CA2059806C - Multistep process for the liquid phase ammoximation of carbonyl compounds - Google Patents

Abstract

A multistep process for the liquid phase ammoximation of carbonyl comp-ounds with H202 and NH3, at 60 - 100°c, at 1.5 - 5 bar and in the presence of a catalyst based on silicon, titanium and oxygen, charact-erized in that:
a) in one or more primary steps, the H202 : carbonyl compound molar ratio ranges from 0.9 to 1.15 by mols and the carbonyl component conversion is carried out up to at least 95%;
b) in a last (exhaustion) step the H202 : carbonyl compound ratio is in a higher range, i.e. from 1.5 to 3.0 by mols.

Description

The invention relates to a multistep process for the liquid phase ammoximation of carbonyl compounds with hydrogen peroxide and ammonia. A typical example is the ammoximation of cyclohexanone to cyclohexanone-oxime, and in the text reference will be almost always made to this particular type of process;
of course, this does not exclude the possibility that the inven-tion may be applicated also to other carbonyl compounds.
European patents 20$,311; 267,362; 299430 and 347,026, in the name of the Applicant teach that said ammoximation can be effectively obtained in the presence of a catalyst based on silicon and titanium. This type of catalysis permits to ob-tain very high conversions and selectivities; nevertheless, a quantitative conversion (which would simplify the oxime separ-ation and recovery) is practically never obtained, particularly in the case of cyclohexanone, and the unreacted carbonyl comp-ound represents a problem not only because it is necessary to recover it, but also owing to the possible secondary reactions which it can give rise to (during the separation and the purific-ation of the oxime). The by-products of these reactions (in the case of cyclohexanone, they are cyclohexyl-cyclohexanone, bis-cyc)ohexe-~o~~~o~
nyl-cyclohexanone and octahydro-phenazine) cause, as is known, a worsening of the quality of the caprolactam obtainable in the subsequent Beckmann rearrangement. In order to complete the oxim-ation of the residual ketone, it would be possible to resort to a reaction with a solution of hydroxyiamine sulphate, under oper-ative conditions known from the art. The problem of the non-quan-titative conversion of the carbonyl compound regards both the ammoximation of cyclo_hexanone to cyclohexannne-oxime and the am-moximation of other ketones (or aldehydes) such as e.g. acetone, methylethyl ketone (2-butanone), acetophenone, cyclododecanone, enantic aldehyde (l-heptanale), etc. But the hydroxylamine sul-phate solutions are obtainable only by complex processes such as, for example, the Raschig process (reduction of the nitrogen ox-ides with ammonium bisulphite).
The Applicant has now set up an ammoximation process which permits to reduce the amount of residual ketone (or of re-sidual aldehyde), contained in the effluents of the primary steps of the ammoximation process, to the same levels which can be reached with the hydroxylamine sulphate, without resorting, however, to the use of said sulphate, which, as already pointed out, can be produced only by means of a considerably complex pro-cess.
In its broadest aspect, the invention relates to a mul-tistep process for the liquid phase ammoximation of carbonyl compounds with H202 and NH3, at 60-100°C, at 1.5-5 bar and in the presence of a (suspended) catalyst based on silicon, tita-nium and oxygen, characterized in that:
a) in one or more primary steps the H202 . carbonyl compound ratio ranges frorn 0.9 to 1.15 by mols (preferably from 1.0 to 1.1) and the carbonyl compound conversion is brought at least up to 95% (preferably up to 96-99%);
b) in a last step (exhaustion step) the H202 . carbonyl compound ratio is at a higher Level, i.e. it ranges from 1.5 to 3.0 by moss (preferably from 1.5 to 2.2).
The above-indicated number ranges are not critical for the purpose of improving the process. The Applicant had previ-ously tried to carry out the quantitative conversion of the car-bonyl compound in a single reactor, without any additional com-pletion (exhaustion) step, and using much higher amounts of ox-idant (H202) since the beginning, as well as much longer re-action times, but it realized that the initial excess of H202 and the too long times could cause the instability of the re-action system. Namely, competitive reactions were observed, which involved the oxime and/or the carbonyl compound and/or the ammonia; the competitiveness, obviously, refers to the am-moximation reaction. Said competitive (secondary) reactions led to a degradation of the quality of 'the produced oxime and to a considerable formation of nitrogen-containing by-products (N20, N2, N02 , N03 etc.): A consequence of this quality degrad-ation was that the quality specifications of the caprolactam ob-2059~0~

