Synthesis of dimethyl ether (DME) by catalytic distillation

Synthesis of dimethyl ether (DME) by catalytic distillation

Chemical Engineering Science 66 (2011) 3195–3203 Contents lists available at ScienceDirect Chemical Engineering Science journal homepage: www.elsevi...

703KB Sizes 0 Downloads 22 Views

Chemical Engineering Science 66 (2011) 3195–3203

Contents lists available at ScienceDirect

Chemical Engineering Science journal homepage: www.elsevier.com/locate/ces

Synthesis of dimethyl ether (DME) by catalytic distillation Zhigang Lei n, Zhiwu Zou, Chengna Dai, Qunsheng Li, Biaohua Chen State Key Laboratory of Chemical Resource Engineering, Beijing University of Chemical Technology, Box 266, Beijing 100029, PR China

a r t i c l e i n f o

abstract

Article history: Received 13 October 2010 Received in revised form 13 February 2011 Accepted 14 February 2011 Available online 22 February 2011

The intrinsic kinetics of liquid phase catalytic dehydration of methanol to dimethyl ether over a macroporous sulphonic acid ion exchange resin was determined in a fixed-bed micro-reactor in the temperature range of 391–423 K and pressures up to 2.0 MPa. The kinetic model based on Eley–Rideal mechanism, as well as the power-rate law model, was adopted for fitting the experimental data. However, the Langmuir–Hinshelwood mechanism is not feasible for describing the dehydration reaction of methanol, as deduced from the macroscopic kinetic data and/or no dependence of methanol conversion on initial methanol concentration in the absence of water at the inlet using acetone as inert solvent. Moreover, an improved process consisting of the combination of a fixed-bed reactor and a catalytic distillation column for the synthesis of DME (Process A) was proposed, and a mathematical model was established, into which the intrinsic kinetics obtained in this work was incorporated. The comparison of operating performance among the improved process, Process B consisting of a fixed-bed reactor and two ordinary distillation columns, and Process C consisting of a catalytic distillation column and an ordinary distillation column was also made. It was found that the improved process is more promising than others in energy consumption, production capacity and column number under the same product purity, and is easy to be implemented based on Process B that is currently used in the actual industrial plants with a long catalyst lifetime. & 2011 Elsevier Ltd. All rights reserved.

Keywords: Catalytic distillation Dimethyl ether (DME) Reaction mechanism Mathematical modeling Multiphase reactors Reaction engineering

1. Introduction In the recent years, dimethyl ether (DME) has been receiving a growing attention as a chemical product having potential advantages in the diminution of global environmental pollution and clean energy supply. At present, DME is used as an aerosol propellant to replace chlorofluorocarbons (CFCs), which was found to harm the ozone layer of the atmosphere (Fortunately, it was reported on September, 2010 that ozone layer depletion has been halted; see http://www.cosmosmagazine.com/news/3742/ ozone-layer-depletion-has-stopped-say-scientists) (Mollavali et al., 2008; Xu et al., 1997). Furthermore, DME has a wide range of applications as a high-quality household fuel in place of liquefied petroleum gas (LPG) because its physical properties are similar to LPG (Song et al., 2008; Varisli and Dogu, 2008). Due to its high cetane number (55–60), no release of sulfur compounds, lower NOx emission, less smoke and engine noise compared with those of traditional diesel fuels, DME is a clean alternative diesel fuel (Liu et al., 2010; Moradi et al., 2008; Ciftci et al., 2010). In addition, DME is also a key intermediate for producing many important chemicals such as dimethyl sulfate, methyl acetate and light olefins (Mao et al.,

n

Corresponding author. Tel.: +86 10 64433695; fax: + 86 10 64419619. E-mail addresses: [email protected], [email protected] (Z. Lei).

0009-2509/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2011.02.034

2006; Pop et al., 2009; Kim et al., 2008). Therefore, the production of DME is now required in large quantities. In general, DME is synthesized via methanol dehydration using solid acid catalysts such as g-alumina (g-Al2O3), modified g-Al2O3, H-ZSM-5, zeolites and ion exchange resins in a fixedbed reactor in a temperature range of 523–673 K and at pressures up to 10 bar, following by several ordinary distillation columns (Bercic and Levec, 1993; Yaripour et al., 2009), which is herein called Process B (see Fig. 1). The main reaction is considered and written as follows: 2CH3OH2CH3OCH3 +H2O

(1)

On the other hand, methanol conversion is limited in the conventional fixed-bed reactor because methanol dehydration is an equilibrium-controlled reaction. Therefore, An et al. (2004) proposed a single catalytic distillation column with pure methanol feeding for the synthesis of DME, which is called Process C together with an ordinary distillation column (see Fig. 1). It is well-known that catalytic distillation is a kind of multifunction reactor (Lei et al., 2005; Grosser et al., 1987; Ramzan et al., 2010), combining catalytic reaction and distillation in a single vessel. Compared with the standard two-step approach involving reaction and distillation, catalytic distillation offers a lot of advantages such as high conversion, high selectivity, energy savings, low consumption and simple operation (Yang et al., 2006; Kamath et al., 2006;

