Process for preparing propylene oxide

Abstract

A continuous process for the preparation of propylene oxide, comprising (i) providing a liquid feed stream comprising propene, hydrogen peroxide, acetonitrile, water, optionally propane, and at least one dissolved potassium salt; (ii) passing the feed stream provided in (i) into an epoxidation reactor comprising a catalyst comprising a titanium zeolite of structure type MWW, and subjecting the feed stream to epoxidation reaction conditions in the epoxidation reactor, obtaining a reaction mixture comprising propylene oxide, acetonitrile, water, the at least one potassium salt, optionally propene, and optionally pane; (iii) removing an effluent stream from the epoxidation reactor, the effluent stream comprising propylene oxide, acetonitrile, water, at least a portion of the at least one potassium salt, optionally propene, and optionally propane.

Claims

1. A continuous process for the preparation of propylene oxide, the process comprising (i) providing a liquid feed stream comprising propene, hydrogen peroxide, acetonitrile, water, optionally propane, and at least one dissolved potassium salt; (ii) passing the feed stream provided in (i) into an epoxidation reactor comprising a catalyst comprising a titanium zeolite of structure MWW, and subjecting the feed stream to epoxidation reaction conditions in the epoxidation reactor, obtaining a reaction mixture comprising propylene oxide, acetonitrile, water, the at least one potassium salt, optionally propene, and optionally propane; and (iii) removing an effluent stream from the epoxidation reactor, the effluent stream comprising propylene oxide, acetonitrile, water, at least a portion of the at least one potassium salt, optionally propene, and optionally propane; wherein at least one dissolved potassium salt is an organic potassium salt.

2. The process of claim 1, wherein the at least one dissolved potassium salt further comprises at least one inorganic potassium salt.

3. The process of claim 1, wherein the at least one organic potassium salt is selected from the group consisting of a potassium salt of an aliphatic saturated monocarboxylic acid, potassium carbonate, and potassium hydrogen carbonate.

4. The process of claim 2, wherein the at least one inorganic potassium salt is selected from the group consisting of potassium hydroxide, a potassium halide, potassium nitrate, potassium sulfate, potassium hydrogen sulfate, and potassium perchlorate.

5. The process of claim 1, wherein the at least one organic potassium salt comprises potassium formate, potassium acetate, or a mixture thereof.

6. The process of claim 1, wherein a concentration of the at least one dissolved potassium salt in the liquid feed stream provided in (i) is at least 10% of the solubility limit of the at least one organic potassium salt in the liquid feed stream provided in (i).

7. The process of claim 1, wherein in (i), a molar ratio of potassium comprised in the at least one dissolved potassium salt relative to hydrogen peroxide comprised in the liquid feed stream is from 5×10.sup.−6:1 to 1000×10.sup.−6:1.

8. The process of claim 2, wherein in (i), a molar ratio of potassium relative to hydrogen peroxide in the liquid feed stream is from 5×10.sup.−6:1 to 1000×10.sup.−6:1.

9. The process of claim 1, wherein the liquid feed stream provided in (i) comprises acetonitrile in an amount of from 60 to 75 weight-%, based on a total weight of the liquid feed stream; hydrogen peroxide in an amount of from 6 to 10 weight-%, based on the total weight of the liquid feed stream; water in a molar ratio of water relative to acetonitrile of at most 1:4; propene with a molar ratio of propene relative to hydrogen peroxide comprised in the liquid feed stream in a range of from 1:1 to 1.5:1; and optionally propane with a molar ratio of propane relative to a sum of propene and propane in a range of from 0.0001:1 to 0.15:1; wherein at least 95 weight-% of the liquid feed stream provided in (i) consists of propene, hydrogen peroxide, acetonitrile, water, the at least one dissolved potassium salt, and optionally propane.

10. The process of claim wherein an amount of ammonium NH.sub.4.sup.+ in the liquid feed stream of (i) is at most 2 weight-ppm.

11. The process of claim 1, wherein a molar ratio of sodium relative to hydrogen peroxide in the liquid teed stream of (i) is from 1×10.sup.−6:1 to 250×10.sup.−6:1.

12. The process of claim 1, wherein the titanium zeolite of framework structure MWW comprises titanium, calculated as elemental titanium, in an amount of from 0.1 to 5 weight-%, based on a total weight of the titanium zeolite of framework structure MWW.

13. The process of claim 1, wherein the titanium zeolite of framework structure MWW comprises zinc, calculated as elemental zinc, in an amount of from 0.1 to 5 weight-%, based on a total weight of the titanium zeolite of framework structure MWW.

14. The process of claim 1, wherein the epoxidation reaction according to (ii) has a propylene oxide selectivity of at least 95%, wherein the propylene oxide selectivity is defined as a molar amount of propylene oxide comprised in the effluent stream removed in (iii) relative to a molar amount of hydrogen peroxide comprised in the liquid feed stream provided in (i).

