ISSN 0023-1584, Kinetics and Catalysis, 2016, Vol. 57, No. 4, pp. 466–473. © Pleiades Publishing, Ltd., 2016. Original Russian Text © A.V. Sulimov, S.M. Danov, A.V. Ovcharova, A.A. Ovcharov, V.R. Flid, 2016, published in Kinetika i Kataliz, 2016, Vol. 57, No. 4, pp. 466–473.
Kinetics of Propylene Epoxidation with Hydrogen Peroxide Catalyzed by Extruded Titanium Silicalite in Methanol A. V. Sulimova, *, S. M. Danova, A. V. Ovcharovaa, A. A. Ovcharova, and V. R. Flidb aNizhny
bMoscow
Novgorod State Technical University, Nizhny Novgorod, 603950 Russia Technological University, Institute of Fine Chemical Technologies, Moscow, 119571 Russia *e-mail:
[email protected] Received November 26, 2015
Abstract—The kinetics of propylene oxidation into propylene oxide in the presence of extruded titanium silicalite was studied. Based on the experimental data, a kinetic model of the process was designed and the activation energies of the target and side reactions, the rate constants, and the adsorption equilibrium constants were determined. The adequacy of the proposed kinetic model was verified on a continuously-operated test bench laboratory unit. Keywords: propylene oxide, hydrogen peroxide, titanium silicalite, epoxidation DOI: 10.1134/S0023158416040121
INTRODUCTION Propylene oxide (PO) is of great industrial importance. In the last twenty years, its annual world production has increased two-fold and, currently, is more than 7.1 million tons. The main application field of PO (65–70%) is the synthesis of polyethers applied for the production of rigid and soft polyurethanes [1–3]. A considerable portion of the produced PO (up to 25%) is used in the preparation of 1,2-propylene glycol and dipropylene glycol [4]. Nonionic surfactants (proxanols and proxamines), allyl alcohol, propylene carbonate, and isopropanolamines are produced from PO on an industrial scale [5, 6]. In some production processes, PO replaces ethylene oxide, since its application is much more favorable from the environmental point of view. For example, a promising field of PO application is preparation of propylene glycol methyl ether, methyl propasol, which is applied in the production of nontoxic technical liquids (domestic heat-transfer media, coolants for food industry) and solvents for paint and varnish materials which can sub-
H2C CH
stitute such toxic products as monoethylene glycol, ethyl cellosolve, butyl cellosolve, etc. [7, 8]. The most promising method for the PO preparation is liquid-phase epoxidation of propylene (Pr) with the environmentally friendly oxidant hydrogen peroxide (HP) in methyl alcohol (MA) in the presence of titanium silicalite [9]: H2C CH CH3 + H2O2
(I)
H2C CH CH3 + H2O. O
Along with the target reaction (I), some side reactions occur in the system to yield propylene glycol (PG), 1-methoxypropan-2-ol (1MP2), 2-methoxylpropan-1-ol (2MP1), and negligible amounts of other compounds:
CH3 + H3C OH
O
H2C CH
CH3 + H2O
O
CH3 , (II)
OH OH
CH2 CH O
H2C CH
CH3 ,
(III)
CH3.
(IV)
OH
CH3 H2C CH
CH3 + H3C OH
O
CH2 CH HO
O CH3
466
KINETICS OF PROPYLENE EPOXIDATION WITH HYDROGEN PEROXIDE
In addition, oxygen was detected among the reaction products, which results from the reaction
2H 2O 2 → 2H 2O+O 2.