tainable from the oxime (in cascade) were not met. In other words, it is the Applicant's merit, that it has surprisingly found that it is possible to operate with a high H202 excess, provided said excess is added beyond a certain conversion level, i.e. provided that the residual carbonyl compound concentrations are very low (more exactly: provided that the carbonyl compound conversion has exceeded 95%). In fact it was observed that under these cond-itions the expected quality worsening did not occur at all and that a practically complete conversion of the carbonyl compound was obtained without carrying out the undersized and complex post-treatments (with hydroxylamine sulphate) and without secondary re-actions. The practically quantitative reactions of the carbonyl compound involved by the process according to the invention are namely accompanied by the production of an oxime of equivalent or higher quality than the one obtainable according to the prior art.
A few preferable operative details are briefly listed hereinafter.
A) RESIDUAL REAGENT IN THE EFFLUENT FROM THE PRIMARY STEPS
Really excellent results, in terms of yield calculated on hydro-gen peroxide and of quality of the producted oxime, are obtain-ed when the concentration of the residual carbonyl compound in the effluent from the primary steps does not exceed 1% (pre°
ferably 0.5%) by weight.
B) OPERATIVE CONDITIONS OF THE PRIMARY STEPS
The ammonia concentration in the liquid reaction medium shall range from 1.0 to 2.5% (preferably from 1.5 to 2.0%) by weight.

20~~~0~
_ 6 _ The H202 . ketone (or aldehyde) feed molar ratio shall preferably range, as already mentioned, from 1.0 to 1.1.
The concentration of the catalyst suspended in the liquid medium shall be such as to have a specific productivity (expressed as parts by weight of produced oxime per part of catalyst and per hour) from 6 to 12, preferably of about 8. The residence time in each of the primary steps shall not exceed 120 minutes, preferably it shall range from 30 to 90 minutes.
C) EXHAUSTION STEP
in the last step (exhaustion step), where a complete con-version of the residual carbonyl compound shall be reach-ed, in particular a concentration of the unreacted comp-ound not higher than 200 ppm (preferably not higher than 100 ppm and, still more preferably, not higher than 50 ppm on the liquid medium), the above-cited variables, as already mentioned, shall be in the following ranges:
- H202 . carbonyl compound molar ratio from 1.5 to 3 (preferably from 1.5 to 2.2);
- residence time from 10 to 60 minutes.
In the exhaustion step it is better not to introduce fur-ther fresh ammonia, since the amount dissolved in the li-quid is sufficient for the purpose; a too high ammonia ex-cess, referred to the carbonyl compound, in the presence of a hydrogen peroxide excess would result in an oxidant 2~59~~~