15. The process of claim 1, wherein the effluent stream removed (iii) comprises propene, optionally propane, and optionally oxygen, the process further comprising (iv) separating propene, optionally together with propane and oxygen, from the effluent stream, obtaining a stream S01 enriched in propylene oxide, acetonitrile, and water; and (v) separating propylene oxide from S01, obtaining a stream comprising propylene oxide and depleted of acetonitrile and water.

Description

EXAMPLES

Reference Example 1

Epoxidation Reaction Setup

(1) A vertically arranged tubular reactor (length: 1.4 m, internal diameter: 7 mm) equipped with a jacket for thermostatization was charged with 15 g of ZnTiMWW catalyst in the form of strands with a diameter of 1.5 mm as described in Reference Example 2 below. The remaining reactor volume was filled with inert material (steatite spheres, 2 mm in diameter) to a height of about 5 cm at the lower end of the reactor and the remainder at the top end of the reactor.

(2) The reactor was thermostatized by passing a mixture of water and ethylene glycol as heat transfer medium through the jacket. The heat transfer medium was fed at the lower end of the jacket, flowing in co-current mode relative to the liquid feed stream passed into the reactor. The temperature of the heat transfer medium at the entrance of the jacket was defined as the reaction temperature, also referred to as T.sub.r. The flow rate of the heat transfer medium was suitably adjusted so that the difference between its temperature at the entrance of the jacket and its temperature at the exit of the jacket was at most 1 K.

(3) The pressure in the reactor was controlled by a pressure control valve and maintained at a constant value of 20 bar.sub.abs.

(4) The reactor feed stream was combined from three separate feed streams which were metered by using separate metering pumps: The first stream consisted either of acetonitrile (Asahi Kasei, chemical grade, acetonitrile content at least 99.9 weight-%, water content less than 500 weight-ppm). This first stream was employed having a flow rate of 68 g/h. The second stream consisted of liquefied polymer grade propene, having a propane content of 99.5 weight-%. This second stream was employed having a flow rate of 10.8 g/h. The third stream consisted of an aqueous hydrogen peroxide solution with a hydrogen peroxide concentration of 40 weight-%. This third stream was employed having a flow rate of 16.8 g/h. Potassium salts used in the experiments as additives were dissolved in the hydrogen peroxide stream in amounts shown below in the Examples.

(5) The three feed streams were premixed before the mixed feed was fed at ambient temperature to the bottom of the tubular reactor as liquid feed stream. Under the conditions the liquid feed stream consisted of one single liquid phase.

(6) The experiments were performed in a continuous manner. At the start of the run (t=0, defined as the point in time at which the hydrogen peroxide metering pump was started), the reaction temperature was set to a value in the range of 30 to 45° C. as shown in the examples. With a fresh catalyst this resulted in an initial 100% conversion of hydrogen peroxide. After a certain period of time, usually within 100 hours on stream, the hydrogen peroxide conversion started to decrease. The temperature was then adjusted, generally once to twice a day, in order to keep the hydrogen peroxide conversion in a range of from 85 to 96%. The average rate at which the temperature was increased in order to keep the hydrogen peroxide conversion essentially constant, referred to hereinbelow as the parameter delta T.sub.r/delta t, is a measure of the rate of catalyst deactivation. This parameter was calculated by dividing the difference between the cooling medium temperature at end of the indicated time period and the starting temperature and dividing it by the total number of hours on stream.

(7) The reactor effluent stream downstream the pressure control valve was collected, weighed and analyzed. Organic components, with the exception of hydroperoxypropanols and oxygen were analyzed in two separate gas-chromatographs. The hydrogen peroxide content was determined colorimetrically using the titanyl sulfate method. The content of hydroperoxypropanols, a mixture of 1-hydroperoxypropanol-2 and 2-hydroperoxypropanol-1, was determined by iodometrically measuring the total peroxide content and then subtracting the hydrogen peroxide content.

(8) The selectivity for propylene oxide given was determined relative to the hydrogen peroxide and was calculated as 100 times the ratio of moles of propylene oxide in the effluent stream divided by the moles of hydrogen peroxide in the feed. The selectivity for propylene glycol given was calculated as 100 times the ratio of moles of propylene glycol in the effluent divided by the moles of hydrogen peroxide in the feed. The selectivity for hydroperoxypropanols given was calculated as 100 times the ratio of twice the number of moles of hydroperoxypropanols in the effluent divided by the moles of hydrogen peroxide in the feed. The selectivity for molecular oxygen given was calculated as 100 times the ratio of twice the number of moles of molecular oxygen in the effluent divided by the moles of hydrogen peroxide in the feed.

