WO2013160485A1 - Process for continuous preparation of lactide directly from concentrated lactic acid - Google Patents

Abstract

The subject-matter of the present invention relates to a process for the continuous, heterogeneously catalysed, direct preparation of L-lactide from concentrated lactic acid. The aim is to circumvent the process steps of the known prepolymerization and subsequent depolymerization, and to reduce them to one synthesis step. Furthermore, the by-products and the racemization are minimized, and so high-purity L-lactide is prepared.

Description

 Process for the continuous production of lactide directly from concentrated lactic acid

The subject of the present invention relates to a process for the continuous, heterogeneously catalyzed direct production of L-lactide from concentrated lactic acid. The aim is to circumvent the process steps of the known prepolymerization and the subsequent depolymerization and to reduce to a synthesis step. Furthermore, by-products and racemization are minimized to produce high purity L-lactide.

To illustrate the present invention, the following nomenclature will be used hereinafter:

LA: aqueous acid

 LD: lactide

 L1A: lactic acid monomer

L2A: lactoyllactic acid, linear lactic acid dimer L3A: lactoyl lactoyl lactic acid, linear lactic acid trimer

 LnA: linear n-lactic acid oligomer

The rising demand for crude oil is driving research into alternatives to conventional plastics. One of the most important bioplastics based on renewable raw materials is polylactic acid (PLA). In order to be competitive with petrochemicals, conventional plastics and other bioplastics, it is necessary to make the process more economical. The interest in resource savings through the reduction of the process steps, the saving of energy and the simplification of the treatment is therefore given.

For the production of high molecular weight polylactic acid, especially PLLA (poly (L-lactide acid)), the intermediate L-lactide is unavoidable according to current state of the art or science. So that the desired mechanical properties are achieved and the already low melting point is not further reduced, the content of meso-lactide must be kept low. In addition, the achievable molecular weight also depends strongly on the purity of the lactide. Each hydroxyl group of one

Monomers provide a start center for the polymerization and reduce the chain length of the polymer. For this reason, the concentration of the monomer must be kept low. Other contaminants, such as water and the linear dimer, are also inhibitors of polymerization.

Lactide is preferably derived from lactic acid for over one hundred years. The dominant in the literature manufacturing process includes two steps, namely the pre- and depolymerization.

The starting material is, according to the current state of the art, obtained by fermentation, optically pure L-lactic acid. The acid is fed in concentrated form to the process. There is a balance of monomer, oligomers and water.

In the first process step, the lactic acid is polycondensed. In this equilibrium reaction, water is continuously withdrawn from the system. This can be realized in different ways; for example, by the usual vacuum distillation, or by means of a heteroazeotropic entraining agent or an osmotic membrane. The average molecular masses of the low molecular weight polylactic acid are in the range of 600 to 5000 and can be obtained by non-catalytic means. Striking are the high reaction times of up to several hours until the desired degree of polymerization is achieved. The subsequent thermal depolymerization is carried out under harsher reaction conditions. The already existing vacuum is reduced to a minimum and at the same time the temperature is increased. The low molecular weight polylactic acid cyclized by a

Transesterification reaction to lactide. The depolymerization is carried out predominantly with the aid of zinc or tin catalysts. These are toxic and, moreover, can not be recovered by regeneration. Furthermore, the separation from the product is complicated and cost-intensive. In general, the long residence times of several hours are disadvantageous in the manufacturing process. Furthermore, two process steps are required for the synthesis of lactic acid from aqueous lactic acid.

Polylactic acid has been used for some time in surgery and wound care. The advantage of biological decomposition, for example, makes one

Second operation superfluous. Due to the comparable properties to conventional plastics, polylactic acid has developed as a renewable alternative to crude oil-based plastics. Due to the adjustable multi-layered properties there is a wide range of applications. For example, PLA is already considered non-toxic today

Food packaging material used. Further, disposable PET cups and PET bottles are substituted by PLA based products. Biodegradable PLA films are also used in the agricultural sector. The continuously increasing demand is driving research into energy-saving, resource-saving and therefore more cost-effective processes. For this reason, some new lactide syntheses have been published in the literature.

The patent DE 69103159T2 describes the production of lactide in two synthesis steps. A commercially available 88% lactic acid is removed by vacuum distillation, the free water. The balance is thereby shifted to the oligomers and lactide is formed. Subsequently, the complex product spectrum, consisting of monomer (UA), dimer (L ₂ A), lactide (LD), oligomers (L _n A, n> 2) and water, separated by a complex vacuum distillation.

Furthermore, in this method, the pre-treatment of the lactic acid is disadvantageous. The degree of polymerization is set between 1.59 and 2.63 and is thus significantly higher than the commercial 88% lactic acid. Furthermore, the process is operated non-catalytic. Although the influence of tin (II) octoate is investigated in Examples 12 to 14, the use of catalysts is considered superfluous. The yields of a trial are unsatisfactory, with no returns. The patent DE 102011088515A1 describes a two-stage homogeneous / heterogeneously catalyzed process for the preparation of D-lactide from aqueous lactic acid. In Example 1, aqueous D-lactic acid (15.W .-% water) at a temperature of 160 ° C and a pressure of 10 Torr (ca.13 mbar) prepolymerized to a molecular weight of 600 g / mol. In Example 2, the produced D-polylactic acid becomes 0.1% by weight more homogeneous

 Zinnoctoatkatalysator mixed and reacted at 200 ° C to a gas stream. The mixture is passed over a heterogeneous alumina catalyst and depolymerized to D-lactide. The patent describes the catalytic effect. The highest yields are 72% and are achieved with 0.2 wt .-% Zinnoctoatkatalysator and the Aluminiumoxidkatalysatorlage.

A differentiation between conversion and selectivity is not made.

In addition to the low degree of polymerization, the patents described above (DE 69103159T2 and DE 102011088515A1) are similar to the conventional process of pre- and depolymerization. From a direct synthesis can not be spoken for this reason.

The patent DE 69105144T2 also describes the synthesis of lactide. The aqueous lactic acid is heated by a nitrogen carrier gas stream and brought into the gas phase, this is at temperatures of 200 "C to 220 ° C rea- lisiert. The process is continuous and at atmospheric pressure. As possible catalysts alumina, silica, alumina-silica combinations, ion exchange resins and zeolites are listed. However, besides the Amberlyst 15, only the amorphous oxides are tested; zeolites are not further specified. With the help of the alumina and silica catalysts, the maximum yields of about 30% to 34% are achieved. The used ion exchange resin catalyzes the formation of carbon monoxide, the selectivity with respect to lactide was estimated at 0%. Overall, the yields achieved are low and not competitive with the existing process.