(V)
In recent years, there emerged quite a lot of publications concerning this method of propylene oxide preparation, which studied different catalytic systems [10], the physicochemical regularities [11, 12], and the kinetics of the process [13]. However, the reaction was performed most often on a fine crystalline heterogeneous catalyst dispersed in the reaction mass, while, under industrial conditions, it is preferable to use molded catalysts capable of functioning in the fixed bed, which excludes the need for their subsequent separation from the reaction mixture. The aim of the present work was to study the kinetics and to develop a mathematical model for the liquid-phase Pr epoxidation with a solution of HP in an organic solvent in the presence of extruded titanium silicalite as a heterogeneous catalyst. EXPERIMENTAL Methanol (analytical grade, Russian Standard GOST 2222–95), propylene oxide (pure grade, GOST 23001–88), 33–34% hydrogen peroxide (highpurity grade, Specifications TU 2611-069-05807977– 2006), propylene (GOST 25043–87), tetrabutoxytitanium (Specifications TU 6-09-2738–89), tetraethoxysilane (Specifications TU 2435-419-05763441–2003), and tetrapropylammonium hydroxide obtained by passing tetrapropylammonium bromide (98%, Acros) through an anion-exchange column were used in this work. Extruded titanium silicalite in the form of cylindrical granules with a diameter of 2 mm and a length of 5 mm was prepared according to a procedure developed by us [14] and was used as a heterogeneous catalyst. A powdered titanium silicalite with a mean particle size of 200–300 nm obtained as described in [15, 16] was subjected to extrusion. The binder was aluminum 5,6-oxynitrate. The resulting powdered and granular catalyst samples were characterized by X-ray powder diffraction and IR spectroscopy. The morphology of samples was studied by scanning electronmicroscopy and low-temperature nitrogen adsorption. The process kinetics was studied in a laboratoryscale continuous flow fixed-bed reactor. A tube-intube reactor with an inner diameter of 15 mm and a height of 250 mm filled with the granular titanium silicalite was made of stainless steel. To ensure a uniform input flow distribution over the reactor cross section, the reactor bottom was filled with Fenske glass helices. The reactor temperature was controlled using a liquid ultathermostat. To control the temperature profile throughout the height of the catalyst bed, thermocouKINETICS AND CATALYSIS
Vol. 57
No. 4
2016
467
ple wells were provided. A pressure of 5–7 atm was maintained throughout the system. The feed flow rate was controlled through electronic control of the pump. After passage of a certain reaction mixture volume required for reaching the steady-state conditions, the reaction mixture was sampled at the outlet of the reactor. The reaction mixture was analyzed on a Khromos GKh-1000 gas chromatorgaph (Khromos, Russia) equipped with a flame ionization detector and a metal column (2 m × 3 mm) packed with 15 wt % Carbowax 6000 supported on Chromaton-N-AW. The flow rate of the carrier gas (nitrogen) in the column was 50 mL/min. The injecvtion port and oven temperatures were maintained around 180 and 130°C, respectively. The detector temperature was 200°C. The composition of the reaction mixture was determined by absolute calibration. The hydrogen peroxide content was measured by iodometric titration. The IR spectra of the catalyst samples in the spectral range from 400 to 4000 s–1 were recorded as KBr pellets in air on an IRAffinity-1 FT-IR spectrometer (Shimadzu, Japan) at room temperature. The analysis showed the complete coincidence between the positions of characteristic bands for the powdered and granular titanium silicalite samples in the region of 540 and 960 cm–1. The XRD patterns of samples prior to and after molding were recorded on a Shimadzu LAB XRD6000 diffractometer (CuKα radiation, nickel filter, scintillation counter, voltage of 30 kV, current of 30 mA) in the 2θ range from 10° to 80°; the scan rate was 2 deg/min, and the step size was 0.02°. The reflections of both samples in the characteristic region 2θ = 23°–25° almost coincide. The intensities of peaks in the XRD pattern of the granular catalyst were slightly lower, which is likely due to the decrease in the amount of the main component, titanium-containing zeolite, as a result of addition of the binder. In addition, the intensities of the reflections at 31.12° and 32.66° in the XRD pattern of the granular sample slightly increased; according to card no. 09-0440 from the JCPDS database, these reflections belong to aluminum oxide. The morphology of the powdered catalyst samples was evaluated by the analysis of photomicrographs recorded on a Hitachi S-2500 scanning electron microscope (Japan) equipped with a JNCA energydispersive X-ray microanalysis attachment (Oxford Instruments, United Kingdom). Based on the data obtained, particle size distribution histograms for the powdered catalyst (more than 250 particles) were plotted. The statistical processing showed that the mean particle size is 255 nm at a distribution width of 25 nm. The specific surface area, total pore volume, and pore size distribution in the powdered and granular samples (Table 1) were measured on a TriStar 3020
468
SULIMOV et al.