Reference Example 2

Preparation of Epoxidation Reaction Catalyst (ZnTiMWW)

(9) 2.1 Preparation of Boron Containing Zeolite of Structure MWW (BMWW)

(10) A 2 m.sup.3 stirred tank reactor was first loaded with 470.4 kg of deionized water. After starting the stirrer at 70 rpm, boric acid (162.5 kg) was added and the suspension was stirred for 3 h. Subsequently, piperidine (272.5 kg) was added at once causing the temperature to rise from 28° C. to 46° C. To this solution colloidal silica (Ludox AS40, 392.0 kg) was added. The reactor was then slowly heated to 170° C. within 5 hours and then kept at this temperature under stirring for 120 hours. The maximum pressure during the reaction was 9.3 bar. Afterwards the reactor was cooled down to 50° C. The gel obtained had a pH of 11.3 and a viscosity of 15 mPa.Math.s at 20° C. The gel was then filtered and the filter cake washed with deionized water until the conductivity of the washings was below 500 microSiemens/cm. The filter cake was then suspended in deionized water and the suspension was spray-dried at 235° C. using nitrogen as the carrier gas. The white powder obtained (174.3 kg) contained 3.5 weight-% water. This white powder was then calcined at 650° C. in a rotary kiln to give 138.2 kg of boron containing zeolite of structure type MWW (BMWW) as a white powder.

(11) 2.2 Deboronation of BMWW with Water

(12) A 5 m.sup.3 stirred tank reactor was loaded with 125 kg of the BMWW obtained according to the previous step and 3750 kg of deionized water. The reactor was then slowly heated to 100° C. within 1 hour under stirring at 70 rpm, and then kept at this temperature for 20 hours and finally cooled to a temperature below 50° C. before it was filtered. The filter cake was then washed with deionized water until the washings had conductivity below 15 microSiemens. The filter cake was then dried for 6 hours under a nitrogen stream. The filter cake was then removed and suspended in 850 kg of deionized water. This suspension was then spray-dried at 235° C. using nitrogen as the carrier gas. The spray dried material weighed 118.5 kg and contained 42.5 weight-% Si, 0.06 weight-% B and 0.23 weight-% C (total organic carbon, TOC).

(13) 2.3 Preparation of Titanium Containing Zeolite of Structure Type MWW (TiMWW)

(14) A 2 m.sup.3 stirred tank reactor was first loaded with 111.2 kg of the spray-dried material from the previous step 2.2. In a separate 2 m.sup.3 stirred tank reactor were placed 400 kg of deionized water. After starting the stirrer at 80 rpm, piperidine (244.0 kg) was added. After the addition of piperidine was finished the mixture was stirred for 5 minutes before tetrabutyl orthotitanate (22.4 kg) was added. The pipe through which the titanate was added was then flushed with 40 kg of deionized water. The mixture was then stirred for 1 hour before being added to the first stirred tank reactor containing the spray-dried powder under stirring (50 rpm). The reactor was then heated to 170° C. and kept at this temperature for 120 h before being cooled to 50° C. The maximum pressure during the reaction was 10.6 bar. The cooled suspension was then filtered and the filter cake was washed with deionized water until the washings had conductivity below 1300 microSiemens/cm and an approximately neutral pH value. The filter cake was then dried under a nitrogen stream for 6 hours. The filter cake containing about 80 weight-% of water was used directly for the next step. The filter cake from the previous step and 1000 kg of deionized water were filled in a 2 m.sup.3 stirred tank reactor. Then 1900 kg of nitric acid (53 weight-% in water) were added under stirring at 70 rpm. The reactor was then heated to 100° C. and kept at this temperature for 20 hours before being cooled to 50° C. The suspension obtained was then filtered and the filter cake was washed with deionized water until the conductivity was below 10 microSiemens/cm and the washings were approximately neutral. Subsequently the filter cake was dried under a stream of nitrogen for 6 hours. This filter cake was then suspended in water and spray-dried at 235° C. using nitrogen as the carrier gas. 96 kg of a spray-dried powder were obtained. This material was then calcined in a rotary kiln at 650° C. 84 kg of titanium zeolite of structure type MWW (TiMWW) were obtained as a powder containing 43 weight-% Si, 2.0 weight-% Ti and 0.2 weight-% C (TOC). The pore volume determined by Hg-porosimetry according to DIN 66133 was 7.3 ml/g and the BET surface area determined according to DIN 66131 was 467 m.sup.2/g.

(15) 2.4 Preparation of a Zinc Containing TiMWW (ZnTiMWW) by Impregnation

(16) A 2 m.sup.3 stirred tank reactor was then loaded with 960 kg of water and 5.83 kg of zinc acetate dihydrate. After stirring for 30 min, TiMWW powder (32.0 kg, obtained according to the previous step) was added. The reactor was then heated to 100° C. and kept at this temperature for 4 hours before being cooled to 50° C. The suspension obtained was then filtered and the filter cake was washed 5 times with 120 liter portions of deionized water. The washed filter cake was then dried under a stream of nitrogen for 6 hours. It was then suspended in deionized water and spray-dried at 235° C. using nitrogen as the carrier gas. 34 kg of spray-dried material were obtained which was then calcined at 650° C. for 30 min in a rotary kiln. 28.5 kg of TiMWW containing zinc (ZnTiMWW) powder were obtained that contained 42 weight-% Si, 1.9 weight-% Ti, 1.6 weight-% Zn and 0.16 weight-% C (TOC). The pore volume determined by Hg-porosimetry according to DIN 66133 was 6.6 ml/g and the BET surface area determined according to DIN 66131 was 335 m.sup.2/g.

