Dimethyl ether synthesis in a multi-stage fluidized bed reactor

Dimethyl ether synthesis in a multi-stage fluidized bed reactor

Accepted Manuscript Title: Dimethyl ether synthesis in a multi-stage fluidized bed reactor Author: M.E.E. Abashar PII: DOI: Reference: S0255-2701(17)...

855KB Sizes 10 Downloads 113 Views

Accepted Manuscript Title: Dimethyl ether synthesis in a multi-stage fluidized bed reactor Author: M.E.E. Abashar PII: DOI: Reference:

S0255-2701(17)30117-4 https://doi.org/10.1016/j.cep.2017.09.018 CEP 7082

To appear in:

Chemical Engineering and Processing

Received date: Revised date: Accepted date:

9-2-2017 17-7-2017 30-9-2017

Please cite this article as: M.E.E.Abashar, Dimethyl ether synthesis in a multi-stage fluidized bed reactor, Chemical Engineering and Processing https://doi.org/10.1016/j.cep.2017.09.018 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Dimethyl ether synthesis in a multi-stage fluidized bed reactor

M. E. E. Abashar Department of Chemical Engineering, College of Engineering, King Saud University, PO Box 800, Riyadh 11421, Saudi Arabia Fax: + 966-1-4678770 E-mail address: [email protected] Graphical abstract

   

Highlights Multi-stage fluidized bed reactors are efficient for DME production. The direct DME synthesis process is simulated. The bifunctional catalyst optimal composition is identified for each bed. DME yield and selectivity are studied for each configuration.

1

Abstract The two-phase model with verified reaction kinetics has been implemented to study numerically dimethyl ether (DME) production in a novel multi-stage fluidized bed catalytic reactor. The single-bed configuration results show that the bifunctional catalyst composition [CuO–ZnO–Al2O3/ HZSM-5] has a pronounced impact on enhancing the reactor performance. Also, the simulation results indicated that the two-bed configuration gives a substantial enhancement of DME yield of 32.67% compared to the single-bed configuration. This yield improvement is also accompanied by significantly improved DME selectivity of almost 100%. The complex interactions of the bifunctional catalyst composition, feed composition and multistage configuration play a central role in the extent of the synergy. Interesting two types of maximum phenomena are observed and explanations have been offered. It seems that synergistic effects have kinetics and thermodynamic strong impact on the development of these phenomena. The

loci of these maxima are significant for optimal design and control of multi-bed configurations. It appears that application of multi-stage fluidized bed catalytic reactors is highly promising in DME industry.

Keywords: Dimethyl ether; Fluidized bed reactor; Modeling; Multi-stage; Simulation

Nomenclature A

cross-sectional of the reactor, m2

Ab

cross-sectional occupied by the bubble phase, m2

Ar

Archimedes number, [-]

Cib, Cid concentration of component i in the bubble and dense phases respectively, kmol m-3 CiF

concentration of component i in the feed, kmol m-3

db

bubble diameter, m

dbm

maximum bubble diameter, m

2

dbo

bubble diameter just above the distributor, m

dP

mean particle diameter, m

D

reactor diameter, m

Dij

molecular diffusivity of component i in component j, m2 s-1

Dij

molecular diffusivity of component i in component j, m2 s-1

Dim

molecular diffusivity of component i in gas mixture, m2 s-1

Fi

molar flow rate of component i in the feed, kmol s-1

g

acceleration of gravity, m s-2

Gb, Gd gas volumetric flow rate in the bubble and dense phase respectively, m3 s-1 h

bed height, m

H

total bed height, m

Hmf

bed height at minimum fluidizing condition, m

-Hi heat of reaction i, kJ mol –1 k1

reaction rate constant of reaction (1), kmol bar-2 kg-1 s-1

k2

reaction rate constant of reaction (2), kmol bar-1 kg-1 s-1

k3

reaction rate constant of reaction (3), kmol bar-1 kg-1 s-1

KCO2, KCO, KH2, adsorption constant for H2 , CO2 and CO respectively, bar-1 K1

equilibrium constant of reaction (1), bar-2

K2

equilibrium constant of reaction (2), [-]

K3

equilibrium constant of reaction (3), bar

(Kbc)i

mass transfer coefficient of component i (bubble phase-cloud phase), s-1

(Kcd)i

mass transfer coefficient of component i (cloud phase-dense phase), s-1

(Kbd)i overall mass transfer coefficient of component i (bubble phase-dense phase) , s-1 Mi

molecular mass of component i, kg/kmol

Pi

partial pressure of component i, bar

3

P

total pressure , bar

R

ideal gas constant, kJ/kmol K

Ri

rate of reaction i , kmol kg-1s-1

T

temperature, K

Ub

velocity of a bubble rising through a bed, m s-1

Uo

superficial gas velocity, m s-1

Umf

superficial gas velocity at minimum fluidizing conditions, m s-1

yi

mole fraction of component i, [-]

Yi

dimensionless concentration of component i, [-]

Z

dimensionless reactor height, [-]

Greek symbols γij

stiochiometric coefficient of component i in the jth reaction

εmf

bed voidage at minimum fluidizing condition, [-]

δ

volume fraction of the catalyst bed occupied by the bubble phase respectively, [-]

θ1, θ2 mass fraction of catalyst 1 and 2 respectively, [-] µg

gas viscosity, kg m-1s-1

ρg

gas density, kg m-3

ρP

solid particle density, kg m-3

1. Introduction Today, the world faces great challenges of energy and environment. These topics are the most concerning problems for scientists, engineers and industry. Now, it becomes evident that dimethyl ether (DME) as a synthetic fuel has a promising potential to substitute the automobiles diesel fuel owing to its attractive characteristics such as [1–7]: it is an oxygenated simple ether (CH3OCH3, 34.78 wt% O2) with a high cetane number (50-