Table 1. Parameters of the catalyst (titanium silicalite) porous structure Form of sample Powder Granules
Specific surface area, m2/g
Pore volume, cm3/g
316.66 268.83
0.182 0.155
automated gas adsorption analyzer (Micromeritics, United States) using the BET method. RESULTS AND DISCUSSION To assess the contribution from the noncatalytic constituent of the process, we performed several experiments in the absence of catalyst in a reactor filled with an inert mass-transfer packing (Fenske helices) at 60°C at the rate of starting mixture feed of 3.33 × 10–5 L/s. The HP conversion did not exceed 0.2–0.3%. Therefore, it was concluded that, under these conditions, the noncatalytic transformations of HP can be neglected. The occurrence of side reactions (II)–(V) in the corresponding systems was also studied under analogous conditions to show the rates of these reactions to be negligible in the absence of catalyst. For this reason, only the catalytic constituent of processes was taken into account hereinafter. The experimental study of propylene epoxidation performed earlier under gradientless conditions [17] showed that the reaction is first order with respect to olefin and fraction order with respect to HP. In addition, the epoxidation products that formed were found to decrease the rate of the target reaction. To assess the effect of the propylene oxide and water concentrations on the kinetic regularities of the reaction catalyzed by the extruded titanium silicalite, an additional set of experiments varying the initial concentrations of these substances were performed. The resulting PO was found to decrease the rate of the target reaction, while the change in the water content within 6.1–8.6 mol/L had no effect on this reaction. The regularities found agree well with our mechanism [18] of olefins epoxidation in the presence of titanium silicalite, according to which the reaction between propylene and hydrogen peroxide occurs on tetracoordinated titanium atoms through the formation of intermediate complex involving H2O2 and solvent. Based on the above-mentioned observations, in the construction of a kinetic model of the process we admitted the following assumptions: (1) the catalyst surface is homogeneous; (2) the adsorption of the starting olefin and water onto the catalyst surface can be neglected; (3) the target reaction occurs between HP adsorbed on the catalyst surface and the olefin being in solution; (4) the rate-limiting step is the surface reaction;
(5) side reactions also proceed on the catalyst surface. Thus, the rate of chemical reactions occurring in Pr epoxidation in the presence of titanium silicalite in MA should be described by the following system of five fractional rational equations:
⎧ ⎪r1 ⎪ ⎪r ⎪2 ⎪ ⎪ ⎨r3 ⎪ ⎪ ⎪r4 ⎪ ⎪r ⎪⎩ 5
k1bHPC HPC Pr , 1 + bHPC HP + bPOC PO k 2bPOC POC W = , 1 + bHPC HP + bPOC PO k3bPOC POC MA , = 1 + bHPC HP + bPOC PO k 4bPOC POC MA , = 1 + bHPC HP + bPOC PO k5bHPC HP , = 1 + bHPC HP + bPOC PO
=
(1) (2) (3) (4) (5)
where r1 is the rate of the target reaction (I) in mol s‒1 g–1; r2, r3, and r4 are the rates of side reactions (II)–(IV), respectively, in mol s–1 g–1; k1 is the rate constant for the target reaction in L s–1 g–1; k2, k3, and k4 are the rate constants of side reactions in L s–1 g–1; bHP and bPO are the coefficients of HP and PO adsorption, respectively, in L/mol; С HP, С Pr , С PO, С MA , and С W are the concentrations of HP, Pr, PO, MA, and water, respectively, in mol/L. To determine the parameters of the model proposed and to verify its adequacy, several sets of experiments were performed under different starting conditions. The catalyst bed height, the volumetric feed flow rate (2.68–6.8) × 10–5 L/s, the reaction temperature (30–60°C), the initial concentrations of HP (0.11–2.28 mol/L) and Pr (2.01–4.76 mol/L), and the concentration of MA (15.40–24.26 mol/L) were varied. To estimate the reproducibility of measurement results, “standard” experiments were performed at regular times to be sure that the catalyst activity remained unchanged during the experiments. Based on the data on the composition of the reaction mixture at the reactor outlet, kinetic curves were plotted. The ratio of the charged catalyst weight (mcat) to the rate of the starting mixture feed (F) was used as conventional residence time (τ). To vary the initial conditions and the residence time, the rate of feed flow into the reaction zone was changed. In reactions on solid catalysts, the transfer effects associated with the heat and weight transports are essential. The effect of these factors becomes noticeable and even dominant, if at least in some steps the transfer rate becomes lower than the rate of chemical reaction itself or comparable with it. Therefore, to obtain reliable kinetic data, a set of experiments was preliminarily performed to determine the way the KINETICS AND CATALYSIS
