PROCESS FOR SYNTHESIZING AMINE FROM NITRILE
The present invention relates to a process for hydrogenating an imine function and a nitrile function. In particular, it relates to the simultaneous hydrogenation (the latter term has the meaning of "per ormed in the same reaction medium"), of at least one imine function and at least one nitrile function carried by the same substrate.
In the industry, there is often a need for aliphatic diamines, including cycloaliphatic diamines. These diamines, which are most often primary ones, are more specifically used for making polycondensates such as polyurea or polyamide or as an intermediate in the formation of isαcyanates, whether masked or not.
The most commonly used syntheses include a step for hydrogenating one or more nitrile f nctions. The reduction, of the nitrile function involves a primary imine intermediate. However, in general, this function is more reducible than the nitrile function, the first hydrogenation being the factor, which limits kinetics, so that the intermediate imine amount is relatively small in the reaction media. The coexistence within the reaction medium of a primary a ine function with an imine leads to the formation of an N-substituted imine (secondary one) which, after reduction, forms a secondary amine making up, in many cases, an impurity both undesirable and expensive to eliminate. This problem becomes even more acute when the geometry of the substrate molecule lends itself to the formation of an internal secondary, imine, for example when an amine function of the substrate being hydroger-ated may react intramolecularly o an
imine, a nitrile, or even an amine function, and will thus form a five-, six- and/or five-membered rings. When the amines and/or the nitrile (s) are carried by a ring, a transition can be made from a monocyclic substrate to a polycyclic secondary amine.
This transimination problem is even more acute if the imine function is not an intermediate function occurring within the reaction medium but exists within the initial substrate which then includes, before any other processing, an imine function.
Moreover, when it is attempted to carry out the reaction in a single step by reducing the imine produced in situ by a ketone ammonia condensation, alcohol corresponding to ketone reduction will often form. In order to remedy these difficulties as far as possible, it is appropriate to make the imine functions as little concentrated as possible. Far that purpose, in the case of the initial imine functions, a two-step hydrogenation technique is used in the art, wherein the first one reduces the initial imine into amine and the second one reduces nitrile into amine. This solution is both costly and poses geometric isomery problems, which will be addressed thereafter.
Another solution would be to use a catalyst for promoting the transition reaction from nitrile to amine, but the catalyst system should allow appropriate hydrogenation kinetics of the nitrile function without degrading the hydrogenation kinetics of the ketimine function. The difficulty to overcome is that these properties are, to a large extent, conflicting.
Also, in the case of geometric isomery, and in particular, of cis/trans isomery, the reaction velocity and temperature play an important part. Specifically, as the temperature and the reduction duration increase,
thermodynamic ratios tend to deviate from the most favorable ones or the required ones. Finally, imine hydrogenation catalysts very often also behave as cis/trans iso erization catalysts. The latter aspect is particularly acute for cis/trans isomeries in cycloaliphatic amines.
Thus, when the imine function is on an intracyclic carbon atom, the hydrogenation leads to a cis/trans isomery with respect to the nitrile function. The ratio between the cis and trans forms is governed, on the one hand, by kinetics factors and on the other hand, by a thermodynamic characteristic related to the nitrile function. During hydrogenation, the system's nitrile function tends to approach the thermodynamic ratio, which does not correspond to the equilibrium of the intermediate amine, or former i ine/nitrile system, but corresponds to the equilibrium of the final polya ine system where nitrile is transformed into an aminomethyl function (NH2-CH2-) - The synthesis of a compound commonly referred to by the abbreviation IPDA (IsoPhoroneDiAmine) is an appropriate example of the type of problems that are faced and may serve as a paradigm (an example for teaching a rule) . Isophorone is a molecule resulting from the three- step condensation of acetone, which after adding a hydrocyanic acid molecule after condensation with ammonia and hydrogenation gives IPDA. The industry desires and uses an IPDA having a cis/trans ratio of about 3. Since it is easier to, decrease than to increase it, it is appropriate to aim at ratios that are greater or equal to this value.
Given below are the cis- and trans- formulae of IPDA, as well as the ratio usually desired. The cis- and
trans- forms are chiral but the racemic forms are conventionally used.