4

60), a clean fuel free from SOx, less engine noise, easy to store and transport. Also, can be used as an alternative fuel for the liquefied petroleum gases (LPG) because its physical properties are almost similar to the LPG. DME has considerable potential to be used as a fuel for fuel cells. Furthermore, DME can be used as a refrigerant and an aerosol propellant instead of ozone harmful chemicals such as chlorofluorocarbons (CFCs). In the chemical industry, DME can be utilized as a feedstock to produce many important chemicals such as hydrogen, methyl acetate, acetic acid, light olefins and dimethyl sulfate. The increased worldwide demand for DME has imposed a wide range of challenges on academia and industry to improve the DME existing technology. Among the key challenges are the reactor design, catalysts and kinetics. In fact, many big companies such as Total, Amoco, Mitsubishi Co, Haldor Topsøe and Navistar International Corp. are interested in the development of DME production technology [3,8]. Two processes are used to produce DME, the indirect and direct process [9]. The indirect process is conventionally used to produce DME and called a two-step process. The first step of this route involves methanol synthesis from syngas. The syngas is usually obtained from natural gas, coal or biomass[10]. However, the natural gas is still the predominant source of the syngas due to its abundance availability. A metallic Cu-based catalyst such as CuO–ZnO– Al2O3 is employed in this step [1,5,11–14]. Two catalytic reactions take place as follows:

Methanol synthesis: CO2  3H 2

C H 3O H  H 2 O

[  H 2 9 8   4 9 .5 0 k J / m o l ]

(1)

Water-gas shift reaction (WGS): CO

 H 2O

H 2  CO2

[  H 2 9 8   4 1 .1 3 k J / m o l ]

5

(2)

And, the net rate of reaction is given by: CO

 2H 2

[  H 2 9 8   9 0 .6 3 k J / m o l ]

C H 3O H

(N1)

It is obvious that, the reactions are dependent. The second step is methanol dehydration to DME using a solid acidic catalyst such as HZSM-5 (Protonated Zeolite Socony Mobil-5) or γ-Al2O3 (silica-alumina)[1,7,15,16]. Zeolite based catalysts due to their tunable acidity and stability have gained numerous advantages over alumina catalysts with respect to catalyst deactivation[4,17,18]. Moreover, zeolites allows low-temperature operation and hence the suppression of coke deposition and catalyst sintering[18]. Also, zeolite based catalysts are found to have good resistance to water poisoning[17,18]. The following catalytic methanol dehydration reaction occurs in this step: Methanol dehydration reaction: 2 C H 3O H 

C H 3O C H 3  H 2 O

[  H 2 9 8   2 3 .5 6 k J / m o l ]

(3)

And , the net rate of reaction is given by: 3C O  3H

2

C H 3O C H

3

 CO

2

[H

298

  2 4 5 .9 5 k J / m o l ]

(N2)

It is obvious reaction (1), (2) and (3) are not independent. The reactions are reversible and exothermic and subjected to thermodynamic limitations. The steps of this method are in series and can be separated or integrated. The separated steps are more costly since the dehydration step is affected by the price and availability of methanol [19,20]. The direct method is called a single-step method. In this method, syngas to dimethyl ether (STD) occurs in a single reactor using a bifunctional catalyst. The methanol synthesis and dehydration catalysts e.g. (CuO–ZnO–Al2O3/ HZSM-5) are combined to form the bifunctional catalyst. Different bifunctional catalyst preparation techniques have been reported by numerous investigators[1,2,12,21–26]. This promising field is still faces many challenges. The two catalytic sites of the bifunctional catalyst allow reactions (1)-(3) to

6

occur simultaneous for in-situ consumption and formation of the components with a strong synergy. This synergy between the reactions plays an important role in the displacement of the thermodynamic equilibriums and gives higher syngas conversion and DME yield compared to the indirect method [1,5,27]. Furthermore, the cost reduction by the direct method is significant. In recent years, the competitive advantages of the direct method have attracted increasing attention to this process. Different types of reactors such as slurry reactors, fixed bed reactors and fluidized bed reactors are involved in DME synthesis by this method. A comprehensive review of DME reactors has been given by Azizi et al. [7]. Lu et al. [1] have reported that the fluidized bed catalytic reactor for DME synthesis shows better performance than the slurry and fixed bed reactors. The comparison based on CO conversion, DME selectivity and productivity (XCO,SDME, PDME), respectively. The results reported as follows: the slurry reactor gives (17%, 70%, 0.2g/g/h)[27]; the fixed bed reactor gives (10.7,91.9,0.5g/g/h)[28]; and fluidized bed reactor gives (48.5%,97%,0.45g/g/h)[1]. As it can be seen that, the superiority of the fluidized is obvious. Moreover, the fluidized bed reactor has attractive features among them: (a) good mixing enables to achieve isothermality and temperature control; (b) negligible diffusion limitations due to fine particles used; (c) low pressure drop; (d) large amount of catalyst can be used; (e) easy to circulate the catalyst for regeneration. To the best of our knowledge, the application potential of multi-stage fluidized bed reactors for production of DME has been scarcely investigated in the literature. Therefore, the purpose of this investigation is to explore by mathematical modeling and simulation the potential and advantages that can be gained by implementing the multi-stage concept to the fluidized bed reactor for direct DME synthesis from syngas. Here, we considered various catalyst compositions at constant total catalyst load to evaluate the performance of the

7

reactor. The reactor performance is analyzed and the distinctive features of different catalyst bed configurations are investigated. It is essential to study these features in order to understand how to properly implement the optimal catalyst bed configurations in the design and operation of these types of important reactors.