Experiments with zeolites were not carried out. A specification of the different structures and acid strengths of zeolites and their influence on the direct synthesis is missing accordingly.

In Example 1, in addition to the 85% lactic acid in addition water for the hydrolysis of the oligomers is added. However, the additional amount of water reduces, compared to the pure 85% lactic acid (Example 2), the yield.

Furthermore, it can not be spoken of a gas phase reaction at the set temperature range. At ambient pressure and a

Evaporator temperature of 150 ° C to 225 ° C are only water and the lactic acid monomer in front gas, but not the other oligomers occurring.

The patent DE 69320467T2 is an extension of the two patents already described above. It includes u.a. two methods of commercial lactic acid to deprive the water. On the one hand water is removed by means of a vacuum distillation, on the other hand an entrainer is used. The separation produces lactide, but both methods are very time-consuming due to the required high reaction times of several hours. In addition, a reaction mixture of lactide, monomer, oligomers and water, which is very expensive to separate.

Furthermore, in this patent, the reaction in the gas phase at Nor- maldruck described in a temperature range of 200 ° C to 220 ° C. In addition to the examples from the patent DE 69105144T2, experiments are additionally carried out with a modified mixture of lactic acid. The mixture is depleted in water and oligomers and mainly contains monomer. The yields do not reach the values of the equivalent experiments with a commercial 85% lactic acid (Example 42).

The patent WO 92/00292 describes a continuous production of lactide. The educt mixture is mixed with an inert carrier gas and mixed at a temperature of 216 ° C in the gas phase. Thereafter, it flows through a reactor which is filled with catalyst.

The disadvantages of the described patents are inter alia in the reaction procedure at ambient pressure. No example was made under negative pressure conditions. As a result of the reaction procedure at reduced pressure, the reaction temperature can be lowered due to the lower boiling temperatures. Mild process conditions such as negative pressure and low temperatures reduce the racemization of both the aqueous and the lactic acid.

In addition, the achieved yields of up to 34% are unsatisfactory and not competitive with conventional pre- and

Depolymerization. Furthermore, only oxides and ion exchange resins are used as catalysts. Examples with zeolites are missing.

The object of the present invention is to minimize the synthesis steps in the production of L-lactide from fermentatively produced, optically pure L-lactic acid. According to the invention, the synthesis is realized by the direct, heterogeneously catalyzed conversion of concentrated aqueous lactic acid to lactide. Compared to the usual methods, the lactic acid is not subjected to an equilibrium reaction, ie, a water separation. An equilibrium shift to higher oligomers or a formation of low molecular weight polylactic acid is avoided. In comparison to the lactide syntheses described so far, the yields are significantly increased by the combination of heterogeneous catalysts with simultaneous vacuum operation in the selected plant construction. As especially effective Zeolites have been found. These achieve a high yield through a well-defined structure and acidity. At the same time, racemization is minimized by a mild reaction procedure in terms of temperature and pressure.

This object is achieved by the method having the features of claim 1. According to the invention, a process is provided for the continuous preparation of optically pure lactide, directly from concentrated lactic acid, which is characterized in that lactic acid is heated under reduced pressure, then passes through a reaction zone filled with a heterogeneous catalyst and then the Product spectrum is cooled and separated.

The advantage of the process according to the invention is the direct production of lactide from concentrated lactic acid. A complex pre-treatment of the aqueous lactic acid is eliminated, thus the fermentatively produced, concentrated lactic acid can be fed directly to the process. Furthermore, the simple system construction is advantageous, this consists of a heating zone, a

Reaction zone, consisting of a simple tubular reactor, and a cooling zone. Continuous operation reduces lactic acid residence times to a few seconds in the presence of lactide. The significantly higher space-time yield of the process according to the invention reduces the energy consumption and reduces the costs incurred. Due to the low residence time and the relatively mild reaction conditions, in particular due to the reduced pressure, the racemization is kept low and is below 9% across all experiments. Another advantage of the invention is the achieved product spectrum consisting of lactide, monomer and water. The direct Lactidsynthese proceeds highly selectively and can be catalytically accelerated. The formation of low molecular weight polylactic acid was not observed, furthermore, the conversion of oligomers is optimal and can be estimated at one hundred percent. In addition, traces of by-products such as acrylic acid, acetone, acetaldehyde, formaldehyde and carbon dioxide have been detected. By means of a simple, commercially available catalyst system high yields can be realized. Furthermore, that is Lactide produced in the presence of water sufficiently stable. Moreover, the separation of the reaction mixture due to the product spectrum is facilitated. According to the invention, the mixture consisting of lactide, lactic acid monomer and water is separated directly. Lactide is obtained as a white solid and contains only monomers. The unreacted lactic acid monomer can be separated from the lactide in a further step and subsequently recycled. To avoid

Hydrolysis effects, the free and the resulting water are condensed separately, consequently, the proportion of water in the solid can be reduced to less than 1%. These and other advantages of the method according to the invention will become apparent to those skilled in the art from the description herein.

 Preferably, the lactic acid is heated to a temperature between 100 ° C and 300 ° C.

A further preferred embodiment provides that there is a negative pressure in the system, preferably from 10 mbar to 200 mbar.

The concentrated lactic acid is preferably a 50% to 100% lactic acid.

The heterogeneous catalyst is preferably a zeolite, an alumina or a silica catalyst. In particular, an L zeolite, a

Mordenite zeolite, a beta zeolite, a B-MFI zeolite, a HY zeolite, an H-ZSM-5 zeolite or a γ-alumina preferred. In this case, the catalyst may preferably have an average acid strength and a small number of acidic centers. The catalyst preferably has a high ratio of Lewis to Bronsted acidic sites. The catalyst preferably has an amorphous or crystalline structure.

A further preferred embodiment provides that the product spectrum, consisting of lactide, monomer and water, is separated directly behind the reaction zone.

Preferably, a separation of lactide and water is realized. It is further preferred that an inert carrier gas stream dilutes the lactic acid.