(17) 2.5 Preparation of Moldings Containing ZnTiMWW and Silica Binder

(18) In a kneader the ZnTiMWW powder from the previous step (27.0 kg) and hydroxymethylcellulose (Walocel™, 2.0 kg) were kneaded for 5 minutes. Then colloidal silica (Ludox® AS 40, 16.9 kg) was added. After kneading for 10 minutes deionized water (57.6 kg) was added and the mixture was kneaded for further 60 minutes. The paste obtained was then extruded through plates with cylindrical holed with 1.5 mm diameter with a pressure of 65-80 bar. The strands obtained were dried for 16 hours at 120° C. and then calcined for 5 hours at 500° C. The strands obtained were then sieved in a 0.8 mm sieve to remove fines. The ZnTiMWW catalyst strands obtained (34.2 kg) had a diameter of 1.5 mm and lengths between 5 and 25 mm. The bulk density of the catalyst was 345 g/I. The pore volume determined by Hg-porosimetry determined according to DIN 66133 was 1.1 ml/g and the BET surface area determined according to DIN 66131 was 371 m.sup.2/g. The elementary analysis showed that the molded ZnTiMWW catalyst contained 41 weight-% Si, 1.4 weight-% Ti and 1.2 weight-% Zn.

Reference Example 3

Preparation of Epoxidation Reaction Catalyst (TiMWW)

(19) 3.1 Preparation of Boron Containing Zeolite of Structure MWW (BMWW)

(20) A 50 liter stirred tank reactor was loaded with 22.05 kg of deionized water and 8.515 kg of piperidine. The mixture was then stirred for a few minutes at 150 rpm before 5.076 kg of boric acid were added. The resulting mixture was stirred for 30 minutes. Pyrogenic silica (Aerosil 200, 4.9 kg) was then added portion wise and the resulting suspension was stirred for 2 hours. The reactor was then heated to 170° C. within 2 hours and kept at this temperature for 120 hours. The maximum pressure during the reaction was 8.9 bar. After cooling to 50° C. the suspension was filtered and the filter cake was washed twice, each washing using 50 liters of deionized water. The filter cake was then dried for 24 hours at 80° C. under a stream of nitrogen, then oven-dried at 100° C. for 16 hours and finally calcined at 600° C. for 10 hours to obtain 4.95 kg of a white BMWW powder containing 1.4 weight-% B.

(21) 3.2 Deboronation of BMWW with Acid

(22) A 200 liter stirred tank reactor was loaded with 150 kg of nitric acid (30 weight-% in water) and the BMWW powder from the previous step and stirred at 100 rpm for 10 minutes. The reactor was then heated to 100° C. and kept at this temperature under stirring for 20 hours. After cooling to 50° C. the suspension was filtered and the filter cake washed with deionized water until the washings were approximately neutral. The filter cake was then dried for 15 hours under a stream of nitrogen and finally oven dried at 120° C. for 16 hours. 4.117 kg of a white powder containing 0.061 weight-% B were obtained.

(23) 3.3 Preparation of Titanium Containing Zeolite of Structure Type MWW (TiMWW)

(24) A 20 liter stirred tank reactor was then loaded with 10.5 kg of deionized water and 5.07 kg of piperidine. The mixture was stirred (170 rpm) for 10 minutes before adding 700 g of tetrabutyl orthotitanate. The mixture was stirred for a further 30 min and then 3.5 kg of the powder obtained from the previous step 3.2 were added. After stirring for 2 hours the reactors was heated to 170° C. and kept at this temperature for 120 hours. The maximum pressure during the reaction was 9.1 bar. After cooling to 50° C. the resulting suspension was filtered and the filter cake was washed with twice with 25 liters of deionized water per washing. The filter cake was then dried at 100° C. for 48 hours. 4.073 kg of a wet white powder containing 2.3 weight-% Ti, 36.0 weight-% Si and 10.4 weight-% C (TOC) were obtained. The powder (4.0 kg) and 120 kg of nitric acid (30 weight-% in water) were then loaded to a 200 liter stirred tank reactor. The suspension was then stirred at 100 rpm and the reactor heated to 100° C. and kept at this temperature for 20 hours. After cooling to 50° C. the resulting suspension was filtered and the filter cake washed with deionized water until the washings were approximately neutral. The filter cake was then dried at 120° C. for 16 hours and finally calcined at 550° C. for 10 hours. 3.185 kg of TiMWW powder with 1.7 weight-% Ti and 45.0 weight % Si were obtained.