loss and in the formation of undesired gaseous by-prodcuts, such as Nz and N20. In the exhaustion step, the specific catalyst productivity decreases to a lower level, i.e. from 0.1 to 5 (preferably from 0.3 to 0.5) due to the different operative conditions, and the temperature is preferably maintained at the same value as specificated for the pri-mary steps.
D) GENERAL CONSIDERATIONS
As already mentioned, the temperature in all the steps shall range from 60° to 100°C (preferably from 70° to 90°C). At lower temperatures, the reaction kinetics is rather slow, while at higher temperatures the negative effect of the parallel reactions as well as of the consecutive reactions (which start from already formed oxime) begins to become noticeable. The pressure in each of the primary steps and exhaustion step shall prevent the reaction liquid to begin boiling and shall maintain the ammonia concentration in the liquid medium from 1 to 2.5% by weight, preferably at a va-lue lower than 2%; the pressure acts also as motive power in the liquid filtration. Generally, values from 1.5 to 5 bar (preferably from 1.8 to 3 bar) are sufficient, with the values decreasing from the first to the last step. The residence time in each step, with exception of the last, shall be such as to have a residual ketone or residual al-dehyde conversion equal to or higher than 95%. The reaction - g time, in each of these steps, is generally not longer than one hour in order to prevent subsequent reaction of the oxime which has formed. Conversely, too short reaction times lead to an un-satisfactory conversion of the carbonyl compound and to a too high concentration of the reagent in the liquid medium, what promotes the formation of by-products through condensation reactions. In the last step (exhaustion step), the reaction time shall be much shorter in consideration of the lower amount of ketone to be converted. The hydrogen peroxide/ketone feed molar ratio in each step, with exception of the exhaustion step, shall preferably be slightly above one, since a little amount of hydrogen peroxide is always consumed, as already mentioned, in parallel reactions (with formation of gaseous products such as N2 and N20, by ammonia oxidation). Furthermore, as already pointed out, the hydrogen peroxide/ketone molar ratio in the last step, where it is no longer advisable to feed ketone and which shall be capable of bringing the ketone concentration to values lower than 200 and preferably than 100 ppm, shall be considerably higher than the one utilized in the preceding steps (from 1.5 to 3 and preferably from 1.5 to 2.2). The productivity of each step is strictly related to the concentration of the ca-talyst suspended in the solution contained in each reactor. The continuous feeding to each step shall be regulated in order to have a specific productivity (expressed in parts of produced oxime per part of catalyst and per hour) within the prefixed ~~~~~U6 _ g values. In order to guarantee an effective dispersion of the catalyst in the liquid medium, the catalyst concen-tration can vary from 1 to 15% by weight. At too low con-centrations, the productivity of each step becomes too low and not profitable in the economic respect, while too high concentrations give rise to problems as regards stirring and/or filtration of the reaction product. Preferably and advantageously said concentration can be maintained from 1 to 60% by weight. As a catalyst it is possible to use a titanium silicalite, as is cited for example in European patents 267,362 and 299,430, or one of the amorphous com-pounds described in European patent 347,926. The average particle size of the catalyst generally ranges from 1 to 100 microns, preferably from 5 to 50 microns.
E) SOLVENTS
Proper solvents for the ammoximation (including the exhaust-ion step) are the usual organic solvents described in the older patent, a few of which have been cited hereinbefore;
said solvents can be water-soluble but also water-insoluble, provided they are stable (under the reaction conditions) to hydrogen peroxide and exhibit a good dissolving power to-wards the oximes, in particular towards cyclohexanone-oxime.
In the case of many oximes it is possible to operate also in an aqueous medium, but cyclohexanone-oxime, owing to its - to -low water-solubility, would tend to depos it onto the cata-lyst, thereby inhibiting the catalyst activity when the saturation limit is reached. Owing to these reasons it is advantageous to use organic solvents for the purpose of obtaining a high specific productivity of the catalyst and of the reactor. Suitable solvents are, for example, tertia-ry alcohols, which are stable to hydrogen peroxide, in part-icular t-butyl alcohol, mixable in any ratios with water, or cyclohexanol, or aromatic compounds such as benzene, to-luene, xylenes, chlorobenzene, mixtures thereof, etc. If water-immiscible solvents are utilized, the presence of the following three phases is observed: an aqueous phase (water is produced by the reaction), an organic phase (which maint-ains in dissolution most of the produced oxime) and a solid phase, which is suspended between the two liquid phases and is composed of the catalytic system. Ail the examples given later on herein concern the use of t-butyl alcohol as a sol-vent; however, that does not exclude the possibility of using other solvents which are stable to hydrogen peroxide (either water-soluble or water-insoluble); particularly ad-vantageous results are obtained, for example, by substitut-ing toluene for t-butanol. Due to the low water solubility of cyclohexanone-oxime, it is advisable to limit the water concentration to the one which forms (during the reaction) and to the one which probably must be recycled with the sol-r 11 .. 2fl~9~fl6~
vent; t-butyl alcohol, for example, which is separated and recycled on conclusion of the reaction, has the composition of the aqueous azeotrope (about 12% by weight of water).
The oxime concentration in each step is gradually rising and its maximum value, when it is operated in an organic solvent, ranges from 10 to 30%, preferably from 20 to 25%
by weight. Although it is economically profitable to oper-ate with an oxime concentration at the maximum values, that is not advisable, as when this concentration exceeds cert-ains values, there is an interference with the consecutive oxime reactions, which lead to the formation of by-prod-ucts, which very badly affect its quality. The ratio bet-ween solvent and carbonyl compound generally ranges from 2.5 to 10 by weight.
F) OPERATIVE DETAILS
The new process according to the invention and the recove-ry of the oxime from the solution leaving the last step (exhaustion step), in which the residual carbonyl compound concentration is reduced to a value lower than 200 ppm and even lower than 100 ppm, can be carried out according to the schemes shown in figure 1 and in figure 2, which are given for merely illustrative purposes, without limiting, however, the scope of the invention.