3196

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

needed to incorporate into the mathematical model. There are generally two kinds of mechanism models, i.e. Eley–Rideal mechanism and Langmuir–Hinshelwood mechanism. In order to decide which one is more suitable, we can deduce the most likely microscopic reaction mechanism from the macroscopic kinetic measurement by comparison of experimental data with the correlated values from the possible intrinsic kinetic models. Moreover, this work will go a further step to propose an improved process for the synthesis of DME, characterized by the combination of a fixed-bed reactor for partial reaction with a catalytic distillation column (see Fig. 1). But the catalytic distillation column, where the solid catalyst particles are often needed to pack into different types of catalyst envelopes (Ellenberger and Krishna, 1999; Xu et al., 1995), is crucial in the whole process. In this regard, a new-type structured catalytic packing which has been described in detail in our previous publication (Lei et al., 2009) can be adopted. Therefore, the objective of this work is to: (i) identify the suitable intrinsic kinetic model of catalytic dehydration of methanol to DME over the ion exchange resin from the viewpoint of chemical reaction engineering; (ii) establish the mathematical model of catalytic distillation column, in combination with the intrinsic kinetic model obtained in this work, and (iii) compare the operating performance among the Processes A, B and C and thus provide useful information for future application in the actual industrial plants.

2. Experimental section 2.1. Materials The catalyst used in the kinetic experiments was a kind of macroporous sulfonic acid ion exchange resin (bulk density of 637.0 kg m  3 and average particle size diameter of 0.60 mm, supplied by Wison International Corporation, China). Prior to use, the catalyst was dried in an oven drier at 353 K for 6 h to remove adsorbed water and then the dry weight was determined. Methanol (HPLC grade with 99.9 wt% purity) and acetone (anhydrous, 99.5 wt%) were purchased from Concord Technology Corporation, Ltd. (Tianjin, PR China) and Beijing Modern Eastern Fine Chemical Corporation, Ltd. (Beijing, PR China), respectively, and they were used as received. 2.2. Apparatus and procedure

Fig. 1. Schematic diagram of the improved process, Process A consisting of a fixedbed reactor and a catalytic distillation column (a), Process B consisting of a fixedbed reactor and two ordinary distillation columns (b) and Process C consisting of a catalytic distillation column and an ordinary distillation column (c).

Malone and Huss, 2003). As an important chemical integration technology, catalytic distillation has been industrialized in the production of methyl acetate, methyl tertiary butyl ether (MTBE), ethylene glycol (EG), and tertiary amyl methyl ether (TAME) (Popken et al., 2001; Bao et al., 2002; Ciric and Miao, 1994; Luyben, 2005). Therefore, the successful applications of catalytic distillation in other fields of chemical industry initiate a research program aiming at the synthesis of DME via catalytic distillation. However, the rigorous mathematical model for catalytic distillation column has to be established to understand the influence of operating and design parameters on the separation and reaction performance. For this purpose, the intrinsic kinetic model is

The schematic diagram of the experimental apparatus for measuring the intrinsic kinetics is illustrated in Fig. 2, which consists of four parts: a liquid feeding system, a preheater, a tubular fixed-bed micro-reactor (10 mm in inner diameter and 420 mm in length) made of stainless steel, and an analytical system. The fixed-bed micro-reactor was operated in continuous mode, which leads to some complexities in the data analysis but experimental data being measured steadily. In a typical experiment, 2.0 g of dry resin was loaded into the fixed-bed micro-reactor, both ends of which were filled with glass wool to support the catalyst bed. The feed (pure methanol, the mixture of methanol and water or the mixture of methanol and acetone) was injected by the syringe pump with a volumetric flowrate of 0.20 ml min  1 to the preheater before entering the reactor. Both the preheater and the reactor were maintained at a constant temperature using temperature controllers. The catalyst bed temperature was measured by means of a type-K thermocouple placed at the center of the catalyst bed with an accuracy of 70.1 K. The reactor outlet products passed through the condenser in order to separate the unconverted methanol and water

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

3197

the chemical and phase are in equilibrium at every stage of the reaction zone and the phase is in equilibrium at every nonreactive stage above and below the reaction zone (Baur et al., 2000; Taylor and Krishna, 2000; Smejkal et al., 2001). As for the fixed-bed reactor, a simpler model RGibbs module was used. The predictive UNIFAC model (Lei et al., 2008) was selected to account for the liquid phase non-idealities in the mixture in the whole process.

4. Results and discussion 4.1. Kinetic study

Fig. 2. Schematic diagram of experimental apparatus for the kinetic measurement of synthesis of DME from methanol. 1—nitrogen cylinder, 2—cylinder regulator, 3—check valve, 4—feed tank, 5—metering pump, 6—three-way valve, 7—preheater, 8—fixed-bed reactor, 9—temperature controller, 10—thermometer, 11—pressure gage, 12—condenser, 13—back pressure regulator, and 14—gas chromatograph.

from the product gases. A small portion of the condenser effluent was subjected to gas chromatograph (GC). The operating pressure of 2.0 MPa was set by the back pressure regulator to guarantee the liquid phase operation at each temperature. Nitrogen was used as a sweep gas.