(25) 3.4 Preparation of Moldings Containing TiMWW and Silica Binder

(26) TiMWW powder (3.0 kg) obtained from the previous step 3.3 and hydroxymethylcellulose (Walocel™, 200 g) were kneaded for 5 minutes. Then colloidal silica (Ludox® AS40, 2.5 kg) were added under continuous kneading. After a further 10 min of kneading deionized water (3.0 kg) was added under kneading. The paste thus obtained was then extruded through plates with cylindrical holed with 1.5 mm diameter with a pressure of 75-85 bar. The strands obtained were dried for 16 hours at 120° C. and then calcined for 5 hours at 500° C. The strands obtained were then sieved using a 0.8 mm sieve to remove fines. The TiMWW catalyst strands obtained (3.88 kg) had a diameter of 1.5 mm and lengths between 5 and 25 mm. The pore volume determined by Hg-porosimetry according to DIN 66133 was 0.7 ml/g and the BET surface area determined according to DIN 66131 was 92 m.sup.2/g. The elementary analysis showed that the molded TiMWW catalyst contained 43.5 weight-% Si and 1.1 weight-% Ti.

Reference Example 4

Characterization of the Catalyst

Reference Example 4.1

Determination of Dv10, Dv50, and Dv90 Values

(27) 1.0 g of the micropowder is suspended in 100 g deionized water and stirred for 1 min. The sample was subjected to the measurement in an apparatus using the following parameters: Mastersizer S long bed version 2.15, ser. No. 33544-325; supplier: Malvern Instruments GmbH, Herrenberg, Germany: focal width 300RF mm; beam length 10.00 mm; module MS17; shadowing 16.9%; dispersion model 3$$D; analysis model polydisperse correction none.

Reference Example 4.2

Determination of the Silanol Concentration of the Moldings of the Present Invention

(28) For the determination of the silanol concentration, the .sup.29Si MAS NMR experiments were carried out at room temperature on a VARIAN Infinityplus-400 spectrometer using 5.0 mm ZrO.sub.2 rotors. The .sup.29Si MAS NMR spectra were collected at 79.5 MHz using a 1.9 μs π/4 (microsecond pi/4) pulse with 10 s recycle delay and 4000 scans. All .sup.29Si spectra were recorded on samples spun at 6 kHz, and chemical shifts were referenced to 4,4-dimethyl-4-silapentane sulfonate sodium (DSS). For the determination of the silanol group concentration, a given .sup.29Si MAS NMR spectrum is deconvolved by the proper Gaussian-Lorentzian line shapes. The concentration of the silanol groups with respect to the total number of Si atoms is obtained by integrating the deconvolved .sup.29Si MAS NMR spectra.

Reference Example 4.3

Determination of the Crush Strength of the Moldings

(29) The crush strength as referred to in the context of the present invention is to be understood as determined via a crush strength test machine Z2.5/TS1S, supplier Zwick GmbH & Co., D-89079 Ulm, Germany. As to fundamentals of this machine and its operation, reference is made to the respective instructions handbook “Register 1: Betriebsanleitung/Sicherheitshandbuch für die Material-Prüfmaschine Z2.5/TS1S”, version 1.5, December 2001 by Zwick GmbH & Co. Technische Dokumentation, August-Nagel-Strasse 11, D-89079 Ulm, Germany. With said machine, a given strand is subjected to an increasing force via a plunger having a diameter of 3 mm until the strand is crushed. The force at which the strand crushes is referred to as the crushing strength of the strand. The machine is equipped with a fixed horizontal table on which the strand is positioned. A plunger which is freely movable in vertical direction actuates the strand against the fixed table. The apparatus was operated with a preliminary force of 0.5 N, a shear rate under preliminary force of 10 mm/min and a subsequent testing rate of 1.6 mm/min. The vertically movable plunger was connected to a load cell for force pick-up and, during the measurement, moved toward the fixed turntable on which the molding (strand) to be investigated is positioned, thus actuating the strand against the table. The plunger was applied to the stands perpendicularly to their longitudinal axis. Controlling the experiment was carried out by means of a computer which registered and evaluated the results of the measurements. The values obtained are the mean value of the measurements for 10 strands in each case.