According to figure l, cyclohexanone (1), hydrogen peroxide (2), ammonia (3) and a t-butanol make-up (not shown in the figure) enter a primary reactor R1, equipped with a stirrer, a filtering element (not shown in the figure) and a vent device (4), where the ammoximation reaction is brought to very high values (up to above 95% of con-version). The reaction mixture (5) flows then into a se-cond step reactor R~ (exhausiton reactor), which too is equipped with a vent (6) and is fed with an excess of hy-drogen peroxide (7). The final effluent (8) (practically free from residual ketone and containing t-butanol, cyclo-hexanone-oxime and ammonia) is sent to a distillation co-lumn C1. From the column top, the ammonia and all the sol-vent (t-butanol in the form of an azeotrope containing 12% by weight of water) are recovered; the ammonia and azeotrope mixture (9) is recycled to the 1st step (prima-ry step). From the bottom of the column, a liquid (10) consisting of water and of cyclohexanone-oxime is recover-ed and is then subjected to extraction in an apparatus E
fed with toluene (11). All the oxime passes to the toluene phase, and from a subsequent separator, not shown in the figure, the toluene phase (12) is withdrawn, which is then sent to a column C2 for the solvent distillation and the oxime dehydration. From said separator (not shown), a water phase (13), containing most of the water-soluble foreign matters, is discharged. From the top of column C2 toluene is recovered in the form of an azeotropic mixture with the ' reaction water; after a demixing (not shown in the figure), toluene (i1) is recycled to 'the extraction section. The an-hydrous oxime (14), which leaves column C2 from the bottom, is sent to the Beckmann rearrangement for the production of caprolactam.
Figure 2 illustrates, by means of analogous symbols, the case in which, instead of 1 primary step, there are 2 pri-may steps; figure 3 and figure 4 concern the results of a few tests and will be discussed in the examples.
G) APPARATUSES
The invention can be advantageously carried into effect in reactors arranged in series and stirred in order to main-tain in suspension the catalyst insoluble in the liquid medium. The most suitable reactor is the one which is known as CSTR (Continuous Stirred Tank Reactor). This type of re-actor guarantees an effective dispersion of the catalyst system and at the same time, on the basis of a proper regul-ation of.the residence times, the desired conversion of the carbonyl compound, the residual concentration of which shall not exceed certain optimum values, beyond which the already cited formation of undesired by-products takes place (which adversely affect the oxime quality and render the oxime not acceptable for the conversion to caprolactam). Ammonia, hy-drogen peroxide and carbonyl compound (in particular cyclo-gexanone) are continuously fed to each reactor of the prima-_ 14 _ ~059~Q~
ry steps, and the temperature is maintained around the desir-ed value (by means of cooling, since the reaction is exotherm-ic). The reaction heat can be removed indirectly, through a heat exchanger arranged inside the reactor, or by causing the reaction liquid to circulate in a refrigerated circuit out-side the reactor (loop reactor). Each reactor shall be equip-ped with a vent for removing little amounts of gases (N2, O2, N20), which form as reaction by-products by direct ammonia oxidation. On said vent it is advisable to mount a scrubber for the little amounts of solvent which could be probably en-trained by the gaseous compounds. In each reactor it is ne-cessary to install also a filtering system and a purge for the exhausted catalyst. The filtering system, which is ar-ranged inside the reactor or on a circuit outside the reactor, permits to separate the liquid phase from the catalyst, which remains in the reactor, while the filtered liquid is sent to another reactor or to a distillation column (in the case of the last step) for the oxime recovery. It is preferable to install - coupled to the filtering system - a device for the discontinuous purging of exhausted catalyst, which shall be replaced by a fresh catalyst make-up in order to maintain un-altered the catalytic activity in each step. The cyclo-hexanone feeding, however, is not provided in the last step, since the specific purpose of this step is a complete con-version of the carbonyl compound. Therefore, the reaction 1i-quid flows directly to the oxime recovery section without undergoing further treatments. The (crystalline or amor-phous) catalyst particle size, of the order of tens of microns, permits, on one side, an easy dispersion in the reaction medium and, on the other side, an easy separation, by means of the usual filtering systems, from the reaction medium. The following examples are given for merely il-lustrative purposes and are not to be considered as a li-mitation of the scope of the invention.
EXAMPLE 1 (COMPARATIVE) - PRIMARY AM MOXIMATION
To a 1 liter reactor, equipped with a stirrer and with continuous feeding and discharge systems, there were con-tinuously fed:
- cyclohexanone = 70.6 g/h;
- t-butyl alcohol (TBA) (containing about 12% by weight of H20) - 232.5 g/h;
- hydrogen peroxide (at 49.7% by weight) - 54.2 g/h (N202 . ketnne feeding molar ratio = 1.10);
gaseous ammonia = an amount sufficient to maintain a constant ---concentration (about 2% by weight calculated on the liquid medium).
The level of the liquid was maintained constant by regulating an average residence time of 72 minutes (+/- 17, and the catalyst concentration was maintained constant around 2% by weight (calculated on the liquid medium). The catalyst consist-- 16 - ~~~9~~~
ed of spheroidal titanium silicalite (suspended in the liquid) having an average particle size of about 20 microns. The re-action temperature was maintained constant at 85°C (+/- 1) by means of thermostatic fluid circulating in the reactor jacket;
the operating pressure was of 2.3 bar. The resulting product was continuously withdrawn through a stainless steel element equipped with a porous baffle and arranged inside the reactor (dimension of the pores = 5 microns), in order to prevent the passage of the catalyst; under regular operating conditions, the product leaving the reactor had the following composition:
- cyclohexanone-oxime 21.0 % by weight - cyclohexanone 0.3 % " "
- water 22.0 % " "
- ammoni a 2.0 % ~~ ..
- solvent (T6A) the balance to 100%
what corresponded to the following results:
- cyclohexanone conversion 98.3 %
- cyclohexanone selectivity to oxime 99.6 %
- H202 conversion 100.0 %
H202 selectivity to oxime 89.i %
Data and results are reported in figure 3 and in Table 1, where also the gaseous by-products (N2 + N20) and the color (APHA) are indicated. Said APHA coloring can be determined, as is known, according to ASTM-D-1209/69 standards.