The conversion of the oligomers is preferably 100%, i. and the product spectrum consists only of lactide, monomer and water.

With reference to the following figure and the examples of the subject invention is to be described in more detail, without wishing to limit this to the specific embodiments shown here.

The direct production of L-lactide from concentrated lactic acid is realized in a continuous plant design. In Fig. 1, the principle of direct Lactidsynthese is shown schematically.

Initially, concentrated lactic acid and an inert carrier gas, namely nitrogen, are fed to the system (101). Alternatively, carbon monoxide, carbon dioxide, methane or noble gases could also be used as inert carriers. The volume flow was between 0 ml / min and 80 ml / min. Furthermore, the composition of the lactic acid can be varied. The range goes from strongly aqueous, monomer dominant 50% lactic acid to highly concentrated anhydrous lactic acid. Preferably, 80% to 100%, especially 90% to 100% lactic acid can be used. The educt stream is heated in the heating zone 102 by a heat exchanger. Due to the heat in this part of the plant, the educt mixture in the

Gas phase to be added, this depends not only on the temperature and pressure, but also on the mixture composition. The system can therefore be operated in the gas or liquid phase. When using a highly concentrated lactic acid, both a gas phase, enriched with water and lactic acid monomer, as well as a liquid phase, enriched with

Oligomers, formed in parallel. The temperature range of 100 ° C to 300 ° C has proven itself. Preferably, a temperature range of 150 ° C to 250 ° C is set, very particularly advantageous is the range of 180 ° C to 240 ° C. In addition, the vacuum mode has proven itself. The optimum pressure range at the fixed catalyst bed is therefore between 10 mbar and

Normal pressure. The pressure range from 10 mbar to 200 mbar has proven to be especially good. Following the heating zone, the heated lactic acid enters the reaction zone (103). This consists of a simple tubular reactor which includes the catalyst bed. The direct conversion of lactic acid to lactide can be carried out in a non-catalytic manner. However, the reaction is advantageously accelerated by means of a catalyst. Zeolites have proven to be particularly effective. In general, zeolites, especially acid, can be used as esterification catalysts. Furthermore, reactions can be controlled in a shape-selective manner by the precisely defined pore sizes of the zeolites. The formation of larger molecules, such as higher oligomers, can thus be prevented.

Generally, zeolites are crystalline aluminosilicates of microporous structure. Their lattices are made up of Si0 ₄ and Al0 ₄ tetrahedra. The crystal lattice forms a strictly regular pore and cavity system characteristic of any zeolite.

Zeolites with different structures such as the B-MFI, the HY, the beta, the L, the mordenite and the H-ZSM-5 allow the direct synthesis of lactic acid to lactide. The acid strengths of the zeolites have a clear influence on the conversion of the reaction and were adjusted by the Si / Al ratio, the modulus. An overview of the zeolites used is given in Table 1.

Table 1: Overview of the zeolites used

Int. Label manufacturer

Zeolite module manufacturer

 Designation

 Own manufacture

B-MFI - HV93 / 05 US Patent 4401637 ment

 Beta 150 HV92 / 62 Uetikon Zeocat PB-1

L 5.5 HV93 / 11 Bayer K zeolite L

Mordenite 15.1 HV94 / 39 Uetikon Zeocat Mordenite

 Zeolyst Int. (Eh.

 HY 80 HV94 / 47 Valfor CBV 780

 PQ)

 H-ZSM-5 25 HV09 / 24 Uetikon Zeocat PZ-2/25 H

H-ZSM-5 300 HV94 / 42 Uetikon Zeocat PZ-2/300 H

H-ZSM-5 1000 HV09 / 22 Uetikon Zeocat PZ-2/1000 H

Furthermore, the synthesis could be represented by an alumina catalyst (D10-10, BASF). The yields obtained are but in comparison to those of the zeolites low. The cost savings due to the simple catalyst, consisting of alumina, is offset by the additional cost of energy, so that a comparable yield as in the zeolites is achieved. For this reason, zeolites are preferably used for the inventive method.

Following the reaction, the mixture is cooled (104). In this case, part of the flow in the first cold trap (106) is condensed. Preferably, there should be no water condense, so that the hydrolysis of lactide to

Lactoyllactic acid, the linear dimer of lactic acid, is prevented. In this cold trap (106) predominantly crude lactide and lactic acid are separated. The water concentration is less than 1%. Depending on the lactide concentration, liquid or crystalline phases could be deposited in the first capacitor. The further, second cold trap (105) represents the total condenser. There condenses the entire remaining mixture, consisting mainly of water and traces of lactic acid monomer from. The temperatures of the two capacitors were set in the range of 0 ° C to 15 ° C. The high temperature gradient between the reactor and the capacitors allows the separation of the product spectrum and suppresses the hydrolysis.

The racemization of educt and product is kept low by the method according to the invention. The mild temperature setting and the underpressure mode are decisive factors for this. To verify the racemization, the samples were analyzed by GC analysis. L-, meso- and D-lactide could be separated from each other. The samples tested gave a racemization of less than 9%, with racemization being calculated using the following formula: racemization = 100% - GC _ARE A, L-LD / (GC _AREA , _L- LD + GC _ARE A, meso- LD + GC _ARE A, D-LD)

In addition to low racemization, desorption of the products from the catalyst in the vacuum is facilitated.

The analysis of the reaction mixtures was implemented by means of the H PLC. In addition, the water content of individual samples with the Karl Fischer Titration performed. The analysis of the exhaust gas flow was carried out with the aid of GC analysis and occasionally showed the formation of volatile by-products, which were continuously discharged through the dry carrier gas stream.

Examples

The following examples illustrate the invention, but without limiting it.

 example 1

The first example serves as a reference experiment and was therefore carried out uncatalyzed. The fixed catalyst bed was filled with glass beads (Roth, 1.25 mm - 1.65 mm). The results should serve as a comparison for the effectiveness of the catalytic experiments in the following examples.

The commercial reactor was pumped with a commercially available 90% lactic acid over four hours, which corresponds to a mass of 46.9 g. The volume flow of the inert gas amounted to 80 ml / min, as the medium nitrogen was used. The total pressure was reduced to 150 mbar on fixed catalyst bed. The temperature in both the evaporator and the reactor was 180 ° C. The residence time in the reactor was calculated at about 2 sec.