4.1.1. Reaction network Under the experimental conditions, only the reaction given in Eq. (1) takes place with the products DME and water, and no sidereaction is expected. This was verified by the detection of FTIR (Fourier transform infrared) spectrometry (Tensor 27, Bruker, Germany) equipped with a CaF2 disk. As shown in Fig. 3, it can be seen that only a new characteristic peak near 1640 cm  1 appears at the outlet of fixed-bed micro-reactor and is assigned to the bending vibration of liquid water, while methanol and methanol plus 10% (mole fraction) acetone are used as reactants at the inlet. But the peaks of C–O in DME and methanol overlap near 1050 cm  1. Fig. 3a and b also shows that both water and acetone (as inert solvent) in the feeding stream have no influence on the reaction network concerned in this work.

2.3. Analysis The composition of the liquid mixture from kinetics experiments was analyzed using a gas chromatography (GC 4000 A Series) equipped with a thermal conductivity detector (TCD) and a Porapak Q column, 3 m in length and 3 mm in diameter. The carrier gas was hydrogen with a flowrate of 30 ml min  1. The detector and column temperatures were set to be 120 1C, and the injector temperature was 150 1C. The electric current of TCD was 130 mA. Methanol conversion is defined as XM ¼

Fin xM,in Fout xM,out  100% Fin xM,in

ð2Þ

where Fin and Fout are the molar flowrates of feeding and product streams, respectively, and xM is the mole fraction of methanol in the feeding stream xM,in or in the product stream xM,out, which are derived from the data obtained from gas chromatography.

3. Mathematical model Experimentally determined reaction kinetic data were used to develop the model for catalytic distillation column. The rigorous equilibrium stage (EQ) model RadFrac has been established to simulate the catalytic distillation column using the commercial simulation software Aspen Plus. The equations that model EQs are known as MESHR equations, into which the kinetic parameters obtained in this work are incorporated. MESHR is an acronym referring to the different types of equations. The M equations are the mass balance, E the phase equilibrium reactions, S the summation equations, H the enthalpy balance, and R the reaction rate equations. The catalytic distillation column is divided into three sections: a central reaction zone where the products are continuously removed from the reaction zone, and thus the reaction equilibrium limitation should be broken; an upper rectification zone where the DME product as light component is purified at the top; and a lower stripping zone where water or the mixture of water and methanol as heavy component is separated at the bottom. The model is based on the assumption that both

4.1.2. Mass transfer resistance The concentration at the reaction sites on the surface of the catalyst may be different from that in the bulk liquid phase because of the transport phenomena intrusions (Zhang and Datta, 1995; Fite et al., 1994). Therefore, it is necessary to verify that both external and internal mass transfer resistances are eliminated. To eliminate the influence of external diffusion, several tests were carried out at different methanol feeding flowrates, keeping liquid–solid contact time (i.e. the ratio of catalyst weight to methanol flowrate, W/F) constant. It was found that the external mass transfer resistance could be excluded when the volumetric flowrate reached 0.20 ml min  1 for 2.0 g of catalyst. Then the particle size of catalyst was varied to investigate the influence of internal diffusion. The measured methanol conversions showed that when the ion exchange resin particle size is in the range of 0.30–0.90 mm (i.e. 20–60 mesh), the influence of internal mass transfer on the kinetics of the studied system is not significant. The plug flow check was done in order to avoid the contribution of axial mixing to mass transport by means of the Mears’ criterion with both the ratio of catalyst bed height to catalyst particle diameter (L/dp)450 (63.3 in this work) and the ratio of tube internal diameter to catalyst particle diameter (D/dp)4 10 (16.7 in this work) (Rase, 1977). Therefore, it can be assumed that the fixed-bed micro-reactor was operated in the manner of plug flow. 4.1.3. Intrinsic kinetics A series of experiment work was performed to examine the effect of such factors as methanol concentration and temperature on methanol conversion. Firstly, kinetic experiments were carried out to determine the dependence of methanol conversion on initial methanol concentration in the absence of water at the temperature of 423 K. The initial mole fraction of methanol varied from 0.20 to 1.00 using acetone as inert solvent. As shown in Fig. 4, it can be seen that methanol conversion has almost no dependence on initial methanol concentration in the absence of water at the inlet.

3198

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

Fig. 4. Effect of initial methanol concentration on methanol conversion in the absence of water at the inlet using acetone as inert solvent.

Fig. 5. Influence of temperature on methanol conversion at different initial mole fractions of methanol: (’) 391 K; (K) 399 K; (m) 407 K; (~) 415 K; and (J) 423 K.

Fig. 3. FTIR (Fourier transform infrared) spectrometry of methanol (a), methanol plus 10% (mole fraction) water (b), methanol plus 10% (mole fraction) acetone (c) as reactants at the inlet of fixed-bed micro-reactor, and their corresponding products at the outlet at 423 K.