Reference Example 4.4

29Si Solid-state NMR Spectra Regarding Q3 and Q4 Structures

(30) The effect of the inventive water treatment on the molding related to Q.sup.3 and Q.sup.4 structures in the material was characterized by comparing the changes in .sup.29Si solid-state NMR spectra under comparable conditions. All .sup.29Si solid-state NMR experiments were performed using a Bruker Advance spectrometer with 300 MHz .sup.1H Larmor frequency (Bruker Biospin, Germany). Samples were packed in 7 mm ZrO.sub.2 rotors, and measured under 5 kHz Magic Angle Spinning at room temperature. .sup.29Si direct polarization spectra were obtained using (pi/2)-pulse excitation with 5 microsecond pulse width, a .sup.29Si carrier frequency corresponding to −65 ppm in the spectrum, and a scan recycle delay of 120 s. Signal was acquired for 25 ms under 45 kHz high-power proton decoupling, and accumulated over 10 to 17 hours. Spectra were processed using Bruker Topspin with 30 Hz exponential line broadening, manual phasing, and manual baseline correction over the full spectrum width. Spectra were referenced with the polymer Q8M8 as an external secondary standard, setting the resonance of the trimethylsilyl M group to 12.5 ppm. The spectra were then fitted with a set of Gaussian line shapes, according to the number of discernable resonances. Relating to the presently assessed spectra, 6 lines in total were used, accounting for the five distinct peak maxima (at approximately −118, −115, −113, −110 and −104 ppm) plus the clearly visible shoulder at −98 ppm. Fitting was performed using DMFit (Massiot et al., Magnetic Resonance in Chemistry, 40 (2002) pp 70-76). Peaks were manually set at the visible peak maxima or shoulder. Both peak position and line width were then left unrestrained, i.e., fit peaks were not fixed at a certain position. The fitting outcome was numerically stable, i.e., distortions in the initial fit setup as described above did lead to similar results. The fitted peak areas were further used normalized as done by DMFit. After the water treatment of the invention, a decrease of signal intensity at the left hand side of the spectrum was observed, a region that includes Q.sup.3 silanol structures (here especially: around and above −104 ppm, i.e. “left” of −104 ppm). Further, an increase of signal at the right hand side of the spectrum (here: below −110 ppm, i.e. “right” of −110 ppm) was observed, which region comprises Q.sup.4 structures exclusively. For the quantification of spectrum changes, a ratio was calculated that reflects changes in the peak areas “left hand” and “right hand”, as follows. The six peaks were labeled with 1, 2, 3, 4, 5, and 6, and the ratio Q was calculated with the formula 100*{[a.sub.1+a.sub.2]/[a.sub.4+a.sub.5+a.sub.6]}/a.sub.3. In this formula, a.sub.i, i=1 . . . 6 represents the area of the fitted peak to which this number was attributed.

Reference Example 4.5

Water Adsorption/Desorption—Water Uptake

(31) The water adsorption/desorption isotherms measurements were performed on a VTI SA instrument from TA Instruments following a step-isotherm program. The experiment consisted of a run or a series of runs performed on a sample material that has been placed on the microbalance pan inside of the instrument. Before the measurement were started, the residual moisture of the sample was removed by heating the sample to 100° C. (heating ramp of 5° C./min) and holding it for 6 h under a N.sub.2 flow. After the drying program, the temperature in the cell was decreased to 25° C. and kept isothermal during the measurements. The microbalance was calibrated, and the weight of the dried sample was balanced (maximum mass deviation 0.01 wt. %). Water uptake by the sample was measured as the increase in weight over that of the dry sample. First, an adsorption curve was measured by increasing the relative humidity (RH) (expressed as weight-% water in the atmosphere inside of the cell) to which the samples was exposed and measuring the water uptake by the sample at equilibrium. The RH was increased with a step of 10 wt. % from 5 to 85% and at each step the system controlled the RH and monitored the sample weight until reaching the equilibrium conditions and recording the weight uptake. The total adsorbed water amount by the sample was taken after the sample was exposed to the 85 weight-% RH. During the desorption measurement the RH was decreased from 85 wt. % to 5 wt. % with a step of 10% and the change in the weight of the sample (water uptake) was monitored and recorded.

Reference Example 4.6

FT-IR Measurements

(32) The FT-IR (Fourier-Transformed-Infrared) measurements were performed on a Nicolet 6700 spectrometer. The molding was powdered and then pressed into a self-supporting pellet without the use of any additives. The pellet was introduced into a high vacuum (HV) cell placed into the FT-IR instrument. Prior to the measurement the sample was pretreated in high vacuum (10.sup.−5 mbar) for 3 h at 300° C. The spectra were collected after cooling the cell to 50° C. The spectra were recorded in the range of 4000 to 800 cm.sup.−1 at a resolution of 2 cm.sup.−1. The obtained spectra are represented in a plot having on the x axis the wavenumber (cm.sup.−1) and on the y axis the absorbance (arbitrary units, a.u.). For the quantitative determination of the peak heights and the ratio between these peaks a baseline correction was carried out. Changes in the 3000-3900 cm.sup.−1 region were analyzed and for comparing multiple samples, as reference the band at 1880±5 cm.sup.−1 was taken.

Reference Example 4.7

Determination of Crystallinity Via XRD

(33) The crystallinity of the zeolitic materials according to the present invention were determined by XRD analysis. The data were collected using a standard Bragg-Brentano diffractometer with a Cu-X-ray source and an energy dispersive point detector. The angular range of 2° to 70° (2 theta) was scanned with a step size of 0.02°, while the variable divergence slit was set to a constant illuminated sample length of 20 mm.