After termination of the experiment, a mass of 39.2 g in the first cold trap and a mass of 3.5 g in the total condenser were recovered. The mass balance was thus 91.1% closed. The analysis of the two reaction mixtures was carried out by HPLC and gave the following results:

Table: Example 1

mass

Pressure temperature reactant conversion selectivity

 Catalyst loss

 (mbar) CO (%) (%) (%) (g)

150 180 LA 90 14.1 6.4 4.2 4.2 The conversion of all lactic acid monomers theoretically contained in the starting mixture is 14.1%. Moreover, a part of the Milchsäu ^¬ re could be directly converted into lactide. The selectivity is 6.4%. These values already show in the noncatalytic experiment the effectiveness of the simple structure and the principal possibility of the direct conversion of lactic acid into lactide.

Moreover, the reaction is much more selective than the numerical value indicated. The chromatograms of the HPLC analysis show, besides the lactide peak and the characteristic peaks of the 90% lactic acid, no further by-products. The reaction is therefore optimal, 100% selective. In the calculation of conversion and selectivity, however, the mass balance influences the results. For this reason, deposits in the plant and losses due to the dry inert gas increase the conversion and reduce the lactide selectivity.

From the supplied lactic acid 0.2 g of lactide could be formed, in contrast, 4.2 g of mass could not be recovered in the capacitors and therefore not evaluated.

In general, even the non-catalytic test with commercially available lactic acid can represent the direct synthesis of concentrated lactic acid to lactide. The reactant stream used, consisting of monomer and oligomers, could be partially reacted. At the same time, no shift in the equilibrium to higher oligomers was observed.

Example 2

The following examples are intended to demonstrate the catalytic effect on the reaction.

In this example a simpler catalyst system is investigated. A commercially available γ-aluminum oxide (D 10-10, BASF) is used. The effectiveness of alumina catalysts has already been demonstrated in the above-described patents, but no tests were carried out under reduced pressure. The investigation of the catalytic effect on the direct conversion of lactide was carried out under the same parameters as in Example 1. The trial runtime was four hours. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The temperature was 180 ° C and the pressure on the fixed catalyst bed at 150 mbar. Further, 5 g of catalyst was charged into the reactor. The catalyst was initially in pellets. For use in the continuous laboratory reactor, the catalyst was placed on the

Particle size spectrum of 1.25 mm to 2.00 mm mortared. The residence time on the catalyst was about 2 sec. The weight hour space velocity (WHSV) was calculated to be 2.3 l / h.

At the end of the experiment, 43.1 g of the total of 49.9 g of lactic acid pumped in were recovered and evaluated. In this case, 39.9 g of reaction mass in the first and 3.2 g in the second capacitor were weighed. Analysis of the reaction masses gave the values shown in Table 2:

Table 2

Massen¬

Pressure Temperature educt WHSV conversion selectivity

 Catalyst loss

 (mbar) CC) (%) (%) (%) (g)

V-Al ₂ 0 ₃ 150 180 LA 90 2.3 23.1 49.5 6.8 The results show a moderate conversion of 23.1% and a selectivity of 49.5%. In addition to lactide and the lactic acid monomer only water and traces of by-products are measurable. The conversion of oligomers (L _n A, n> 2) is correspondingly one hundred percent. The conversion and the selectivity are influenced by the non-evaluable mass, which can be quantified to 6.8 g. One fifth, approx. 1.4 g, of the mass that can not be evaluated, is caused by the mass increase of the catalyst. The remaining proportion is in the range of the mass loss of Example 1 and can be explained by the problems already described, which comprises, for example, the continuous discharge of the water of reaction.

From the pumped aqueous 90% lactic acid, 3.1 g of lactide could be directly synthesized. The amorphous γ-alumina catalyst can in spite of its simple structure, represent the direct conversion and is more than ten times as effective (Y _D io-io = H, 4%) in terms of yield as the uncatalyzed experiment of the first example (YBiank = 0.9%). Example 3 - 4

In the following examples (Examples 3 - 8) the individual parameters are optimized to optimize the yield. The catalyst used is again the y-alumina from Example 2.

The selected parameters of the previous examples have the direct ones

Lactide production from commercial IVlacchsäure already shown. In these examples, the influence of the pressure on the reaction will be examined below. In addition to the pressure increase on the catalyst fixed bed

400 mbar or at atmospheric pressure, all other parameters remain constant compared to the above examples. The trial period was another four hours. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The temperature was 180 ° C. The weight hour space velocity (WHSV) was calculated at 2.3 1 / h.

During the four hour run, a total of 40.9 grams of the 90% lactic acid was pumped into the reactor at the 400 mbar pressure setting. In the first condenser, 32.4 g of reaction mass was collected. In addition, 2.8 grams were in the total condenser. The mass balance is accordingly 85.8%.

Much less satisfactory is the mass balance of the experiment at atmospheric pressure. This could be closed under the chosen conditions only 67.2%. All in all, only 23.1 g of the total educt mass of 34.3 g in total could be recovered and evaluated.

The evaluation of the respective reaction masses that the investigation of the pressure variation gave the results shown in Table 3: Table 3

Massen¬

Pressure Temperature educt WHSV conversion selectivity

 Catalyst loss

(mbar) CO (%) (h ¹ ) (%) (%) (g)

V-Al ₂ 0 ₃ 400 180 LA 90 2.3 40.0 19.0 5.7 y-AI ₂ 0 ₃ 1000 180 LA 90 2.3 48.1 9.6 11.2

Striking are the comparatively high conversions of the reactions, which are directly related to the high proportion of unexplainable mass. At the set reduced pressure of 400 mbar 5.7 g can not be recovered and therefore not evaluated. Under atmospheric pressure, the value is even 11.2 g.

Most of these non-evaluable masses can be explained by the weight gain of the catalyst. This can be estimated as 4.5 g (400 mbar) or 6.0 g (1000 mbar).

These observations can be explained on the one hand by the presumed mixed phases in the rector. When increasing the pressure at the fixed catalyst bed to 400 mbar or 1000 mbar and at the same time constant temperature setting of 180 ° C, can not from a total evaporation of the

Eduktgemisches, especially the oligomers are assumed. For this reason, under these conditions, there is a gaseous and liquid phase in the laboratory reactor. The mass in the reactor, as well as the residence time increase. After termination of the test are therefore more deposits in the system, which can not be evaluated. On the other hand, the desorption of the products is made difficult by the catalyst. A significant increase in weight, amounting to 4.5 g or 6.0 g, is the result.