Then, kinetic experiments were carried out at the temperatures of 391, 399, 407, 415 and 423 K in the fixed-bed microreactor. At each temperature, five experimental series, in which the initial mole fraction of methanol varied from 0.80 to 1.00 in methanol/water mixture, were performed. The experimental results are shown in Fig. 5. It can be seen that methanol conversion strongly depends on the operating temperature. Methanol conversion is the lowest at 391 K, while it increases rapidly when the reaction temperature rises from 391 to 423 K. Water inhibits the catalytic dehydration of methanol to DME.

Therefore, increasing the initial mole fraction of methanol will result in a higher methanol conversion. But the inhibiting effect of water on the methanol conversion is slightly weakened at high temperatures. Experiments at 2.0, 2.5 and 3.0 MPa showed that there is no pressure dependence for methanol conversion because in this case all reactants are in the liquid state. Besides, no catalyst activity loss was observed during the whole experiments.

4.1.4. Kinetic model The intrinsic kinetics of methanol dehydration has ever been investigated over a series of solid acid catalysts like acidic zeolites (Raoof et al., 2008; Jiang et al., 2004; Yaripour et al., 2005; Blaszkowski and van Santen, 1996). However, ion exchange resins are preferred over acidic zeolites in that they require relatively low temperatures and thus the temperatures of reaction and distillation can match each other in the catalytic distillation column. However, few works on the ion exchange resins as catalysts were published except that An et al. (2004) reported

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

two intrinsic rate equations in the form of rD ¼

2 kCM ðKW =KM ÞCW þ CM

B0 ¼ ð3Þ

and rD ¼

2 kCM

ð4Þ

ððKW =KM ÞCW þCM Þ2

which is derived based on the Eley–Rideal and Langmuir– Hinshelwood mechanisms, respectively. Both are assumed that surface reaction is the controlling step, and DME does not adsorb on the surface because adsorption of the more polar components (e.g. methanol and water) is much stronger than that of the less polar components (e.g. DME). In the Eley–Rideal mechanism, methanol adsorbed at catalytic active sites reacts with methanol from the bulk liquid phase to yield DME and water, while in the Langmuir–Hinshelwood mechanism, the surface reaction takes place between two methanol molecules adsorbed at catalytic active sites. The temperature dependencies of rate constant and adsorption equilibrium constants were expressed by the Arrhenius and Van’t Hoff equations, respectively:   Ea k ¼ A0 exp  , ð5Þ RT   DHi , Ki ¼ Ai0 exp  RT

i ¼ M,W

ð6Þ

In this work, one-dimensional heterogeneous plug flow model was employed to describe the experimental fixed-bed microreactor, and a differential mole balance on the catalyst bed yielded FM0

dXM ¼ 2rD dW

ð7Þ

subject to XM ¼0 at W¼0. Depending on the form of rD, different integral forms of Eq. (7) may be derived. The integral expressions corresponding to Eqs. (3) and (4) are, respectively, given as follows:   2CM0 W 1 KW CM0 þ 2CW0 1 1 ¼ 1 FM0 lnð1XM Þ k 2KM 1XM lnð1XM Þ CM0   1 KW þ 1 ð8Þ k 2KM 2

2

 1

2CM0 W 1 KW ðCM0 þ 2CW0 Þ 1 ¼ FM0 k 2KM 1XM CM0   1 KW KW 1 ðCM0 þ 2CW0 Þlnð1XM Þ þ k KM 2KM  2 1 KW þ 1 CM0 XM k 2KM which, in an easy-to-understand form, can be rewritten as   2CM0 W CM0 þ 2CW0 1 1 ¼A þB 1 FM0 lnð1XM Þ 1XM lnð1XM Þ CM0

   2 1 KW KW 1 KW 1 and C 0 ¼ 1 : k KM 2KM k 2KM

The resulting terms except A, B, A0 , B0 and C0 could be derived from the experimental data, and their values were fitted to determine A, B, A0 , B0 and C0 (from which k and KW/KM were derived directly) at each temperature using multiple linear regression (MLR). But for Eq. (11) it was found that the coefficients of the first and third terms on the right-hand (i.e. A0 and C0 ) are not kept positive and thus the correlation results are meaningless. Moreover, it is straightforward from Eq. (10) that methanol conversion has no dependence on initial methanol concentration CM0 at a given CM0/FM0 ( ¼5.0 min ml  1) in the absence of water at the inlet (i.e. CW0 ¼0), which is consistent with the trend shown in Fig. 4, while Eq. (11) does reflect the dependence. Therefore, the Langmuir–Hinshelwood mechanism is contradictory with the experimental data and cannot be used for describing the catalytic dehydration of methanol to DME. For Eley–Rideal model, an activation energy of 68.7 kJ mol  1 and a pre-exponential factor A0 ¼120.7 m3 kgcat  1 s  1 were obtained from the fitness of the Arrhenius plot (see Fig. 6). This activation energy is a little higher than that of Amberlyst 35 (51.7 kJ mol  1), indicating that Amberlyst 35 is more active for the synthesis of DME. According to Eq. (6), the linear regression of ln(KW/KM) versus 1/T gives the following expression:   KW 14439 ð12Þ ¼ exp 33:85 þ T KM The ratio of adsorption equilibrium constant of water to methanol is 21.7 at 391 K. This means that water has an advantage over methanol in the competition of adsorption at catalytic sites on the surface of ion exchange resin. But this advantage is not so obvious at high temperatures because the ratio of adsorption equilibrium constant of water to methanol reduces to 1.3 at 423 K. On the other hand, the empirical power-rate law model is also used to correlate the experimental data since its form is simple and the model parameters can be directly input to the modern chemical process simulation software such as Aspen Plus, PROII, CFX, Fluent, etc. It can be written as m n rD0 ¼ k0 CM CW