Vol. 57
No. 4
2016
KINETICS OF PROPYLENE EPOXIDATION WITH HYDROGEN PEROXIDE
epoxidation reaction is controlled and to rule out diffusion limitations. The effect of internal diffusion limitations was studied in a set of experiments on two catalysts with a granule size of 1.5–1.7 and 2.1–2.3 mm. The experiments were performed at identical initial concentra0 0 tions of components (С Pr = 2.83 mol/L, С HP = 0 0.98 mol/L, С MA = 16.79 mol/L, and τ = 1.46 × 10 4 g s L–1) and different temperatures: 30, 40, 50, and 60°C. The resulting data are given in Table 2. As seen, under the experimental conditions the increase in the grain size results in no change in the HP conversion, which suggests the absence of internal diffusion limitations. For this reason, all subsequent studies were performed on granules with a diameter of 2.1–2.3 mm. To assess the effect of hydrodynamics on the process rate and to exclude the external diffusion limitations, the method described in a monograph by Kafarov [19], which is widely used in similar works [20, 21], was employed. For this purpose, two sets of experiments were performed using different catalyst weights (different heights of the catalyst bed) at different feed space velocities and constant composition of the starting mixture at the reactor inlet (F = (2.68– 0 0 = 2.60 mol/L, С HP = 16.8) × 10–5 L/s, С Pr 0 0.94 mol/L, and С MA = 16.87 mol/L). Since the conditions under which kinetic control is reached change with temperature [22], the experiments were performed at the maximum possible temperature of 60°C. For each set of experiments, the HP conversion was plotted against the conventional residence time (Fig. 1). It can be seen from Fig. 2 that the experimental data obtained in both experimental sets fall satis-
469
Table 2. HP conversion in relation to the catalyst granule size HP conversion, %
Granule size, mm
30°C
40°C
50°C
60°C
1.5–1.7 2.1–2.3
8.3 8.2
13.5 13.4
22.0 21.9
33.1 33.2
factorily on one curve. Consequently, the external diffusion limitations can be neglected under the considered conditions, since the HP conversion does not depend on the rate of reaction mixture feeding. For this reason, in all subsequent kinetic experiments, the feed rate was maintained no lower than the minimum value, 2.68 × 10–5 L/s (τ = 18.7 × 10 4 g s L–1). If the external diffusion limitations are absent at 60°C, they can be neglected at lower temperatures as well. According to the available recommendations [23], the unknown parameters k1, k2, k3, k4, k5, bHP, and bPO appearing in the kinetic model under consideration were determined in two steps. At the first step, based on the data obtained by isolation of reactions in the differential flow reactor at low conversions of the key reactants (5–10%), we preliminary checked how the experimental data fit with the proposed model and performed the primary evaluation of unknown parameters from the kinetic equations (1)–(5). At this step, the kinetics of each simple reaction (I)– (V) at key reactant conversions of 5–10% (HP in reactions (I) and (V) and PO in reactions (II), (III), and (IV)) was considered separately. It was assumed that HP undergoes no decomposition (V) during the formation of PO by reaction (I).
xHP, %
CHPСPr/(r1 × 105), mol s g L–2
90
7
80
6
70
5
60
4
1
mcat = 5.00 g mcat = 10.03 g
50
2
3
3
40
2
4
30
1 0
20 0
5
10
15
20
τ × 10 , g s/L
0.2
0.4
0.6 0.8 CHP, mol/L
–4
Fig. 1. Effect of the conventional residence time (τ) and catalyst weight (mcat) on the hydrogen peroxide conversion (хHP) in the synthesis of PO. KINETICS AND CATALYSIS
Vol. 57
No. 4
2016
Fig. 2. Kinetics of the initial step of Pr epoxidation at different concentrations of HP and different temperatures: (1) 30, 0 (2) 40, (3) 50, and (4) 60°C. С Pr = 3.50–4.76 mol/L, 0 = 0.095–0.70 mol/L, and F = 1.33 × 10–4 L/s. С HP
470
SULIMOV et al.
To estimate the unknown parameters (k1 and bHP), a set of experiments was performed at different initial concentrations of HP in excess Pr in the MA medium. If Eq. (7) really describes the kinetics of the initial step of Pr epoxidation, the average concentrations С HP, and С P r and the corresponding rates r1 calculated from Eq. (8) at constant temperature must fall on a straight line in the coordinates of the equation C HPC Pr r1 = f ( C HP ). The experimental dependences shown in Fig. 2 show that the calculated values fall satisfactorily on such straight line, the mean relative deviation being 1.29%.