In addition to the first two examples, the direct production of lactide from lactic acid could also be achieved at the selected pressure setting of 400 mbar and 1000 mbar. At a pressure of 400 mbar, 1.7 g of lactide could be obtained from 40.9 g of commercially available, aqueous lactic acid. At 1000 mbar, 0.9 g of lactide were obtained from 34.4 g of lactic acid. However, the conspicuously high mass of deposits in the laboratory reactor and the large increase in weight of the catalyst speak in favor of a reaction mode of operation in the higher negative pressure.

Example 5 - 6

In addition to pressure, temperature plays a decisive role. The range is limited on the one hand by the decomposition temperature of lactic acid upwards. On the other hand, the reaction requires a minimum temperature, so that sales in the first place is possible.

The parameters were chosen for Examples 5 and 6 as follows: The runtime was four hours. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The pressure was 150 mbar. The catalyst still used was the amorphous γ-aluminum oxide. The catalyst mass was 5 g with a particle size spectrum of 1.25 mm to 2.00 mm. Furthermore, the temperature was reduced to 120 ° C. or increased to 240 ° C. The weight hour space velocity (WHSV) was calculated to be 2.3 l / h.

At the lower temperature of 120 ° C, 48.3 g of 90% lactic acid was pumped in. After termination of the experiment, 34.5 g of reaction mass could be taken from the first condenser and 3.6 g were weighed out in the total condenser.

The total mass of the pumped 90% lactic acid was 47.7 g at the temperature setting of 240 ° C. Of these, 37.7 g of reaction mass was recovered in the first condenser. This subdivided for the first time in a crystalline, white and a liquid, milky phase. The solid phase has a very low water content of less than 1% and consists mainly of lactide and monomer. In addition, 4.7 g could be weighed in total condenser. The evaluation of the individual reaction masses gave the values shown in Table 4: IS

Table 4

Massen¬

Pressure Temperature educt WHSV conversion selectivity

 loss

catalyst

(mbar) CO (96) (%) (%) (B) v-AI ₂ 0 ₃ 150 120 LA 90 2.3 8.5 13.1 10.2

V-Al ₂ 0 ₃ 150 240 LA 90 2.3 15.0 98.9 5.3 Analogously to the tests at elevated pressure from Examples 3 and 4, at a temperature of 120 ° C., it is not possible to obtain a total evaporation of the Lactic acid mixtures are assumed. Accordingly, the mass loss due to deposits in the tube and on the catalyst (5.4 g) is very high. The selectivity of lactide is therefore narrowed and is at a low 13.1%. However, no other by-products could be detected, the reaction proceeds highly selectively under the set mild conditions. Typical by-products such as acrylic acid, acetone, acetaldehyde, formaldehyde and carbon dioxide were not detected.

At the elevated temperature of 240 ° C, a large part of the educt mixture is evaporated. Sales are comparatively low at 15.0%. By contrast, the selectivity is almost optimal at 98.9%. At the elevated temperature, the deposits in the system are reduced to a minimum. The non-evaluable mass thus consists almost exclusively of water and is u.a. explained by the dry inert gas stream. The conversion and the selectivity are thus not influenced by the 5.2 g of the mass difference. Furthermore, the desorption of the catalyst is additionally positively influenced by the elevated temperature. The weight gain has dropped to 0.2 g. The catalyst is not stuck together and also has no coking. Consequently, a longer term or a multiple use of y-alumina harmless. In total, 3.8 g of lactide could be prepared directly from the 47.7 g of aqueous 90% lactic acid.

Example 7 - 8

The advantage of the method according to the invention is the possibility of variation of the educt mixture. Both from weakly concentrated, oligomer-deprived lactic acid, as well as from highly concentrated lactic acid, the direct production of lactide is possible.

Examples 7 and 8 illustrate the wide range of applications of the structure according to the invention. From a commercial 90% lactic acid, two different batches were prepared. On the one hand, additional water was added to the mixture so that a 77% lactic acid can be used. On the other hand, the mixture was removed by heating with water until a 97% lactic acid was formed. The two mixtures prepared were analyzed for their composition by HPLC analysis and Karl Fischer titration. The 77% lactic acid contains about 23.8% water and is depleted of oligomers. The degree of polymerization is about 1.17. The 97% lactic acid, however, has a water content of 7.8% and includes, for example, a significantly higher proportion by weight of lactoyl lactic acid (L ₂ A = 25%). The degree of polymerization is about 1.55. Despite the water separation, the degrees of polymerization achieved are well below the usual degrees of polymerization of a PLA

Prepolymer.

The parameters evaluated from the above examples were adopted for Examples 7 and 8. The reactor was filled with 5 g of the y-alumina catalyst. The particle size was between 1.25 mm and 2.00 mm. The temperature was adjusted to 240 ° C and the pressure on the catalyst fixed bed was reduced to 150 mbar. The run time was four hours and the nitrogen flow rate was again set to 80 ml / min. The weight hour space velocity (WHSV) was calculated at 2.3 1 / h.

Of the dilute, 77% lactic acid, 44.6 g was pumped into the reactor. Of these, 33.8 g were recovered in the first condenser and 5.5 g in the total condenser.

Of the highly concentrated, 97% lactic acid, 31.0 g were pumped in. Of these, 20.3 g were recovered in the first condenser and 3.8 g in the total condenser. The evaluation of the examples with differently concentrated lactic acid gave the results shown in Table 5:

Table 5

Massen¬

Pressure Temperature educt WHSV conversion selectivity

 Catalyst loss

(mbar) (° c) (%) (rr ¹ ) (%) (%) (g)

Y-AIA, 150 240 LA 77 2.3 49.8 52.9 5.3

_V -AI ₂ 0 ₃ 150 240 LA 97 2.3 53.6 95.0 6.9

The results show a turnover of 49.8% and 53.6%. When using the concentrated (97%) lactic acid, the optimal selectivity can be demonstrated by the high numerical value. This is 95.0%. The dilute (77%) lactic acid achieves a selectivity of 52.9%.

For the selected parameters, the temperature-dependent desorption of the products is favored. Regardless of the concentration, a slight increase in the weight of the catalyst was found to be 0.2 g and 0.3 g.