  E0 k0 ¼ A00 exp  a RT

ð13Þ ð14Þ

The DME concentration does not appear in Eq. (13) due to the high volatility of DME, and thus it is very low in the liquid phase.

and 

3199

ð9Þ

ð10Þ

and   2CM0 W ðCM0 þ 2CW0 Þ2 1 ¼ A0 1 FM0 1XM CM0 þ B0 ðCM0 þ 2CW0 Þlnð1XM Þ þ C 0 CM0 XM respectively, where   1 KW 1 KW , B¼ 1 , A¼ k 2KM k 2KM

A0 ¼

  1 KW 2 , k 2KM

ð11Þ

Fig. 6. Arrhenius’s plot and ratio of adsorption equilibrium constants of water to methanol as a function of temperature.

3200

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

Table 1 Optimized specifications and operating conditions for Processes A, B and C. Contents

Specifications and conditions Process A

Process B

Process C

Feed stream Temperature (K) Pressure (MPa) Flowrate (mol s  1)

298 2.0 2.5

298 2.0 2.5

298 0.9 2.5

Composition (mole fraction) Methanol Water DME

1.0 0 0

1.0 0 0

1.0 0 0

RGibbs (fixed-bed reactor) Temperature (K) Pressure (MPa)

403 2.0

403 2.0

RadFrac (catalytic distillation column) Total stages 15 Rectification zone 1–2 Reactive zone 3–9 Stripping zone 10–15 Feed stage 3 Catalyst loading (kg/stage) 15.0 Column pressure (MPa) 0.9 Reflux ratio (mole) 3.5 1 Distillate rate (kmol h ) 4.50 RadFrac (ordinary distillation column) Total stages Feed stage Column pressure (MPa) Reflux ratio (mole) Distillate rate (kmol h  1)

Fig. 7. Comparison of methanol conversion between the experimental data (Xexp) and the calculated results (Xcal) from the Eley–Rideal model (a) and power-rate law model (b).

The Marquardt method as in Press et al. (1992) was used for data correlation, and the estimated model parameters are: A00 ¼ 5.19  109 m3 kgcat  1 s  1, E0a ¼133.8 kJ mol  1, m¼1.51 and n ¼  0.51. Fig. 7 shows the comparison of methanol conversion between experimental data and calculated results from the Eley–Rideal and power-rate law models. It can be seen that both agree very well. The average relative deviations (ARD) for methanol conversion are 5.33% and 9.03% for the Eley–Rideal and power-rate law models, respectively. But it seems that the Eley–Rideal model obtained in this work is more accurate than the power-rate law model, and thus incorporated into the mathematical model of catalytic distillation column. Therefore, we focused on the Eley– Rideal model rather than the power-rate law model in the simulation.

4.2. Comparison among Processes A, B and C The optimized specifications and operating conditions for Processes A, B and C are listed in Table 1. An optimized set of input parameters for Aspen Plus software (version 11.1) in the simulation was determined using Design Specs module of Aspen Plus with the target of 99.55% of DME product purity and 99.50%

30 1–7 8–25 26–30 8 15.0 0.9 3.5 2.245 No. 1

No. 2

15 8 0.9 3.5 4.20

15 10 0.1a 5 0.60

15 10 0.1 5 4.51

of final water purity at the bottom, and the optimized parameters are consistent among these three processes. As shown in Fig. 1, pure methanol was fed into the fixed-bed reactor, where the chemical reaction is assumed to reach equilibrium (in Processes A and B). The equilibrium mixture or pure methanol was sent to the upper part of the reaction zone of the catalytic distillation column (in Processes A and C). It was assumed that pressure drop along the column is within the permitted operating region. The catalytic distillation column (in Process A) or the ordinary distillation columns (in Processes B and C) consists of 15 stages including a total condenser and a reboiler, while the catalytic distillation column in Process C has 30 stages. The stages are numbered from the top to the bottom, with stage 1 as the condenser and stage 15 as the reboiler. Fig. 8 shows the profiles of temperature, composition in the vapor and liquid phases along the catalytic distillation column for Process A (i.e. the improved process). DME is collected as the top product, while water is collected as the bottom product. The profiles of composition in the vapor and liquid phases exhibit the similar trend. Within the reaction zone, there is an extremum point of the mole fraction of methanol, but methanol is finally converted into DME and water. Meanwhile, the mole fraction of DME in the liquid phase is the lowest because of the volatilization from liquid to vapor phases. As a result, the equilibrium limitation in Eq. (1) is overcome and thus methanol conversion increases. But there is an abrupt change of temperature near the stage no. 4, which means that it is a sensitive plate and should be paid more attention in the design and control of catalytic distillation column. Comparison of the operating performance among Processes A, B and C is given in Table 2. It can be seen that compared with Process B, the heat duty of Process A on reboiler and condenser