- l~ _ ~~5~~0~-' EXAMPLE 1/BIS (COMPARATIVE) Example 1 was repeated, increasing the (H202:ketone) feed molar ratio up to a value of i.15. The results, which are reported in Table 1 and (graphically) in figure 3, prove that an increase in the hydrogen peroxide amount involves a little reduction of the product color, while it causes a not allowable increase of the (gaseous) by-products deriving from ammonia oxidation, in particular of NZ and N20. Conversely, if said ratio is reduced below 1.10, lower amounts of gaseous by-prod-ucts, but also much higher (APHA) coloring values are obtained, as is shown in figure 3, where the by-products amount is ex-pressed as N 1./mole (normal litres per mole of oxime present in the reaction system).
EXAMPLE 2 (COMPARATIVE) - COMPLETION OF AMMOXIMATION WITH
HYDROXYLAMINE SULPHATE; INTEGRATED AMMOXIMATION
Example 1 was repeated and the effluent from the re-actor was directly fed to an azeotropic distillation column, from the top of which the solvent (t-butanol containing about 12% by weight of H20) was recovered; from the bottom there was recovered a mixture having the following composition:
- cyclohexanone-oxime 58.6 % by weight - cyclohexanone 0.85% "
- water X0.5 % "
Said tail mixture was continuously fed to a second (stirred) CSTR reactor, to which there was fed also an aqueous ~o~o~os solution of hydroxylamine sulphate of formula (NH30H)2504, here-inafter referred to as NYXAS, at a concentration of 10% by weight.
The hydroxylamine amount was such as to maintain a NH2UH/cyclo-hexanone molar ratio equal to 2. The pH was constantly maintain-ed at about 4 (+/- 0.1) by adding an ammonia aqueous solution (at 15% by weight). The temperature was maintained at 90°C (+/
- 1). The average residence time was of 15 minutes (+/- 1), so obtaining a cyclohexanone-oxime having a maximum concentration of residual cyclohexanone lower than 100 ppm. The effluent from the reactor was sent into a phase separator, where (after a re-sidence time sufficient to obtain a sharp phase separation) a cyclohexanone-oxime molten phase, containing 6.5% by weight of water, and a saline aqueous phase were obtained. Said oxime was then dehydrated and sent to the Backmann rearrangement; data and results are reported in Table 1.
rvn~m c ~
To a first step (primary ammoximation step) there were fed, under the operative conditions of example 1 .
- cyclohexanone 70.6 g/h - TBA (12% of 1120) 232.5 g/h - hydrogen peroxide (49.7%) 54.2 g/h ( H202:ketone feed molar ratio = 1.10) - ammonia . an amount sufficient to maintain a steady concen-tration (about Z% by weight in the liquid medium).
The effluent from this first step, equal to 300 g/h, 2059~~6 _ 1g _ having the following composition:
- cyclohexanone-oxime 21.0 % by weight - cyclohexanone 0.30% " "
- water 22.0 % .. ., - ammonia 2.0 % "
was fed to a second reactor (exhaustion reactor) similar to the first and maintained at a steady operation pressure of 1.8 bar and at a temperature of 85°C, to which reactor also an aqueous solution of hydrogen peroxide at 50% by weight was fed. The sol-ution amount was equal to 1.6 g/hour, corresponding to a H202/
residual ketone molar ratio (exhaustion ratio) equal to 2.02.
Also in this second step (exhaustion step) it was operated with a catalyst in suspension (titanium silicalite) in an amount equal to about 2% by weight of the solution contained in the reactor.
The average residence time was of 30 minutes (+/- 1). The re-action product leaving the second step (through the filtering element) exhibited the following composition:
- cyclohexanone-oxime 21.3 % by weight - cyclohexanone less than 100 ppm - water 23.2 % by weight - ammonia 1,7 % ~~ ..
- solvent the balance to 100%
Considering the globally fed amounts of reagents, the H202/cyclohexanone total molar ratio was equal to 1.i3. The cyclohexanone conversion was equal to 99.95%, the selectivity of ~0~9~0~