In comparison of the two examples, a highly concentrated lactic acid shows better properties for the direct production of lactide than a dilute one. The most important reason is the high dimer concentration in the starting mixture. The linear dimer of lactic acid is considered to be the starting point for the cyclization of lactide.

Furthermore, all oligomers are hydrolyzed and thus form smaller molecules, such as dimers and monomers. The newly formed dimers can form additional lactide.

Example 9-12

The foregoing Examples 2 to 8 have demonstrated the known effectiveness of a simple amorphous γ-alumina catalyst. Compared to the above patent DE 69105144T2 significantly higher yields could be achieved by reducing the pressure. The maximum yield is 50.9% and is produced by a highly concentrated lactic acid at a achieved by 240 ° C and a pressure of 150 mbar.

Mild process conditions, especially temperature and pressure settings, not only affect the process costs but also the racemization of the reactant and the product. To avoid the latter and to reduce the process temperature zeolites are investigated in the following examples. Although these are more expensive compared to amorphous alumina catalysts, higher yields can be achieved. The following examples therefore include different zeolites with different structures and acid strengths.

Except for the B-MFI zeolite, which was prepared according to Example No. 3 of the patent US4401637 itself, all of the following catalysts were present in powder form. For this reason, the respective zeolite powder was pressed into tablets for continuous use in the tubular reactor and then comminuted to the desired particle size of 1.25 mm to 2.00 mm.

In order to compare the results of the following examples with those of Examples 1 (without catalyst) and 2 (alumina catalyst), the parameters from these examples are reused. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The temperature was 180 ° C and the pressure on the fixed catalyst bed at 150 mbar. The weight hour space velocity (WHSV) was calculated at 2.3 1 / h. Further, 5 g of catalyst having a particle size range of 1.25 mm to 2.00 mm was charged into the reactor. The residence time on the catalyst was about 2 seconds and is almost identical due to the approximately same density of alumina catalyst and zeolites. The educt mixture used again is the commercially available 90% lactic acid.

In the following four zeolites are examined in more detail, using the self-made B-MFI zeolite, in addition a beta, an L and a mordenite zeolite. These catalysts differed in their structure and in their acid strength.

The B-MFI zeolite has an average acid strength due to ion exchange and has many weak acid sites. In the attempt collected during the four-hour run, 33.6 g of reaction mass in the first condenser. In the total condenser, 5.6 g of the total 46.3 g of pumped mass could be recovered.

Beta zeolite (Zeocat PB-1, Uetikon) has a modulus, i. H. a silicon-aluminum ratio of 150. The plant was pumped 47.4 g of 90% lactic acid. Reaction masses of 37.0 g in the first condenser and 4.3 g in the total condenser were recovered.

The L zeolite (K zeolite L, Bayer) has a modulus (Si / Al) of 5.5. In this experiment, 36.1 g of the total of 45.1 g of 90% lactic acid mass to be pumped in the first condenser and 5.5 g in the total condenser were recovered after termination of the experiment.

The zeolite with the mordenite structure (Zeocat Mordenit, Uetikon) has a modulus (Si / Al) of 15.1. A total of 46.0 g of 90% lactic acid were pumped into the laboratory. In the individual capacitors, masses of 37.0 g and 3.3 g were weighed out.

The analysis of the individual zeolites is listed in Table 6 below: TABLE 6

Massen¬

Pressure Temperature educt WHSV conversion selectivity

Zeolite loss

(mbar) CC) (%) (h- ¹ ) (%) (%) (g)

B-MFI 150 180 LA 90 2.3 27.1 14.7 7.1 g

Beta 150 180 LA 90 2.3 14.8 46.7 6.1 g

L 150 180 LA 90 2.3 60.9 10.9 3.5 g

Mordenite 150 180 LA 90 2.3 20.2 36.7 5.7 g The obtained values of the different zeolites again show the basic possibility of the direct synthesis of lactide from aqueous lactic acid.

The B-MFI zeolite used achieves a moderate conversion of 27.1% at a low selectivity of 14.7%. The mass loss again affects the results obtained, the sales increase and the selectivity decreases. The Not analyzable mass is due to, inter alia, deposits and decomposition reactions in the evaporation of lactic acid and can be quantified to 7.1 g. Only the zeolite knows a mass increase in the amount of

1.3 g. The B-MFI pentasil zeolite catalyzes large amounts of by-products and shows a broad product spectrum. The yield is correspondingly low and is only 4.0%. In comparison, the yield of the alumina catalyst is 11.4%.

The beta zeolite achieved a turnover of 14.8%. The mass loss (6.1 g) is similar to the loss of the pentasil zeolite (7.1 g), but the beta zeolite catalyzes selective lactide. In addition to lactide, lactic acid and water, the product range consists only of traces of by-products. Accordingly spreader ^¬ accordingly is reduced selectivity mainly by the loss of mass and 46.7% consistently higher than in the preceding example, the B-MFI zeolite (S _{B ~} FI = 14.7%). The catalyst weight gain was

1.4 g

In the experiment with the L zeolite, a conspicuously high conversion is observed. At the same time, the mass loss of 3.5 g is very low. The L zeolite does not catalyze the reaction selectively. The lactide selectivity is as low as 10.9% and is greatly reduced by by-products and decomposition of the lactic acid.

The mordenite zeolite catalyzes a moderate turnover of 20.2%. Part of the turnover is due to the mass loss. The loss can be estimated at 5.7 g. Part of this mass (0.8 g) settles on the catalyst. Furthermore, a lactide selectivity of 36.7% is achieved.

In addition to the alumina catalyst, the beta and the mordenite zeolite achieve similarly selective product spectra. These consist only of lactide,

Lactic acid and water. However, with 6.9% (beta) and 7.4% (mordenite) the zeolites have lower yields than the alumina catalyst (YDIO- io = ll, 4%). In addition, the chromatograms of the zeolites do not show 100% conversion of the oligomers. Example 13

The used, wide-pored HY zeolite has a particularly large cavity with a diameter of 7.4 Ä. This commercially available

 (Valfor CBV 780, Zeolyst International, formerly PQ Corporation), powdered catalyst was used in the particle size of 1.25 mm to 2.00 mm. The modulus or the Si / Al ratio of the HY zeolite is 80. The experimental procedure and the catalyst preparation were carried out as in the above examples. The trial runtime was four hours. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The temperature was 180 ° C and the pressure at 150 mbar. The reactor was filled with 5 g of catalyst particles. The residence time on the catalyst was about 2 sec. The weight hour space velocity (WHSV) was calculated to be 2.3 l / h.

