Isophorone hydrogenation

Figure 2 Experimentally determined phase boundaries of four mixtures of isophorone/C02/H2 of varying composition (isophorone 5-22% wt.% isophorone/hydrogen molar ratio was fixed at 1 1.7). N.B. 100 bar = 10 MPa.

In the multistep production of IPDI, isophorone is first converted to 3-cyano-3,5,5-trknethylcyclohexanone (231—235), then hydrogenated and ammoniated to 3-aminomethyl-3,5,5-trknethyl-l-aminocyclohexane (1) (236,237), also known as isophorone diamine (IPDA). In the final step IPDA is phosgenated to yield IPDI (2) (238). Commercial production of IPDI began in the United States in 1992 with the startup of Olin s 7000 t/yr plant at Lake Charles, Louisiana (239), and the startup of Hbls integrated isophorone derivatives plant in Theodore, Alabama (240). Hbls has a worldwide capacity for IPDA of 40,000 t/yr. 

Iododesilylation, 41 of aryltrimethylsilanes, 42 Iodomethyltrimethylsilane, 27 4-Iodophenylalanine, 42 Ionic hydrogenation, 136 Ireland-Claisen rearrangement, 112-14 ftwtf-2 Isocyanocyclohexanol, 137 Isophorone, 52 Isophorone dienol ether, 135 (Isopropoxydimethylsilyl)methyl magnesium chloride, 58... 

Epoxyeyclohexanone has been prepared in 30% yield4 by epoxi-dation of 2-cyclohexen-l-one with alkaline hydrogen peroxide, using a procedure described for isophorone oxide (4,4,6-trimethyl-7-oxabicyclo[4.1.0]heptan-2-one).5 A better yield (66%) was obtained using f r/-butyl hydroperoxide (1,1-dimethylethylhydroperoxide) and Triton B in benzene solution.6 The procedure described here is simple and rapid. 

The reaction mechanism is shown in Figure 4 and is adapted from work by Fiego et al. [9] on the acid catalysed condensation of acetone by basic molecular sieves. The scheme has been modified to include the hydrogenation of mesityl oxide to MIBK. The scheme begins with the self-condensation of acetone to form diacetone alcohol as the primary product. The dehydration of DAA forms mesityl oxide, which undergoes addition of an addition acetone to form phorone that then can cyclise, via a 1,6-Michael addition to produce isophorone. Alternatively, the mesityl oxide can hydrogenate to form MIBK. 

Highly mesoporous carbon supported Pd catalysts were prepared using sodium formate and hydrogen for the reduction of the catalyst precursors. These catalysts were tested in the enantioselective hydrogenation of isophorone and of 2-benzylidene-l-benzosuberone. The support and the catalysts were characterized by different methods such as nitrogen adsorption, hydrogen chemisorption, SEM, XPS and TPD. 

The type of catalyst strongly influences the enantioselectivity of heterogeneous catalytic hydrogenations (1). In the enantioselective saturation of the C=C bond of isophorone over (-)-dihydroapovincaminic acid ethyl ester ((-)-DHVIN) modified Pd catalysts (scheme 1) the optical purity strongly depended on the type and properties of the support used (2, 3, 4). 

In the enantioselective hydrogenation of isophorone in the presence of (-)-DHVIN modifier the best optical purity was afforded by small dispersion (<0,05) Pd black catalyst (up to 55%) (7). The influence of the preparation method of Pd black on the optical yield was reported (8). A correlation was found between the oxidation state of the metal surface and the enantioselectivity, the catalyst having more oxidised species on its surface giving higher enantiomeric excess, while the Pd black with lower surface area was more enantioselective. 

In the hydrogenation of isophorone the catalyst type 1 of smaller dispersion resulted in higher enantiomeric excesses especially with DPPM modifier. In the hydrogenation of 2-benzylidene-l-benzosuberone the catalyst type 2 of higher dispersion was more enantioselective. These reverse tendencies or smaller relative difference in e.e. for the latter reaction can be attributed to the use of modifiers with totally different structure and working mode. 

Table 4 Enantiomeric excesses in the enantioselective hydrogenation of isophorone and 2-benzyl- 1-benzosuberone on highly mesoporous carbon supported Pd catalysts.

New modifiers have traditionally been discovered by the trial-and-error method. Many naturally occurring chiral compounds (the chiral pool38) have been screened as possible modifiers. Thus, the hydrogenation product of the synthetic drug vinpocetine was discovered to be a moderately effective modifier of Pt and Pd for the enantioselective hydrogenation of ethyl pyruvate and isophorone.39 Likewise, ephedrine, emetine, strychnine, brucine, sparteine, various amino acids and hydroxy acids, have been identified as chiral modifiers of heterogeneous catalysts.38... 