2. Reaction kinetics Despite the fact that extensive literature has been published on methanol synthesis over the copper-zinc oxide catalysts, the investigators overwhelmed by the complexity of the subject and fail to provide an articulated unified kinetic model. The kinetics models reported in the literature show different views for the methanol synthesis from H2: CO mixture. Klier et al. [29] considered CO is the main source of carbon in methanol and small concentrations of CO2 have promotion effect attributed to its ability to oxidize Cu0 to Cu+, increasing catalyst activity. And, the inhibitory effect at high concentrations of CO2 owing to its strong adsorption on the active sites, decreasing catalyst activity. Graaf et al. [13,30] considered a scheme of three reactions for the three-phase methanol synthesis: the hydrogenation of CO, CO2 and the water gas shift reactions. They found that the hydrogenation of CO2 is the most important reaction. Bussche and Froment [14] considered only CO2 as a precursor to methanol and CO is converted to CO2 by the water gas shift reaction. The enhancement and inhibitory effects of CO2 ascribed to water formed by hydrogenation of CO2. At low concentrations of CO2, the water consumed by the water gas shift reaction favors both reactions. At high concentrations of CO2 the excessive water causes site blockage, decreasing catalyst activity. This view is supported by two reaction mechanisms of Lu et al. [1] and Bussche & Froment [14] and presented in Table 1. Surprisingly, the kinetic model based on this view able to predict satisfactory the experimental data reported by Klier et al. [29]. Despite differences in views, the reported results agreed on the observed phenomena of the enhancement effect of methanol synthesis

8

rate at small concentrations of CO2 and the inhibitory effect of water at high concentrations [3]. The reaction kinetics developed by Lu et al. [1] over a bifunctional catalyst composed of catalyst 1 (CuO-ZnO–Al2O3) and catalyst 2 (HZSM-5) for DME synthesis in a bubbling fluidized bed reactor is adopted in this study for many reasons among them: the kinetic parameters are obtained from the fluidized bed reactor experiments; satisfactory predication and based on a strong view. The rate expressions for reaction (1)-(3) are given as follows:

R1 

R2 



k 1  PC O PH  PC H O H PH O 2 2 3 2  1  K CO 

k 1  K 

2

2

/K

PC O  K C O PC O  2

K

1

2 PH  2 

H2

PH

2

 

(4)

3

 P H O  PC O P H / K 2 PC O  2 2 2  

CO 2

PC O  K 2

CO

 PC H O H PC H O C H 3 3 3 R3  k3    P K H 2O 3  2

2

PC O 

K

(5)

PH  H 2 2 

   

(6)

Where: k 1  3 5 .4 5 e x p   1 .7 0 6 9  1 0 / R / T



4

k

k

,

k g s b ar

2

 7 .3 9 7 6 e x p   2 .0 4 3 6  1 0 / R / T

3

 8 .2 8 9 4  1 0 e x p   5 .2 9 4 0  1 0 / R / T

4

4

4

k m ol



,

k m ol

2

(7)

(8)

k g s b ar



,

k m ol

(9)

k g s b ar

The adsorption equilibrium constants for CO2, CO and H2 are given by the following equations [1,14,27]:

9

K

K

K

CO 2

 1 .0 2  1 0

7

7

e x p  6 .7 4 0 0  1 0 / R / T



,

b ar

e x p  5 .8 1 0 0  1 0 / R / T



,

b ar

,

b ar

4

CO

 7 .9 9  1 0

H 2

 0 .2 4 9 e x p  3 .4 3 9 4  1 0 / R / T

4

4



1

(10)

1

(11)

1

(12)

and the equilibrium constants K1, K2 and K3 are given by [1,27] : ln K 1  4 2 1 3 / T  5 .7 5 2 ln T  1 .7 0 7  1 0

3

T  2 .6 8 2  1 0

6

T

2

 7 .2 3 2  1 0

10

T

3

 1 7 .6

(13) lo g 1 0 K 2  2 1 6 7 / T  0 .5 1 9 4 lo g 1 0 T  1 .0 3 7  1 0 ln K 3  4 0 1 9 / T  3 .7 0 7 ln T  2 .7 8 3  1 0

3

3

T  2 .3 3 1  1 0

T  3 .8  1 0

7

T

2

7

T

2

 1 .2 7 7 7

 6 .5 6 1  1 0 / T 4

3

 2 6 .6 4

(14) (15)

It is worth mentioning that many investigators [1,5,31] have implemented the DME production scheme of reaction (1)-(3).

3. Mathematical modeling The two-phase model is well established and the most reliable model for the fluidized bed reactor. This model has been used widely used to model successfully important industrial processes such as steam reforming of methane. For the first time, Lu et al. [1] have been able to develop a new mechanism and reaction kinetics for DME synthesis using experimental data obtained from a laboratory fluidized bed reactor. Furthermore, the developed kinetics was implemented in the two-phase model to simulate the fluidized bed reactor based on the assumption that the dense phase is fully back-mixed (P-M model). Table 2 shows the prediction accuracy of the model compared to experimental results as reported by Lu et al. [1] at SV=3000 ml/gcat/h and H2/CO=1.0. The experimental data estimated from the points on the plots provided by Lu et al. [1].

10

In this study the reliable two-phase model is implemented for the numerical simulation of the multi-stage fluidized bed reactor. Unfortunately there is not yet experimental data available for this new novel configuration as a basis for the comparison. A schematic diagram of the two-bed freely bubbling fluidized bed reactor is shown in Fig. 1(a). Also, a schematic representation of the bubble and dense phases of the two-phase model are depicted in Fig.1(b). The model equations are derived based on the following main assumptions: an isothermal steady state operation, the bubbling bed is composed of the dense and bubble phases, the bubble phase in plug flow, the bubbles are spherical, the dense phase is well mixed, ideal gas behavior, there is no catalyst deactivation, and the intraparticle diffusion resistance is neglected due to the small size of the catalyst particle. 3.1 Conservation equations for the bubble phase DME reaction scheme involves 6-components: CO2, H2, CH3OH, H2O, CO, and CH3OCH3. A differential molar balance on the ith component in the bubble phase gives: dC

ib

 

 K bd  i

A b C

dh

ib

C

id



i  1, ..., 6

,

(16)

Gb

With the initial conditions: C

ib

(0)  C

id

(0)  C

(17)

iF

where the subscript b and d refer to bubble and dense phases, respectively. Defining the following dimensionless state variables and parameters: Y ib  C ib / C r e f , Y id  C id / C r e f , Z  h / H ,

 i   K bd

i

Y if  C iF / C r e f ,

Ab H / G b ,

 i   K bd

i

(18) Ab H / G d

And, the average dimensionless concentration of component i in the bubble phase: H