A total of 45.8 g of the 90% lactic acid was pumped into the laboratory reactor during the four hour experiment. The nearly anhydrous first condenser contained 30.5 g of reaction mass. This subdivided into a, at room temperature, solid phase and into a liquid, milky phase. The mass ratio was about 2 to 1. Furthermore, 8.0 g of additional reaction mass could be recovered in the total condenser. The evaluation of the individual masses gave the values shown in Table 7:

Table 7

Massen¬

Pressure Temperature educt WHSV conversion selectivity

Zeolite loss

(mbar) (° C) (%) (h ¹ ) (%) (%) (g)

HY 150 180 LA 90 2.3 60.2 43.2 7.3

Sales were more than quadrupled compared to example 1. In addition, the selectivity could be increased. In addition to the product, which can be found preferably in the solid phase, only

Traces of by-products are detected. Especially in the aqueous phase of the total condenser minimal amounts of acrylic acid were found. Overall, 7.3 g of mass was lost due to the dry inert gas flow and could not be evaluated for this reason. Among other things, a weight gain of the catalyst was measured at 2.5 g. Without this mass increase of the catalyst, the non-evaluable mass amounts to the value from Example 1 and confirms the results obtained there.

The analysis of the individual reaction masses shows a complete conversion of the oligomers (L _n A, n> 2). This behavior was not observed in Example 1. There was a slight shift in the individual concentrations, but the turnover of the oligomers was not nearly as high. In addition to complete conversion of the oligomers, the monomer could not be directly converted to lactide. Monomer and lactide dominate in the first cold trap, while in the second cold trap mainly water and traces of monomer and acrylic acid are found.

The results suggest that formation of the lactide is preferred

Close oligomers. From the literature, the backbiting principle for the production of lactide is known. According to this principle, the lactide could also form from smaller oligomers, as they occur directly in the commercially available 90% lactic acid. Furthermore, the oligomers could hydrolyze to the dimer or to the monomer, with simultaneous conversion of the linear dimer into the cyclic dimer.

In the direct synthesis of lactide from lactic acid, the use of a catalyst is of great advantage, according to the results shown above. Compared to the experiment without catalyst

%), the yield of the catalyzed experiment (Υ _Η γ = 26.0%) was significantly increased. The yield of

In addition, HY zeolites are more than twice as large as the comparison test with the alumina catalyst (YDIO-IO = H _/ 4%).

Example 14 - 16

In the following, the influence of the H-ZSM-5 zeolite is investigated. This has a medium pore structure, which is in a diameter of the pore of 5.6 Ä. The different modules of the zeolites used are used to investigate the influence of the acidity. H-ZSM-5 zeolites with the modules 25 (Zeocat PZ-2/25 H, Uetikon), 300 (Zeocat PZ-2/300 H, Uetikon) and 1000 (Zeocat PZ-2/1000 H, Ueticon) are used. , The silicon-aluminum ratio is a measure of the acidity. The more aluminum in the catalyst, the lower the modulus and the more weakly bound protons are. These form many weak acid centers and the zeolite becomes more acidic overall. Accordingly, the H-ZSM-5 zeolite with a modulus of 25 is strongly acidic, while a modulus of 1000 represents a weakly acidic catalyst with few but strongly acidic centers.

The preparation of the catalyst was carried out as described above, so that in each case 5 g of zeolite were used with a particle size of 1.25 mm to 2.00 mm. The residence time on the catalyst was about 2 sec. The parameters were taken from the above examples again. The trial runtime was four hours. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The temperature was 180 ° C and the pressure at 150 mbar. The weight hour space velocity (WHSV) was calculated at 2.3 1 / h.

In the experiment with the strongly acidic H-ZSM-5 zeolite with a modulus of 25, a lactic acid mass of 39.5 g was pumped into the laboratory plant. 17.6 g were recovered in the first condenser and 11.1 g of reaction mass in the total condenser.

The medium acid H-ZSM-5 zeolite with a modulus of 300 again catalyzed two phases in the first condenser. A mass ratio of solid to liquid phase of 3 to 1 was measured. Overall, a mass of 31.6 g of 45.2 g of pumped 90% lactic acid was found there. The total condenser contained a further 5.9 g of reaction mass.

In the experiment with the weakly acidic H-ZSM-5 zeolite with a modulus of 1000, again only one liquid phase could be observed in the first condenser. The mass amounted to 37.6 g. The total condenser recovered 4.2 g of reaction mass. During the four-hour run, 48.7 g of 90% lactic acid were added to the system. guided.

The analysis of each experiment is listed in Table 8 below:

Table 8

Massen¬

Pressure Temperature educt WHSV conversion selectivity

H-ZSM-5 loss

(mbar) CC) (%) (h- ¹ ) (%) (%) (g)

Module 25 150 180 LA 90 2.3 54.0 18.0 10.8

Module 300 150 180 LA 90 2.3 52.3 71.7 7.7

Module 1000 150 180 LA 90 2.3 33.7 52.8 6.9

The results again demonstrate the effectiveness of a catalyst in the direct synthesis of lactide from lactic acid. All three modules of H-ZSM-5 zeolites allow direct synthesis.

The lowest of the tested modules shows an increased mass loss. This can be calculated to 10.8 g, it already contains the mass increase of the catalyst in the amount of 1.4 g. Further, in addition to lactide, the chromatograms show other by-products which reduce selectivity in addition to mass loss. The lactide selectivity is 18.0%, with the selected modulus of 25. The acidity of the zeolite used is too high. Therefore, lactide can not be selectively produced and much of the lactic acid is decomposed.