(34) The data were then analyzed using TOPAS V4 software, wherein the sharp diffraction peaks were modeled using a Pawley fit containing a unit cell with the following starting parameters: a=14.4 Angstrom (1 Angstrom=10.sup.−10 m) and c=25.2 Angstrom in the space group P6/mmm. These were refined to fit the data. Independent peaks were inserted at the following positions. 8.4°, 22.4°, 28.2° and 43°. These were used to describe the amorphous content. The crystalline content describes the intensity of the crystalline signal to the total scattered intensity. Included in the model were also a linear background, Lorentz and polarization corrections, lattice parameters, space group and crystallite size.

Reference Example 5

Definition and Determination of the Octanol-water Partition Coefficient KOW

(35) The octanol-water partition coefficient K.sub.OW of a given compound is defined as the ratio of said compound's chemical concentration in the octanol phase relative to said compound's chemical concentration in the aqueous phase in a two-phase system of 1-octanol and water at a temperature of 25° C. The octanol-water partition coefficient K.sub.OW of a given compound is determined using the shake-flask method which consists of dissolving the compound in a volume of high-purity 1-octanol and deionized water (pre-mixed and calibrated for at least 24 h) and measuring the concentration of the compound in each the 1-octanol phase and the water phase by a sufficiently exact method, preferably via UV/VIS spectroscopy. This method is described in the OECD Guideline for the testing of chemicals, number 107, adopted on Jul. 27, 1995.

Examples

Potassium Salts as Additives

Example 1

(36) In a first run, potassium dihydrogen phosphate was dissolved in the third stream, the aqueous hydrogen peroxide solution described above in Reference Example 1. The concentration of the potassium dihydrogen phosphate was chosen so that the molar ratio of potassium relative to the hydrogen peroxide in the liquid feed stream was 130×10.sup.−6:1. In particular, the concentration of the potassium dihydrogen phosphate in the liquid feed stream was 130 micromol per mol hydrogen peroxide. The epoxdation reaction was carried out as described above in Reference Example 1.

(37) After a certain time on stream the reaction was stopped. At this point in time, the hydrogen peroxide conversion had reached a value of 90%. Since for this additive and this catalytic system, comprising a ZnTiMWW catalyst and potassium dihydrogen phosphate, excellent propylene oxide selectivities were observed, this system was set as reference system for all other systems below.

Examples 2 to 6

(38) Since the use of potassium dihydrogen phosphate as additive led to such good results, the following examples and comparative examples were carried out by starting the continuous epoxidation reaction with potassium dihydrogen phosphate as additive. Then, the reaction was continued until the hydrogen peroxide conversion reached a value of 90%.

(39) At this point in time, the addition of potassium dihydrogen phosphate via the third stream described in Reference Example 1 was stopped and other potassium salts were used instead as indicated in Table 1 below.

(40) For every potassium salt used, the concentration of the salt was chosen so that the potassium concentration in the liquid feed stream passed to the epoxidation reactor remained at the value of example 1 (molar ratio of potassium relative to the hydrogen peroxide in the liquid feed stream 130×10.sup.−6:1).

(41) For every system, the epoxidation reaction was continued until, after a first deviation of the hydrogen peroxide conversion, this conversion reached the value of 90% again, i.e. the value at which the additive had already been changed from potassium dihydrogen phosphate to another additive. At a point, when this value of 90% conversion could be kept constant for 48 hours without changing the reaction parameters, the epoxidation reaction was stopped.

(42) The deactivation rate was then determined as the difference in the reaction temperature (delta Tr) divided by the period of time (delta t). The difference in the reaction temperature is defined as the reaction temperature at the end of the epoxidation reaction (i.e. the reaction temperature at which the hydrogen peroxide conversion had reached the value of 90% and could be kept constant for another 48 hours without any further parameter changes within the system after having replaced potassium dihydrogen phosphate by the new additive) minus the reaction temperature at the point in time where potassium dihydrogen phosphate had been used as additive, the hydrogen conversion had reached the value of 90% and potassium dihydrogen phosphate had been replaced. The period of time (delta t) is defined by the time on stream from the change in additive until the end of the epoxidation reaction, i.e. the period of time during which the epoxidiation reaction was carried out in the presence of the additive replacing potassium dihydrogen phosphate.

Example 2a

(43) As shown in Table 1 below, potassium formate was found to be the most preferred additive among all potassium salts tested. Therefore, a further run was carried out in the same manner as the run according to Example 2, with the difference that potassium formate was used as additive from the very beginning of the epoxidation reaction. The deactivation rate was then determined as the difference in the reaction temperature (delta Tr) divided by the period of time (delta t). The difference in the reaction temperature is defined as the reaction temperature at the beginning of the epoxidation reaction (i.e. the reaction temperature at the beginning of the experiment was identical to the ones used in Examples 1 to 6 and Comparatives Examples 1 to 3) minus the reaction temperature at the point in time the hydrogen peroxide conversion had reached the value of 90% and could be kept constant without any further parameter changes for 48 h.