Y ib 

Y 0

1 ib

dh / H 

Y

ib

(19)

dZ

0

11

Hence, equation (16) gives: Y

ib

Y

id

 Y

iF

Y

id

 1 

exp(  i ) /  i

(20)

3.2 Conservation equations for the dense phase The molar balance on the ith component in the dense phase gives: G d C id  G d C iF 

H

 K  bd

0

i

A b  C ib  C id  d h 



2



j 1



 P (1   m f ) (1   ) A H   1   ij R j   2  i 3 R 3  ,

(21) i  1, ..., 6



where  i j is the stoichiometric coefficient of component i in the jth reaction (positive for products, negative for reactants and zero for inerts) and θ1 and θ2 are the mass fraction of catalyst 1 and 2, respectively. Defining the dimensionless rate of reaction as: R    P (1   m f )(1   ) A H R / C

re f

(22)

Gd

Substituting equation (20) into the dimensionless form of equation (21), we obtain: 2

(Y

iF

Y

id

)   i (Y

ib

Y

id

)   1   ij R j  2  i 3 R 3  0 ,

i  1, ..., 6

(23)

j 1

Total methanol and actual methanol yield are defined by the following relations [1,5,27]: T o ta l m e th a n o l y ie ld (% ) 

T o ta l m e th a n o l fo r m e d

 100

T o ta l c a r b o n in 

M e th a n o l o u tp u t  M e th a n o l c o n v e r te d to D M E

F

CO2

 FC O



in

12

 100 

FC H

3O H

F

 2 FC H

CO2

 FC O

(24) 3O C H 3



in

 100

A c tu a l m e th a n o l y ie ld ( % ) 

M e th a n o l o u tp u t

 100 

T o ta l c a r b o n in

FC H

F

CO2

3O H

 FC O



 100 in

(25) and DME yield and selectivity are defined as: D M E y ie ld ( % ) 

M e th a n o l c o n v e r te d to D M E

 100 

to ta l c a r b o n in

D M E s e le c tiv ity ( % ) 

M e th a n o l c o n v e r te d to D M E

2 FC H

F

CO2

3O C H 3

 FC O

100

FC H

3O H

(26)

in

2 FC H

 100 

to ta l m e th a n o l fo r m e d



3O C H 3

 2 FC H

 100

(27)

3O C H 3

The set on of non-linear algebraic equations (23) are solved by an IMSL subroutine called ZSPOW. The Golden search method is used to calculate the locus of the maxima.

4. Results and Discussion The hydrodynamic parameters of the bubbling fluidized reactor are given in Table 3 [1,32], and the data used for the simulation are given in Table 4 [1]. The content of CO2 and CO in the feed is an important factor for the production of DME because these components are the carbon sources for the reactions and also used to characterize the feed. We utilized the content of these components in the feed as a key parameter. In this study, we implemented the concentration range of Cox in the feed [CO2+CO=0-25 mol%] used by several investigators [3,14,29,31]. Also, we denote the methanol synthesis and dehydration catalysts as catalyst 1 and 2, respectively. Different bed configurations are considered and all beds have the same amount of catalyst and the total amount of the catalyst is kept constant to ensure fair comparison.

dP

150.0 µm

13

.

4.1 Single-bed reactor configuration The fundamental understanding of the single-bed configuration is essential for the study of the two-bed configuration because this stage feeds the second stage. Here, we considered various catalyst compositions to evaluate the performance of the fluidized bed reactor. A limiting case of catalyst 1 alone [A(θ1=100%, θ2=0%)] is considered to represent methanol production by methanol synthesis reactions. In addition to three bifunctional catalyst compositions as follows: a bifunctional catalyst rich in catalyst 1 and lean in catalyst 2 [B (θ1=80%, θ2=20%)], a bifunctional catalyst with equal catalysts loading [C (θ1=50%, θ2=50%)], a bifunctional catalyst lean in catalyst 1 and rich in catalyst 2 [D (θ1=20%, θ2=80%)]. Beside another limiting case of catalyst 2 alone [E (θ1=0%, θ2=100%)]. This limiting case of methanol dehydration has no practical meaning for DME production in the single-bed configuration, since the methanol dehydration reaction cannot occur without methanol, but can be considered as a frame of reference. The total methanol yield produced by these catalyst configurations as a function of CO2 concentration in the feed is depicted in Fig. 2. As it can be seen that for a feed with low CO2 concentration, a significant increase of the total methanol yield is obtained by implementing the bifunctional catalyst compared to only methanol synthesis process, A(100%, 0%). The total methanol yield decreases steadily as CO2 content in the feed is increased.

It is interesting to note that two types of maxima

phenomena can be identified. The first maximum point is clearly shown in the profile of the limiting case A(100%, 0%). This maximum point has been developed as the total methanol yield varies with CO2 concentration in the feed. The maximum point appears due to the increase and decrease of rate of reactions as demonstrated by Fig.3. The rate of reactions

14

increase and decrease owing to conflicting effects of positive and negative synergic coupled with water inhibition effect. According to the Le Châtelier's principle of equilibrium, reaction (1) and (2) are reversible with a strong synergy between them. The positive synergetic effect promotes kinetically and thermodynamically reaction (1) and (2) in order to improve the total methanol yield as follows: water formed by reaction (1) is consumed by reaction (2), and H2 and CO2 produced by reaction (2) favor reaction (1). The negative synergic effect becomes significant when the prevailing reaction conditions favor backward rates, causing excessive water formation and decreasing the rate of reactions. It is clearly shown in Fig. 3 that the increase of CO2 in the feed has a more pronounced effect on the water-gas shift reaction (2), and the reaction is completely reversed after the equilibrium point (EP). This maximum phenomenon type was also observed by Ng et al. [5] in an internal recycle reactor. The second maximum phenomenon is not clearly observed in Fig. 2. As it can be seen in Fig. 2 that, the total methanol yield increases and decreases by increasing the methanol dehydration catalyst (catalyst 2) load in the bifunctional catalyst composition. The maximum is principally formed when the methanol dehydration catalyst (catalyst 2) increases at the expense of the methanol synthesis catalyst (catalyst 1). In fact the formation of this maximum is more complex than the first case because the composition of the bifunctional catalyst and the methanol dehydration reaction (3) become additional effective dimensions on the synergy between the reactions. Two examples of the effect of the feed composition on the total methanol yield as a function of bifunctional catalyst composition are depicted in Fig. 4. The maxima are clearly shown and their values decrease with the increase of CO2 content in the feed. To follow this behavior for a wide range of the feed composition, the locus of the maxima is also shown as a dotted line. It is interesting that the maxima are confined to a small range of bifunctional catalyst composition, θ2=48.5%-50.4%.