Cis (e/e) / 75% Trans (a/e) / 25%
It is therefore an object of the present invention to provide a process for obtaining, in a single step (that is, in a single reaction medium and with the same catalyst) , a diamine or a poly mine by hydrogenating a substrate or an intermediate compound comprising at least one nitrile function and at least one imine function, even transiently.
Another object of the present invention s to provide a process of the above-mentioned kind for obtaining an IPDA yield (RR relative to isophorone nitrile) of at least 80%, advantageously 85%, and preferably 90%.
Another object of the present invention is to provide a process of the above-mentioned type, which allows, when applied to the synthesis of a cycloaliphatic polyamine, a cis/trans isomeric ratio to be obtained, which differs by at least 30%, and advantageously, by at least 100%, from the thermodynamic equilibrium at the reaction temperature.
Another object of the present invention is to provide a catalyst system, which enables accelerated kinetics of the hydrogenation reaction of nitrile functions by maintaining fast kinetics for hydrogenating the imine function.
Another object of the present invention is to provide a catalyst system of the above mentioned type, which allows treatment of a substrate carrying a nitrile function and an imine function, wherein the latter is formed in situ through the action of ammonia on ketone.
These and other objects of this invention will become clear in the following and are achieved by means of a hydrogenation process in a condensed, and advantageously liquid phase, of a substrate carrying a nitrile function and an imine function, which includes a step involving a reagent comprising at least:
- a catalyst based on a metal of group VIII in the periodic table of the elements;
- water; and - ammonia.
The effect is sensitive for substrate/catalyst ratios at least equal to 0,02 mole par gram of catalyst, advantageously from 0.03 to 0.3, preferably from 0.05 to 0.2 The preferred processing conditions are as follows:
Temperature; from 40°C to 150DC, and advantageously from 60° to 120°C (temperature is not required to be kept permanently at a single value chosen within this range and it may advantageously be programmed so as to increase while the reaction proceeds either continuously or stepwise) ;
Total pressure: from 10s to lQ7 Pa, advantageously, from 2.10s to 8.10e Pa, and preferably from 2.10s to 6.1C6 Pa; - ρH2: from 5.105 to 9.10° Pa, and preferably from 3.10G to 5.1015 Pa;
- Ammonia/imine mole ratio: at least two, and the upper limit is economic rather than technical but an upper limit for this ratio would be 100.
- Catalyst/substrate weight ratio: at least 0.5%, and the upper limit is economic rather than technical, but an upper limit for this ratio would be of at most 50% (in particular for a continuous reaction) . The reaction is performed in a liquid phase, advantageously in a solvent or a solvent system, the solvents being chosen from common solvents for nitrile hydrogenation .
In particular, these common solvents include alcohols, preferably primary ones, including in particular methanol, ethers, particularly cyclic ethers
(such as dioxane, THF ana tetrahydropyrane) and mixtures thereof.
The presence of water is very important.' In this regard, it is desirable that the [H20] / [substrate] ratio of water, expressed in moles, to the imine function, expressed as an equivalent, be at least 1, advantageously 2, and preferably 3.
However, this presence should be moderate. Thus, it is desirable that the [H20] / [imine] ratio of water, expressed in moles, to the imine function, expressed as an equivalent, be at most 10, and advantageously 8.
It is desirable that for said hydrogenation step, the ratio of ammonia, expressed in moles, to the imine functions, expressed as an equivalent
( [NH3] / [substrate] ) , be at least 2, advantageously 4, and preferably 6.
In order to obtain a good yield, it is preferable that the ratio ( [NH3] / [imine] ) of ammonia, expressed in moles, to the imine functions, expressed as an equivalent, be at most 50, and advantageously 25.
It is advisable to choose as the ratio of ammonia, expressed in moles, to water, expressed in moles
( [NH3] / [H20] ) , a value of at least 1, advantageously 2, and preferably 5.
Advantageously, said group VIII element-based catalyst (the periodic table of elements used in the present application is that given in the addition to the
"Bulletin de la Societέ Chimique de France", January
1966, No.l) is chcsen from those referred to as "Raney© type metals".
Said Raney® type metal catalyst is a catalyst whose group VIII metal is advantageously cobalt or nickel, preferably cobalt.