Comparative Examples

Salts Other than Potassium Salts as Additive; No Additive

Comparative Example 1

(44) The Comparative Example 1 was carried out as Example 2. However, after the hydrogen peroxide conversion had reached a value of 90% with potassium dihydrogen phosphate as additive, the addition of potassium dihydrogen phosphate was stopped, and via the third stream described in Reference Example 1, no additive was used anymore.

Comparative Examples 2 to 3

(45) The Comparative Examples 2 and 3 were carried out as Example 2. However, after the hydrogen peroxide conversion had reached a value of 90% with potassium dihydrogen phosphate as additive, the addition of potassium dihydrogen phosphate was stopped, and via the third stream described in Reference Example 1, the additives according to Table 1 below were used instead. Regarding Comparative Example 2, the concentration of NH.sub.4H.sub.2PO.sub.4 was identical to the concentration of KH.sub.2PO.sub.4 according to Example 1, i.e. 130 micromol per mol hydrogen peroxide. Regarding Comparative Example 3, the concentration of NaH.sub.2PO.sub.4 was identical to the concentration of KH.sub.2PO.sub.4 according to Example 1, i.e. 130 micromol per mol hydrogen peroxide.

(46) TABLE-US-00001 TABLE 1 Results of the Examples and the Comparative Examples deactivation catalytic rate/ propylene oxide example additive system K/d selectivity/% 1 KH.sub.2PO.sub.4 ZnTiMWW 0.sup.1) 98.8 KH.sub.2PO.sub.4 2 HCOOK ZnTiMWW −1.14 98.8 HCOOK 2a.sup.2) HCOOK ZnTiMWW −2.18 99.1 HCOOK 3 K.sub.2CO.sub.3 ZnTiMWW −0.67 98.4 K.sub.2CO.sub.3 4 KHCO.sub.3 ZnTiMWW −0.48 98.7 KHCO.sub.3 5 KOH ZnTiMWW +0.13 98.4 KOH 6 KCl ZnTiMWW +0.19 98.4 KCl comp. 1 — ZnTiMWW +1.23 97.1 comp. 2 NH.sub.4H.sub.2PO.sub.4 ZnTiMWW +0.83 97.8 NH.sub.4H.sub.2PO.sub.4 comp. 3 NaH.sub.2PO.sub.4 ZnTiMWW +0.58 97.4 NaH.sub.2PO.sub.4 .sup.1)reference value .sup.2)no replacement of additive; the potassium formate was used as additive from the very beginning of the epoxidation reaction
Results

(47) Compared to the epoxidation reactions where either no additive or an additive other than a potassium salt was employed, all epoxidation reactions which were carried out with a potassium salt as additive showed higher propylene oxide selectivities at the same hydrogen peroxide selectivity of 90%. Reference is made to Table 1, showing that in all comparative examples, the propylene oxide selectivity is below or significantly below 98% whereas in the Examples, all respective values are significantly above 98% and, for potassium formate as additive, even above 99%. Thus, these reactions where potassium salts were employed as additives lead to a higher yield in propylene oxide as valuable product, and the selectivity with regard to undesired by-products or side-products is decreased.

(48) Further, it was found that these higher propylene oxide selectivites at constant hydrogen peroxide conversion can be realized at advantageous deactivation rate. Compared to the epoxidation reaction with potassium dihydrogen phosphate as additive, the deactivation rate at the end of a run where no potassium salt was employed was at least +0.58 K/d, i.e. significantly higher than the maximum value of +0.19 K/d obtained for reactions where potassium salts were employed as additives. In this context, it is noted that a positive value indicates that compared to potassium dihydrogen phosphate, a respective additive leads to an increase in reaction temperature which in turn means that more energy has to be provided to realize a given hydrogen peroxide conversion. Therefore, a low positive value and in particular a negative value are preferred.

(49) Therefore, on particular the organic potassium salts, represented by potassium formate, potassium carbonate and potassium hydrogen carbonate in the Examples, and the respective catalytic systems show excellent suitability for the epoxidation reaction since they all lead to a negative relative deactivation rate and, at the same time, to very high propylene oxide selectivities. Especially potassium formate was found to be of particular suitability since this additive showed the lowest relative deactivation rate and, at the same time, the highest propylene selectivity.

CITED LITERATURE

(50) WO 2011/006990 WO 2009/008493 US 2007043226 A1 U.S. Pat. No. 6,114,551 Wu et al., “Hydrothermal Synthesis of a novel Titanosilicate with MWW Topology”, Chemistry Letters (2000), pp. 774-775 Ullmann's Encyclopedia of Industrial Chemistry, 5.sup.th edition, volume A 13 (1989) pp. 443-466 EP 1 122 249 A1 EP 0 427 062 A2 U.S. Pat. No. 5,194,675 WO 2012/074118 WO 2011/152268 WO 2012/157473 Xiangqing Fang et al., Phys. Chem. Chem. Phys. 2013, 15, 4930-4938 US 2010/0234623 Lihao Tang et al., Macromolecules, 2008, 41, 7306-7315