15

Fig. 5. Single bed configuration: actual methanol yield as a function of CO2 concentration in the feed for various catalyst compositions. In Fig. 5 is shown the influence of the bifunctional catalyst composition on the actual methanol yield as a function of the feed composition. The results indicate that for a specific reactor feed concentration the actual methanol yield decreases as the concentration of the methanol synthesis catalyst (catalyst 1) decreases in the bifunctional catalyst composition. As one can see, for a certain catalyst bed composition that the actual methanol yield profiles assume maximum values with the increase of CO2 concentration in the feed. This maximum is formed as a balance between the rates of methanol produced by reaction (1) and consumed by reaction (3). The locus of maxima shows that, the locations of the maxima are shifted to the right as the concentration of catalyst 1 decreases and CO2 concentration in the feed increases. This shift can be thought of as impact of compensation of CO2 concentration on reaction (1) due to the decrease of catalyst 1.

Fig. 6 shows DME yield versus CO2 content in the feed for various bifunctional catalyst compositions. For all bifunctional catalyst compositions, the DME yield decreases as CO 2 concentration in the feed is increased. It is interesting to note that at a specific feed composition the DME yield increases and decreases as the dehydration catalyst loading is increased leading to the appearance of the maximum phenomenon. This maximum phenomenon is exemplified by two feed compositions as shown in Fig. 7. The maxima are formed due to the fact that the increase of catalyst 2 loading will be at the expense of catalyst 1 loading. The locus of maxima is presented for the whole range of the feed compositions. The corresponding DME selectivity is presented in Fig. 8. As it can be seen that the DME selectivity is favored at a high CO concentration in the feed and the bifunctional catalyst is rich

16

in the methanol dehydration catalyst because at these conditions the water-gas shift reaction has a strong positive effective impact on the reaction synergy.

4.2. Two-bed configuration The fluidized bed reactor with two-bed configuration has a very large possible number of permutations of the bifunctional catalyst loadings. In this study, the first bed composition is kept constant at an optimum value of (θ11= 52%, θ21=48%) obtained from the results of the above single bed investigation while the second bed composition is varied. The same three bifunctional catalyst compositions used in the single bed are utilized in the study of the second bed as follows: a bifunctional catalyst rich in the methanol synthesis catalyst (80%, 20%), an equal loading bifunctional catalyst composition (50%, 50%), a bifunctional catalyst rich in the methanol dehydration catalyst (20%, 80%). Fig. 9 shows the influence of the bifunctional catalyst composition in bed2 on the actual methanol yield as a function of CO2 concentration in the feed. The results clearly show that the second bed exhibits the maximum phenomenon in a similar way to the first bed. It is once more revealed that the role of the strong synergistic effect between the reactions is responsible for the maxima formation as explained earlier. However, the maxima are shifted to the right in the second bed case. The corresponding DME yield and selectivity obtained from the second bed are presented in Figs. 10 and 11, respectively. It is evident that the addition of the second bed (bed2) gives a substantial increase

of the DME yield compared to the first bed (bed1) and in particular for CO rich feeds. About 32.67% increase in DME yield is achieved by the second bed (bed2) at a feed concentration of [CO2/(CO2+CO)=0.01] and a catalyst composition of (50%,50%). This improvement is achieved owing to the fact that the second bed has different feed and bed composition compared to the first bed and single bed configuration. It is seen that the enhancement of the DME yield decreases steadily with the increase of CO2 concentration in the feed attributed to

17

the increasing of the water inhibition effect. Also, a significant enhancement of the DME selectivity is achieved by the second bed (bed2). Almost 100% DME selectivity is obtained at a feed of low CO2 concentration and bed composition of (50%, 50%). The DME yield and selectivity are favored at a low CO2 concentration in the feed owing to the effective positive synergy as discussed earlier. It is interesting to observe that at a specific feed composition the DME yield and selectivity increases and decreases as the concentration of the catalyst 2 increases in the composition of the bifunctional catalyst. This behavior leads to the appearance of the maxima phenomena of the DME yield and selectivity as shown in Figs. 12 and 13, respectively. The reasons for the formation of such maxima have been offered in details earlier.

5. Conclusions A mathematical model has been utilized to explore the potential application of a multi-stage fluidized bed catalytic reactor for efficient production of DME. The simulation results indicated that the two-bed configuration gives a substantial enhancement of DME yield (32.67%) compared to the yield from the single-bed. Also, DME selectivity of almost 100% is obtained. This significant improvement achieved, because the second bed has different feed condition and bifunational catalyst composition. It has been shown that the performance of the second bed is very sensitive to the catalyst bed composition. Interesting maxima phenomena are observed. The actual methanol yield assumes a maximum value due to the balance between methanol consumption and production rates. The location of the maxima is affected by the catalyst bed and feed composition. Also, DME yield shows optimal values due to the balance between the amount of catalyst 1 and 2 in the bifunctional catalyst composition. The promising results of this study suggest that the multi-stage fluidized bed catalytic reactor can favorably be employed in industrial application to enhance DME yield and

18

selectivity by the direct process. Finally, future research should focus on the rigorous optimization and control of the process and application of various bifunctional catalysts patterns. Also, coupling the exothermic reactions in this study with appropriate endothermic reactions for energy integration, displacement of the thermodynamic barriers and product variety.

Acknowledgements This project was supported by King Saud University, Deanship of Scientific Research, College of Engineering Research Center.