cyclohexanone to cyclohexanone-oxime was higher than 99%. The hydrogen peroxide conversion was practically quantitative and the selectivity of hydrogen peroxide to oxime was of 87.4%.
After separation of the solvent by distillation and after de-hydration of the resulting oxime, there was obtained, by Beckmann rearrangement, a caprolactam corrisponding (after purification) to the quality characteristics required by the market (optical density, at 290 nanometers, lower than 0.05;
permanganate number higher than 20,000 seconds; volatile bases below 0.5 milliequivalents/kg). It is evident that by operat-ing according to the invention it is possible to obtain excel-lent results without having to utilize a new reagent alien to the ammoximation reaction (for example hydroxylamine sulphate).
The very little amount of gaseous by-products (0.41 N i./mole) and the final color (about 180 APHA) are reported as a diagram in figure 4~, which permits to immediately realize the technical importance of the invention. The crossing point of the two curves in figure 3 practically indicates an optimum conversion level, beyond which it is advisable to pass to the exhaustion step (with very high H202:ketone ratios),It was virtually im-possible to foresee the range corresponding to the best re-suits (95-99%).

The reactor of example 1 was fed with - cyclohexanone 35.3 g/h - 21 _ - TBA (12% H20) 232.5 g/h - hydrogen peroxide (50%) 25.5 g/h (H202:ketone feed ratio = 1.04) gaseous ammonia: an amount sufficient to maintain constant its concentration (about 2% by weight calculated on the liquid medium).
The liquid level in the reactor was maintained cons-taut and the average residence time was of 60 minutes (+/- 1).
The catalyst (titanium silicalite) concentration in the re-actor was maintained constant (about 2 by weight calculated on the reaction medium). Also the reaction temperature was maintained constant at 85°C (+/- 1) by means of a thermostat-is fluid circulating in the reactor jacket; the pressure was equal to 2.8 bar. The cyclohexanone conversion was of 97.8%.
The composition of the effluent from the first step was as follows:
- cyclohexanone-oxime 13.0 % by weight cyciohexanone 0,26 % « ..
- ammonia 2.0 % "
- water 18.2 % "
The product flowing from this first step passed to a second reactor, identical with the preceding one, and simulta-neously there were fed:
- cyclahexanone 35.3 g/h - hydrogen peroxide (at 50% b.wg.) 26.0 g/.h - 2z -(H202:ketone feed ratio = 1.06) - gaseous ammonia: an amount sufficient to maintain constant its concentration (about 2% by weight).
The operative conditions of the second step were .
- temperature 85°C (+/- 1) - pressure 2.3 bar catalyst in suspension 2 % by weight - average residence time 60 minutes (+/- 1).
The effluent from the second step, equal to 373 g/h, had the following composition:
- cyclohexanone-oxime 21.4 % by weight - cyclohexanone 0.30% " "
- ammonia 2.05% " "
- water 22.1 % " "
Said effluent (from the second step) was fed to a third reactor (exhaustion reactor), similar to the reactors of the first and second steps, operating under the following conditions:
- temperature 85°C (+/- 1) pressure 1.8 bar - concentration of the cata-lyst in suspension about 2% by weight - averge residence time 30 minutes.
Said third reactor was fed with 8 g/h of hydrogen peroxide (at 10% by weight), what was corresponding to an ex-haustion ratio equal to 2.06.

2~~9~~fi The product flowing out from the third step, equal to 380 g/h, had the following composition:
- cyclohexanone-oxime 21.3 % by weight - residual cyclohexanone less than 100 ppm The hydrogen peroxide/cyclohexanone total molar ratio was equal to 1.08.
The cyclohexanone conversion was higher than 99.9%.
The cyclohexanone selectivity to cyclohexanone-axime was of 99.4%.
The hydrogen peroxide conversion was quantitative.
The hydrogen peroxide selectivity to oxime was of 91.7%.