CHPСPr/(r1 × 106), mol s g L–2 6 1
5 4 3
2
2
3 4
1 0 0.2
0.4
0.6
0.8
The effect of the PO concentration on the epoxidation kinetics and bPO was evaluated in the set of experiments performed at constant contents of Pr and HP with addition of PO in different amounts. Under these conditions, the kinetics of PO formation is described by Eq. (1), which in the linearized representation takes the form
1.0 1.2 CPO, mol/L
Fig. 3. Kinetics of the initial step of Pr epoxidation at different concentrations of PO and different temperatures: 0 30 (1), 40 (2), 50 (3), and 60°C (4). С Pr = 2.01– 0 0 2.65 mol/L, С HP = 0.66–0.89 mol/L, С PO = 0.27– –5 L/s. 1.20 mol/L, F = 6.67 × 10
b С HPС Pr 1 + bHPС HP = + PO С PO. r1 k1bHP k1bHP
Under the conditions when the process is performed in the differential reactor at low HP conversions, the product concentrations can be taken equal to zero and Eq. (1) appears as
r1 =
k1bHPC HPC Pr , 1 + bHPC HP
(6)
or, in linearized form,
С HPС Pr = 1 + 1 С HP. r1 k1bHP k1
(7)
It is reasonable to assume that, in the differential reactor at low conversions of starting substances, the reaction rates and the substance concentrations along the catalyst bed must change linearly [23]. Therefore, to determine the reaction rates, we used the following equation [22]:
ri =
C i0 хF . mcat
(8)
(9)
If Eq. (9) describes the kinetics of the initial epoxidation step, the concentrations С HP, С Pr , and С PO and the corresponding rates r1 calculated from Eq. (8) at constant temperature must fall on a straight line in the coordinates of Eq. (7). This is observed in Fig. 3. The departure from linearity is small to be 5.23%. Based on the data shown in Figs. 2 and 3, the parameters of the kinetic equation (1) k1, bHP, and bPO were calculated. It follows from Fig. 2 that k1 = 1 tanα , and bHP = tanα L , where α is the slope angle of the straight line with respect to the X axis and L is the Y-intercept. According to Fig. 3, bPO = k1bHP tanβ, where β is the slope angle of straight line to the X axis. The parameters given in the first row of Table 3 were calculated from the temperature dependences of the rate constant k1 and the adsorption coefficients bHP and bPO in the coordinates of the Arrhenius and Van’t Hoff equations. The kinetics of side reactions (II)–(V) was studied in the same manner; the results from the
Table 3. Parameters of the Arrhenius and Van’t Hoff equations for propylene epoxidation Reaction
Rate constant*
Ea, J/mol
0 bHP , L/mol
0 bPO , L/mol
QHP, J/mol
QPO, J/mol
(I) (II) (III) (IV) (V)
3.21 × 102 1.40 × 102 2.21 × 102 2.91 × 102 5.11
46.36 × 103 61.92 × 103 63.86 × 103 62.28 × 103 50.19 × 103
4.82 × 10–4 4.96 × 10–4 5.80 × 10–4 4.29 × 10–4 5.70 × 10–4
1.93 × 10–4 1.93 × 10–3 1.69 × 10–3 1.87 × 10–3 1.62 × 10–3
18.34 × 103 18.56 × 103 17.76 × 103 19.09 × 103 19.45 × 103
21.06 × 103 20.76 × 103 20.86 × 103 19.87 × 103 20.46 × 103
*The constants k10, k 20, k30, k 40 are measured in L s–1 g–1, and the constant k50 is measured in mol s–1 g–1. KINETICS AND CATALYSIS
Vol. 57
No. 4
2016
KINETICS OF PROPYLENE EPOXIDATION WITH HYDROGEN PEROXIDE
471
Table 4. Kinetic parameters of Pr epoxidation Parameter
Value
Parameter
Value, J/mol
k10
2 (3.64 ± 0.44) × 10 L s–1 g–1
Ea1
(45.03 ± 3.15) × 10
k 20
2 (1.31 ± 0.11) × 10 L s–1 g–1
Ea2
(60.59 ± 4.54) × 10
3
k30
2 (2.35 ± 0.19) × 10 L s–1 g–1
Ea3
(63.45 ± 5.65) × 10
3
k 40
2 (2.66 ± 0.27) × 10 L s–1 g–1
Ea4
(62.97 ± 5.86) × 10
3
k50
(4.96 ± 0.35) mol s–1 g–1
Ea5
(49.81 ± 3.44) × 10