References [1]

W.-Z. Lu, L.-H. Teng, W.-D. Xiao, Simulation and experiment study of dimethyl ether synthesis from syngas in a fluidized-bed reactor, Chem. Eng. Sci. 59 (2004) 5455–5464. doi:10.1016/j.ces.2004.07.031.

[2]

S. Asthana, C. Samanta, A. Bhaumik, B. Banerjee, R.K. Voolapalli, B. Saha, Direct synthesis of dimethyl ether from syngas over Cu-based catalysts: Enhanced selectivity in the presence of MgO, J. Catal. 334 (2016) 89–101. doi:10.1016/j.jcat.2015.10.020.

[3]

F. Dadgar, R. Myrstad, P. Pfeifer, A. Holmen, H.J. Venvik, Direct dimethyl ether synthesis from synthesis gas: The influence of methanol dehydration on methanol synthesis reaction, Catal. Today. (2015). doi:10.1016/j.cattod.2015.09.024.

[4]

M. Cai, A. Palčić, V. Subramanian, S. Moldovan, O. Ersen, V. Valtchev, V.V. Ordomsky, A.Y. Khodakov, Direct dimethyl ether synthesis from syngas on copper– zeolite hybrid catalysts with a wide range of zeolite particle sizes, J. Catal. 338 (2016)

19

227–238. doi:10.1016/j.jcat.2016.02.025. [5]

K.L. Ng, D. Chadwick, B.A. Toseland, Kinetics and modelling of dimethyl ether synthesis from synthesis gas, Chem. Eng. Sci. 54 (1999) 3587–3592. doi:10.1016/S0009-2509(98)00514-4.

[6]

M. Alavi, H. Jazayeri-rad, R.M. Behbahani, Optimizing the Feed Conditions in a DME Production Process to Maximize the Methanol Conversion, Sci. Technol. 3 (2013) 61– 66. doi:10.5923/j.scit.20130302.03.

[7]

Z. Azizi, M. Rezaeimanesh, T. Tohidian, M.R. Rahimpour, Dimethyl ether: A review of technologies and production challenges, Chem. Eng. Process. Process Intensif. 82 (2014) 150–172. doi:10.1016/j.cep.2014.06.007.

[8]

F. Manenti, A.R. Leon-Garzon, Z. Ravaghi-Ardebili, C. Pirola, Systematic staging design applied to the fixed-bed reactor series for methanol and one-step methanol/dimethyl ether synthesis, Appl. Therm. Eng. 70 (2014) 1228–1237. doi:10.1016/j.applthermaleng.2014.04.011.

[9]

J. Xia, D. Mao, B. Zhang, Q. Chen, Y. Tang, One-step synthesis of dimethyl ether from syngas with Fe-modified zeolite ZSM-5 as dehydration catalyst, 98 (2004) 1–6.

[10] S. Hosseininejad, A. Afacan, R.E. Hayes, Catalytic and kinetic study of methanol dehydration to dimethyl ether, Chem. Eng. Res. Des. 90 (2012) 825–833. doi:10.1016/j.cherd.2011.10.007. [11] N. Diban, A.M. Urtiaga, I. Ortiz, J. Ereña, J. Bilbao, A.T. Aguayo, Influence of the membrane properties on the catalytic production of dimethyl ether with in situ water removal for the successful capture of co2, Chem. Eng. J. 234 (2013) 140–148. doi:10.1016/j.cej.2013.08.062. [12] J. Ereña, I. Sierra, A.T.A.T. Aguayo, A. Ateka, M. Olazar, J. Bilbao, J. Ere??a, I. Sierra, A.T.A.T. Aguayo, A. Ateka, M. Olazar, J. Bilbao, Kinetic modelling of dimethyl ether

20

synthesis from (H2+CO2) by considering catalyst deactivation, Chem. Eng. J. 174 (2011) 660–667. doi:10.1016/j.cej.2011.09.067. [13] G.H. Graaf, P.J.J.M. Sijtsema, E.J. Stamhuis, G.E.H. Joosten, Chemical Equilibria in Methanol Synthesis, Chem. Eng. Sci. 41 (1986) 2883–2890. [14] K.M. Vanden Bussche, G.F. Froment, A Steady-State Kinetic Model for Methanol Synthesis and the Water Gas Shift Reaction on a Commercial Cu / ZnO / Al 2 O 3 Catalyst, J. Catal. 10 (1996) 1–10. doi:10.1006/jcat.1996.0156. [15] I. Sierra, J. Ere??a, A.T. Aguayo, A. Ateka, J. Bilbao, Kinetic modelling for the dehydration of methanol to dimethyl ether over ??-Al2O3, Chem. Eng. Trans. 32 (2013) 613–618. doi:10.3303/CET1332103. [16] M. Fazlollahnejad, M. Taghizadeh, A. Eliassi, G. Bakeri, Experimental Study and Modeling of an Adiabatic Fixed-bed Reactor for Methanol Dehydration to Dimethyl Ether, Chinese J. Chem. Eng. 17 (2009) 630–634. doi:10.1016/S1004-9541(08)60255-4. [17] V. V. Ordomsky, M. Cai, V. Sushkevich, S. Moldovan, O. Ersen, C. Lancelot, V. Valtchev, A.Y. Khodakov, The role of external acid sites of ZSM-5 in deactivation of hybrid CuZnAl/ZSM-5 catalyst for direct dimethyl ether synthesis from syngas, Appl. Catal. A Gen. 486 (2014) 266–275. doi:10.1016/j.apcata.2014.08.030. [18] A. García-Trenco, A. Martínez, A rational strategy for preparing Cu-ZnO/H-ZSM-5 hybrid catalysts with enhanced stability during the one-step conversion of syngas to dimethyl ether (DME), Appl. Catal. A Gen. 493 (2015) 40–49. doi:10.1016/j.apcata.2015.01.007. [19] F. Hayer, H. Bakhtiary-Davijany, R. Myrstad, A. Holmen, P. Pfeifer, H.J. Venvik, Synthesis of dimethyl ether from syngas in a microchannel reactor-Simulation and experimental study, Chem. Eng. J. 167 (2011) 610–615. doi:10.1016/j.cej.2010.09.080. [20] M. De Falco, M. Capocelli, G. Centi, Dimethyl ether production from CO2 rich