The catalyst further contains aluminum, which originates from the starting mother alloy (it should be recalled that Raney® type catalysts are obtained through alkaline leaching of aluminum alloys, of the catalyst base metal and possibly, of all or part of the potentiator metals described below) ; it is alsc preferable that the potentiator metals having a redox potential, under the processing conditions of the invention, close to (that is in the inclusive range of
±0.3V) or greater than that of aluminum, be introduced in the initial aluminum alloy) .
Aluminum is not considered as a potentiator agent in the present invention. In general, the atomic ratio of aluminum to the base metal in the catalysts is at most 1/3, and advantageously at most 1/4, preferably at most 1/5, and more preferably at most 10%.
Iron is often present as impurity in the mother alloy. It has no detrimental effect on the property of the catalyst; on the contrary, when used according to present invention, it enhances the property of the catalyst at least when the content expressed in atomic percentage towards cobalt is at most equal to 5% advantageously 2.5%.
Advantageously, the catalyst is potentiated by at least one, and preferably two element (s) other than the metal on which the catalyst is based. These elements are used or serve as co-catalysts and are chosen from the following group of elements: gallium, chromium, nickel, platinum group metals (in particular rhodium and iridiurr.) , hafnium, zirconium, lead, and tin.
Thus, said catalyst is then used in the presence of at least one element serving as a co-catalyst chosen from the above elements.
It is desirable that the atomic ratio [co- cat] /[cat] of each of the elements present as the co- catalyst to the base metal in the catalyst be at least 1/1000, and advantageously at least 5/1000. It is also desirable that the [co-cat] / [cat] atomic ratio of each of the elements present as the co-catalyst to the base metal in the catalyst be at most 1/10, and preferably, at most 1/20.
Desirably, the [∑co-cat] / [cat] atomic ratio of the sum of the elements present as the co-catalyst (s) to the metal in the catalyst will be at least 2/1000, advantageously at -least 5/1000, and preferably at least
1%.
It is also desirable that the [∑co-cat] / [cat] atcmic ratio of the sum of the elements present as the
■ co-catalyst (s) to the metal in the catalyst be at most
1/3, advantageously at most 1/5, preferably at most 1/10 and more preferably at most 7%.
The preferred catalysts will include cobalt-based ones, among which those: containing at least chromium as a potentiator element; containing at least nickel as a potentiator element;
and advantageously, containing at least both nickel and chromium as the potentiator elements . According to the present invention, it is suggested to use at least two distinct elements serving as the co- catalysts .
The co-catalyst elements may be introduced into the aluminum alloy comprising the raw material of the Raney© type metal, before the alkaline leaching. They may also be introduced as salts, which should advantageously be soluble, into the initial reaction medium but also as salts during the alkaline leaching of the alloy.
Said process is particularly well suited to substrates which have at least one nitrile function carried by an aliphatic ring, especially when said sxibstrate initially has an imine function carried by an aliphatic ring.
The method is particularly useful when said substrate is such that the final diamine has a cis/trans isomery with respect to a ring carrying the amine and aminomethyl functions (-CH2-NH2) .
In this case, it is preferable that said hydrogenation step be stopped before the ratio corresponding to the cis/trans thermodynamic equilibrium is reached.
Advantageously, in the substrate, the imine and nitrile functions of the substrate are carried by an aliphatic isocyclic radical, wherein the nitrile function is carried by an intracyclic sp3 hybridization carbon atom and the imine function carrying carbon atom is intracyclic and is linked to sp3 hybridization carbon atoms .
The obtained results are of particular interest when the process is used for synthesizing cycloaliphatic amines, which are usually employed as the raw material for cycloaliphatic isocyanate (NBDI, H12MDI) and especially, for synthesizing IPDA.
The process according to the present invention leads to cis/trans ratios greater than 3 (which may range up to 5 or even about 15) . When needed, a simple heating step allows this ratio to be reduced to the desired value (in general, about 3, as mentioned above).
The reaction time generally ranges from 1 hour to
10 hours, and advantageously, from 3 to 8 hours. It is good practice to interrupt the reaction when the conversion rate of the nitrile function and the intermediate amidine reaches a predetermined value ranging from 90% to 99%, and advantageously from 95% to
98%. The catalyst amount and hydrogen partial pressure are then chosen for the conversion rate to reach the chosen value after a time duration ranging from 1 hour to 10 hours.