3
L/mol
QHP
(18.71 ± 0.86) × 10
3
L/mol
QPO
(20.48 ± 1.07) × 10
0 bHP
(5.00 ± 0.71) × 10
−4
0 bPO
(1.83 ± 0.15) × 10
−3
mathematical treatment of resulting data are also given in Table 3. Thus, the primary evaluation of the kinetic equations allowing describing the initial step of Pr epoxidation and corresponding side reactions was performed. The data obtained by an independent study of individual reactions validate the admitted assumptions and the trueness of the determined parameters. In order to refine the parameters of the kinetic model (1)–(5) and to assess whether it can be used for the description of Pr epoxidation up to high HP conversions, at the second step of operation in the integral flow-type reactor we performed several sets of experiments under different initial conditions at different conventional residence times. The experimental data obtained were processed mathematically to refine the parameters of the process. The earlier determined
3
constants given in Table 3 serves as the initial estimates of the corresponding parameters. The kinetic equations were integrated by the Euler method with varying step of integration, the constants of these equations were determined by the nonlinear Newton method minimizing the sum of squared deviations of the calculated and experimental concentrations. The results are given in Table 4. Statistical processing of the proposed kinetic model using the F-test showed that, at a significance level of 0.05, the model describes adequately the experimental data and allows predicting the composition of the reaction mixture in a wide variation range of different factors at different HP conversions (Figs. 4–9). In these figures, the lines correspond to Сi, mol/L 0.08
СHP, СPO, mol/L 1.2
3
0.06
PG 1MP2 2MP1
, , ,
, , ,
40°С 50°С 60°С
1.0 0.04
HP PO
0.8
, , , ,
0.6 0.4
30°С 40°С 50°С 60°С
0.02
0.2 0 0
20
40
60
80
100
τ × 10 , g s L –4
–1
Fig. 4. Concentrations of HP and PO as a function of the conventional residence time at different temperatures. 0 С Pr
0 = 3.54 mol/L, С Pr
0 С HP
= 2.75, and mcat = 15 g.
KINETICS AND CATALYSIS
Vol. 57
No. 4
2016
20
40
60
80
100
τ × 10 , g s L–1 –4
Fig. 5. Concentrations of by-products of epoxidation as a function of the conventional residence time at different 0 0 0 temperatures. С Pr = 3.54 mol/L, С Pr = 2.75, and С HP mcat = 15 g.
472
SULIMOV et al.
Сi, mol/L
СHP, СPO, mol/L
0.020
0 0 / C HP HP PO СPr
1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0
, , , ,
0
0
PG 1MP2 2MP1 CPr / С HP
1.99 3.02 4.14 5.04
0.016
, , ,
, , ,
3.02 4.14 5.04
0.012 0.008 0.004
20
40
60
80
100
20
0
40
60
τ × 10 , g s L –4
Fig. 6. Concentrations of HP and PO as a function of the conventional residence time at different initial concentra0 tions of HP. The temperature was 40°C, С Pr = 3.06 mol/L, and mcat = 15 g.
Fig. 7. Concentrations of by-products in epoxidation as a function of the conventional residence time at different initial concentrations of HP. The temperature was 40°C, 0 С Pr = 3.06 mol/L, and mcat = 15 g.
Сi, mol/L
1.0
0.020 PG 1MP2 2MP1
0.016
0.6
0 0 HP PO CPr / С HP
, , , ,
0.4 0.2
0
20
40
60
2.12 2.94 4.05 4.98 80
100 –4
СHP, СPO, mol/L
0.8
80
τ × 10 , g s L–1
–1
, , ,
, , ,
0 0 / СHP CPr
2.94 4.05 4.98
0.012 0.008 0.004
100
0
20
40
60
τ × 10–4, g s L–1 Fig. 8. Concentrations of HP and PO as a function of the conventional residence time at different initial concentra0 tions of Pr. The temperature was 40°C, С HP = 0.86 mol/L, and mcat = 15 g.