Isophorone diamine is synthesized traditionally by aminoreduction of iso-phoronenitrile. Raney cobalt was used for this process. More recently, a new two-step process was patented. The first step consists of synthesizing the imine and the second one of hydrogenating the latter. Ra-Ni was used as catalyst at 150°C and 60 bar hydrogen pressure. Under these conditions, the catalyst reduces the nitrile groups and is able to cleave the N-N bonds, too. Ammonia is required to promote primary amine formation during nitrile hydrogenation (Scheme 4.151).554... 

Isophorone oxide has been prepared by the epoxidation of isophorone with alkaline hydrogen peroxide.2 3... 

The bicyclic tropane ring of cocaine of course presented serious synthetic difficulties. In one attempt to find an appropriate substitute for this structural unit, a piperidine was prepared that contained methyl groups at the point of attachment of the deleted ring. Condensation of acetone with ammonia affords the piperidone, 17. Isophorone (15) may well be an intermediate in this process conjugate addition of ammonia would then give the aminoketone, 16. Further aldol reaction followed by ammonolysis would afford the observed product. Hydrogenation of the piperidone (18) followed then by reaction with benzoyl chloride gives the ester, 19. Ethanolysis of the nitrile (20) affords alpha-eucaine (21), an effective, albeit somewhat toxic, local anesthetic. 

Isomerization of jS-isophorone to a-isophorone has been represented as a model reaction for the characterization of solid bases 106,107). The reaction involves the loss of a hydrogen atom from the position a to the carbonyl group, giving an allylic carbanion stabilized by conjugation, which can isomerize to a species corresponding to the carbanion of a-isophorone (Scheme 9). In this reaction, zero-order kinetics has been observed at 308 K for many bases, and consequently the initial rate of the reaction is equal to the rate constant. The rate of isomerization has been used to measure the total number of active sites on a series of solid bases. Figueras et al. (106,107) showed that the number of basic sites determined by CO2 adsorption on various calcined double-layered hydroxides was proportional to the rate constants for S-isophorone isomerization (Fig. 3), confirming that the reaction can be used as a useful tool for the determination of acid-base characteristics of oxide catalysts. 

The possible reactions that can take place are outlined in Figure 1 (6). Typicai by-products of the aldol reaction are phorone and isophorone, however as can be seen in the figure other by-products can be formed, especially in the presence of a hydrogenating functionality. 

At 673 K the product distribution over the palladium catalysts (Figure 4) was still highly selective to MIBK (> 80%) however further aldol condensation of acetone with MO to form the three intermediates, phorone, 4,4 -dimethyl hepta-2,6-dione and 2,4-dimethyl hept-2,4-dien-6-one was observed. These species were formed by aldol condensation of acetone with MO at different points in the molecule (10). All can continue to react either by subsequent aldol condensation, Michael addition, or hydrogenation as per Figure 1. There was no detectable isophorone produced, the product of internal 1,6-aldol reaction of... 

It is noticeable that over the nickel catalysts, isophorone was a major product. The production of isophorone and the small quantities of other byproducts once again revealed formation of the three intermediate aldol products of the reaction between MO and acetone. However it is clear that hydrogenation was not as facile over the nickel catalysts as it was over the palladium catalysts. Hence there was more secondary and tertiary aldol condensation. To further investigate this a reaction was performed at 573 K using the Ni-KOH/silica catalyst with nitrogen as the carrier gas rather than hydrogen. The results are shown in Table 1. 

Clearly there is a switch from MO to MIBK when hydrogen was made the carrier gas. The yield of isophorone dropped as at this temperature the hydrogenation intercepted the MO before it could undergo another aldol condensation. At 673 K the aldol condensation reaction became kinetically more competitive with the hydrogenation reaction. 

In conclusion, four catalyst systems, Pd-KOH/silica, Pd-CsOH/silica, Ni-KOH/silica, and Ni-CsOH/silica, have been investigated for the conversion of acetone to MIBK. Systems highly selective to MIBK have been obtained (Pd-CsOH/silica, 100 % at 473 K). Both the metal and the base affected the product distributions. The nickel catalysts were generally less selective, with MIBK being further hydrogenated to MIBC and isophorone becoming a major product at high temperatures. Over Ni-CsOH/silica, a selectivity of around 30% was... 

Tests of the best hydrogen bonding agents in ether-alcohol showed the following compds trimethyl-, trimethallyl-, tributyl phosphates, isophorone, dimethyl acetamide, and dibutyl tartrate were superior to dibutyl phthalate or triacetin in the amts reqd for gelatinization and in the viscosities of the resulting solns... 

---