With a high ratio of the catalyst to the substrate the isomeric ratio approaches the kinetic ratio and conversely, a small ratio will drive it closer to the thermodynamic ratio. Thus, in the case of IPDA originating from isophorone nitrile, the cis/trans ratio increases with the catalyst/substrate ratio.
It should be recalled that iminated isophorone nitrile has the following formula:
and is sometimes introduced as a mixture with the following equilibrium:
One of the advantages of the present invention is to be able to dispense with the presence of a hydroxide (or any compound leading thereto in the reaction medium) , notably a monovalent metal or a quaternary ammonium- These hydroxides are objectionable according to the present invention.
Thus, according to the present invention, the content of these compounds is more related to the impurity of the components of the reaction medium than to a possible intentional addition.
It should therefore be mentioned that, preferably, the reaction mixture should contain at most 1%, advantageously at most I'ss and preferably 100 ppm in weight of such hydroxides, in particular alkaline metals .
The definitions of the various yield concepts will be summarized below:
TT = Transformation rate of the starting material WITH ττ_.number of moles of converted material number of moles of initial material P,R = yield over the initial material WITH number of moles of end material
RR= number of moles of initial material
RT = yield of converted material WITH
number of moles of final material number of moles of converted material CATALYSTS TESTED IN EXAMPLES :
Grace Davison 2700 Catalyst: 97.51% Cobalt (94.0% minimum, 100% maximum) 1.85% Aluminum (5.0% maximum) 0.3% Iron (0.5% maximum) 0.34% Nickel (0.5% maximum) pH in water 9.2 (10.0 maximum) 13 ppm (50 maximum) of Sodium Oxide
0.1 ppm (10 maximum) of Alurrάnum Oxide
Partide Size <10th % 7.33 microns (5.0 minimum)
Particle Size <50th s: 30.08 microns (20.0 minimum,
40. 0 maximum)
Particle Size <90th %: 81.79 microns (110.0 maximum)
Grace Davison 2724 Catalyst:
91.7% Cobalt (89.1% minimum, 92.8% maximum)
3.27% Aluminum (3.0% minimum, 5.0% maximum)
0.3% Iron (0.7% maximum)
2.13% Chromium (1.8% minimum, 2.4% maximum)
2.58% Nickel (2,4% minimum, 2.8% maximum) pH 9.5 in water (8.7 minimum, 10.0 maximum)
12.5 ppm (50 maximum) of Sodium Oxide
0.1 ppm (10 maximum) of Aluminum Oxide
Particle Size <10ch%: 5.98 microns (5.0 minimum)
Particle Size <50th %: 26.94 microns (25.0 minimum,
45.0 maximum)
Particle Size <90r"r' %: 71.37 microns (110.0 maximum)
Hydrcgen Desorption: 11 ml/g (9.3 minimum)
Activated Metals Corporation A-8B46 Catalyst:
90.7% or 86.75% Cobalt (83% minimum) 3.55% or 4.05% Aluminum (8% maximum) 0.33% Iron (0.7% maximum)
Chromium + Nickel - 5% (analyzed 2.1% Cr and 2.4% Ni) Average Particle Size; 55 or 35 microns
pH in water 11.5 (13 maximum)
PROCEDURE FOR EXPERIMENTS:
Preparation of the stock solutions 50 grams each of the catalysts were thoroughly washed inside a glove box with water, then methanol, then diethyl ether, and finally blown dry with a stream of nitrogen. The oxygen level was kept at 0.5 ppm or less . Stock solutions of the substrates dissolved ir. various solvents (here methanol if not otherwise mentioned) containing an internal standard, ammonia, and water were prepared and stored at -4°C.
For these examples, the substrate was isophorone nitrile (IPN) ; the solvent was methanol; the internal standard was triethylene glycol dimethyl ether (triglyme) ; and the ammonia to substrate molar ratio was approximately 20.
Isophorone imine nitrile (IPIN) and water were formed in situ and were stable in the stock solution.