the calculation data and the points correspond to the experimental data. CONCLUSIONS As can be seen from Figs. 4–9, the calculated and experimental data are in good agreement. Consequently, the kinetic model (1)–(5) describes quite completely the real process, which is confirmed by the adequacy of the mechanism proposed for the liquidphase epoxidation of Pr with a solution of HP in the
80
100
τ × 10 , g s L–1 –4
Fig. 9. Concentrations of by-products of epoxidation as a function of the conventional residence time at different initial concentrations of Pr. The temperature was 40°C, 0 = 0.86 mol/L, and mcat = 15 g. С HP
presence of the granular titanium silicalite. The results of this work can be used for the engineering design of a tubular reactor for Pr epoxidation and for further optimization of the process using computer simulation of chemical-engineering systems. ACKNOWLEDGMENTS This work was supported by the Ministry of Education and Science of the Russian Federation (grant agreeKINETICS AND CATALYSIS
Vol. 57
No. 4
2016
KINETICS OF PROPYLENE EPOXIDATION WITH HYDROGEN PEROXIDE
ment no. 14.577.21.0093; unique identifier of applied scientific research (project), RFMEFI57714X0093). REFERENCES 1. Tsuji, J., Yamamoto, J., Ishino, M., and Oku, N., Sumitomo Kagaku, 2006, no. 1, p. 1. 2. Brasse, C., Degussa Sci. Newsl., 2004, no. 6, p. 12. 3. The Evonik-Uhde HPPO Technology, Press release, 2009. 4. Ashpina, O. and Kim, S., Chem. J., 2007, no. 10, p. 21. 5. Serebryakov, B.R., Masagutov, R.M., and Pravdin, V.G., Novye protsessy organicheskogo sinteza (New Organic Synthesis Processes), Moscow: Khimiya, 1989. 6. Shvets, V.F., Safin, D.Kh., and Petukhov, A.A., Khim. Prom–St. Segodnya, 2005, no. 8, p. 45. 7. Plate, N. and Slivinskii, E., Osnovy khimii i tekhnologii monomerov (Fundamentals of Monomer Chemistry and Technology), Moscow: Nauka, 2002. 8. Busygin, V.M., Safin, D.Kh., Ashikhmin, G.P., Puchkina, O.N., and Chebareva, A.I., Khim. Prom– St. Segodnya, 2005, no. 1, p. 21. 9. Clerici, M.G., Bellussi, G., and Romano, U., J. Catal., 1991, vol. 129, p. 159. 10. Bassler, P., Weidenbach, M., and Goebbel, H., Chem. Eng. Trans., 2010, vol. 21, p. 571. 11. Shin, S.B. and Chadwick, D., Ind. Eng. Chem. Res., 2010, vol. 49, p. 8125.
KINETICS AND CATALYSIS
Vol. 57
No. 4
2016
473
12. Russo, V., Tesser, R., Santacesaria, E., and Di Serio, M., Ind. Eng. Chem. Res., 2013, vol. 52, p. 1168. 13. Russo, V., Tesser, R., Santacesaria, E., and Di Serio, Ind. Eng. Chem. Res., 2014, vol. 53, p. 6274. 14. RF Patent 2422360, 2011. 15. US Patent 4410501, 1983. 16. Sulimov, A.V., Danov, S.M., and Ovcharov, A.A., Khim. Prom–st. Segodnya, 2010, no. 4, p. 17. 17. Sulimov, A.V., Danov, S.M., and Ovcharov, A.A., Khim. Interesah Ustoich. Razvit., 2012, no. 20, p. 693. 18. Sulimov, A.V., Danov, S.M., Krasnov, V.L., and Ovcharova, A.V., Russ. J. Phys. Chem. A, 2013, vol. 87, no. 11, p. 1809. 19. Kafarov, V.V., Metody kibernetiki v khimii i khimicheskoi tekhnologii (Methods of Cybernetics in Chemistry and Chemical Engineering), Moscow: Khimiya, 1985. 20. Bottino, A., Capannelli, G., and Comite, A., Chem. Eng. J., 2003, vol. 94, no. 1, p. 11. 21. Sridevi, U., Bhaskar Rao, B.K., and Pradhan, N.C., Chem. Eng. J., 2001, vol. 83, no. 3, p. 185. 22. Lebedev, N.N., Manakov, M.N., and Shvets, V.F., Teoriya khimicheskikh protsessov osnovnogo organicheskogo i neftekhimicheskogo sinteza (Theory of Chemical Processes in Basic Organic and Petrochemical Syntheses), Moscow: Khimiya, 1984. 23. Zhorov, Yu.M., Kinetika promyshlennykh organicheskikh reaktsii (Kinetics of Industrially Important Organic Reactions), Moscow: Khimiya, 1989.
Translated by K. Utegenov