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

3201

Table 2 Comparison of the operating performance among Processes A, B and C using the mathematical model established in this work. Contents DME product purity (mol%) DME mass flowrate (kg h  1) Total methanol conversion (%) Water purity at the bottom (mol%)

per unit amount of DME decreases 26.22% and 26.85%, respectively, while compared with Process C, it decreases 87.66% and 84.06%, respectively. Moreover, under the same DME product

Process B

Process C

99.55 206.27 99.55 99.50

99.55 192.62 92.91 99.50

99.55 102.96 49.66 99.50

24.42  100.27 78.23

24.42  127.93 98.96

0.00  313.85 316.33

Heat duty per unit amount of DME (kJ kg  1) Fixed-bed reactor 425.97 Condenser  1749.06 Reboiler 1364.61

456.40  2390.97 1849.53

0.00  10,973.78 11,060.49

Heat duty (kW) Fixed-bed reactor Condenser Reboiler

Fig. 8. Profiles for temperature (a), vapor phase composition (b) and liquid phase composition (c): (J) temperature along the catalytic distillation column; (K) methanol; (m) DME; and (’) water.

Process A

purity, production capacity and total methanol conversion in Process A is the highest, but the column number is the least. Therefore, Process A is superior to Processes B and C. The reason is attributed to two aspects: one is that the dehydration of methanol is an exothermic reaction, and the reaction heat is given off for the distillation in the catalytic distillation column; the other is that more amount of DME can be produced from both the fixed-bed reactor and catalytic distillation column. Application of catalytic distillation technology offers advantages in increasing quality and quantity of product, while the fixed-bed reactor contributes to a large part of methanol conversion (up to 92.91% methanol conversion). Another advantage of Process A is that the acidic ion exchange filled in fixed-bed reactor serves not only as catalyst for the reaction, but also as a guard bed for metallic ions contained in the feeding stream which could otherwise deactivate the catalyst in the catalytic distillation column, since the replacement of catalyst material in the fixed-bed reactor is less time-consuming and expensive than the replacement within the catalytic distillation column. That is to say, fixed-bed reactor and catalytic distillation column never repel each other, and should be utilized together. In particular, almost all of methanol in the feeding is completely converted (up to 99.55%) in the improved process so that the ordinary distillation column for separating methanol and water is not required. However, in principle, the combination of fixed-bed reactor and catalytic distillation column can be replaced by only a single catalytic distillation column. In this case only a single catalytic distillation column in Process C is adequate for the synthesis of DME with high-quality and quantity. It was found that when the number of theoretical stages in reaction zone increases from 18 to 37, Process C consisting of only a catalytic distillation column has almost the same DME product purity, DME quantity and heat duty per unit amount of DME as does Process A. But the structured catalytic packings containing solid catalyst particles loaded into the reaction zone can bring about an one or two orders of magnitude higher pressure drop per meter than common structured packings only for distillation, and their normal operation region with respect to liquid and vapor velocities becomes narrower (Hoffmann et al., 2004; Li et al., 2009). The number of structured catalytic packings should be restricted in reaction zone. Otherwise, liquid flood may be induced in the column. Therefore, Process A can be operated in a safer mode than Process C consisting of only a catalytic distillation column, although extra fixed-bed reactor, which will counteract some height of the reaction zone of catalytic distillation column, is added.

3202

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

5. Conclusions Since DME has been receiving a growing attention all over the world in the recent years, the study on the synthesis of DME is important and urgent from the viewpoint of chemical reaction engineering. In the engineering aspect, an improved process consisting of a fixed-bed reactor and a catalytic distillation column was proposed in this work. Based on the experimentally determined kinetic data, the rigorous equilibrium (EQ) stage model RadFrac was used to simulate the catalytic distillation column. It was found that the improved process is more promising in energy consumption, production capacity and column number than the existing technology. However, it should be mentioned that there is no extra streams and equipments added when revamping Process B that is currently used in industry. The only change is that the traditional structured distillation packings in the middle section of the first ordinary distillation column are replaced by structured catalytic packings, and the second ordinary distillation column may be shut down. But if the first ordinary distillation column in Process B is replaced by a single catalytic distillation column having rectification zone, reaction zone and stripping zone, the height of reaction zone may be not enough to achieve a high methanol conversion and in this case the fixed-bed reactor plays a decisive role in increasing methanol conversion and relieving the burden of catalytic distillation column. Therefore, the improved process is easy to be implemented in the actual industrial plants. In the science aspect, the kinetic model for dehydration of methanol to DME over a macroporous sulphonic acid ion exchange resin was established based on the Eley–Rideal mechanism and the empirical power-rate law model. Moreover, it was found that the Langmuir–Hinshelwood mechanism is not feasible for describing the dehydration reaction of methanol, as deduced from the macroscopic kinetic data and/or no dependence of methanol conversion on initial methanol concentration in the absence of water at the inlet using acetone as inert solvent. In addition, methanol conversion strongly depends on the reaction temperature and initial concentration of methanol in the presence of water which, however, inhibits the dehydration of methanol.