Dry catalyst powders were added to arrays of 96 glass vials (1ml each) in common headspace parallel batch reactors for high-throughput testing (see European, Patent Application No. EP 1174185. Stock solutions were added into vials (total amount 380mg ±5 mg) and the reactors were sealed before removing them from the glove box. The reactors were placed in a vortex-heating unit and a gas line was attached that was evacuated and purged several times at 25°C before pressurizing with hydrogen (35 bar except if mentioned otherwise (one bar is 100 kPa) (500 psi) for these examples). The vortεxing speed was increased to 800 rpm before the reactors were heated to the reaction temperature. The maximum total pressure varied between
40 bar (590 psi) ar-d 50 bar (740 psi) for these experiments. After set period of time (6 hours except if mentioned otherwise) , the vortexing was stopped and the reactors cooled to 25 °C or less before transferring them to the glove box. Liquid samples were withdrawn and analyzed by GC/MS and calibrated GC methods. Example 1: effect of water; test in Parr Bomb
The tests were carried under hydrogen partial pressure of 35 bar with a molar ratio NB3/IPN equal to 20 with a IPN/Cat ratio (by weight) equal to 8,8 p/p. in all of the following test IPN conversion was complete but the intermediate conversion is only partial. Thus the reactions were not finished. These conditions allow having good approach of the effect of the parameters.
Table: Effect of water content on the efficiency of the catalyst (at time = 6 )
Operating conditions :
Effect of the potentiator elements
The hereinafter examples show the effects of different catalysts and additives on the IPDA Yield and Cis/Trans Ratio-
Example 2 : ef ect of the potentiator elements introduced in the alloy precursor of the Raney® type cobalt In these experiments, a 2700 or 2724 Cobalt catalyst from Grace Davison was used; the reaction temperature was 80°C. The water content corresponds to the water formed by the ammonia ketone condensation. The catalyst to substrate weight ratio the same in the two experiments. Without additives, the 2724 catalyst gave a higher yield and lower Cis/Trans ratio than the 2700 catalyst, 89.2% and 3.2 versus 85% and 3.5%.
Catalyst Content % by weight in IPDA Cis/Trans used Yield Ratio
Further experiments were carried out to compare Grace Davison 2700 catalyst with A8-B46 the reaction temoeraturε was 80°C
The substrate to catalyst weight ratio (IPN/cat)was 8.2 ± 0.2; and the reaction time was 6 hours. The results is given in the following table:
5 Example 3 : e ect of potentiator elements introduced through liquid phase on Raney® type cobalt with low nickel content
The reaction temperature was Θ0°C. The water content corresponds to the water formed by the ammonia 10 ketone condensation.
The catalyst to substrate weight ratio wasthe same in the experiments and the reaction time was 6 hours.
Trie amount of additive used (1.3 mole % of the active metal) was calculated by assuming that the
15 catalyst was pure cobalt and that only 10 wt. % was active as a catalyst. Additives dissolved in methanol were added to the catalysts in vials that were agitated then allowed to stand overnight. The samples were then dried by gently blowing off the solvent with a stream of
20 pure nitrogen gas inside a glove box. Acetylacetonates of hafnium, chromium, and iridium as well as tin acetate
-- decreased the Cis/Trans ratio and/or increased the IPDA yield of the 2700 catalyst.
Effect of one potentiator element on Raney® type cobalt with low nickel content Operating parameters :
Catalyst: Grace Co 2700
IPN/Cat, mg/mg: the same for all the tests Total estimated Maximum Pressure, psi: 652
NH3/IPN, mole/mole: 20,00
Total Solvent/IPN, mg/mg: 17,87
Solventl: MeOH
Active Metal, moles: 3, 8181Ξ-06
Approximate Solventl, mg: 310,0
Solvent2 : None
IPN, mg: 17,4
Triglyma (Internal Standard), mg: 18,7
NH3, mg: 35,8
Temperaturel , βC: 80
Hold Ti el, hrs: 6
H2 Pill Pressure at 25° C, psi: 500
Maximum Gas Pressure, psi: 595,3 Estimated Maximum Vapor Pressure, psi: 57,0
Speed, rp : 800
The reaction temperature was 80 °C. The water content corresponds to the water formed by the ammonia ketcne condensation.