21

feedstocks in a one-step process: Thermodynamic evaluation and reactor simulation, Chem. Eng. J. 294 (2016) 400–409. doi:10.1016/j.cej.2016.03.009. [21] F. Song, Y. Tan, H. Xie, Q. Zhang, Y. Han, Direct synthesis of dimethyl ether from biomass-derived syngas over Cu–ZnO–Al2O3–ZrO2(x)/γ-Al2O3 bifunctional catalysts: Effect of Zr-loading, Fuel Process. Technol. 126 (2014) 88–94. doi:10.1016/j.fuproc.2014.04.021. [22] Y.J. Lee, M.H. Jung, J.B. Lee, K.E. Jeong, H.S. Roh, Y.W. Suh, J.W. Bae, Single-step synthesis of dimethyl ether from syngas on Al2O 3-modified CuO-ZnO-Al2O3/ferrierite catalysts: Effects of Al2O3 content, Catal. Today. 228 (2014) 175–182. doi:10.1016/j.cattod.2013.11.015. [23] S. Allahyari, M. Haghighi, A. Ebadi, H. Qavam Saeedi, Direct synthesis of dimethyl ether as a green fuel from syngas over nanostructured CuO-ZnO-Al2O3/HZSM-5 catalyst: Influence of irradiation time on nanocatalyst properties and catalytic performance, J. Power Sources. 272 (2014) 929–939. doi:10.1016/j.jpowsour.2014.07.152. [24] S. Papari, M. Kazemeini, M. Fattahi, Modelling-based optimisation of the direct synthesis of dimethyl ether from syngas in a commercial slurry reactor, Chinese J. Chem. Eng. 21 (2013) 611–621. doi:10.1016/S1004-9541(13)60505-4. [25] Z. Li, J. Li, M. Dai, Y. Liu, D. Han, J. Wu, The effect of preparation method of the Cu– La2O3–ZrO2/γ-Al2O3 hybrid catalysts on one-step synthesis of dimethyl ether from syngas, Fuel. 121 (2014) 173–177. doi:10.1016/j.fuel.2013.12.050. [26] N. Khandan, M. Kazemeini, M. Aghaziarati, Dehydration of Methanol to Dimethyl Ether Employing Modified H-ZSM-5 Catalysts, Iran. J. Chem. Eng. 6 (2009) 3–11. [27] Z.L. Wang, J.F. Wang, J. Diao, Jin Y., The synergy effect of process coupling for dimethy; ether synthesis in slurry reactors, Chem. Eng. Technol. 24 (2001) 507–511.

22

[28] W.Z. Lu, L.H. Teng, W.D. Xiao, Simulation analysis of fixed-bed reactor for dimethyl ether synthesis, Nat. Gas Chem. Ind. 27 (2002) 53–61. https://www.mendeley.com/library/. [29] K. Klier, V. Chatikavanu, R.G. Herman, G.W. Simmons, Catalytic Synthesis of Methanol from CO / H2. IV. The Effects of Carbon Dioxide, 360 (1982) 343–360. [30] G.H. Graaf, J.G.M. Winkelman, E.J. Stamhuis, A.A.C.M. Beenackers, Kinetics of the three phase methanol synthesis, Chem. Eng. Sci. 43 (1988) 2161–2168. https://www.mendeley.com/library/. [31] I. Iliuta, M.C. Iliuta, F. Larachi, Sorption-enhanced dimethyl ether synthesis— Multiscale reactor modeling, Chem. Eng. Sci. 66 (2011) 2241–2251. https://www.mendeley.com/library/. [32] D. Kunii, O. Levenspiel, Fluidization Engineering, Butterworth-Heinemann, USA, 1991.

23

Fig. 1.Schematic diagram of the multi-stage catalytic fluidized bed reactor: (a) two beds, and (b) representation of the two-phase model.

H2/COx=3: CO2+CO=25%, H2=75%

thanol yield (%)

50.00

Single bed A (100%, 0%) B (80%, 20%) C (50%, 50%) D (20%, 80%) E (0%, 100%)

C B

40.00 D 30.00 20.00

A

24

Limiting case

Fig. 2. Single bed configuration: total methanol yield as a function of CO2 concentration in the feed for various catalyst compositions.

25

H2/COx=3: CO2+CO=25%, H2=75% Single bed: A(100% , 0%)

Reaction rate106, kmol/kg/s

3.00

Limiting case

2.00 R1 R2

1.00 0.00 -1.00 -2.00 0.00

EP

0.20

0.40

0.60

0.80

1.00

[CO2/(CO2+CO)]inlet

Fig. 3. Single bed configuration: rates of reaction (1) and (2) as a function of CO2 concentration in the feed.

[CO2/(CO2+CO)]inlet

Single bed Locus of maxima

methanol yield (%)

50.00 40.00

a

30.00 20.00

26 b

a=0.2 b=0.6

Fig. 4. Single bed configuration: total methanol yield as a function of catalyst 2 compositions for various feed compositions.

Locus of maxima

Actual methanol yield (%)

10.00 B 8.00 C 6.00 4.00 2.00

27 D

Single bed B (80%, 20%) C (50%, 50%) D (20%, 80%)

Fig. 5. Single bed configuration: actual methanol yield as a function of CO2 concentration in the feed for various catalyst compositions.

Single bed B (80%, 20%) C (50%, 50%) D (20%, 80%)

50.00

DME yield (%)

40.00 30.00

C B D

20.00

28 10.00 0.00

Fig. 6. Single bed configuration: DME yield as a function of CO2 concentration in the feed for various catalyst compositions.

Single bed Locus of maxima

50.00

DME yield (%)

40.00

a

30.00 20.00 b

10.00 0.00

29

[CO2/(CO2+CO)]inlet a=0.2 b=0.6

Fig. 7. Single bed configuration: maximum phenomena of DME yield.