Nomenclature A0 Ci D dp Ea F DHi k Ki L m, n P r R T W xi Xi

pre-exponential factor (m3 kgcat  1 s  1) concentration of component i (mol m  3) internal diameter of the tube (mm) catalyst particle diameter(mm) activation energy (kJ mol  1) molar flowrate of methanol (mol s  1) heat of adsorption of component i (kJ mol  1) reaction rate constant (m3 kgcat  1 s  1) adsorption equilibrium constant of component (m3 mol  1) height of the catalyst bed (mm) reaction order (dimensionless) operating pressure (Pa) reaction rate (mol kgcat  1 s  1) universal gas constant (J mol  1 K  1) reaction temperature (K) catalyst mass (kg) mole fraction of component i (dimensionless) conversion of component i (dimensionless)

Subscripts 0 cal

inlet conditions calculated values

i

D exp in M out W

dimethyl ether experimental data feeding stream methanol product stream water

Acknowledgements This work is financially supported by the National Nature Science Foundation of China under Grant (Nos. 20821004 and 21076008), and the Fundamental Research Funds for the Central Universities.

References An, W.Z., Chuang, K.T., Sanger, A.R., 2004. Dehydration of methanol to dimethyl ether by catalytic distillation. The Canadian Journal of Chemical Engineering 82, 948–955. Bao, J., Gao, B.L., Wu, X.Q., Yoshimoto, M., Nakao, K., 2002. Simulation of industrial catalytic-distillation process for production of methyl tert-butyl ether by developing user’s model on Aspen plus platform. Chemical Engineering Journal 90, 253–266. Baur, R., Higler, A.P., Taylor, R., Krishna, R., 2000. Comparison of equilibrium stage and nonequilibrium stage models for reactive distillation. Chemical Engineering Journal 76, 33–47. Bercic, G., Levec, J., 1993. Catalytic dehydration of methanol to dimethyl ether. kinetic investigation and reactor simulation. Industrial and Engineering Chemistry Research 32, 2478–2484. Blaszkowski, S.R., van Santen, R.A., 1996. The mechanism of dimethyl ether formation from methanol catalyzed by zeolitic protons. Journal of the American Chemical Society 118, 5152–5153. Ciftci, A., Sezgi, N.A., Dogu, T., 2010. Nafion-incorporated silicate structured nanocomposite mesoporous catalysts for dimethyl ether synthesis. Industrial and Engineering Chemistry Research 49, 6753–6762. Ciric, A.R., Miao, P.Z., 1994. Steady state multiplicities in an ethylene glycol reactive distillation column. Industrial and Engineering Chemistry Research 33, 2738–2748. Ellenberger, J., Krishna, R., 1999. Counter-current operation of structured catalytically packed distillation columns: pressure drop, holdup and mixing. Chemical Engineering Science 54, 1339–1345. Fite, C., Iborra, M., Tejero, J., Izquierdo, J.F., Cunill, F., 1994. Kinetics of the liquidphase synthesis of ethyl tert-butyl ether (ETBE). Industrial and Engineering Chemistry Research 33, 581–591. Grosser, J.H., Doherty, M.F., Malone, M.F., 1987. Modeling of reactive distillation systems. Industrial and Engineering Chemistry Research 26, 983–989. Hoffmann, A., Noeres, C., Gorak, A., 2004. Scale-up of reactive distillation columns with catalytic packings. Chemical Engineering and Processing 43, 383–395. Jiang, S., Hwang, J.S., Jin, T.H., Cai, T.X., Cho, W., Baek, Y.S., Park, S.E., 2004. Dehydration of methanol to dimethyl ether over ZSM-5 zeolite. Bulletin of the Korean Chemical Society 25, 185–189. Kamath, R.S., Qi, Z.W., Sundmacher, K., Aghalayam, P., Mahajani, S.M., 2006. Comparison of reactive distillation with process alternatives for the isobutene dimerization reaction. Industrial and Engineering Chemistry Research 45, 2707–2714. Kim, S.M., Lee, Y.J., Bae, J.W., Potdar, H.S., Jun, K.W., 2008. Synthesis and characterization of a highly active alumina catalyst for methanol dehydration to dimethyl ether. Applied Catalysis A: General 348, 113–120. Lei, Z.G., Chen, B.H., Ding, Z.W., 2005. Special Distillation Processes. Elsevier, Amsterdam. Lei, Z.G., Chen, B.H., Li, C.Y., Liu, H., 2008. Predictive molecular thermodynamic models for liquids, solid salts, polymers, and ionic liquids. Chemical Reviews 108, 1419–1455. Lei, Z.G., Yang, Y., Li, Q.S., Chen, B.H., 2009. Catalytic distillation for the synthesis of tert-butyl alcohol with structured catalytic packing. Catalysis Today 147, 352–356. Li, Q., Chang, Q.L., Tian, Y.M., Liu, H., 2009. Cold model test and industrial applications of high geometrical area packings for separation intensification. Chemical Engineering and Processing 48, 389–395. Liu, J., Liu, S.H., Li, Y., Wei, Y.J., Li, G.L., Zhu, Z., 2010. Regulated and nonregulated emissions from a dimethyl ether powered compression ignition engine. Energy and Fuels 24, 2465–2469. Luyben, W.L., 2005. Comparison of pressure-swing and extractive-distillation methods for methanol-recovery systems in the TAME reactive-distillation process. Industrial and Engineering Chemistry Research 44, 5715–5725. Malone, M.F., Huss, R.S., 2003. Green chemical engineering aspects of reactive distillation. Environmental Science and Technology 37, 5325–5329.

