i