The catalyst to substrate weight ratio was 0.13; and the reaction time was 6 hours.
The amount of additive used (1.3 mole % of the active metal) was calculated by assuming that the catalyst was pure cobalt and that only 10 wt- % was active as a catalyst. Additives dissolved in methanol were added to the catalysts in vials that were agitated then allowed to stand overnight. The samples were then dried by gently blowing off the solvent with a stream of pure nitrogen gas inside a glove box. Palladium acetylacetonate slightly increased the IPDA yield of the 2724 catalyst.
Additiv , Additive Additive Additive, IPDA Cis/Trans mol % of MW, g/mole mg yield, Ratio
Active
Metal
0 None 0 0 58.0 12.9
6 Nb(0Et)5 318.22 0.97184388 52.0 12.1
6 Ni (acac) 2 256.93 0.78466422 41.0 11.6
6 Zr (acac) 4 487.66 1.4Θ931364 64.0 12.0
6 Re207 484.4 1.4793576 0.0 0.0
6 Ru (acac) 3 398.4 1.2167136 64.0 12.1
6 TiO(acac) 2 262.12 0.80051448 38.0 10.8
6 V (acac) 3 348.27 1.06361658 19.0 7.9
6 Zn(0Ac)2 183.46 0.56028634 25.0 8.6
Rh(acac) 3 400.24 1.22233296 63.0 12.4
When introduced through liquid phase:
- The nickel is a bad potentiator,
- The platinum group metals are good potentiators. The niobium, rhenium, titanium, vanadium and zinc are comparative .
Example 5: temperature effect in both cases the conditions are
- Catalyst used: A-8B46 Cobalt;
- Catalyst to Substrate Weight Ratio: 0.48; the reaction time: 6 hours. Temperatures tested: 60°C and lOO'C. The water content corresponds to the water formed by the ammonia ketone condensation.
The experimental parameters are otherwise identical.
For reaction temperature of 60°C, and for an average of 12 tests, the results read as follows:
IPDA Yield: 87.0% (standard deviation 0.4) and Cis/Trans Ratio 4.3 (standard deviation 0.2).
For reaction temperature of 100°C, and for an average of 8 tests, the results read as follows:
- IPDA Yield: 89.1% (standard deviation 0.3)
- and Cis/Trans Ratio 4.3 (standard deviation 0.1). The IPDA yield increased and the Cis/Trans Ratio decreased as the Temperature was increased.
Example 6 : e ct of the Catalyst to Substrate Weight Ratio on the IPDA Yield and Cis/Trans ratio Two groups of experiment were tested.
In both groups of experiments, an A-8B46 Cobalt catalyst from Activated Metals Corporation was used; the reaction temperature was 100 °C; and the reaction time was 6 hours. The catalyst to substrate ratio for the first group of experiments (8 tests) was 0.48, and 0.11 for the second group (six tests) of experiments. The water content corresponds to the water formed by the ammonia ketone condensation. The samples were otherwise identical- ^ The results read as follows:
At these reaction conditions, the IPDA Yield increased and the Cis/Trans Ratio remained approximately the same as the catalyst to substrate weight ratio was increased.
Example 7 : the effect of the Reaction Time
The effect of the Reaction Time on the IPDA Yield and Cis/Trans Ratio is illustrated hereinafter.
The average of three Data Set are compared:
<• 8 Experiments for a duration of 3 hours;
<* 4 Experiments for a duration of 6 hours;
♦t* 10 Experiments for -a duration of 12 hours.
The water content corresponds to the water formed by the ammonia ketone condensation.
In all cases, an A-8B46 Cobalt catalyst from Activated Metals Corporation was used; the Catalyst to Substrate Weight Ratio was 0.48; and the reaction temperature was 80 °C. The reaction time was changed from 3 to 6 and then 12 hours. They were otherwise identical.
The average values read as follows:
Example 8 : effect of the alkaline hydroxides
catalyst ratios Reaction Conditions :
3.8 mg IPN per vial 0 μl SM/IS/MeOH 30 μl 7N NH
3/MeOH 00 μl of H
20 + MeOH (variable) 0C, 6 hrs, 500 psi H
2 at 25
αC, 800 rpm
The results are given in the following table