Z. Lei et al. / Chemical Engineering Science 66 (2011) 3195–3203

Mao, D.S., Yang, W.M., Xia, J.H., Zhang, B., Lu, G.Z., 2006. The direct synthesis of dimethyl ether from syngas over hybrid catalysts with sulfate-modified galumina as methanol dehydration components. Journal of Molecular Catalysis A: Chemical 250, 138–144. Mollavali, M., Yaripour, F., Atashi, H., Sahebdelfar, S., 2008. Intrinsic kinetics study of dimethyl ether synthesis from methanol on g-Al2O3 catalysts. Industrial and Engineering Chemistry Research 47, 3265–3273. Moradi, G., Ahmadpour, J., Nazari, M., Yaripour, F., 2008. Effects of feed composition and space velocity on direct synthesis of dimethyl ether from syngas. Industrial and Engineering Chemistry Research 47, 7672–7679. Pop, G., Bozga, G., Ganea, R., Natu, N., 2009. Methanol conversion to dimethyl ether over H-SAPO-34 catalyst. Industrial and Engineering Chemistry Research 48, 7065–7071. Popken, T., Steinigeweg, S., Gmehling, J., 2001. Synthesis and hydrolysis of methyl acetate by reactive distillation using structured catalytic packing: experiments and Simulation. Industrial and Engineering Chemistry Research 40, 1566–1574. Press, W.H., Flannery, B.P., Teukolsky, S.A., Vetterling, W.T., 1992. Numerical Recipes in FORTRAN: The Art of Scientific Computing, 2nd. Cambridge University Press, Cambridge, England. Ramzan, N., Faheem, M., Gani, R., Witt, W., 2010. Multiple steady states detection in a packed-bed reactive distillation column using bifurcation analysis. Computers and Chemical Engineering 34, 460–466. Raoof, F., Taghizadeh, M., Eliassi, A., Yaripour, F., 2008. Effects of temperature and feed composition on catalytic dehydration of methanol to dimethyl ether over g-alumina. Fuel 87, 2967–2971. Rase, H.F., 1977. Chemical Reactor Design for Process Plants: Principles and Techniques.. John Wiley & Sons, New York.

3203

Smejkal, Q., Hanika, J., Kolena, J., 2001. 2-Methylpropylacetate synthesis in a system of equilibrium reactor and reactive distillation column. Chemical Engineering Science 56, 365–370. Song, D., Cho, W., Lee, G., Park, D.K., Yoon, E.S., 2008. Numerical analysis of a pilotscale fixed-bed reactor for dimethyl ether (DME) synthesis. Industrial and Engineering Chemistry Research 47, 4553–4559. Taylor, R., Krishna, R., 2000. Modelling reactive distillation. Chemical Engineering Science 55, 5183–5229. Varisli, D., Dogu, T., 2008. Production of clean transportation fuel dimethylether by dehydration of methanol over Nafion catalyst. Gazi University Journal of Science 21, 37–41. Xu, M.T., Lunsford, J.H., Goodman, D.W., Bhattacharyya, A., 1997. Synthesis of dimethyl ether (DME) from methanol over solid-acid catalysts. Applied Catalysis A: General 149, 289–301. Xu, X., Zheng, Y.X., Zheng, G.W., 1995. Kinetics and effectiveness of catalyst for synthesis of methyl tert-butyl ether in catalytic distillation. Industrial and Engineering Chemistry Research 34, 2232–2236. Yang, B.L., Wu, J., Zhao, G.S., Wang, H.J., Lu, S.Q., 2006. Multiplicity analysis in reactive distillation column using ASPEN PLUS. Chinese Journal of Chemical Engineering 14, 301–308. Yaripour, F., Baghaei, F., Schmidt, I., Perregaard, J., 2005. Synthesis of dimethyl ether from methanol over aluminium phosphate and silica–titania catalysts. Catalysis Communications 6, 542–549. Yaripour, F., Mollavali, M., Jam, S.H., Atsshi, H., 2009. Catalytic dehydration of methanol to dimethyl ether catalyzed by aluminum phosphate catalysts. Energy and Fuels 23, 1896–1900. Zhang, T.J., Datta, R., 1995. Integral analysis of methyl tert-butyl ether synthesis kinetics. Industrial and Engineering Chemistry Research 34, 730–740.