Single bed B (80%, 20%) C (50%, 50%) D (20%, 80%)

100.00

DME selectivity (%)

D 80.00

C B

60.00 40.00 20.00 0.00 0.00

30 0.20

0.40

0.60

[CO2/(CO2+CO)]inlet

0.80

1.00

Fig. 8. Single bed configuration: DME selectivity as a function of CO2 concentration in the feed for various catalyst compositions.

Bed2 (80%, 20%) (50%, 50%) (20%, 80%)

Bed1 (52% , 48%)

Actual methanol yield (%)

10.00

8.00

6.00

4.00

31 Locus of maxima

2.00 0.00

0.20

0.40

0.60

[CO /(CO +CO)]

0.80

1.00

Fig. 9. Two-bed configuration: actual methanol yield as a function of CO2 concentration in the feed for various catalyst compositions.

Bed1 (52% , 48%)

Bed2 (80%, 20%) (50%, 50%) (20%, 80%)

60.00

DME yield (%)

50.00 40.00 30.00 20.00 10.00 0.00 0.00

32 0.20

0.40

0.60

[CO2/(CO2+CO)]inlet

0.80

1.00

Fig. 10. Two-bed configuration: DME yield as a function of CO2 concentration in the feed for various catalyst compositions.

Bed1 (52% , 48%)

Bed2 (80%, 20%) (50%, 50%) (20%, 80%)

DME selectivity (%)

100.00

80.00

60.00

40.00

20.00 0.00

33 0.20

0.40

0.60

0.80

1.00

Fig. 11. Two-bed configuration: DME selectivity as a function of CO2 concentration in the feed for various catalyst compositions.

Bed2

[CO2/(CO2+CO)]inlet

60.00

a=0.2 b=0.6 Locus of maxima

DME yield (%)

50.00 a

40.00 30.00 b

20.00 10.00 0.00

34 25.00 50.00 75.00 100.00

Fig. 12.Two-bed configuration: DME yield maximum phenomena.

35

Bed2

DME selectivity(%)

100.00

[CO2/(CO2+CO)]inlet

a

a=0.2 b=0.6

b

90.00

80.00

Locus of maxima

70.00 0.00

25.00 50.00 75.00 100.00

(%)

Fig. 13. Two-bed configuration: DME selectivity maximum phenomena.

Bussche & Froment [14]

Lu et al. [1]

H 2( g)  2s

2 H .s

(1a)

2C u2  H 2

2C u2 H

CO2( g)  s

O .s  C O ( g )

(2a)

C O2  C u2

O C O

C O 2 ( g )  O .s  s

C O 3 .2 s

|

(3a)

(1b)

|

Cu  Cu

36

(2b)

C O 3 .2 s  H . s

H C O 3 .2 s  s

H C O 3 .2 s  s

(4a)

H C O 2 .2 s  O . s

H C O 2 .2 s  H . s

|

(5a)

|

 2C u2 H

C u 2C H 2O  C u 2O  C u 2

H 2 C O 2 .2 s  s

(6a) (7a)

C u 2C H 2O  C u 2 H

C u 2C H 3O  C u 2

(4b)

C u 2C H 3O H  C u 2

(5b)

H 2C O .s  H .s

H 3C O .s  s

(8a)

C u2C H 3O  C u2 H

H 3C O .s  H .s

C H 3O H ( g )  2 s

(9a)

C u2C H 3O H

(10a)

C u 2O  2C u 2 H

(11a)

C u2 H 2O

(12a)

H 2O  C u2

C u2O  H 2

C O  C u2O

C u2  C O 2

O .s  H .s

O H .s  s

O H .s  H .s

H 2O .s  s

H 2O ( g )  s

H 2O .s

(3b)

Cu  Cu

H 2C O .s  O .s

H 2 C O 2 .2 s

O C O

(6b)

C u2  C H 3O H 2C u 2  C u 2 H 2O

C u2  H 2O

(7b) (8b) (9b) (10b)

Table 1: Reaction mechanisms for methanol synthesis. ____________________________________________________________________ Pressure (MPa)

Temperature (oC)

2.0

250.0 260.0 27.0

3.0

4.0 CO conversion (%)

Experimental

28.7

40.0

48.6

30.0

40.0

45.1

Model

28.9

39.1

46.9

31.1

39.1

44.6

______________________________________________________________________

Table 2: Comparison between experimental and the two-phase model prediction [1].

Ar   g

 mf



p

 g

 1 .0   0 .5 8 6    Ar 

g

d

3 p

(c1)

g

0 .0 2 9

2

 g     p

   

0 .0 2 1

(c2)

37

U

mf

g



 

gd p

Gd  U

mf

G b  U

o

2 5 .2 5  0 .0 6 5 1 A r  2 5 .2 5   2

(c3)

(c4)

A

U

mf

A

(c5)

d b m  0 .6 5 2  A  U 

U

o



d b o  0 .3 4 7 7 .8 5  1 0

5

mf

U

  o

0 .4

(c6)

U



mf

0 .4

(c7)

  0 .3 h  d b  d bm   d bm  d bo  e x p    D 

(c8) Ub U

 

U

o

U

o

U U

H 

H

mf

 0 .7 1 1  g d b

(c10)

b

mf

(c11)

1    T P V i n

im

(c9)

mf

D ij  0 .0 4 3 5 7

D



0 .5

 1  y

i

/ j 1 j i

3/2

V

1/3

y

j

D

ij

1 1/3 j



2

M

(c12) j

 m f D im U b

 6 .7 8

 K bc i

 U mf   D im g  4 .5    5 .8 5  5/4  db   db

(c17)

3

db

1/ 2

 K bd i

M

i

(c13)

 K cd  i

1

1





1

 K cd  i



1/ 4

  

(c18)

1

(c19)

 K bc i

Table 3: Hydrodynamic parameters for the bubbling fluidized bed reactor [1,32]. D

0.026 m

Hmf T P Uo

1.0 m 533.15 K 50.0 bar 0.06243 m/s

Umf

0.02193 m/s

38

ρp

Table 4.

1982.5 kg/m3

Data for the fluidized bed reactor [1]

39