Under the operative conditions of example l, a 2-li-ter reactor was fed with:
- cyclohexanone 133.75 g/h - t-butyl alcohol (12% H20) 491.2 g/h - hydrogen peroxide (50% by wg.) 90 g/h (hydrogen peroxide/ketone feed molar ratio = 0.97) - gaseous ammonia: an amount sufficient to maintain constant the concentration (about 2% by weight on the liquid medium).
The effluent, equal to 752 g/h, having the following composition:
- cyclohexanone-oxime 19.4 % by weight - cyclohexanone 0.9 % " "

20~~~0~

- water 20.1 % by weight - ammoni a 2.0 ~ .. ..

was fed to an exhaustion reactor of 1-liter volume, operat-ing under the following conditions:
- temperature 85°C
- pressure 1.8 bar residence time 32 minutes - suspended catalyst (calcul-ated on the reaction medium) 2 % by weight.
To the exhaustion reactor there were fed also 44 g/h of hydrogen peroxide at 10% by weight (hydrogen peroxide/
ketone molar ratio = 1.87). The product leaving the reactor had the following composition:
- cyclohexanone-oxime 19.3 % by weight - cyclohexanone 200 ppm - water ~n r, ~ ~". ...... _~...
- ammonia 1.5 % " "
The hydrogen peroxidelcyclohexanone total molar ratio was 1.6.
The cyclohexanone conversion was higher than 99.9%
The ketone selectivity to oxime was equal to 99.3%.
The hydrogen peroxide conversion was quantitative.
The hydrogen peroxide selectivity to oxime was equal to 93%.

~05J~06 i ~ o 0 o a ~ ~

r. a ..- .- C

U

O

N 1c M p Z * M tn 'O

t * . 4A

N is O O C s' Z O

+-~

X

O _ ~ ~ ~ -~

N of M p . C

C i .O O
.-.

O N CO Ol j dE

+> > C11 01 C
~

V
N C

Y O

V Q7 i a - C

C O al O ~ ~ +~ s r- C p .D OJ +~
i~ ~G
O-2f O M . C C

C +, C vi W7 r '- ~ O V7 C
O

4. ~C 4J np 'O
O

r r~

-- - _~ -~ C

33 r O r9 O i~

lJ r C

C V

-! ~ O U

r O O N N

CO ~ ~ C C ~ 'O ~J

>> r Q = ~ X r +,O X

C L O

-~ - --- -- N 4J

> d 4-O

OO G7 ~ ~ O C

7 C p ~ ..-. VfCJ

ch 'fl CO a~ --'D O *

~ ~ N O O
r ~

.CL
O C O

N a~b Y

.

4-s ro _~ ~ -- ~- O i C C~1 C ~

O N S

r ~ v r 3Z 'C r +~6-sd X N N

O C ~ .~ N

r N

~ r O L

> .1.1r i->

- .~ r In.p.,1r ~ ~

~i, r (O r O i. .-C. r ~O L L r0 d 4J~O E

O +~ E .,.~r- s.

+.r o m o 4-0 0 ro, ~

L * ~ a U d ~G * .. , ~

r r-N ~ r ...

O

a - - *

-~c -ki~ *

* .r ..r v .n i~

..r N
-x ' w -- v-- -- -- -.k O O

O * w II II w w -jcIIr r r +~ +~

O ~o !- L

. - ._ - . - - --- N U
- -C C U

to n~ E

v .G .~ >Z E .,.a .--_ a..~ .a-~ rx C a. a r N O p, M G1. O ~p C 'I'~ 1 N II ~
en t0 th O N O p O O

Lt. .Y N N r-- O T9 +..r O

r r N p r-- - _ ..- .~ .-. s l!--N C O

O .r lty t'~ L

+~ +-~ N O ~O ~ T

N . VI O II II Ii r .. ~

O ~ O * II N T~ .C

N _C * N U t6 O X r N N -!~ O

x -- ~v 3 r c ~p O Q.1 N

r '- M N M 91 C C
s t~

~ II ~ II O O
- O O O O

O O C N N

r O Q1 ~ x .

IA .1-1 ~
O O

N- d SQ
s r ~ ~G - ! ..v. C C
+~ V- -_--- ~ b ro N O ~

O O

4- F~ O M O O

W N N C ~ ~ O ~ ~' P
M M

E ~ P 1' r- N N N M r X W - ... ..- .-- ~ N
N N "'.

O i.~ 3~ 'C7 'b ~6-a -~-- -_ r,~ - x v o O

O O L L

O '- v v yJ 'I'~ O_ O.

N '. ~' to I
~3 ,C L O O ~ C

r w ~ ~ II O N
II

.. ~ w 11 r- v O ~ O

'pp ~ ~ O O
P

* L L

N 4- -k 'C3 'C9 _ -. ~ ._, - .~ .C
_ -.. a .,O ~ C C

N '~ O QJ

'- N ~ t>7 !L 4D
O O

'fl O w w ~ ~y 3 x V cv L ~ ~.r r,~ C~ +~ +~
O t0 '~
CO

Lt- 'C ~ t~ N N 01 N Ul L ~' \ ~7 .8 S d.