The conversion of the medium-acid H-ZSM-5 zeolite with a modulus of 300 is 52.3% and is thus significantly higher than the turnover of the

uncatalyzed experiment from Example 1. The selectivity is even more than eleven times higher and reaches a value of 71.7%. However, the values are again influenced by the mass balance. From the pumped mass 7.8 g can not be evaluated at the end of the experiment. The conversion is thereby increased and the selectivity lowered. The zeolite selectively catalyzes lactide and undergoes a 0.7 g mass increase. The remaining 7.1 g of unexplainable mass comes again through the difficult to steaming lactic acid and dead volume in the plant. Another part of the loss leaves the plant via the dry nitrogen stream, which can accumulate with water and other reactants. The product spectrum of the H-ZSM-5 catalyst with a modulus of 300 is identical to that of the HY zeolite. In addition to lactide are only water, monomer and traces of by-products in the capacitors. The oligomers contained in the educt mixture of the aqueous 90% lactic acid have completely reacted. This further indicates formation of lactide from lactoyl lactic acid and hydrolyzed oligomers. The formation of the lactic acid monomer can be largely excluded. The yield of the experiment with the H-ZSM-5 zeolite (module 300) is again significantly higher at 37.5% than that of the uncatalyzed experiment of Example 1

 %). In addition, the value obtained is again significantly higher than the examples shown from the ambient pressure tests of the patent DE 69105144T2.

The employed H-ZSM-5 zeolite with a modulus of 1000 catalyzes the highly selective direct synthesis of lactide from aqueous lactic acid. Sales are at 33.7% and again impacted by mass loss. This also reduces the lactide selectivity to 52.8%. However, apart from lactide, the product spectrum shows only lactic acid and water.

Example 17-20

In Examples 13 and 15, the use of the HY zeolite and the H-ZSM-5 zeolite showed the highest lactide selectivities compared to the zeolites used. In the following, the achieved values of 71.1% and 43.2% are to be maximized with the help of parameter optimization. For this purpose, the temperature is increased to 210 ° C for better desorption of the products from the catalyst surface. Furthermore, oligomer-rich 97% lactic acid is used.

The zeolite powder was tableted and ground to a particle size range of 1.25 mm to 2.00 mm. The residence time on the catalyst was about 2 sec. The parameters were taken from the above examples again. The trial runtime was four hours. The inert gas stream (nitrogen) was adjusted to 80 ml / min. The pressure on the fixed catalyst bed was on 150 mbar regulated. To increase the selectivity further experiments were carried out with different residence times. These were realized by catalyst masses of 3 g and 5 g. The weight hour space velocity (WHSV) was thus set to 3.8 h ^"1 and 2.3 h ^{" 1} .

In the experiments with the H-ZSM-5 zeolite, a lactic acid mass of 46.5 g (WHSV 2.3 h ^-1 ) or 45.6 g (WHSV 3.8 h ^-1 ) was pumped into the system within the test runtime , In the first condenser, depending on the test, 22.9 g and 31.7 g and in the total condenser 10.8 g and 5.5 g of reaction mass could be recovered.

When using the HY zeolites, the pumped lactic acid masses amounted to 43.5 g (WHSV 2.3 h ^-1 ) and 44.5 g (WHSV 3.8 h ^-1 ). In the first cold trap, 20.5 g or 31.6 g could be condensed to ground depending on the test. The masses of the aqueous phases in the second condenser were 5.5 g and 6.7 g.

The evaluation of the individual masses gave the values shown in Table 9:

Table 9

Massen¬

Pressure Temperature educt WHSV conversion selectivity

 Zeolite loss

(mbar) CO (%) (h- ¹ ) (%) (%) (g)

H-ZSM-5 150 210 LA 97 2.3 70.8 34.9 11.8

H-ZSM-5 150 210 LA 97 3.8 62.1 30.7 8.4

 HY 150 210 LA 97 2.3 91.2 20.5 17.5

HY 150 210 LA 97 3.8 61.8 56.4 6.2

The tabulated results show a higher conversion of 70.8% (H-ZSM-5) and 91.2% (HY) as a result of the increase in temperature. Despite the better

Desorption deteriorates the mass balance compared to the described Examples 13 and 15. The losses are 10.1 g and 16.0 g and are due to the increased formation of volatile by-products. The catalyst weight increases are 1.7 g and 1.5 g. Out For this reason, the selectivity drops to 34.9% (H-ZSM-5) and 20.5% (HY).

The results with an increased WHSV show an expected reduced conversion as a result of the higher loading of the zeolites and the shortened residence time. When using the H-ZSM-5 zeolite 62.1% of the aqueous 97% lactic acid were reacted. The HY zeolite reached a value of 61.8%. In parallel, the mass loss also dropped to 4.8 g and 7.3 g. The catalyst weight gain was 1.1 g and 1.3 g. Catalysis of the H-ZSM-5 zeolite gave a selectivity of 30.7%. The residence time thus has no great influence on the formation of by-products. At the

Using the HY zeolites, the higher WHSV shows an increase in selectivity to 56.4%.

Claims

claims

1. A process for the continuous production of optically pure lactide, directly from concentrated lactic acid, characterized in that lactic acid is heated under reduced pressure, then passes through a reaction zone filled with a heterogeneous catalyst and then the product spectrum is cooled and separated.

2. The method according to claim 1, characterized in that the lactic acid is heated to a temperature between 100 ° C and 300 ° C.

3. The method according to any one of the preceding claims, characterized in that there is a negative pressure in the system, preferably from 10 mbar to 200 mbar.

4. The method according to any one of the preceding claims, characterized in that the concentrated lactic acid is a 50% to 100% lactic acid.

5. The method according to any one of the preceding claims, wherein the heterogeneous catalyst is a zeolite, an alumina or a silica catalyst.

A process according to claim 5, wherein an L zeolite, a mordenite zeolite, a beta zeolite, a B-MFI zeolite, a HY zeolite, an H-ZSM-5 zeolite or a

 y-alumina is used.

A process according to any one of the preceding claims, wherein the catalyst has an average acid strength and a small number of acidic sites.

A process according to any one of the preceding claims wherein the catalyst has a high Lewis to Bronsted acid site ratio.

9. The method according to any one of the preceding claims, wherein the catalyst has an amorphous or crystalline structure.

A process according to any one of the preceding claims, wherein the product spectrum consisting of lactide, monomer and water is separated directly behind the reaction zone.

11. The method according to any one of the preceding claims, wherein the separation of lactide and water is realized.

A process according to any one of the preceding claims, wherein an inert carrier gas stream dilutes the lactic acid.

13. The method according to any one of the preceding claims, wherein the turnover of the

Oligomere is 100% and the product spectrum consists only of lactide, monomer and water.