O O

_-_~ ~r s-+, tC7 ra rtf O ~p M M ~ L L

C ~ II

b O II M L L N
IC ~ O I,n t17 ~ ~ C r. M M M II ~ ~ D.
O

,.._ r r O
. ~

\ O O +~
+.~

~ N ~ V1 0.~ ~0. 0. ~
0. 0. 0.
+,a 0.' +~
+~

LJL tyJ w IrJ ,y k w uJ N
cn V.!
v1 N !-f0 N I-r f ' F- I-- f *

3 ~

4J N E 'N cn ~ N * .
N N rO ,.
rtf r .P .C -. ~-' ~ 'o x a +~L .+wv '~
-v rs c x x w a v1 c r N
c L v ~

Q .a rN M
N ~
i - _,. -_ - _._ w M

Claims (22)

1. A multistep process for the liquid phase ammoximation of a carbonyl compound with H2O2 and NH3, at 60° - 100°C, at 1.5 - 5 bar and in the presence of a catalyst based on silicon, titanium and oxygen, characterized in that a) in one or more primary steps, the H2O2:
carbonyl compound molar ratio ranges from 0.9 to 1.15 and the carbonyl compound conversion is carried out at least up to 95%;
b) in a last step (exhaustion step), the H2O2:
carbonyl compound molar ratio is at a higher level than in the primary steps.

2. The process of claim 1 in which, in the primary steps, the H2O2: carbonyl compound molar ratio ranges from 1.0 to 1.1.

3. The process of claim 1 or 2 in which, in the primary steps, the carbonyl compound conversion is carried out at 96 to 99%.

4. The process of any one of claims 1 to 3 in which, in the exhaustion step, the H2O2: carbonyl compound molar ratio ranges from 1.5 to 3Ø

5. The process of claim 4 in which, in the exhaustion step, the H2O2: carbonyl compound molar ratio ranges from 1.5 to 2.2.

6. The process of any one of claims 1 to 5, in which:
- the number of the primary steps is selected from 1 and 2 and the catalyst is titanium silicalite;
- the residual carbonyl compound concentration in the effluent from the primary steps is equal to ar lower than to by weight, and the ammonia concentration in the liquid reaction medium in all the step; ranges from 1.0 to 2.5% by weight;
- the specific productivity in the primary steps ranges form 6 to 12 parts by weight of oxime per part of catalyst and per hour.

7. The process of claim 6, in which the residual carbonyl compound concentration in the effluent from the primary steps is equal to or lower than 0.5 % by weight.

8. The process of claim 6 or 7 in which the ammonia concentration in the liquid reaction medium in all the steps ranges from 1.5 to 2.0% in weight.

9. The process of any one of claims 1 to 8, in which the temperature ranges from 70° to 90°C, the pressure ranges from 1.8 to 4 bar and the catalyst concentration ranges from 1% to 15%, by weight.

10. The process of claim 9 in which the catalyst concentration ranges from 1 to 6% by weight.

11. The process of any one of claims 1 to 10, in which the specific productivity in the exhaustion step ranges from 0.1 to 5 parts by weight of oxime per parts of catalyst and per hour.

12. The process of claim 11 in which the specific productivity in the exhaustion step ranges from 0.3 to 0.6 parts by weight of oxime per parts of catalyst and per hour.

13. The process of any one of claims 1 to 12, conducted in the presence of an organic solvent, the ratio between said solvent and said carbonyl compound ranging from 2.5 to 10 by weight.

14. The process of claim 13 in which the organic solvent is selected from t-butanol and toluene.

15. The process of any one of claims 1 to 14, in which the maximum oxime concentration in the liquid reaction medium ranges from 10 to 30% by weight.

16. The process of claim 15 in which the maximum oxime concentration in the liquid reaction medium ranges from 20 to 25% by weighed.

17. The process of any one of claims 1 to 16, in which the catalyst particles suspended in the reaction liquid have an average of 1 to 100 microns.

18. The process of claim 17 in which the catalyst particles suspended in the reaction liquid have an average size of 5 to 50 microns.

19. The process of any one of claims 1 to 18, wherein the carbonyl compound is selected from cyclohexanone, acetone, methyl ethyl ketone, acetophenone, cyclododecanone and enantic aldehyde.

20. The process of any one of claims 1 to 19, wherein oxime is recovered from the reaction liquid after the exhaustion step by means of an azeotropic distillation, followed by an extraction with organic solvents.

21. The process of claim 20 in which the organic solvent used for the extraction is toluene.

22. The process of any one of claims 1 to 21, wherein the reactor for each step is of the CSTR type and is equipped with a porous filtering element, the pores of which have an average size lower than the average size of catalyst particles.